The Quality of Care Among Older Adults with Diabetes Comorbid with Other Chronic Conditions

The quality of care among older adults with diabetes comorbid with other chronic conditions

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy (PhD) Institute of Health Policy, Management & Evaluation University of Toronto

The quality of care among older adults with diabetes comorbid with other chronic conditions

Yelena Petrosyan Doctor of Philosophy Institute of Health Policy, Management and Evaluation University of Toronto

2017

Abstract

The management of people with multiple chronic conditions requires understanding the extent to which concurrent chronic conditions contribute and interact to affect the patient’s health status, as well as assessing the risk and benefits of various strategies for the treatment of complex needs in patients. A single condition focus in both clinical care and research remains and limits the assessment of care for people with multiple chronic conditions.

The overall aim of this project was to evaluate the quality of overall care for older adults with selected disease combinations in ambulatory care settings. The first study aimed to identify a set of evidence-based and valid quality indicators for evaluating ambulatory care for older adults with selected chronic conditions, including diabetes, hypertension, chronic ischemic heart disease, major depression and osteoarthritis. The second study aimed to critically appraise the identified quality indicators and select a set of indicators for evaluating the quality of care for older adults with diabetes with comorbid concordant and discordant chronic conditions. The third study aimed to examine the difference in the quality of care between patients with 2 vs. 1 selected concordant vs. discordant comorbid conditions, and to examine associations of quality of care and hospitalizations among older adults with selected disease combinations.

The study findings suggest that older adults with diabetes are at risk of suboptimal care with additional selected comorbid conditions, especially those with discordant comorbid

ii conditions. The study findings also support the importance of continuity of care for older diabetes patients with comorbid chronic conditions. The study findings suggest that the likelihood of hospitalizations increases with the number of prescribed drugs among older adults with comorbidities.

There is a need for a holistic approach in education and clinical care of older adults with diabetes taking into account concomitant conditions that affect patient’s health status. Future research is needed for measuring the quality of care in the larger diabetes population and reporting by different stratifications, including age, sex, primary care models to see if there are any patterns in certain groups and target the interventions towards improving practices for specific sub groups.

iii

Acknowledgments

I would like to dedicate this work in memory of my mother, Hripsime Baghdasaryan. This work is also dedicated to my father, Rafael Petrosyan and brother, Armen Petrosyan for always supporting, helping, and encouraging me. Despite the fact that we have been living far apart during my graduate studies, their endless support and unconditional love have always been with me through all the ups and downs of this long journey.

I would like to sincerely thank my supervisor, Dr. Walter Wodchis for his guidance and mentorship over the past four years. I am also grateful for all of the research opportunities that he exposed me to during past four years. I would like to express my appreciation to my thesis committee members, Dr. Jan Barnsley, Dr. Kerry Kuluski, and Dr. Barbara Liu for providing insightful comments and questions on my research as it developed.

I would like to extend my sincere thanks to Yeva Sahakyan, who helped me to accomplish the systematic reviews. I would like to express thanks to Chris Bai and Andrew Calzavara, ICES UofT, for their continuous support. I would like to thank all of the health care providers/researchers who participated in this research.

I would like to express sincere thanks to my colleagues and friends from the Toronto Health Economics and Technology Assessment (THETA) Collaborative, UofT, for nurturing environment and support over the past two years.

I want to thank the Health System Performance Research Network (HSPRN) for providing me with financial support during my study years.

Table of Contents

Acknowledgments ...... iv List of Tables ...... ix List of Figures ...... xii List of Appendices ...... xiii

CHAPTER 1: Introduction ...... 1 1.1 Chronic conditions, aging and multimorbidity ...... 1 1.2 Multimorbidity among patients with diabetes ...... 2 1.3 Multimorbiditiy’s issues/challenges ...... 5 1.3.1 Multimorbidity: high health care utilization ...... 5 1.3.2 Management of older adults with multiple chronic conditions ...... 7 1.3.3 Multimorbidity: implications for primary care ...... 9 1.3.4 Continuity of care ...... 11 1.3.4 Quality indicators in health care ...... 13 1.4 Performance Measurement for People with Multimorbidity: Conceptual model ...... 15 1.5 Summary ...... 19 1.6 Thesis project outline ...... 20 1.7 Broad project objectives ...... 21

CHAPTER 2: Systematic review of quality indicators for care for older adults with diabetes, hypertension, chronic ischemic heart disease, major depression and osteoarthritis ...... 24 Abstract ...... 24 2.1 Introduction ...... 26 2.2 Rationale for the study ...... 28 2.3 Study objectives ...... 30 2.4 Methods ...... 30 2.4.1 Search strategy ...... 30 2.4.2 Study selection ...... 32 2.4.3 Methodological assessment ...... 34 2.4.4 Data extraction ...... 36 2.5 Results ...... 37 2.5.1 Systematic review of quality indicators for care for older adults with diabetes ...... 37 2.5.1.1 Search results ...... 37

2.5.1.2 Study characteristics ...... 38 2.5.1.3 Methodological quality ...... 40 2.5.1.4 Quality indicators ...... 41 2.5.2 Systematic review of quality indicators for care for older adults with hypertension, and chronic ischemic heart disease ...... 45 2.5.2.1 Search results ...... 45 2.5.2.2 Study characteristics ...... 47 2.5.2.3 Methodological quality ...... 49 2.5.2.4 Quality indicators ...... 50 2.5.2.4.1 Quality indicators for hypertension ...... 50 2.5.2.4.2 Quality indicators for chronic ischemic heart disease ...... 53 2.5.3 Systematic review of quality indicators for care for older adults with major depression ...... 55 2.5.3.1 Search results ...... 55 2.5.3.2 Study characteristics ...... 57 2.5.3.3 Methodological quality ...... 58 2.5.3.4 Quality indicators ...... 60 2.5.4 Systematic review of quality indicators for care for older adults with osteoarthritis ...... 62 2.5.4.1 Search results ...... 62 2.5.4.2 Study characteristics ...... 64 2.5.4.3 Methodological quality of indicators for osteoarthritis care among older adults . 65 2.5.4.4 Quality indicators for osteoarthritis care among older adults ...... 66 2.6 Discussion ...... 68 2.6.1 Limitations ...... 71 2.7 Conclusions ...... 72

CHAPTER 3: Development of quality indicators for ambulatory care for older adults with concurrent chronic conditions: a Delphi study ...... 74 Abstract ...... 74 3.1 Introduction ...... 75 3.1.1 Rationale for the study ...... 77 3.1.2 The study objectives: ...... 80 3.2 Methods...... 80 3.2.1 Assembly of Delphi panel ...... 81

3.2.2 Questionnaire preparation ...... 81 3.2.3 Delphi Round I ...... 83 3.2.4 Delphi Round II ...... 87 3.3 Results ...... 89 3.3.1 Quality indicators for primary care for older adults with controlled diabetes with comorbid hypertension ...... 90 3.3.1.1 Delphi Round I ...... 90 3.3.1.2 Delphi Round II ...... 91 3.3.2 Selection of quality indicators for primary care for older adults with controlled diabetes with comorbid hypertension and chronic ischemic heart disease ...... 95 3.3.2.1 Delphi Round I ...... 95 3.3.2.2 Delphi Round II ...... 97 3.3.3 Selection of quality indicators for primary care for older adults with controlled diabetes with comorbid osteoarthritis ...... 100 3.3.3.1 Delphi Round I ...... 101 3.3.3.2 Delphi Round II ...... 102 3.3.4 Quality indicators for primary care for older adults with controlled diabetes with comorbid osteoarthritis and moderate depression...... 106 3.3.4.1 Delphi Round I ...... 106 3.3.4.2 Delphi Round II ...... 109 3.3.5 Quality indicators for primary care for older adults with controlled diabetes with comorbid osteoarthritis and hypertension ...... 114 3.3.5.1 Delphi Round I ...... 114 3.3.5.2 Delphi Round II ...... 116 3.4 Discussion ...... 121 3.4.1 Limitations ...... 127 3.5 Conclusions ...... 130

CHAPTER 4: Evaluating quality of overall care among older adults with diabetes with comorbid chronic conditions: a retrospective cohort study ...... 133 Abstract ...... 133 4.1 Introduction ...... 135 4.1.1 Rationale for the study ...... 139 4.1.2 Research questions ...... 141 4.2 Methods ...... 142 4.2.1 Study design and study participants ...... 142

vii

4.2.2 Measures ...... 145 4.2.2.1 Dependent variables ...... 146 4.2.2.2 Independent variables/ Process measures ...... 148 4.2.2.3 Covariates ...... 152 4.2.2.4 Data linkage and data collection ...... 153 4.2.3 Statistical analysis ...... 153 4.2.3.1 Regression models ...... 155 4.3 Results ...... 156 4.3.1. Quality of overall care and hospitalizations among patients with concordant disease combinations ...... 156 4.3.1.1 Quality of overall care and hospitalizations among older adults with diabetes with comorbid hypertension ...... 157 4.3.1.2 Quality of overall care and hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease ...... 165 4.3.2 Quality of overall care and hospitalizations among older adults with diabetes with discordant comorbid conditions ...... 174 4.3.2.1 Quality of overall care and hospitalizations among older adults with diabetes with comorbid osteoarthritis and without major depression ...... 174 4.3.2.2 Quality of overall care and hospitalizations among older adults with diabetes with comorbid osteoarthritis and major depression ...... 181 4.3.3 Sensitivity analysis results ...... 190 4.4 Discussion ...... 190 4.4.1 Strengths and limitations...... 202 4.5 Conclusions ...... 204

CHAPTER 5. Synthesis ...... 207 5.1 Study overview ...... 207 5.2 Main findings ...... 208 5.3 Implications for practice ...... 212 5.4 Implications for policy ...... 215 5.5 Implications for research...... 217 References ...... 221

viii

List of Tables

Table 1.1 Concurrent disease combinations, by comorbidity……………………………….... 21 Table 2.1 Characteristics of included studies for diabetes care indicators …………………… 39 Table 2.2 Methodological quality of studies for diabetes care indicators……………...... 41 Table 2.3 Quality indicators for diabetes care among older adults…………………………..... 43 Table 2.4 Characteristics of included studies for hypertension and chronic ischemic heart disease care………………………………………………………………………………...... 48 Table 2.5 Methodological quality of included studies for indicators for care for hypertension and chronic ischemic heart disease……………………………………………………………. 50 Table 2.6 Quality indicators for hypertension care among older adults…………………...... 51 Table 2.7 Quality indicators for care for chronic ischemic heart disease among older adults…………………………………………………………………………………………… 54 Table 2.8 Characteristics of studies included for quality indicators for depression care……… 57 Table 2.9 Methodological quality of studies for depression care indicators…………………... 59 Table 2.10 Quality indicators for care for older adults with major depression…………...... 61 Table 2.11 Characteristics of included studies for quality indicators for osteoarthritis care…………………………………………………………………………………………….. 64 Table 2.12 Methodological quality of studies for osteoarthritis care indicators………………. 66 Table 2.13 Quality indicators for osteoarthritis care among older adults……………………... 67 Table 3.1 Main characteristics of the Expert panel………………………………………...... 89 Table 3.2 Round I – Indicators for care for older adults with diabetes with comorbid hypertension ………………………………………………………………………...... 90 Table 3.3 Round II – Indicators for care for older adults with diabetes with comorbid hypertension………………………………………………………………………...... 92 Table 3.4 High and moderate consensus indicators for care for older adults with diabetes with comorbid hypertension………………………………………………………...... 94 Table 3.5 Round I – Quality indicators for care for older adults with diabetes with comorbid hypertension and chronic ischemic heart disease……………………………………………… 96 Table 3.6 Round II – Quality indicators for care for older adults with diabetes with comorbid 98

ix hypertension and chronic ischemic heart disease……………………………………………… Table 3.7 High and moderate consensus indicators for care for older adults with diabetes with comorbid hypertension and chronic ischemic heart disease……………………………...... 99 Table 3.8 Round I – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis…………………………………………………………………………………….. 101 Table 3.9 Round II – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis…………………………………………………………………………………….. 103 Table 3.10 High and moderate consensus indicators for care for older adults with diabetes with comorbid osteoarthritis…………………………………………...... 105 Table 3.11 Round I – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and depression………………………………………………...... 107 Table 3.12 Round II – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and depression…………………………………………………………………… 110 Table 3.13 High and moderate consensus indicators for care for older adults with diabetes with comorbid osteoarthritis and depression…………………………………………………… 113 Table 3.14 Round I – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and hypertension………………………………………………...... 114 Table 3.15 Round II – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and hypertension……………………………………………...... 117 Table 3.16 High and moderate consensus indicators for care for older adults with diabetes with comorbid osteoarthritis and hypertension……………………………………...... 120 Table 4.1 Process and outcome measures for ambulatory care for older adults with four selected disease combinations ……………………………...... 145 Table 4.2 Baseline characteristics of older diabetes patients with comorbid hypertension……. 157 Table 4.3 Distribution of process measures and hospitalizations among adults with diabetes with comorbid hypertension………………………….………………………………………… 160 Table 4.4 Bivariate associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension……………..... 161 Table 4.5 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension, adjusted for other covariates………………………………………………………………………………………. 163

Table 4.6 Baseline characteristics of older adults with diabetes with comorbid hypertension and chronic ischemic heart disease…………………………………………………………..... 166 Table 4.7 Distribution of process of care and outcome measures among older adults with diabetes comorbid with hypertension and chronic ischemic heart disease……………………. 168 Table 4.8 Bivariate associations between process of care measures and hospitalizations, among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease………………………………………………………………………………………..... 169 Table 4.9 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease, adjusted for other covariates………………………………………………………...... 172 Table 4.10 Baseline characteristics of older adults with diabetes with comorbid osteoarthritis………………………………………………………………………………….... 175 Table 4.11 Distribution of process of care and outcome measures among older adults with diabetes with comorbid osteoarthritis………………………………………………………...... 174 Table 4.12 Bivariate associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid osteoarthritis…………….... 178 Table 4.13 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid osteoarthritis, adjusted for other covariates………………………………………………………………………………………. 180 Table 4.14 Baseline characteristics of older diabetes patients with comorbid osteoarthritis and major depression…………………………………………………………………………...... 182 Table 4.15 Distribution of process of care and outcome measures among older adults with diabetes with comorbid osteoarthritis and major depression………………………………...... 185 Table 4.16 Bivariate associations between process measures and the likelihood of hospitalizations among older adults with diabetes with osteoarthritis and major depression...... 186 Table 4.17 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with osteoarthritis and major depression, adjusted for other covariates……………………………………………………………………………………….. 188

List of Figures

Figure 1.1 Performance Measurement for People with Multiple Chronic Conditions Conceptual model……………………………………………………………………………. 15 Figure 1.2 Thesis project framework………………………………………………………… 22 Figure 2.1 Flow diagram for selection of studies for identifying indicators for diabetes care among older adults…………………………………………………………………………… 37 Figure 2.2 Flow diagram for selection of studies for identifying indicators for hypertension/chronic ischemic heart disease care among older adults………………………. 46 Figure 2.3 Flow diagram for selection of studies for identifying indicators for major depression care among older adults………………………………………………………….. 56 Figure 2.4 Flow diagram for selection of studies for identifying indicators for osteoarthritis care among older adults……………………………………………………………………… 63

xii

List of Appendices

Appendix 1. Search strategies……………………………………………………………… 240 Appendix 2. The AIRE instrument: domains and items…………………………………… 244 Appendix 3. Full list of extracted indicators………………………………………………. 246 Appendix 4. Email text – Delphi Round I………………………………………………….. 256 Appendix 5. Participant information sheet – Delphi Round I……………………………… 258 Appendix 6. Survey questionnaire, Delphi Round I………………………………………... 264 Appendix 7. Criteria for rating quality indicators………………………………………….. 282 Appendix 8. Detailed information related to candidate indicators…………………………. 284 Appendix 9. Respondent comments………………………………………………………... 286 Appendix 10. Email text – Delphi Round II………………………………………………... 307 Appendix 11. Feedback material for Round II……………………………………………... 309 Appendix 12. Quality indicators that did not reach consensus, or were not included in this study………………………………………………………………………………………… 310 Appendix 13. Study Time Frame Definitions……………………………………………… 314 Appendix 14. Comorbid chronic conditions……………………………………………….. 315 Appendix 15. Sensitivity analyses………………………………………………………….. 317 Appendix 16. DIN/PIN codes of drugs…………………………………………………….. 320

xiii

CHAPTER 1

Introduction

1.1 Chronic conditions, aging and multimorbidity

Chronic conditions are the leading cause of death, and disability and are estimated to account for two thirds of the overall burden of disease (1). In 2011, approximately 5 million

Canadians were 65 years of age or older accounting for 15% of the total population (2, 3).

For many chronic conditions, prevalence increases with age, causing a disproportionate health burden for people aged 65 and over (4, 5). Improvements in survival and an aging population contribute to the increasing prevalence of multiple chronic conditions or multimorbidity (6). Multimorbidity is defined as simultaneous co-existence of two or more chronic conditions in a single individual (6-8).

According to the Canadian Institutes of Health Information (CIHI) data, adults aged 65 and over are almost four times more likely to report having a chronic condition compared to those aged 18 to 24 (4). According to the Ontario Medical Association (OMA), chronic conditions affect about 81% of Ontario adults aged 65 or over (9). Data from the Canadian

Community Health Survey showed that 71% of Canadians of 65 to 79 years old and 82% of adults aged 80 and older have two or more chronic conditions (4, 10).

A population-based study conducted in Ontario demonstrates that 81% of adults aged 75 years or more had at least two chronic conditions (5). Moreover, data from several community surveys across Canada indicate that multimorbidity affects the most vulnerable

1 groups in society, including those who are less educated, have low incomes, and are living in rural communities (10-12).

Diabetes mellitus is one of the most common chronic diseases, being one of the leading causes of death and disability in Canada (13). In 2014, the number of prevalent cases of diabetes in Canada was 3.3 million (14, 15). Results from a recent study show that individuals with diabetes aged 65 and over have the highest rates of hyper- or hypoglycemic crises, cardiovascular complications, major lower-extremity amputation, end-stage renal disease, and visual impairment (16).

Diabetes was identified as one of the most costly health conditions, mostly because of the severity of its complications and the treatments required for controlling them (17-19).

Evidence suggests that diabetes healthcare costs in Canada exceeded $4.6 billion in 2000 and that costs were predicted to rise to $8.1 billion by 2016 (20). Between 2004 and 2012, the overall per-person attributable costs associated with diabetes in Ontario were $10,315 for males and C$9,731 for women (21).

1.2 Multimorbidity among patients with diabetes

Several studies show that diabetes occurs mostly in conjunction with other chronic conditions, in particular 85% of adults with diabetes have at least one comorbid chronic condition, and as many as 40% have three or more (22, 23). A recent population-based cohort study conducted in Ontario demonstrates that over 90% of community-living older

2 adults with diabetes had at least one comorbid condition; and more than 40% of the cohort had 5 or more comorbidities (24).

The literature suggests that the most common conditions among older adults with diabetes are hypertension, arthritis, ischemic heart disease and anxiety (23, 24). A recent population- based study demonstrated that hypertension was the most common condition in older adults with diabetes, affecting 79.1% of the entire cohort, arthritis affected 59.6% of the cohort, while ischemic heart disease and anxiety affected 59.3% and 37.6% of the cohort, respectively (24).

Evidence shows that comorbid chronic conditions in patients with diabetes significantly impact their self-management ability, quality of life and healthcare utilization (22, 25). Such chronic conditions as depression and arthritis impair patients’ functioning and pose significant barriers to lifestyle changes and regimen adherence among diabetes patients (26).

Moreover, such disabling conditions as advanced heart failure and dementia may make standard diabetes self-care goals impossible to reach (22).

According to the theoretical framework proposed by Piette and Kerr (22), there are several typologies of comorbid chronic conditions in diabetes patients, including diabetes- concordant, such as cardiovascular conditions and stroke, and diabetes-discordant, such as musculoskeletal conditions and depression. Diabetes-concordant conditions have similarities with diabetes with respect to their pathogenesis and the disease management plan, while diabetes-discordant conditions are not directly related to diabetes in their pathogenesis and

3 are more likely to add to the complexity of clinical decision-making in management of people with diabetes (22, 25, 27).

There is conflicting evidence that the presence of different types of comorbid conditions among people with diabetes leads to differences in the quality of diabetes care (27-31).

Quality of diabetes care is commonly assessed by receipt of evidence-based monitoring tests for diabetes care, including blood pressure measurement, glycated hemoglobin (HbA1c) and low-density lipoprotein cholesterol (LDL-C) tests and eye examination (27-31).

Using these quality measures, Pentakota and colleagues (28) found that the quality of diabetes care varies by comorbidity type, in particular, the presence of diabetes-concordant comorbid conditions was associated with better diabetes care, while having diabetes- discordant conditions was associated with diminished diabetes care. Another study found that having more concordant conditions makes diabetes care goal achievement more likely, while number of discordant conditions had no significant impact on diabetes goal achievement

(30).

Woodard and colleagues (29) found that patients with diabetes with greater complexity received higher quality diabetes care compared to less complex patients, regardless of comorbidity type, while in another study the same research team suggested that co-existing discordant comorbid conditions among patients with diabetes may increase the overall number of primary care visits and, as a result, opportunities for monitoring of diabetes may be greater (27).

A focus of both clinical care and research still remains on the quality of a single condition - index disease, such as diabetes care quality, and limits the assessment of care for the whole person with multiple chronic conditions (27-30).

1.3 Multimorbiditiy’s issues/challenges

Having multiple chronic conditions is associated with a number of issues, including poor quality of life, psychological distress, increased rate of hospital admissions and emergency department visits, high use of ambulatory services, polypharmacy, difficulty in applying guidelines, and fragmented and ineffective care, as well as challenging organizational problems such as coordination of care and accessibility (11, 32-38).

1.3.1 Multimorbidity: high health care utilization

Co-existence of multiple chronic conditions places a heavy burden on the health system in terms of healthcare utilization and costs. Adults with multiple chronic conditions, in particular those aged 65 and over, are significant users of healthcare services, and account for more than two-thirds of healthcare spending (37, 38).

Recent research results demonstrate that the addition of each chronic condition in seniors led to an increase in primary care visits, avoidable emergency department (ED) and hospital admissions, and total health care costs (35). Iron and colleagues (36) examined the prevalence of six common chronic conditions in Ontario: hypertension, asthma, diabetes,

5 chronic obstructive pulmonary disease, congestive heart failure, and acute myocardial infarction. They found that Ontarians with three or more diagnoses had 57% more primary care visits, 76% more specialist visits, 11% more emergency department visits, 256% more inpatient hospital stays, and 68% more prescriptions, as compared to those with a single chronic condition (36).

The literature suggests that presence of concurrent chronic conditions in patients with diabetes is associated with considerable consequences for quality of care, higher healthcare utilizations and related costs (39-42). A recent population-based study conducted in Ontario found that a proportion of community-living older adults with diabetes with at least one hospital admission or emergency department visit was 3-4 times higher among those with five or more comorbid conditions compared to those with no comorbidity (24).

Older adults with diabetes are at substantial risk for both diabetes-related long-term complications, including micro- and macrovascular complications, and diabetes-related short-term complications, including acute metabolic decompensation with subsequent hospitalizations (43). The literature suggests that patients with diabetes aged 75 years and over have double the rate of emergency department visits for metabolic decompensations, including severe hypo- and hyperglycemia, than the general population with diabetes (16,

44).

1.3.2 Management of older adults with multiple chronic conditions

Clinical management of people with multiple chronic conditions is defined as prevention, on- going management and treatment, and rehabilitation (45). Providing optimal care for the older adults with multiple chronic conditions becomes difficult and complex in terms of the magnitude of the tasks, administration of multiple drugs, and cost (46). Clinical practice guidelines aim to assist physicians in providing the best care using the most recent evidence

(34, 47). However, current single disease management approaches are not suitable for people with multiple chronic conditions (8, 46, 48). Fortin et al. found that only 18.8% of the

Canadian clinical guidelines provide specific treatment recommendations for specific pairs of chronic conditions, and even less for patients with more than two co-existing chronic conditions (34).

Clinical practice guidelines are usually based on randomized clinical trials that have excluded people with multiple chronic conditions as well as older adults, thereby limiting their applicability to those populations (34, 46). There is evidence that older adults are heterogeneous in terms of functional and cognitive status, severity of different concurrent conditions, as well as risk of adverse events. The dosages of medications in younger patients may result in adverse effects in older adults due to age-related changes in pharmacokinetics and pharmacodynamics as well as coexistence of multiple conditions (49).

Moreover, disease-specific guidelines may lead to complex and sometimes contradictory treatment for people with multiple chronic conditions (34, 47, 50). Thus, the management of people with multiple chronic conditions requires an understanding of the extent to which the

7 concurrent chronic conditions contribute and interact to affect the patient’s health status, as well as assessing the risks and benefits of various treatment options in this complex population (50).

Having multiple chronic conditions is commonly associated with the use of multiple drugs

(51). This condition is known as polypharmacy which has been defined as the administration of more medications than are clinically indicated, representing unnecessary drug use, and affects many people aged 65 and over (51). Other authors have defined polypharmacy as a concurrent use of six or more medications with different active compounds (49, 51, 52). The main risk factors for polypharmacy that have been identified in the literature include age, education, multiple chronic conditions, number of health care visits and multiple providers

(51, 53, 54).

Polypharmacy increases the risk of inappropriate prescribing. Among seniors with multiple chronic conditions about 40% reported a medication error, including wrong dosage or wrong medication (33, 51, 55). Polypharmacy is particularly relevant for people with diabetes and possibly would vary depending on the concordance or discordance of other comorbid conditions.

People with diabetes may be prescribed more than a dozen of different classes of drugs to treat diabetes, its complications and other comorbid conditions, and being asked to swallow over 20 tablets every day (56). Recent research found that poor adherence to drug therapy in people with diabetes is mostly related to the prescribed combination of oral antidiabetic drug

8 therapy and other co-medications (56). Moreover, disease-specific clinical guidelines may not be appropriate for treating diabetes patients with comorbidities, especially when comorbid conditions are discordant (22, 48). For instance, administration of steroids for treating inflammatory and autoimmune conditions leads to an increase in glucose level among diabetes patients (57).

1.3.3 Multimorbidity: implications for primary care

Primary care providers play a central role in preventing and managing chronic conditions, often in collaboration with specialised services. Primary care is well placed to have an important impact on outcomes of care for people with multiple chronic conditions, including patients with diabetes comorbid with other chronic conditions (58). The Institute of Medicine defines primary care as “the provision of integrated, accessible health care services by clinicians that are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context of family and community”(59).

In Ontario, primary care includes many different practice models, ranging from small one- physician practices to large, regionally organized interdisciplinary teams (60). These models differ mainly by physician compensation methods (e.g. fee-for-service, capitation) and several organizational elements, including patient enrollment, quality incentives, provision of after-hour services, and inter-disciplinary teams (61, 62).

Previous research revealed that as compared with practices in the fee-for-service model, those in the capitation model had patients who had lower morbidity and comorbidity indices, and had more emergency department visits (63). Several studies found that salaried models had the highest overall performance of chronic disease management (60, 62). Recent research revealed that people with diabetes who were not enrolled in one of these models were least likely to receive recommended diabetes testing, while the highest likelihood of receiving diabetes recommended tests was observed in patients enrolled in a blended capitation model

(64).

Primary care physicians face difficulties in addressing the complex multifarious needs of older adults with multiple chronic conditions. Actually, treatment of people with multiple chronic conditions often requires “trade-off” decisions, because current clinical guidelines are often impractical in the presence of multiple chronic conditions (65).

Given that not all concurrent chronic conditions in diabetes patients have the same impact, they may not require the same degree of management (22, 66). In particular, diabetes or some co-existing conditions may be ignored or under-treated given the competing treatment priorities because of severity and complexity of other co-existing conditions (22, 66).

Moreover, concordant co-existing conditions may be managed synergistically, while discordant co-existing conditions are not directly related in the disease management plan

(22).

1.3.4 Continuity of care

Healthcare providers, policy-makers and patients are increasingly expressing concern about fragmentation of care. People with multiple chronic conditions are likely to have frequent consultations with various health providers with cross-referral related to the co-existing conditions (67). Therefore, continuity and coordination of care are very important.

Continuity of care is defined by two core elements: 1) the experience of care by a single patient with his or her provider care and 2) the care continues over time (68).

There are three main types of continuity of care: relational, management and informational

(68). Relational continuity refers to an ongoing relationship between provider and client and consistency of personnel (68). Consistency seeing the same care provider is crucial even in a setting where there is little expectation of establishing long-term relationships between clients and providers, such as acute care (69).

Management continuity refers to consistency of care or creating management plans to insure consistency during the treatment (68). People with multiple chronic conditions require flexible care plans to allow for changes in their needs and circumstances (70). Informational continuity refers to the transfer of documented information from one provider to another and accumulated knowledge of the patient to bridge separate care events (68, 69).

In the last two decades, over a dozen indices have been developed to assess how care is concentrated among the different providers that a patient sees. The most widely used measures of continuity of care include team-based continuity of care, Usual Provider of Care

(UPC) and the Continuity of Care (COC) indices (68). Saultz at. al. (71) defined team-based continuity as “care that allows previous knowledge of the patient to be available when the patient requires a range of services spanning the medical specialties”. Calculation of team- based continuity includes number of patient’s visits that occurred with any member as the numerator, and empaneled patient visits as the denominator (71). The UPC index measures the proportion of visits with a usual provider over a given period of time.

COC index include: 1) widely used measure, permitting comparison between studies, 2) may be adapted to measure concentration of care, 3) accounts for number of different providers seen, and 4) sensitive to shifts in the distribution of visits among providers (68, 72). At the same time, the COC index does not account for other aspects of continuity of care, including informational or management continuity (73).

The literature provides mixed results on the level of continuity of care among people with multiple chronic conditions. Previous research found that people with multiple chronic conditions had higher consultation rates and less continuity of care, compared to those with single conditions (67, 72). The results of another population-based study conducted in

Ontario demonstrate that continuity of care was relatively stable among people with one or more concomitant conditions; and the risk of hospitalizations was reduced in people with greater continuity of care (74).

1.3.4 Quality indicators in health care

Quality indicators are extensively used for measuring and monitoring the quality of care and provided services (75). The Institute of Medicine (IOM) defined quality of care as the “the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge” (76). The

Canadian Council on Health Services Accreditation (CCHSA) defines an indicator as “a measurement tool, screen or flag that is used as a guide to monitor, evaluate and improve the quality of client care, clinical services, support services and organizational functions that affect patient outcomes” (77).

Health care quality indicators are powerful tools that can guide efforts to improve patient care (78). However, strict adherence to disease-specific measures for older adults with multiple chronic conditions may lead to the unintended consequence of delivering inappropriate care that is not aligned with the patient’s goals and preferences. Moreover, applying numerous measures targeting a variety of diseases among older adults could lead to the high measurement burden, often without attaining better outcomes. Therefore, it is crucial to identify measures of the quality of care among older adults with multiple chronic conditions to improve their care.

The key characteristics of a good quality indicator include: 1) indicator is defined in detail, with explicit data specifications in order to be specific and sensitive, 2) indicator

13 discriminates well, 3) indicator is valid and reliable; 3) indicator intended to capture a specific aspect of quality; 4) indicator permits useful comparisons; 5) indicator is evidence- based, and 6) it is feasible to collect and analyse data for this indicator (75, 79).

Evidence-based clinical indicators are true measures of quality and predict patient outcomes, although indicators based on professional consensus may be all that feasible for certain conditions or patient populations (75). Research methods for development of quality indicators include: 1) non-systematic approaches do not tap in to the evidence base of an aspect of health care; they are based on the availability of data (e.g. quality improvement projects based on one case study), 2) systematic: evidence based, where indicators are based directly upon scientific evidence such as rigorously conducted clinical trials or empirical studies, 3) systematic: evidence combined with consensus, when scientific evidence is lacking, quality indicators can be defined by an expert panel of professionals by means of consensus technique based on their experience, and 4) systematic: guideline driven indicators

(75, 78, 79).

The literature suggests the most rigorous way of developing quality indicators as combined systematic literature search with consensus techniques (78, 80, 81). Campbell and colleagues

(78) defined consensus techniques as “group facilitation techniques, which explore the level of consensus among a group of experts by synthesising and clarifying expert opinion in order to derive a consensus opinion from a group with individual opinions combined into a refined aggregated opinion” (p.360).

1.4 Performance Measurement for People with Multimorbidity: Conceptual model

A Performance Measurement for People with Multiple Chronic Conditions (PM-MCC)

Conceptual model (70) was utilized for the purpose of this project. It has been developed recently by Giovannetti and colleagues (70) based on existing conceptual models and reviews of the literature.

Figure 1.1 Performance Measurements for People with Multiple Chronic Conditions Conceptual Model

*Adapted from “Performance Measurement for People with Multiple Chronic Conditions: Conceptual Model”, Giovannetti, E.R. and colleagues, 2013, the American Journal of Managed Care.

This conceptual model aims to address the scope and heterogeneity of care for people with multiple chronic conditions, in particular, issues related to the risk of fragmented and uncoordinated care across multiple settings, as well as the risk of adverse health outcomes.

This conceptual model is coherent with a consensus both within Canada and internationally that primary care and the entire healthcare system should aim to be: accessible, effective, efficient, safe, coordinated, and patient and family centred (82).

At the centre of this framework is a patient with multiple conditions with a family and preferences for care. Any given comorbid condition may affect the patient with diabetes at any one time, and these conditions can be concordant or discordant with diabetes care, and may be a dominant condition, for example, class IV heart failure, so complex and serious that it eclipses the management of other co-existing conditions (22). Depending on their co- existing conditions, including type and seriousness, and patient’s/family goals and care preferences, people with diabetes can receive care in various sites from different healthcare providers.

The circles surrounding the patient represent shifting sites and providers that support and care for people with multiple chronic conditions. The most relevant sites for people with multiple conditions included in the conceptual model are: primary care, specialty care, home- cased care, community care, hospital and post-acute care, nursing home and pharmacy. At any given time, multiple types of providers may provide care to the patient with multiple chronic conditions.

The types of care that patients with multiple chronic conditions receive are not necessarily linear or mutually exclusive, and they include: screening, prevention, diagnosis, treatment and management, community services, acute exacerbation, rehabilitation, end-of-life care and palliation. For example, an individual may be seen in the hospital for an acute exacerbation of heart failure, but this patient may also need continuing treatment and management of co-existing diabetes and asthma.

The outer circle highlights the priority domains of measurement that apply across sites and types of care and appropriate for use with patients with multiple chronic conditions. These domains are not mutually exclusive, and a given measure could fall into multiple domains.

The measures in the framework are categorized under six priority areas, including 1) health and well-being, such as quality of life, screening , 2) effective prevention and treatment, such as diabetes on-going monitoring, pharmacological treatment, 3) effective communication and coordination of care, such as continuity of care, shared decision making, 4) patient safety, such as inappropriate treatment, polypharmacy issue, healthcare utilization, and 5) affordable care, such as total healthcare costs, and 6) person-and-family-centered care.

Some measurement domains may be specific to a site or type of care, such as glycated hemoglobin testing, while others, such as effective coordination of care, apply across all types and sites of care. Moreover, in each measurement domain can be non-disease-specific measures such as pain screening, and disease-specific measures such as HbA1c tests in people with diabetes may be applicable. Each priority domain of measurement included in the model

17 addressed using different types of measures including structure, process, outcome, cost and composite measures.

The current project focused on evaluating the quality of care for older adults with multiple chronic conditions in ambulatory care settings using an administrative database. The following modifiable measurement domains of the proposed conceptual model were used for the purpose of this project: effective prevention and monitoring, such as frequency of diagnostic or monitoring tests; pharmacological treatment of co-existing conditions; patient safety, such as appropriateness of the prescribed medications, healthcare utilization; as well as clinical effectiveness, and care continuity, using the continuity of care (COC) index (72).

The quality measures of all defined domains were categorized according to Donabedian’s structure- process-outcome framework (83, 84). Structure indicators denote the attributes of the settings where care is delivered, including adequate facilities, qualification of care providers or patients, or methods of reimbursement. Process indicators examine how care has been provided in terms of appropriateness, acceptability, completeness or competency. These commonly include diagnosis, treatment, preventive care and continuity of care. Outcome indicators attempt to describe the effects of care on the health status of patients or population, such as healthcare utilization or mortality. This model proposes that each component has a direct effect on the next one, such as structure or process of care determines outcomes of care.

1.5 SUMMARY

While diabetes occurs mostly in conjunction with other comorbid conditions, most diabetes management programs and guidelines are entirely focused on a single condition and do not address the challenges that primary care/specialist physicians and patients face when managing multiple concomitant conditions (28, 64, 85). Having multiple chronic conditions is associated with a number of issues, including poor quality of life, increased healthcare utilization, polypharmacy, difficulty in applying guidelines, and fragmented and ineffective care, as well as challenging organizational problems such as coordination of care and accessibility (11, 32, 33). Issues associated with multiple chronic conditions are relevant to diabetes management since more than 90% of older adults with diabetes have at least one comorbid condition (22, 23).

The management of patients with multiple chronic conditions requires understanding of the extent to which the concurrent chronic conditions contribute and interact to affect the patient’s health status. Quality indicators can be used as a tool to assess ambulatory care, including primary and specialty care, delivered to patients with multiple chronic conditions, as well as to identify potentially modifiable gaps in care. This may inform areas for future intervention aimed at improving management for patients with multiple chronic conditions by addressing gaps in health care delivery.

Various quality measures have been developed for assessing the care for single diseases that may lead to inappropriate judgment of the care provided to older adults with diabetes and complex comorbidities (46). Therefore, it is crucial to identify measures of the quality of overall care among older diabetes patients with comorbidities to improve their care. Most existing studies examine the influence of different types of comorbid conditions on the quality of diabetes care (27, 28, 64). There remain gaps in assessing of care for whole person with diabetes with comorbidities, including both process and outcomes of care.

1.6 Thesis project outline

This project aims to evaluate the quality of overall care among older adults with diabetes comorbid with other chronic conditions. This project is focused on older adults because they are more likely than younger individuals to have comorbid chronic conditions that can be complex and difficult to manage (5, 46). Diabetes has been chosen as a condition of interest due to the high burden of co-existing chronic conditions in this group of patients (5, 24).

Four chronic conditions among diabetes patients, including hypertension, chronic ischemic heart disease, major depression and osteoarthritis, were selected for the purpose of this project, since they have been identified as the most common and costly conditions among people diagnosed with diabetes (23, 24). In particular, hypertension was the most common conditions affecting 79.1% of the cohort of older adults with diabetes, arthritis affected

59.6% of the cohort, ischemic heart disease and anxiety affected 37.6% and 36.9% of the cohort, respectively (24).

These five chronic conditions were grouped in five disease combinations (Table 1.1). Prior population-based epidemiological research in Ontario showed that these five disease combinations represent the most prevalent clusters of concurrent conditions among people with diabetes and other concurrent conditions (5, 24). The defined five disease combinations are also representative of different comorbidity types, including diabetes-concordant (have similarities with diabetes with respect to the pathogenesis and management plan), diabetes- discordant (are not directly related to diabetes in their pathogenesis) and both types of comorbid conditions (22).

Table 1.1 Concurrent disease combinations, by comorbidity

DM +hypertension Diabetes-concordant conditions DM +hypertension + chronic ischemic heart disease

DM +osteoarthritis Diabetes-discordant conditions DM +osteoarthritis +major depression

Both concordant and discordant DM +osteoarthritis +hypertension conditions

1.7 Broad project objectives

This thesis addresses the following three objectives:

1. To identify a set of evidence-based and valid quality indicators for evaluating care

for older adults with selected chronic conditions, including diabetes,

hypertension, chronic ischemic heart disease, major depression and osteoarthritis,

as well as quality indicators for five selected disease combinations in ambulatory

care settings;

2. To critically appraise and select a set of appropriate quality indicators by means

of a Delphi panel for evaluating care for older adults with selected disease

combinations in ambulatory care settings;

3. To apply newly developed quality indicators to evaluate the quality of overall

care for older adults with five selected disease combinations in ambulatory care

settings.

Figure 1.2 Thesis project framework

Project I Project II Project III Systematic review Methodological assessment of Population-based quality indicators using a retrospective cohort study: assessing quality of care for A systematic review of Delphi technique existing quality indicators, older adults with selected disease combinations by selected disease Critical appraisal of identified categories: diabetes, quality indicators in the context  Screening, on-going hypertension, chronic of five selected disease monitoring ischemic heart disease, combinations using a Delphi  Treatment osteoarthritis, and major technique  Patient safety depression  Clinical effectiveness

To accomplish the overall study objectives several steps were required. First, a systematic review was conducted to identify a set of published quality indicators for assessing care for older adults in ambulatory care settings for diabetes and each of the four selected comorbid diseases. From these reviews, indicators that are measureable with existing health administrative databases were extracted as candidate quality indicators. Then, the identified candidate indicators were critically appraised and a set of quality indicators was developed for each particular disease combination by means of the Delphi technique. The final set of appropriate indicators, as judged by the Delphi Panel, was used in empirical analyses to assess quality of care for older adults with selected combinations of the concurrent chronic conditions using health administrative data from 2010 through 2014 in Ontario.

CHAPTER 2

Systematic review of quality indicators for care for older adults with diabetes, hypertension, chronic ischemic heart disease, major depression and osteoarthritis

Abstract

Background: Despite the growing interest in assessing the quality of care for chronic conditions, there has been little evaluation of the quality of care provided for older adults with chronic conditions in ambulatory care settings.

Purpose: This study aimed to identify evidence-based and valid quality indicators feasible for monitoring, evaluating and improving the quality of care for chronic conditions, including diabetes, hypertension, chronic ischemic heart disease, major depression and osteoarthritis, as well as indicators addressing comorbidity among older adults in ambulatory care settings.

Methods: Ovid MEDLINE, Ovid EMBASE and Ovid PsycINFO databases, and gray literature, including relevant organizational websites, were searched from 2000 to 2014. Two reviewers independently selected studies if 1) the study methodology combined a systematic literature search with assessment of quality indicators by an expert panel and 2) quality indicators were applicable to assessment of care for the five selected chronic conditions in ambulatory care settings for older adults. Included studies were appraised using the Appraisal of Indicators through Research and Evaluation (AIRE) instrument.

Results: We identified quality indicators for care for older adults with selected chronic conditions, including 6 indicators for major depression care, 3 indicators for osteoarthritis care, 12 indicators for diabetes care, 7 indicators for hypertension care, and 6 indicators for chronic ischemic heart disease care. All identified quality indicators assess clinical aspects of ambulatory care for people with the selected chronic conditions. Most quality indicators referred to the processes of care. The methodological characteristics of the quality indicators varied considerably.

Conclusions: Relatively little research has been done to develop indicators that assess the quality of care of chronic conditions for older adults in ambulatory care settings. The identified sets of quality indicators with high AIRE scores might well be suitable for use in daily practice for monitoring and assessing care for older adults with selected chronic

24 conditions. Despite the growing prevalence of multimorbidiy, we found only a few published indicators for care for older adults with comorbidities. Therefore, there is a need to develop quality indicators for care for older adults with various disease combinations with an awareness of how the treatment of one disease can affect the treatment of any or all other conditions.

2.1 Introduction

For many chronic conditions, prevalence increases with age, causing a disproportionate health burden on people aged 65 and over (5). According to the Global Burden of Disease estimated for 2010, about 25% of total disease burden is attributable to diseases in people aged 65 and over, including cardiovascular conditions, musculoskeletal diseases, diabetes and mental diseases (86). Recent population-based study conducted in Ontario identified that the most prevalent chronic conditions include diabetes, osteoarthritis, mood disorders, coronary artery syndrome, asthma and cancer (5).

Diabetes is a global health concern, with worldwide prevalence estimated to increase from

2.8% in the year 2,000 to 4.4% by 2030 (87). For many chronic conditions, such as diabetes, prevalence increases with age, causing a disproportionate health burden on people aged 65 and over (4, 5). In Canada, the prevalence of diabetes among people in the 75- to 79-year-old age group ranges from 20 to 23% (88). Recent research has found that more than 90% of older adults with diabetes had at least one comorbid condition (24).

Depression is a common mental health problem for older adults associated with significant functional decline, family stress, increased utilization and cost of healthcare, and increased suicide and all-cause mortality (89-91). Depressive disorders are most often managed by a primary care physician, unless the severity of depression is such that care is required from a specialist (78, 92). However, poor levels of detection, treatment and monitoring of depression have been highlighted in ambulatory care settings (93). Osteoarthritis is a common chronic condition and one of the leading causes of disability and poor quality of life

26 in older adults (5, 94, 95). Osteoarthritis is frequently diagnosed and treated in primary care settings; it is the second most common diagnosis among older adults leading to consultations with their general practitioners (94, 95).

The literature suggests that hypertension is the most important preventable cause of morbidity and mortality globally (96). Hypertension is a well-established risk factor for cardiovascular disease (96). The prevalence of cardiovascular diseases, including hypertension and ischemic heart disease, increases from about 40% in individuals 40-59 years of age, to 70-75% in people 60-79 years of age, and to 79-86% among those aged 80 years or older (97). Primary care providers play a central role in preventing and managing cardiovascular disease, including early detection, follow-up and monitoring of conditions

(98).

The World Health Organization (WHO) defined chronic conditions as requiring “ongoing management over a period of years or decades” (p.24) (99). The literature suggests that chronic conditions frequently are poorly controlled until acute complications arise (100). The

Institute of Medicine (IOM) defined quality of health care as “the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge” (101). The assessment and monitoring of care quality can be achieved by using quality indicators which are based on standards of care and the best available evidence (75). Data generated from these measures can be used to assess past performance, identify suboptimal practices, and plan improvements.

Over the last few decades, much effort has gone into developing and applying measures of quality of care. Most indicators have been developed to assess or improve care in hospitals but, increasingly, indicators are being developed for ambulatory care in most countries, including Europe and the United States (78, 102). Quality indicators indicate potential problems that might need addressing, usually manifested by perceived unacceptable variation in care or statistical outliers (78).

Assessment and monitoring of the quality of care by quality indicators has become crucial for the health care system for numerous reasons: enhancing the accountability of healthcare providers and making comparisons, resource allocation efficiency, identifying and minimizing medical errors, and improving health outcomes, as well as supporting patient choice of provider (75, 78).

2.2 Rationale for the study

Despite the growing interest in assessing the quality of care for chronic conditions, there has been little evaluation of the quality of care provided for older adults with chronic conditions, including diabetes, hypertension, chronic ischemic heart disease, major depression and osteoarthritis in ambulatory care settings. The current systematic review aimed to identify evidence-based and valid quality indicators feasible for monitoring, evaluating and improving the quality of care for these five chronic conditions among older adults in

28 ambulatory care settings, as well as for care for older diabetes patients with comorbid hypertension, chronic ischemic heart disease, osteoarthritis and major depression (Table 1.1).

This study is focused on clinical indicators that describe the performance of care and related outcomes (75). The key characteristics of a good quality indicator include: 1) indicator is defined in detail, with explicit data specifications in order to be specific and sensitive, 2) indicator has discriminative power, 3) indicator is valid and reliable; 3) indicator intended to capture a specific aspect of quality; 4) indicator permits useful comparisons; 5) indicator is evidence-based, and 6) it is feasible to collect and analyse data for this indicator (75, 79).

This study is focused on older adults because they are more likely than younger individuals to have comorbid chronic conditions that can be complex and difficult to manage (5, 46).

Diabetes has been chosen as a condition of interest due to the high burden of co-existing chronic conditions in this group of patients (5, 24). Almost 80% of the care of people with diabetes takes place in ambulatory care settings (86). With increasing prevalence of diabetes among older adults and high rate of comorbidities, family physicians and primary care providers face challenges in terms of providing optimal care for people with diabetes (22).

Previous research showed that there is a care gap between the recommendations outlined in clinical practice guidelines for diabetes management and real life clinical practice (87, 88).

2.3 Study objectives

1. To conduct a systematic review of published quality indicators that can be measured

using Ontario administrative data by each disease category, including diabetes,

hypertension, chronic ischemic heart disease, major depression, and osteoarthritis

among older adults, as well as quality indicators for care for five selected disease

combinations, including:

 Diabetes with comorbid hypertension;

 Diabetes with comorbid hypertension and chronic ischemic heart disease;

 Diabetes with comorbid osteoarthritis;

 Diabetes with comorbid osteoarthritis and major depression;

 Diabetes with comorbid osteoarthritis and hypertension.

2. To critically appraise a set of identified quality indicators, using AIRE (Appraisal of

Indicators through Research and Evaluation) instrument;

3. To prepare a summary of the published quality indicators by each disease category in

ambulatory care settings that can be measured using Ontario administrative data.

2.4 METHODS

2.4.1 Search strategy

A search strategy was developed to identify articles concerning the development, testing or implementation of indicators of the quality of care for diabetes, hypertension, chronic

30 ischemic heart disease, osteoarthritis and major depression applicable to older adults in ambulatory care settings. A literature search was conducted using Ovid MEDLINE and Ovid

EMBASE for identifying indicators for diabetes, hypertension, chronic ischemic heart disease and osteoarthritis, and Ovid MEDLINE and PsycINFO for identifying indicators for major depression, from 2000 to 2014. The search was restricted to English articles of human studies, and when the participants consisted of adults, including people aged 65 years or older.

To identify studies related to quality indicator development the following search terms were used: “performance indicator(s)”, “quality indicator(s)”, “performance measure(s)”, “quality measure(s)”, “benchmark”, “report card”, “quality of health care”, “clinical guideline”, or

“quality assurance”. To identify studies related to diabetes the following search term were used: “diabetes mellitus”, “type 1 diabetes mellitus”, “type 2 diabetes mellitus”, “non-insulin dependent diabetes”, or “diabetes complications”. To identify studies related to hypertension and chronic ischemic heart disease the following search terms were used: “hypertension”,

“essential hypertension”, “high blood pressure”, “ischemic heart disease”, “coronary artery disease”, or “coronary heart disease”.

To identify studies related to depression the following search term were used: “depression or

“depressive disorder”. To identify studies related to osteoarthritis the following search term were used: “osteoarthritis”, “arthritis, degenerative”, “osteoarthritis, hip”, “osteoarthritis, knee”. Then, we combined the search terms results for quality indicators with search terms results for each disease category, respectively. (Appendix 1)

In addition, we conducted a gray literature search to find information about quality indicators development initiatives, by each disease category that were not published in peer-reviewed journals. For that purpose, we searched available public repositories such as the National

Quality Measures Clearinghouse (NQMC, http://www.qualitymeasures.ahrq.gov) and the

National Quality Forum (NQF, http://www.qualityforum.org).

Additionally, we looked for existing indicators on web sites of major organizations involved in quality measurement and reporting of indicators for assessing the quality of care among older adults by each disease category, including the RAND Health Corporation/Assessing

Care of Vulnerable Elders (ACOVE), Health Quality Ontario (HQO), National Diabetes

Quality Improvement Alliance (NDQIA), European Practice Assessment (EPA) Cardio project, Canadian Cardiovascular Outcomes Research Team (CCORT), Performance

Measurement Canadian Mental Health Association (www.cmha.ca), the online inventory maintained by the Center for Quality Assessment in Mental Health (www.cqaimh.org),

American College of Rheumatology (ACR), and the Canadian Rheumatology Association

(CRA).

2.4.2 Study selection

The literature suggests the most rigorous way of developing quality indicators is a combined systematic literature search with consensus techniques (78, 80, 81). Where possible, quality indicators should be derived from scientific evidence (78, 81). It has been noted that the stronger the evidence, the stronger the potential benefit of clinical indicators in terms of

32 improved quality of care (78, 81). The main reasons for developing measures using consensus techniques include synthesizing accumulated expert opinion, enhancing decision- making, facilitating development of indicators where evidence alone is insufficient, and identifying areas of care where there is controversy or uncertainty (78).

Therefore, articles were included for the purpose of this study if all three of the following criteria were met:

 The study methodology combined a systematic literature search/development of

indicators from clinical guidelines with assessment of quality indicators by an

expert panel;

 The study provided quality indicators applicable to the provision of ambulatory

care for chronic conditions, including diabetes, hypertension, chronic ischemic

heart disease, major depression and osteoarthritis among older adults;

 Quality indicators can be measured using Ontario administrative data.

The identified titles were entered in a bibliographical database and duplicates were removed.

One of the reviewers (YP) checked for the selected keywords in the title, abstract, and subject heading of the articles. The resulting abstracts were included for full text review. Two reviewers (YP and YS) independently conducted full text review according to the inclusion criteria. Also, the references of selected articles were screened for other relevant studies that had not been found in the electronic search. The resulting set of articles was included in the methodological assessment process. The level of consensus between reviewers evaluating

33 studies for inclusion and undertaking methodological assessments was assessed using the kappa statistic (103).

2.4.3 Methodological assessment

We used the Appraisal of Indicators through Research and Evaluation (AIRE) instrument for the methodological assessment of the quality of the included articles (104). It is a validated instrument that has been used previously in similar peer-reviewed studies (105-107). The

AIRE instrument contains 20 items, subdivided into four domains: a) purpose, relevance and organizational context, b) stakeholder involvement, c) scientific evidence, and d) additional evidence, formulation, usage. (Appendix 2)

The first domain (purpose, relevance and organizational context) aims to examine whether there was a clear statement of the purpose of the indicator development project such as quality assessment, quality improvement, research or other, as well as to assess whether the target population was clearly described (e.g., what medical condition they had). The second domain (stakeholder involvement) aims to assess appropriate stakeholder involvement in the development of the indicator in accordance with the overall aim of the quality indicator development project, and whether the indicator has been formally endorsed.

According to the Institute of Medicine (IOM), it is important for the appropriate high-level leadership, organization or expertise, rather than those who developed the measure, has to review and endorse measures of quality intended for population health improvement (108).

To this end, we identified which of the extracted quality indicators were endorsed by

National Quality Forum (NQF) and which were not. The National Quality Forum (NQF) is a voluntary not-for-profit membership organization created to develop and implement a national strategy for health care quality measurement and reporting. NQF-endorsed consensus standards are now widely viewed as the "gold standard" for measurement of healthcare quality, and NQF-endorsed measures are deemed to be evidence-based and valid

(109).

The third domain (scientific evidence) aims to assess whether a systematic literature search was conducted to summarize the evidence relevant to the clinical area for which the indicators were developed, whether the authors used explicit criteria for rating the indicators for final selection, as well as to evaluate the description of methods used to formulate the final indicator sets (e.g., Delphi process, project specific consensus process).

The final domain (additional evidence, formulation, and usage) evaluates the extent of supporting evidence provided by the authors for the final list of indicators, assesses whether explicit inclusion and exclusion criteria and the numerator and denominator were provided for each indicator in the final set. It also evaluates whether the information was provided about use of defined indicators with existing data and instructions for presenting and interpreting results.

Two authors (YP and YS) independently appraised the included studies with the AIRE instrument. The AIRE instrument was applied to each completed set of quality indicators, because publications usually gave information about the development and scientific evidence of the total set of indicators instead of for each indicator separately (105-107). Each item of

35 the AIRE instrument has a score ranging from 1 to 4 with: 1- strongly disagree (confident that the criterion has not been fulfilled or no information was available); 2/3 - disagree/agree

(unsure whether the criterion has been fulfilled; and 4 - strongly agree (confident that the criterion has been fulfilled) (104).

Scores for each of the four domains were calculated by summing up all the scores of the individual items in a domain and standardizing the total as a percentage of the maximum possible score for that domain. The maximum possible score for each domain was calculated by multiplying the maximum score per item (4) by the number of items in that domain (5, 3,

3, 9) and the number of appraisers (2). The minimum possible score was calculated by using the minimum score per item (1).

The standardized domain score was calculated as the total score per domain, minus the minimum possible score for that domain, divided by the maximum possible score, minus the minimum possible score * 100% (107). The standardized score ranges between 0% and

100%, and a score of 50% and higher indicates a higher methodological quality for each domain of the instrument (106). We conducted and reported this study according to the

Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) statement.

2.4.4 Data extraction

A structured data extraction form was used to describe the selected studies with respect to the quality of chronic disease care among older adults in ambulatory care settings. The extraction information consisted of title of the study, the publication date; summary of the indicator

36 selection process; description of indicators applicable to the ambulatory care for selected chronic conditions among older adults, including type, numerator and denominator of each quality indicator. The identified quality indicators were categorized according to

Donabedian’s “structure- process-outcome” framework (84, 110).

2.5 Results

2.5.1 Systematic review of quality indicators for care for older adults with diabetes

2.5.1.1 Search results

The systematic literature review identified 4,895 potentially relevant studies from OVID

MEDLINE and OVID EMBASE. (Figure 2.1) Three additional publications were identified through gray literature searching. After the review of titles/abstracts, only 37 studies were deemed potentially relevant. The full texts of these abstracts were obtained for the review.

One publication was derived after tracking the references. Of these, 32 publications were excluded primarily for the inability to meet inclusion criteria for a combined systematic literature search/ development of indicators from clinical guidelines and expert panel opinion. Finally, 5 publications were included in the review (kappa=0.93; very good agreement) (111).

Figure 2.1 Flow diagram for selection of studies for identifying indicators for diabetes

care among older adults

Records identified through database Additional records identified through search “gray” literature search (n = 4,895) (n = 3)

entification Identification

Records screened for duplicates (n = 4,898) Duplicates removed (n = 1,689)

Records screened on title and abstract Screening (n = 3,209) Records excluded (n = 3,172)

Full-text articles assessed for eligibility, including reference-checking Full-text articles excluded: (n = 37) (n = 32)

Eligibility  no indicator development  not on diabetes  no indicators applicable to ambulatory care settings Studies included in review  no indicators applicable to

(n =5) administrative data



Included Extracted quality indicators (n=12)

2.5.1.2 Study characteristics

The studies included in the review are summarized in Table 2.1. All included articles used a combination of literature review/development of quality indicators from clinical guidelines

38 and some form of consensus technique to derive a final set of quality indicators. One of the studies, obtained from the United States, aimed to develop a set of quality indicators for the care of diabetes mellitus in vulnerable older adults (112). In the context of the Organization for Economic Cooperation and Development (OECD) Quality Indicators Project, a set of quality indicators was developed for diabetes care at the health systems level (113). The study obtained from the Netherlands aimed to develop quality indicators of type 2 diabetes ambulatory care (114).

Another included study aimed to develop quality indicators for type 2 diabetes using electronic health records (115). One of the studies, obtained from Canada, aimed to rigorously develop and validate a set of quality indicators for type 2 diabetes to measure quality of diabetes care using standardized metric (116). All included studies identified quality indicators that are designed for use by physicians who manage the ongoing care of patients with diabetes, including primary care and specialist physicians.

Table 2.1 Characteristics of included studies for diabetes care indicators

Country/ First author/ Organization/ Year Study design Organization Initiative

Literature review for identifying RAND/ United States, candidate indicators; Shekelle (112) ACOVE 2007 RAND/UCLA Appropriateness Method for critical appraisal of indicators Literature review for identifying OECD Quality candidate indicators; indicators OECD, 2006 RAND/UCLA Appropriateness Method Nicolucci (113) Project for critical appraisal of indicators

Clinical guidelines review for identifying The candidate indicators; ______Netherlands, RAND/UCLA Appropriateness Martirosyan (114) 2006 Method for critical appraisal of indicators Literature review for identifying candidate indicators; Perez-Cuevas (115) ______Mexico, 2012 RAND/UCLA Appropriateness Method for critical appraisal of indicators Literature/Clinical guidelines review for identifying candidate indicators; Majumdar (116) ______Canada, 2005 Modified 3-stage Delphi technique for critical appraisal of indicators RAND, Research and Development; ACOVE, Assessing the Care of Vulnerable Elders Project; OECD, Organisation for Economic Co-operation and Development.

2.5.1.3 Methodological quality

The methodological quality of the included studies varied according to the AIRE instrument domain scores (Table 2.2). All studies were clear on the first AIRE instrument domain, demonstrating good evidence for describing the purpose of quality indicators development and the patient population to whom they were meant to apply, as well as good scientific evidence for indicator development process.

All five included studies received relatively low scores for the second AIRE domain due to lack of information regarding the formal endorsement of quality indicators. All included studies received high scores on the domains “Scientific evidence” and “Additional evidence, formulation, usage”; some initiatives reported numerator and denominator, while others only provided the list of indicators. In the included studies, the quality indicators were appraised for different criteria, including importance of the quality indicators to be scientifically sound, valid, reliable, and acceptable. The information regarding the use of defined indicators using

40 existing data and instructions for presenting results were described in three of the selected studies (114-116).

Table 2.2 Methodological quality of studies for diabetes care indicators

AIRE Instrument-Standardized Score (%)

Additional Purpose, evidence, First author relevance and Stakeholder Scientific formulation and organizational involvement evidence usage context

Shekelle (112) 94 73 94 88

Nicolucci (113) 92 72 83 85

Martirosyan (114) 85 70 81 83

Perez-Cuevas 89 66 83 85 (115)

Majumdar (116) 92 72 83 85

2.5.1.4 Quality indicators

Quality indicators were extracted only if they were relevant to ambulatory care for diabetes among older adults. For the purpose of this study, the target population was defined as patients aged 65 and older with a diagnosis of diabetes. Most of the included studies identified quality indicators for type 2 diabetes. A total of 12 quality indicators were identified from the included studies, many of which overlap conceptually or in content

(Table 2.3). Several quality indicators were represented in multiple articles and this may be a reflection of the attention to these areas of diabetes care in daily ambulatory care settings.

Eight out of 12 quality indicators were endorsed by NQF. All identified quality indicators have been categorized into groups: process and outcome. A full list of extracted indicators for diabetes care is presented in Appendix 3a.

Process indicators reflected the process of diabetes care in terms of its appropriateness and completeness. The identified process indicators were categorized into the following common groups: frequency of assessment of laboratory or screening tests, including HbA1c, LDL- cholesterol, microalbumin tests, and eye examination, as well as pharmacological treatment of diabetes. Several quality indicators were identified for care for older adults with diabetes with comorbid hypertension or chronic ischemic heart disease, including use of angiotensin- converting enzyme (ACE) inhibitors, statins, beta blockers and acetyl salicylic acid.

We identified no “do not do” or “negative” indicators for diabetes care in older adults.

All identified outcome indicators refer to the development of diabetes-related complications, including short-term (hypoglycemia, hypoglycemia, ketoacidosis, hyperosmolarity, or coma), and long-term complications (micro- and macrovascular complications). Two outcome indicators related to hospitalizations for diabetes-related short- or long-term complications

(117, 118) were developed by the Agency for Healthcare Research and Quality and endorsed by the National Quality Forum (NQF).

A number of identified quality indicators were not included for the purpose of this study because they are not amenable to measurement using Ontario administrative data, including

hemoglobin or cholesterol level, BP measurement and control, foot checking, patient

education, and assessment of self-management skills, etc.

Table 2.3 Quality indicators for diabetes care among older adults

Theme Indicator Source(s) Description and/ or Numerator, Denominator of indicator

PROCESS INDICATORS

Frequency of Majumdar, 2005 *Numerator: Number of patients who had glycated Nicolucci, 2006 2 measurements of HbA1c during the hemoglobin Perez-Cuevas, measurement period (HbA1c) tests 2012 Denominator: Number of patients aged 65 Frequency of Shekelle, 2007 years and older with a diagnosis of diabetes assessment of Health Quality during the measurement period laboratory/ Ontario screening NQF-endorsed tests Frequency of LDL- Majumdar, 2005 Numerator: Number of patients who had cholesterol tests Nicolucci, 2006 annual LDL-cholesterol testing Perez-Cuevas, Denominator: Number of patients aged 65 2012 years and older with a diagnosis of diabetes Health Quality during the measurement period Ontario Frequency of Majumdar, 2005 Numerator: Number of patients who had miroalbumin tests Nicolucci, 2006 annual microalbumin testing Perez-Cuevas, Denominator: Number of patients aged 65 2012 years and older with a diagnosis of diabetes Shekelle, 2007 during the measurement period

Frequency of eye Majumdar, 2005 Numerator: Number of patients who had examination Nicolucci, 2006 annual delated eye examination NQF-endorsed Denominator: Number of patients aged 65 Perez-Cuevas, years and older with a diagnosis of diabetes 2012 during the measurement period Shekelle, 2007 Health Quality Ontario NQF-endorsed

Use of angiotensin- Martirosyan, Numerator: Number of patients who were converting-enzyme 2008 prescribed an ACE inhibitor or ARB as a (ACE) inhibitors or Majumdar, 2005 first-choice drug angiotensin II Perez-Cuevas, Denominator: Number of patients aged 65 receptor blockers 2012 years and older diagnosed with diabetes and Treatment (ARBs) Health Quality hypertension during the measurement period Ontario NQF-endorsed

Use of beta Martirosyan, Numerator: Number of patients who were blockers 2008 prescribed a beta-blocker Majumdar, 2005 Denominator: Number of patients aged 65 years and older diagnosed with diabetes and ischemic heart disease during the measurement period Use of statins Martirosyan, Numerator: Number of patients who were 2008 prescribed a statin Majumdar, 2005 Denominator: Number of patients aged 65 Shekelle, 2007 years and older diagnosed with diabetes and Health Quality history of cardiovascular disease during the Ontario measurement period NQF-endorsed

Use of acetyl Martirosyan, Numerator: Number of patients who were salicylic acid 2008 prescribed acetyl salicylic acid (ASA) Majumdar, 2005 Denominator: Number of patients aged 65 Shekelle, 2007 years and older diagnosed with diabetes and NQF-endorsed history of cardiovascular disease during the measurement period OUTCOME INDICATORS

Diabetes- Lower-extremity Nicolucci, 2006 Numerator: Number of patients who had related amputation Majumdar, 2005 lower-extremity amputation during the microvascular NQF-endorsed measurement period complications Denominator: Number of patients aged 65 years and older with a diagnosis of diabetes during the measurement period Renal disease Nicolucci, 2006 Numerator: Number of patients who had Majumdar, 2005 end-stage renal disease during the measurement period Denominator: Number of patients aged 65 years and older with a diagnosis of diabetes during the measurement period Cardiovascula Cardiovascular Nicolucci, 2006 Numerator: Number of patients who had r mortality mortality died from cardiovascular disease (heart disease, hypertension, cerebrovascular

disease, peripheral vascular disease) Denominator: Number of patients aged 65 years and older with a diagnosis of diabetes during the measurement period Diabetes- Hospitalizations for Agency for Numerator: Number of patients who had related long- diabetes-related Healthcare been discharged with a principal code for term long-term Research and diabetes-related long-term complications complications complications Quality, 2014 (renal, eye, neurological, circulatory, or Health Quality complications not otherwise specified) Ontario Denominator: Number of patients aged 65 NQF-endorsed years and older with a diagnosis of diabetes during the measurement period Diabetes – Hospitalizations for Majumdar, 2005 Numerator: Number of patients who had related short- diabetes-related Agency for been discharged with a principal code for term short-term Healthcare diabetes-related short-term complications complications complications Research and (hypoglycemia, ketoacidosis, Quality, 2014 hyperosmolarity, or coma). Health Quality Denominator: Number of patients aged 65 Ontario years and older with a diagnosis of diabetes NQF-endorsed during the measurement period *Number of patients from the denominator

2.5.2 Systematic review of quality indicators for care for older adults with hypertension and chronic ischemic heart disease

2.5.2.1 Search results

Prior research results showed that, mostly, quality indicators for various cardiovascular

conditions are being presented in a single study (119); thus, we combined the search

strategies for identifying quality indicators for hypertension and chronic ischemic heart

disease. The systematic literature review identified 7,992 potentially relevant studies from

OVID MEDLINE and OVID EMBASE (Figure 2.2). Four additional publications were

identified through gray literature searching.

After the review of titles/abstracts, only 40 studies were deemed potentially relevant. The full texts of these abstracts were obtained for the review. Two publications were derived after tracking the references. Of these, 35 publications were excluded primarily for the inability to meet inclusion criteria for a combined systematic literature search/ development of indicators from clinical guidelines and expert panel opinion. Finally, 7 publications were included in the review (kappa=0.93; very good agreement) (111).

Figure 2.2 Flow diagram for selection of studies for identifying indicators for

hypertension/chronic ischemic heart disease care among older adults

Records identified through database Additional records identified through search “gray” literature search (n = 7,988) (n = 4)

entification Identification

Records screened for duplicates (n = 7,992) Duplicates removed (n = 2,683)

Records screened on title and abstract Screening (n = 5,309) Records excluded (n = 5,268)

Full-text articles assessed for eligibility,

including reference-checking (n = 42) Full-text articles excluded: (n = 35)

 no indicator development Eligibility  not on hypertension/ chronic IHD Studies included in review  no indicators applicable to (n =7) ambulatory care settings

 no indicators applicable to

administrative data  Extracted quality46 indicators Included (n=7 – for hypertension) (n=6 – for chronic IHD)

2.5.2.2 Study characteristics

The studies included in the review are summarized in Table 2.4. Four of the included studies were obtained from the United States, one study presented the European Union European

Practice Assessment Cardio Project quality indicators project, and one study was obtained from Mexico. All included articles used a combination of literature review/development of quality indicators from clinical guidelines and some form of consensus technique (Delphi,

Modified Delphi methodology) to derive a final set of quality indicators. Included studies differ in terms of target population and settings, but all presented quality indicators for care of hypertension and/or chronic ischemic heart disease in older adults in ambulatory care settings and were measurable using Ontario administrative data.

Two of the studies, obtained from the United States, aimed to develop a set of quality indicators for the care of hypertension and acute/ chronic ischemic heart disease in vulnerable elders (120, 121). Another study obtained from the United States aimed to develop a set of quality indicators for cardiopulmonary conditions, including hypertension and ischemic heart disease (122). The Canadian Cardiovascular Outcomes Research Team

(CCORT) aimed to develop a set of quality indicators for ambulatory care practice for the primary prevention and chronic disease management of hypertension, ischemic heart disease, hyperlipidemia and heart failure (123).

One of the studies obtained from the United States aimed to develop a set of quality indicators for the care of hypertension and acute/ chronic ischemic heart disease (124).

Another study aimed to develop a set of quality indicators for the prevention and

47 management of cardiovascular disease, including ischemic heart disease, in nine European countries (125).

Table 2.4 Characteristics of included studies for hypertension and chronic ischemic heart disease care

Organization/ First author/ Initiative Country/ Year Study design Organization

Kerr (122) RAND United States, 2000 Literature review for identifying candidate indicators; RAND/UCLA Appropriateness Method for critical appraisal of indicators Doubova (126) ------Mexico, 2013 Literature review for identifying candidate indicators; RAND/UCLA Appropriateness Method for critical appraisal of indicators Min (120) RAND/ACOVE United States, 2007 Literature review for identifying candidate indicators; RAND/UCLA Appropriateness Method for critical appraisal of indicators Burge (123) CCORT Canada, 2007 Literature/clinical guidelines review for identifying candidate indicators; 4-stage modified Delphi approach for critical appraisal of indicators Watson (121) RAND/ACOVE United States, 2007 Literature review for identifying candidate indicators; RAND/UCLA Appropriateness Method for critical appraisal of indicators Drozda (124) ACCF/AHA/AM United States, 2011 Literature/clinical guidelines review A-PCPI for identifying candidate indicators; Delphi technique for critical appraisal of indicators Campbell (125) EPA Cardio European Union*, Literature/clinical guidelines review Project 2008 for identifying candidate indicators; Delphi technique for critical appraisal of indicators RAND, Research and Development; ACOVE, Assessing the Care of Vulnerable Elders Project; CCORT, Canadian Cardiovascular Outcomes Research Team; ACCF, American College of Cardiology Foundation;

AHA, American Heart Association, AMA, American Medical Association, PCPI, Physician Consortium for Performance Improvement; EPA, European Practice Assessment. * Nine European countries participated in the panel: Austria, Belgium, Finland, France, Germany, The Netherlands, Slovenia, United Kingdom, and Switzerland.

2.5.2.3 Methodological quality

The methodological quality of the included studies varied according to the AIRE instrument domains’ scores (Table 2.5). All studies were demonstrating good evidence for describing the purpose of quality indicators development and patient population to whom they meant to apply, as well as good scientific evidence for indicator development process.

Only one of the included studies, presented information regarding the formal endorsement of developed quality indicators (124). The remaining six studies received relatively low scores for the second AIRE domain due to lack of information regarding the formally endorsement of quality indicators. All included studies received high scores on the domains “Purpose, relevance and organizational context“ and “Additional evidence, formulation, usage”; some initiatives reported numerator and denominator, while others only provided the list of indicators. In the included studies, the quality indicators were appraised for different criteria, including importance of the quality indicators to be scientifically sound, valid, reliable, and acceptable.

One of the studies (126) provided scarce information regarding the systematic methods that were used to search for scientific evidence and critical appraisal of the supporting evidence.

The information regarding the piloting indicators on practice and instructions for presenting results were described in six selected studies.

Table 2.5 Methodological quality of included studies for indicators for care for hypertension and chronic ischemic heart disease

AIRE Instrument-Standardized Score (%)

Purpose, Additional First author relevance and Stakeholder Scientific evidence, organizational involvement evidence formulation context and usage Kerr (122) 92 78 83 85

Doubova (126) 87 56 56 88

Min (120) 92 74 84 83

Burge (123) 90 76 83 86

Watson (121) 92 74 84 83

Drozda (124) 88 93 83 87

Campbell (125) 73 67 55 62

2.5.2.4 Quality indicators

2.5.2.4.1 Quality indicators for hypertension

Quality indicators were extracted only if they were relevant to ambulatory care for hypertension for older adults. A total of 7 quality indicators were identified from the included studies, many of which overlap conceptually or in content (Table 2.6). Three out of 7 quality

50 indicators were endorsed by the NQF. All identified quality indicators relate to process indicators.

Process indicators reflected the process of diabetes care in terms of its appropriateness and completeness. The identified process indicators have been categorized into the following common groups: frequency of assessment of laboratory or screening tests, including LDL- cholesterol/total cholesterol tests, serum creatinine among patients with hypertension, and microalbumin and HbA1c tests in hypertensive patients with diabetes, as well as treatment of hypertensive patients with and without comorbid conditions, including diabetes and ischemic heart disease. We identified no “do not do” or “negative” indicators for care for older adults with hypertension.

Quality indicators for hypertension care that are not amenable to measurement using Ontario administrative data were not included in this study, e.g. blood pressure measurement and control, education and follow-up of care, care documentation, or non-pharmacological interventions, including diet, exercise, weight loss, alcohol use. A full list of extracted indicators for hypertension care is presented in Appendix 3b.

Table 2.6 Quality indicators for hypertension care among older adults

Process Source(s) Description and/ or Numerator, Denominator of Indicator indicator

Frequency of Doubova, 2013 *Numerator: Number of patients who had a total/LDL- assessment of Kerr, 2000 cholesterol test during the measurement period laboratory/ Denominator: Number of patients aged 65 years and older

screening tests with a diagnosis of hypertension during the measurement period Doubova, 2013 Numerator: Number of patients who had a serum Kerr, 2000 creatinine test during the measurement period NFF-endorsed Denominator: Number of patients aged 65 years and older with a diagnosis of hypertension during the measurement period Doubova, 2013 Numerator: Number of patients who had a HbA1c Kerr, 2000 (glycated hemoglobin) test during the measurement period Denominator: Number of patients aged 65 years and older with a diagnosis of hypertension during the measurement period Burge, 2007 Numerator: Number of patients who had a microalbumin test during the measurement period Denominator: Number of patients aged 65 years and older with a diagnosis of hypertension and diabetes during the measurement period Doubova, 2013 Numerator: Number of patients who were prescribed an NQF-endorsed angiotensin converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) during the measurement period Denominator: Number of patients aged 65 years and older with hypertension and diabetes during the measurement period Min, 2007 Numerator: Number of patients who were prescribed an

NQF-endorsed angiotensin converting enzyme (ACE) inhibitors or

angiotensin receptor blockers (ARBs) during the

measurement period Treatment Denominator: Number of patients aged 65 years and older with a diagnosis of hypertension and ischemic heart disease during the measurement period Min, 2007 Numerator: Number of patients who were prescribed a beta-blocker during the measurement period Denominator: Number of patients aged 65 years and older with a diagnosis of hypertension and ischemic heart disease during the measurement period *Number of patients from the denominator

2.5.2.4.2 Quality indicators for chronic ischemic heart disease

Quality indicators were extracted only if they were relevant to ambulatory care for chronic ischemic heart disease. Chronic ischemic heart disease is defined as an established pattern of angina pectoris, a history of myocardial infarction, or the presence of plaque documented by catheterization (97). A total of 6 quality indicators were identified from the included studies, many of which overlap conceptually or in content (Table 2.7). Several quality indicators were represented in multiple articles that may be a reflection of the attention to these areas of care for chronic ischemic heart disease in daily ambulatory care settings. Two out of 6 identified quality indicators were endorsed by the NQF. All identified quality indicators relate to process indicators.

Process indicators reflected the process of chronic ischemic heart disease care in terms of its appropriateness and completeness. The identified process indicators have been categorized into the following common groups: frequency of assessment of laboratory or screening tests, including, LDL-cholesterol and HbA1c tests, as well as treatment of ischemic heart disease, including antiplatelet therapy, use of ACE inhibitor/ARBs, beta blockers and statins. We identified no “do not do” or “negative” indicators for care for older adults with chronic ischemic heart disease.

Quality indicators for ischemic heart disease care that are not amenable to measurement using Ontario administrative data were not included in this study, e.g. BMI assessment,

53 smoking cessation counseling, lipid or glucose level control, or exercise testing. A full list of extracted indicators for care for chronic ischemic heart disease is presented in Appendix 3c.

Table 2.7 Quality indicators for care for chronic ischemic heart disease among older adults

Process Source(s) Description and/ or Numerator, Denominator of Indicator indicator

Burge, 2007 *Numerator: Number of patients who had a LDL- Campbell, cholesterol test in the last 15 months 2008 Denominator: Number of patients aged 65 years and older Frequency of Watson, 2007 with a diagnosis of chronic ischemic heart disease during the assessment of NQF-endorsed measurement period laboratory/ Burge, 2007 Numerator: Number of patients who had a HbA1c test screening tests during the measurement year Denominator: Number of patients aged 65 years and older with a diagnosis of chronic ischemic heart disease during the measurement period Burge, 2007 Numerator: Number of patients who were prescribed Watson, 2007 aspirin or other antiplatelet therapy during the Campbell, measurement period 2008 Denominator: Number of patients aged 65 years and older Drozda, 2011 with a diagnosis of chronic ischemic heart disease during the Kerr, 2000 measurement period NQF-endorsed Burge, 2007 Numerator: Number of patients who were prescribed a Watson, 2007 beta blocker during the measurement period Campbell, Denominator: Number of patients aged 65 years and older 2008 with a diagnosis of chronic ischemic heart disease during the Treatment Drozda, 2011 measurement period Burge, 2007 Numerator: Number of patients who were prescribed an Watson, 2007 angiotensin converting enzyme (ACE) inhibitors or Drozda, 2011 angiotensin receptor blockers (ARBs) during the measurement period Denominator: Number of patients aged 65 years and older with a diagnosis of chronic ischemic heart disease during the measurement period Campbell, Numerator: Number of patients who were prescribed 2008 statins during the measurement period Denominator: Number of patients aged 65 years and older

with a diagnosis of chronic ischemic heart disease during the measurement period *Number of patients from the denominator

2.5.3 Systematic review of quality indicators for care for older adults with major depression

2.5.3.1 Search results

The systematic literature review identified 3,838 potentially relevant studies from OVID

MEDLINE and OVID PsycINFO (Figure 2.3). Two additional publications were identified through gray literature searching. After the review of titles/abstracts, 38 studies were deemed potentially relevant. One publication was derived after tracking the references. The full texts of these abstracts were obtained for the review. Of these, 32 publications were excluded primarily because they failed to combine a systematic literature search or development of indicators based on clinical guidelines and expert panel opinion. Two more publications were excluded because a set of provided quality indicators applied to the organizational aspects of depression care and could not be measured using Ontario administrative data. Finally, four publications were included in the review (kappa=0.89; very good agreement) (111).

Depression in this systematic review connotes major depression and dysthymia, since most clinical practice guidelines only address treatment of major depression (89). Previous studies demonstrate that treatments for major depression also apply to dysthymic disorders (127-

129). Major depression disorder is defined “as a period lasting at least 2 weeks characterized either by depressed mood (most of the day, nearly every day) and/or markedly diminished

55 interest or pleasure in all, or almost all, activities (most of the day, nearly every day), during which a patient experiences five or more symptoms”(130). Dysthymic Disorder is

“characterized by a chronically depressed mood that occurs most of the day, more days than not, for at least 2 years”(130).

Figure 2.3 Flow diagram for selection of studies for identifying indicators for depression care among older adults

Records identified through database Additional records identified through search “gray” literature search (n = 3,838) (n = 4)

entification Identification

Records screened for duplicates (n = 3,842) Duplicates removed (n = 306) Records screened on title and abstract

Screening (n = 3,536) Records excluded (n = 3,497)

Full-text articles assessed for eligibility, including reference-checking (n = 39) Full-text articles excluded:

Eligibility (n = 35)  no indicator development  not on depression Studies included in review  no indicators applicable to (n = 4) ambulatory care settings

 no indicators applicable to administrative data

Included Extracted quality indicators (n=6)

2.5.3.2 Study characteristics

The studies included in the review are summarized in Table 2.8. Three of the included studies were obtained from the United States, and one study presented the OECD quality indicators project. All included articles (131-134) used a combination of literature review/development of quality indicators from clinical guidelines and some form of consensus technique (Delphi, Modified Delphi methodology, or “constituency approach”) to derive a final set of quality indicators. Included studies differed in terms of target population and settings, but all of them provided quality indicators measurable using administrative data and applicable to older adults in ambulatory care settings.

One of the studies was focused on assessing care for vulnerable elders with major depression or dysthymia in both outpatient and inpatient settings (131). Two studies (132, 134) aimed to develop a set of evidence-based indicators for mental health services, including those for depression and delivered within ambulatory care settings for older adults. Another study aimed to develop quality indicators for the management of major depressive disorders among adults in ambulatory care settings (133).

Table 2.8 Characteristics of studies included for quality indicators for depression care

Country/ First author Organization/ Year Study design /Organization Initiative

Nakajima (131) RAND/ACOVE United States, Literature review for identifying 2007 candidate indicators; RAND/UCLA Appropriateness Methodology for critical appraisal of indicators Hermann (132) United States, Literature review for identifying 2004 candidate indicators; Two-stage Modified Delphi method for critical appraisal of indicators Veterans Health VHA/DOD United States, Development of candidate indicators from Administration 2000 guidelines; Department of Delphi methodology for critical appraisal Defence (133) of indicators Hermann (134) OECD HCQI OECD, 2006 Candidate indicators were dawn form Project OECD member countries quality initiatives; Modified Delphi Methodology for critical appraisal of indicators RAND, Research and Development; ACOVE, Assessing the Care of Vulnerable Elders Project; VHA/DOD, Veterans Health Administration Department of Defence; OECD, Organisation for Economic Co-operation and Development.

2.5.3.3 Methodological quality

The methodological quality of the included studies varied according to the AIRE instrument domains’ scores (Table 2.9). All studies were clear on the first AIRE instrument domain, demonstrating good evidence for describing the purpose of quality indicators development and the patient population to whom they were meant to apply, as well as presenting the indicator selection criteria and applicability of measures. Two studies (132, 133) received low scores for the second AIRE domain due to lack of information regarding the relevant stakeholders’ involvement at some stage of the indicator development process, such as

58 quality indicator development and review, as well as lack of formal endorsement of the indicators.

The scientific evidence for the quality indicator development process was scarcely described in two articles (133, 134). Finally, most studies received high scores on the domain

“Additional evidence, formulation, usage”; some initiatives reported numerator and denominator, while others only provided the list of indicators. In the included studies, the quality indicators were appraised for different criteria, including importance of the quality indicators to be scientifically sound, valid, reliable and feasible to collect.

In most studies information regarding the piloting indicators in practice and instructions for presenting and interpreting results were scarcely described. Feasibility of data collection was assessed in three studies (132-134). In most studies information regarding the use of defined indicators with existing data and instructions for presenting and interpreting results were scarcely described.

Table 2.9 Methodological quality of studies for depression care indicators

AIRE Instrument-Standardized Score (%)

Purpose, Additional First author Scientific relevance and Stakeholder evidence, evidence organizational involvement formulation and

context usage

Nakajima (131) 88 73 94 88

Hermann (132) 78 55 61 75

VHA/DOD (133) 73 54 55 78

Hermann (134) 78 83 60 72

2.5.3.4 Quality indicators

Quality indicators were extracted only if they were relevant to ambulatory care for major depression (Table 2.10). For the purpose of this study, the target population was defined as patients aged 65 and older with a diagnosis of major depression or dysthymia. The literature suggests that treatment of major depressive disorder also applies for dysthymia disorder

(127-129, 135), and, mostly, quality indicators related to the management of major depression apply to that of dysthymia.

A total of 6 quality indicators were identified from the included studies, many of which overlap conceptually or in content (see Table 3). Several quality indicators were represented in multiple articles that may be a reflection of the attention to these areas of depression care in daily ambulatory care settings. All resulting quality indicators reflected the process of care for major depression in terms of its appropriateness and completeness. Only two quality indicators were endorsed by the NQF.

The identified process indicators have been categorized into the following common groups: depression assessment/diagnosis, initiating and continuing of depression treatment, and antidepressant choice. We identified a few “negative” process indicators, including use of tertiary amine tricyclics, MAO inhibitors, or benzodiazepines as first- or second-line therapy in older adults with major depression. There were no indicators identified for care for older adults with diabetes with comorbid major depression in ambulatory care settings.

Quality indicators for depression care that are not amenable to measurement using Ontario administrative data were not included in this study, e.g. patient∕ caregiver education, medication review, access to care, documenting depression symptoms or assessing depression remission score. A full list of extracted indicators for major depression care is presented in Appendix 3d.

Table 2.10 Quality indicators for care for older adults with major depression/dysthymia

Process Source(s) Description and/ or Numerator, Denominator of Indicator indicator

Assessment/ VHA/DOD, *Numerator: Number of patients with a diagnosis of major Diagnosis 2000 depression or dysthymia during the previous 12 months. Denominator: Number of patients aged 65 and older seen in a general medicine, primary care, women’s or mental health primary care clinic, during the previous 12 months. Antidepressant Nakajima, 2007 Numerator: Number of patients who were prescribed choice (“Negative antidepressants using tertiary amine tricyclics, MAOIs

indicator”) (unless atypical depression is present), benzodiazepines, or stimulants (except methylphenidate) as first- or second-line therapy Denominator: Number of patients aged 65 and older with a diagnosis of major depression or dysthymia during the measurement period Hermann, 2006 Numerator: Number of patients who were prescribed Hermann, 2004 anticholinergic antidepressants as first- or second-line (“Negative therapy indicator”) Denominator: Number of patients aged 65 and older with a diagnosis of major depression or dysthymia during the measurement period Interactions with Nakajima, 2007 Description: If a patient with a diagnosis of major depression MAOIs or dysthymia is taking an SSRI, then an MAOI should not be used for at least 2 weeks after termination of the SSRI, and vice versa Continuing Hermann, 2004 Numerator: Number of patients who responded to antidepressant Hermann, 2006 antidepressant medication and remained on an medication VHA/DOD, antidepressant treatment for at least 3 months (12 treatment in acute 2000 weeks) phase (NQF-endorsed) Denominator: Number of patients aged 65 and older who diagnosed with a new episode of major depression or dysthymia during the measurement period Continuing Nakajima, 2007 Numerator: Number of patients who responded to depression Hermann, 2006 antidepressant medication, remained on the drug at the therapy in Hermann, 2004 same dose for at least 6 months continuation (NQF-endorsed) Denominator: Number of patients aged 65 and older phase newly diagnosed with and treated for major depression or dysthymia, during the measurement period *Number of patients from the denominator

2.5.4 Systematic review of quality indicators for care for older adults with osteoarthritis

2.5.4.1 Search results

The systematic literature review identified 4,524 potentially relevant studies from OVID

MEDLINE and OVID EMBASE (Figure 2.4). Two additional publications were identified

62 through gray literature searching. After the review of titles/abstracts, only 31 studies were deemed potentially relevant. The full texts of these abstracts were obtained for the review.

One publication was identified after tracking the references. Of these, 28 publications were excluded primarily for the inability to meet inclusion criteria for a combined systematic literature search/ development of indicators from clinical guidelines and expert panel opinion. One more publication was excluded because the provided quality indicators can’t be measured using administrative data. Finally, 3 publications were included in the review

(kappa=0.91; very good agreement) (111).

Figure 2.4 Flow diagram for selection of studies for identifying indicators for osteoarthritis care among older adults

Records identified through database Additional records identified through “gray” literature search search (n = 4,524) (n = 2)

entification Identification

Records screened for duplicates (n = 4,526) Duplicates removed (n = 1,636)

Records screened on title and abstract Screening (n = 2,890) Records excluded (n = 2,858)

Full-text articles assessed for eligibility, including reference-checking Full-text articles excluded: (n = 32) (n = 29) Eligibility  no indicator development  not on osteoarthritis  no indicators applicable to Studies included in review ambulatory care settings (n =363)  no indicators applicable to administrative data



Extracted quality indicators Included (n=3)

2.5.4.2 Study characteristics

The studies included in the review are summarized in Table 2.11. All three included studies were obtained from the United States. All three articles used combination of literature review/ development of quality indicators from clinical guidelines and some form of consensus technique to derive a final set of quality indicators.

One of the studies aimed to develop a set of evidence-based indicators for osteoarthritis among vulnerable older adults (136), while the others focused on the care of osteoarthritis among adults (137, 138). All included studies provide quality indicators that are designed for use by physicians who manage the ongoing care of patients with osteoarthritis, including primary/specialty care physicians.

Table 2.11 Characteristics of included studies for quality indicators for osteoarthritis care

First author/ Organization/ Country/ Organization Study design Initiative Year

United States, Literature review for identifying MacLean (136) RAND/ ACOVE 2007 candidate indicators;

RAND/UCLA Appropriateness Method for critical appraisal of indicators Literature review for identifying candidate indicators; United States, Pencharz (137) Arthritis Foundation Modification of the RAND/UCLA 2004 Appropriateness Method for critical appraisal of indicators Literature /clinical guidelines review for identifying candidate United States, PCPI (138) AAOS/ PCPI indicators; 2006 Consensus technique for critical appraisal of indicators RAND, Research and Development; ACOVE, Assessing the Care of Vulnerable Elders Project; AAOS, American Academy of Orthopaedic Surgeons; PCPI, Physician Consortium for Performance Improvement; EPA, European Practice Assessment.

2.5.4.3 Methodological quality of indicators for osteoarthritis care among older adults

The methodological quality of the included studies varied according to the AIRE instrument domains’ scores (Table 2.12). All studies were demonstrating good evidence for describing the purpose of quality indicators development and patient population to whom they meant to apply, as well as good scientific evidence for indicator development process.

Two studies (136, 137) received low scores for the second AIRE domain due to lack of information regarding the formally endorsement of quality indicators. The information regarding use of indicators and instructions for presenting and interpreting results were scarcely described in one of the selected studies (137). In the included studies, the quality indicators have been appraised for different criteria, including importance of the indicators to

65 be scientifically sound, valid, and reliable. Feasibility of data collection was assessed in one of the selected studies (138).

Table 2.12 Methodological quality of studies for osteoarthritis care indicators

AIRE Instrument-Standardized Score (%)

Purpose, Additional First author relevance and Stakeholder Scientific evidence, organizational involvement evidence formulation and context usage

MacLean (136) 88 73 94 88

78 Pencharz (137) 67 94 67

86 94 78 87 PCPI (138)

2.5.4.4 Quality indicators for osteoarthritis care among older adults

From the included studies, quality indicators were extracted only if they were relevant to the ambulatory care for osteoarthritis among older adults in ambulatory care settings that can be measured using administrative data. A total 3 quality indicators were identified from the included studies that partially overlap conceptually or in content (Table 2.13).

All three studies concerned the process indicators for osteoarthritis care (136-138). Process indicators represent the way osteoarthritis care is delivered. They include pharmacological treatment: use of acetaminophen as a first-line therapy, use of use of anti-inflammatory medications that were endorsed by the NQF, and use of combination of non-steroidal anti- inflammatory medications and misoprostol or proton pump inhibitor. There were no indicators identified for care for older adults with diabetes with comorbid osteoarthritis in ambulatory care settings.

A number of identified quality indicators were not included for the purpose of this study because they are not amenable to measurement using Ontario administrative data, e.g. physical examination, functional pain assessment, patient∕ caregiver education, BMI control, or pain reduction assessment. A full list of extracted indicators for osteoarthritis care is presented in Appendix 3e.

Table 2.13 Quality indicators for osteoarthritis care among older adults

Process Source(s) Description and/ or Numerator, Denominator of Indicator indicator

Use of anti- PCPI, 2006 *Numerator: Number of patients who were prescribed an inflammatory NQF-endorsed non-steroidal anti-inflammatory medication (NSAID) medications Denominator: Number of patients aged 65 years and older with a diagnosis of osteoarthritis during the measurement period First-line MacLean, 2007 Numerator: Number of patients who were prescribed pharmacological Pencharz, 2004 acetaminophen first, unless there is a documented

67 therapy contraindication to use Denominator: Number of patients aged 65 years and older with a diagnosis of osteoarthritis who were started on pharmacological therapy, during the measurement period Gastroprotection PCPI, 2006 Numerator: Number of patients taking a NSAID in MacLean, 2007 combination with misoprostol or proton pump inhibitor Denominator: Number of patients aged 65 years and older with a diagnosis of osteoarthritis during the measurement period *Number of patients from the denominator

2.6 Discussion

This systematic review was conducted to identify evidence-based and valid quality indicators for evaluating ambulatory care for people aged 65 and older with diabetes, major depression, osteoarthritis, hypertension, and chronic ischemic heart disease. All identified quality indicators assess clinical aspects of ambulatory care, including primary and specialty care, for people with selected chronic conditions.

Mainz (75) stated that clinical indicators create the basis for quality improvement and prioritization in the health care system. All included studies used the most rigorous way of developing quality indicators by combining a systematic literature search/developing indicators from clinical guidelines with appraisal of candidate indicators using expert panel opinion (78, 80, 81). Despite the importance of care for selected chronic diseases among older adults, relatively few of the identified quality indicators are formally indorsed as legitimate measures of quality of care by the NQF. The studies included in the present systematic review mainly represent the views of healthcare providers and researchers, while it is recommended to include perspectives of all potential end users including patients/

68 caregivers, health professionals and managers in the process of developing quality indicators

(139).

Most quality indicators identified in this study focused on the process of ambulatory care for selected chronic conditions. Process indicators measure the activities and tasks in ambulatory care for chronic conditions among older adults, including frequency of laboratory/screening tests, initiating, implementing and monitoring chronic disease treatment. We identified a few

“negative” indicators for care for older adults with major depression, including use of anticholinergic antidepressants, tetracyclic antidepressants, benzodiazepines and monoamine oxidase inhibitors as first- or second-line therapy.

The advantages of process indicators for the assessment of the quality of provided care include: 1) providing feedback for quality improvement initiatives; 2) most process indicators do not require risk adjustment; and 3) they can easily be assessed using various data sources, including medical records or administrative data. Meanwhile, the literature suggests that although the providers of care might need detailed information about the process of care for quality improvement purposes, process indicators do not signify quality of care until they are validated by detecting their relationship to desirable outcomes (75).

Mainz et al. stated that, although the healthcare providers might need detailed information about the process of care for quality improvement purposes, financiers of the care and consumers might be more interested in outcomes of the care (75). Therefore, a combination

69 of process and outcome indicators might be most suitable for measuring the quality of care for chronic diseases.

Outcome indicators were identified only for ambulatory care of diabetes among older adults, including diabetes-related hospitalizations, cardiovascular mortality and renal disease rate.

There were no outcome indicators identified for care for hypertension or chronic ischemic heart disease from the studies included in this review. Several outcome indicators were identified for care for major depression and osteoarthritis, including depression remission and treatment response scores, or functional improvement and pain reduction in patients with osteoarthritis. However, they were not included in this study since they are not amenable to measurement using Ontario administrative data.

Overall, reasons for the small number of outcome indicators for chronic disease management may be the limited scientific evidence linking structure and process to outcomes of chronic disease care, or perhaps the length of time it takes to assess outcomes due to the often long- term and fluctuating nature of chronic conditions or lack of understanding of what meaningful outcomes could/should be measured.

As it was mentioned above, this systematic review aimed to identify evidence-based and valid quality indicators for care for selected single chronic conditions or combinations of them.

The current systematic review results demonstrate that there is a limited number of published quality indicators addressing comorbidity issue among older adults. In particular, we

70 identified only a few indicators for care for diabetes patients with comorbid hypertension and ischemic heart disease, including use of antihypertensive medication, statins and antiplatelet therapy.

There were no published quality indicators for care for people with discordant or both types of comorbid conditions, including care for diabetes with comorbid osteoarthritis or major depression, or care for diabetes with comorbid osteoarthritis and hypertension. The possible explanation can be that there is lack of scientific evidence on which to build quality indicators for people with multiple chronic conditions. Most single disease clinical practice guidelines fail to give adequate guidance for care for older adults with multiple chronic conditions (46, 48). Moreover, most clinical trials tend to exclude people with mutimorbidity, and most observational studies rarely include analyses addressing treatment effectiveness or interactions in patients with multimorbidity (48).

2.6.1 Limitations

As demonstrated in this study, relatively little research has been done to develop quality indicators to assess the quality of ambulatory care for older adults with selected chronic conditions and combinations of them. However, our intention to include only indicators that have been developed through an evidence-based approach, including a combination of literature search with expert panel opinion, may have led to the exclusion of some indicators that were developed using other approaches.

The AIRE instrument that was used to assess the quality of the included studies is mainly focused on the indicator development process. Thus, we may underestimate the methodological quality of some studies, because the information related to the indicator development process was not always described within the articles. We tried to track down additional information in the literature about the development process of quality indicators, but we were able to obtain relevant additional information only for a few sets of quality indicators. Due to time constraints, we did not contact any organization/author to elicit any additional information. This indicates the need for further development of quality indicators with detailed methodological specifications for monitoring and accurate assessment of the care for older adults with selected chronic conditions.

The current systematic review aimed to identify quality indicators only if they can be measured using Ontario administrative data. Thus, we did not include a number quality indicators related to various aspects of care for chronic conditions among older adults, including patient self-management and education, patient satisfaction and quality of life, non- pharmacological treatment, patient-physician relationships, or documentation of care. Since the literature search was restricted to studies published in English potentially relevant publications in other languages may have been omitted.

2.7 Conclusions

Relatively little research has been done to develop quality indicators that assess the quality of care for older adults with selected chronic conditions or combinations of them in ambulatory

72 care settings. As the burden of chronic conditions, including diabetes, hypertension, chronic ischemic heart disease, osteoarthritis and major depression, is high among older adults, and much of it is presented clinically to primary care/specialist physicians, incorporation of these indicators to routine ambulatory care practice is recommended.

The study results demonstrate that there are only a few published indicators for care for older adults with diabetes comorbid with concordant conditions, including hypertension and ischemic heart disease. Therefore, there is need for developing quality indicators for care for older adults with various disease combinations with an awareness of how the treatment of one disease can affect the treatment of any or all other conditions.

CHAPTER 3

Development of quality indicators for ambulatory care for older adults with concurrent chronic conditions: a Delphi study

Abstract

Background: While disease-specific indicators are important, they may not be appropriate to evaluate the care for people with multiple chronic conditions, since they do not address several multimorbidity issues, including treatment interactions and disease interactions. Therefore, it is necessary to balance the clinical benefits and risks of disease-specific quality standards in managing people with multiple chronic conditions.

Purpose: 1) to critically appraise and select the most appropriate set of quality indicators for ambulatory care for older adults with five selected disease combinations that are amenable to measurement using administrative data; 2) to prepare a summary of defined quality indicators by each selected disease combination.

Methods: We used a two-stage web-based modified Delphi process to critically appraise and select the quality indicators for care for older adults with diabetes with comorbidities. A fifteen-member Canadian expert panel with broad geographical and clinical representation participated in this study. A 73% agreement threshold (11 of 15 panelists) was required for high consensus, and 60-72% for moderate consensus as measured on a 5-point Likert type scale. The panel evaluated process indicators for meaningfulness, potential for improvements in clinical practices, and overall value of inclusion, while outcome indicators for importance, modifiability and overall value of inclusion.

Findings: Panelists reached rapid consensus on quality indicators for care for older adults with diabetes with concordant comorbid conditions, unlike on those for care for diabetes patients with discordant or both types of comorbid conditions. Overall, twenty four high- consensus and twenty four medium-consensus indicators were selected for assessing care for older adults with five selected disease combinations. All selected indicators assess clinical aspects of the performance of ambulatory care and related outcomes for older adults with selected disease combinations.

Conclusions: The developed quality indicators are not intended to provide a comprehensive tool set for measuring quality of care for older adults with diabetes with comorbidities. Rather, they address clinical aspects of care and can be used as a starting point for further development and use in ambulatory care settings. There is a need to develop specific outcome measures for older people with diabetes with comorbidities to reflect what matters most to patients, which is the effect of all their conditions on their health status.

3.1 Introduction

Care for people with multiple chronic conditions includes multiple domains, such as screening and prevention, management and secondary prevention, community services, rehabilitation and end-of-life care (70). Evidence shows that the majority of care for adults with chronic conditions is provided in ambulatory care settings, and therefore this should be the main setting for approaches to improve the management of people with multiple chronic conditions (140).

Recent population-based study conducted in Ontario identified that the most prevalent chronic conditions include diabetes, osteoarthritis, depression, coronary artery syndrome, asthma and cancer (5). Another study demonstrated that more than 90% of older adults with diabetes in Ontario had at least one comorbid condition, and the number of comorbid conditions increases with age (24). The researchers found that the most common conditions among patients with diabetes included hypertension (79.1%), ischemic heart disease (37.6%), arthritis (59.3%), and anxiety (36.9%) (24).

The IOM defined quality of care as the “the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge”(p. 4 ) (76). Using quality indicators is common way of measuring and monitoring the quality of care and services (75). To improve the quality of care for people with multimorbidity, quality indicators should be sensitive to the heterogeneity and scope of care for a particular individual with particular types of co-existing conditions, as well as the patient’s care preferences, and the risks of uncoordinated care

75 across various settings and adverse events in the presence of comorbidities. Quality indicators should reflect explicit actions, performed for carefully specified and easily identified patients with multiple chronic conditions (141).

Data generated from these measures can be used to document the quality of care, make judgments and set priorities, make comparisons (benchmarking), support quality improvement, and support accountability and regulation (75). For instance, measuring the appropriateness and number of prescribed medications among people with diabetes comorbid with different types of chronic conditions would facilitate monitoring, and thereby, reduce undesired consequences from polypharmacy.

The literature suggests that quality indicators should be evidence-based and be derived from the academic literature. However, when scientific evidence is lacking, quality indicators can be defined by an expert panel of professionals by means of consensus techniques based on their experience (75). Campbell et al. (78) defined consensus techniques as “group facilitation techniques, which explore the level of consensus among a group of experts by synthesising and clarifying expert opinion in order to derive a consensus opinion from a group with individual opinions combined into a refined aggregated opinion” (p.360).

Evidence suggests that the systematic method of combining scientific evidence and expert opinion is the most rigorous way of developing quality indicators (78, 102). A Delphi technique is widely used for developing quality indicators in healthcare. Alder et al. (139) defined a Delphi technique as “an exercise in group communication that brings together and

76 synthesizes the knowledge of group of geographically scattered participants who never meet”

(p.257).

3.1.1 Rationale for the study

Diabetes has been chosen as a condition of interest due to the high burden of co-existing chronic conditions in this group of patients (5, 24). Recent research found that that more than

90% of older adults with diabetes in Ontario had at least one comorbid condition, and the number of comorbid conditions increased with age (24). This study is focused on older adults because they are more likely than younger individuals to have comorbid chronic conditions that can be complex and difficult to manage (5, 41).

Given that diabetes mostly occurs in conjunction with other chronic conditions, it is important to evaluate the quality of care in diabetes patients with multiple chronic conditions

(23, 142). However, most existing research tends to focus on a specific condition. For example, there are several studies that aim to assess the quality of diabetes care against a background of other comorbid conditions; thus, diabetes is being considered as an index disease (28, 30, 31).

A growing body of evidence suggests that individuals with multiple chronic conditions, including those with diabetes, have different care needs as compared to those with single chronic conditions; therefore, processes of care addressing patients’ single complains are not ideal for those with multiple chronic conditions (32, 143-145). The literature suggests that adherence to disease-specific measures for patients with multiple chronic conditions may

77 lead to the unintended consequence of delivering inappropriate care that can do harm and/or is misaligned with the patient’s goals and preferences (70).

Four chronic conditions among diabetes patients, including hypertension, chronic ischemic heart disease, depression and osteoarthritis, were selected for the purpose of this study, since they have been identified as the most common conditions among people diagnosed with diabetes (23, 24). Recent research demonstrated that hypertension consistently appeared in the majority of the top ten condition clusters for older adults with 2 and 3 comorbid conditions (24). Other chronic conditions that appeared in many of the clusters for older diabetes patients with 2 and 3 comorbidities were arthritis, other cardiovascular conditions and anxiety (24).

These chronic conditions were grouped in five disease combinations that represented the most prevalent clusters of concurrent conditions across multimorbidity groupings shown in previous research (5, 24). Finally, the defined five disease combinations were categorized into three groups by comorbidity type (22), including diabetes-concordant, diabetes- discordant and both types of comorbid conditions (Table 1.1).

The results of the systematic review (Chapter 2) that aimed to identify existing evidence- based quality indicators for ambulatory care for older adults with five selected chronic conditions demonstrate that there are only a few published indicators for care for older adults with diabetes comorbid with concordant conditions, including hypertension and chronic ischemic heart disease. Conversely, there were no indicators identified for care for older

78 adults with diabetes with comorbid discordant conditions, including osteoarthritis and major depression. The methodology to measure quality of care for diabetes patients with multiple chronic conditions has been poorly developed.

While disease-specific indicators are important, they can be used but with limitations and caution on the part of the clinician to evaluate the appropriateness of care for people with multiple chronic conditions, since they may not be applicable for individuals with several multimorbidity issues, including treatment interactions and disease interactions (34, 46, 70).

It is necessary to balance the clinical benefits and risks of disease-specific quality standards in managing people with multiple chronic conditions (146, 147). There is a need for valid quality indicators that provide an objective comprehensive assessment of care for people with multiple chronic conditions.

A Delphi technique was used to develop a set of appropriate quality indicators for ambulatory care for older adults with five selected disease combinations, using a set of indicators identified after the 1st study as a starting point, including:

 Diabetes with concordant comorbid conditions:

o Diabetes with comorbid hypertension

o Diabetes with comorbid hypertension and chronic ischemic heart disease

 Diabetes with discordant comorbid conditions:

o Diabetes with comorbid osteoarthritis;

o Diabetes with comorbid osteoarthritis and major depression;

 Diabetes with both types of chronic conditions:

 Diabetes with comorbid osteoarthritis and hypertension.

3.1.2 The study objectives:

 To critically appraise and select the most appropriate set of quality indicators for

ambulatory care for older adults with five selected disease combinations that are

amenable to measurement using Ontario administrative data, by means of a

Delphi study;

 To prepare a summary of defined quality indicators by each selected disease

combination.

3.2 Methods

The Delphi technique was used for the purpose of this study. The Delphi method aims to gather consensus of opinion and choice about different types of quality indicators from a selected panel (148). It is a structured iterative process that uses repetitive administration of questionnaires to gather information (78, 139). The main stages in the Delphi methodology include: 1) identifying a research problem, selecting appropriate Expert panelists, developing the questionnaire, conducting anonymous Delphi rounds, providing feedback to the panelists between rounds, and summarizing the findings (78, 80).

3.2.1 Assembly of Delphi panel

Panelists were selected with the goal of having a heterogeneous group of experts to better reflect the variety of specialities that are actually involved in multimorbid patient treatment decisions. The panel consisted of geriatricians, primary care specialists, clinical pharmacists, and a senior methodologist from a quality measurement organization. Names of possible panelists were obtained from study team members. There was diverse geographical representation among the panel members, with representation from various provinces of

Canada, including Ontario, British Columbia and Quebec. A key component of the Delphi technique is the anonymity of the expert panel members. Thus, no panelist knew the identity of the other panel members. This allowed a true consensus that was reached among the panel members and eliminates some problems that may arise from peer influence among the panelists.

3.2.2 Questionnaire preparation

First, a systematic review was conducted to identify a set of evidence-based and valid quality indicators for assessing care for older adults in ambulatory care settings by each selected disease category, including diabetes, major depression, hypertension, chronic ischemic heart disease and osteoarthritis, as well as selected disease combinations (Chapter 2). The results

81 showed that most of the existing indicators have been developed for assessing care for single diseases.

These indicators were then sorted into those potentially measurable with the administrative data, and those that required other sources of data. The latter group was excluded from the study (Chapter 2). A refined list of potential indicators was then included in the first-round questionnaire for each selected disease combination. In particular, the initial set of candidate indicators was compiled by combining the identified indicators for care for the relevant single conditions for each disease combination, accordingly. For example, the set of candidate indicators for care for older adults with diabetes with comorbid hypertension was compiled by combining the identified indicators for diabetes care with those for hypertension care. These indicators were divided into ambulatory process of care indicators and outcome indicators.

The initial questionnaire with candidate indicators was pre-tested with three physicians (two

GPs and a geriatrician) to anticipate the average completion time, and for clarity. As a result of the pre-test, modifications were made to the questionnaire with respect to definition of several quality indicators.

The initial questionnaire for assessing ambulatory care for older adults with five selected disease combinations included:

 11 candidate quality indicators for assessing care for older adults with diabetes with

comorbid hypertension;

 11 candidate quality indicators for assessing care for older adults with diabetes with

comorbid hypertension and chronic ischemic heart disease;

 12 candidate quality indicators for assessing care for older adults with diabetes with

comorbid osteoarthritis;

 16 candidate quality indicators for assessing care for older adults with diabetes with

comorbid osteoarthritis and major depression;

 16 candidate quality indicators for assessing care for older adults with diabetes with

comorbid hypertension and osteoarthritis.

Once the list of candidate panelists was formed, each person was contacted via e-mail

(Appendix 4). We sent to all candidates the invitations to participate with a description of the study, its objectives, the number of Delphi rounds to be included, the promise of anonymity, as well as benefits from participation, and a confirmation of the panelist’s acceptance

(Appendix 5)

3.2.3 Delphi Round I

The first round was performed from October 1 to December 11, 2015. The panelists, who confirmed his/her participation in the study, received the first-round questionnaire (Appendix

6) by electronic mail along with the appraisal tool criteria and instructions for rating

(Appendix 7). The panelists were also provided with detailed information related to the

83 candidate indicators selected during the systematic review, including numerator/denominator, inclusion/exclusion criteria, data source, and rationale/supporting evidence (Appendix 8).

Since we were not able to account for severity of conditions due to limitation of the administrative data, we asked panelists to consider all diseases presented in selected combinations as if they are at moderate severity, such as controlled diabetes, moderate depression, osteoarthritis with moderate pain and moderate functional disability. The panel was asked to rate each potential indicator, on a five-point scale, according to the appraisal criteria as adapted from the methodology for eliciting expert opinion using the Delphi technique and used in prior research (149, 150).

Criteria for rating process indicators included (Appendix 7):

 Meaningfulness: This is a meaningful measure of the quality of care we deliver to the

patient aged 65 and older diagnosed with this disease combination;

 Potential for improvements in clinical practices: It is possible to improve the care that

impacts this indicator in patients aged 65 and older diagnosed with this disease

combination;

 Overall value of inclusion: Overall Assessment – considering your ratings on all

dimensions, rate this process measure overall for inclusion in the context of this

disease combination.

Criteria for rating outcome indicators included (Appendix 7):

 Importance: This outcome is an important indicator of the quality of care of the

patient aged 65 and older diagnosed with this disease combination;

 Modifiability: This outcome is potentially modifiable by improvements in patient’s

care;

 Overall value of inclusion: Overall assessment – considering your ratings on all

dimensions, rate this outcome measure overall for inclusion in the context of this

disease combination.

A score of one indicated the lowest rating and a score of five indicated the highest possible rating. For the criterion of “overall value of inclusion” of a quality indicator, the dimensions of the scale were:

 Do not include (rating=1);

 Little reason to include (rating=2);

 Could include (rating=3);

 Should include (rating=4);

 Must include (rating=5).

Moreover, panel members were also invited to comment on each item using a dedicated

"comment box", and/or to add items considered as important with respect to each particular disease combination selected for the purpose of this study. Then, the responses were reviewed and summarized, including comments and suggestions added to the “comment box” of the questionnaire. The panel members were encouraged to draw upon their

85 experience as well as to use any research or other available resources to select an appropriate set of indicators for each particular disease combination.

Consensus level was defined based on the RAND/UCLA Appropriateness Methodology agreement definition for panel size of 15 – at least 11 of 15 panelists rated an indicator a 4 or

5, or 1 or 2 (80). The extent of disagreement in ratings was defined by the panel’s mean absolute deviation from the median (MADM) –the average distance of the panelists’ ratings from the panel’s median rating (151). This was categorised into low and high variation according to the observed MADM scores (low ≤ 1.03 and high > 1.03) (151).

Thus, consensus for Round I was defined based on three selection criteria: 1) panel median score of 4 or 5, or 1 or 2; 2) having at least 11 of 15 panelists (73%) rated a given indicator a

4 or 5 – “should include or must include”, or 1 or 2 – “do not include or little reason to include” (80), 3) mean absolute deviation from the median (MADM) less or equal to 1.03

(151).

The main statistics used in Delphi studies are measures of central tendency, including means, median, or mode, and level of dispersion for presenting information concerning the judgments of respondents (152). The evidence suggests that the use of median score is preferred, based on Likert-type scale and considering the skewed expectation of panelists’ responses (152).

The panelists’ comments and suggestions were analyzed to modify indicators and/or create a list of new quality indicators for each particular disease combination (Appendix 9). After

Round I, several indicators were modified accordingly, and a list of new quality indicators was included in the second-round questionnaire.

3.2.4 Delphi Round II

The second round was performed from December 14, 2015 to February 29, 2016. Quality indicators that reached consensus in Round I were not represented in Round II (153). The remaining indicators were included in the Round II of the study together with the modified or/and new indicators suggested by the panelists. Only indicators that were amenable for measuring using Ontario administrative data were included in the second-round questionnaire.

All panelists who had participated in Round I were sent an email (Appendix 10) with the second-round questionnaire along with the results of the first round including median panel rating for each indicator, their individual rating from the first round, the frequency distribution of all panelists ratings (Appendix 11), as well as comments (80). The panelists were asked to re-score each quality indicator using the same criteria for rating based on their own opinion, the panel responses obtained during the first round, and comments or suggestion of other panelists.

To be included in the final list of quality indicators, the items were selected by two levels of consensus for each particular disease combination (154):

 High consensus: a minimum of 11 of 15 panelists (73%) rated a given indicator a 4 or

5 – “should include or must include”, or 1 or 2 - “do not include or little reason to

include”, and the mean absolute deviation from the median was less or equal to 1.03;

 Moderate consensus: 9 or 10 of 15 panelists (60-67%) rated a given indicator a 4 or 5

– “should include or must include”, or 1 or 2 - “do not include or little reason to

include”, and the mean absolute deviation from the median was less or equal to 1.03.

Quality indicators that did not meet the criteria for high or moderate consensus were not included in the final list. The final list of indicators was categorized into groups according to the categories of the Conceptual Model of Performance Measurement for People with

Multiple Chronic Conditions (70), including screening, on-going monitoring, pharmacological treatment, patient safety, and healthcare utilization/clinical effectiveness.

Evidence suggests that there are no strict guidelines on the optimal number of Delphi rounds

(139, 155). In the literature the number of rounds varies between two and four (139). This study was conducted in two rounds for several reasons: 1) there were no new indicators suggested by the panelists after the second round, 2) there were no comments or suggestions to modify any indicator after the second round, 3) too many rounds may lead to fatigue among panelists, and 4) difficult to maintain high response rate. Financial incentives were offered for participating in both rounds of the study. For each round, three reminders were sent to non-respondents via email.

Ethics approval for this study was obtained from the University of Toronto Research Ethics

Board.

3.3 Results

The characteristics of the panelists who participated in the study/non-respondents are presented in Table 1. They were selected because they had a well-known recognition in their field of expertise. Of 30 panelists contacted to participate in the Delphi study, 15 (50%) accepted: six geriatricians, four primary care specialists, two general internists, two clinical pharmacists, and a clinician and senior methodologist from a quality improvement organization. All participants completed both rounds of the Delphi survey.

Table 3.1 Main characteristics of the Expert panel

Characteristics Expert panel Non-respondents (n=15) (n=15) Sex, n (%) Male 9 (60%) (65%) Female 6 (40%) (35%) Specialty, n (%) Geriatrics 6 (40%) 6 (40%) Primary care 4 (27%) 8 (53%) General internal medicine 2 (13%) 0 Clinical pharmacy 2 (13%) 0 Senior methodologist from a quality 1 (7%) 1 (6%) improvement organization Provinces of Canada, n (%) Ontario 11 (74%) 11 (74%) British Columbia 2 (13%) 4 (26%) Quebec 2 (13%) 0

3.3.1 Quality indicators for primary care for older adults with diabetes

with comorbid hypertension

3.3.1.1 Delphi Round I

Of the 11 potential quality indicators for ambulatory care for older adults with diabetes with

comorbid hypertension, 3 process and 1 outcome indicators reached the first level of

consensus (at least 11 of 15 panelists (73%) rated an indicator a 4 or 5, and MADM was less

or equal to 1.03). Four indicators did not reach the consensus: three indicators because of the

median rating was 3, and four indicators because less than 11 of 15 panelists (73%) of

panelists rated it as 4. The levels of consensus and median values for each indicator for

Round I are summarized in Table 2.

Table 3.2 Round I – Indicators for care for older adults with diabetes with comorbid hypertension

Potential for Overall value of Meaningfulness improvements in inclusion Process Indicator clinical practices Consensus Median Median Median (%) MADM (min; max) (min; max) (min; max) *HbA1c testing Consensus to 4 (2; 5) 4 (2; 5) 4 (3: 5) 0.36 80% every 6 month include **LDL-cholesterol No consensus 3 (1: 5) 3 (1; 5) 3 (1; 5) 1.11 testing once per year Eye examination Consensus to 5 (1; 5) 4 (1; 5) 5 (1;5) 0.64 86% every 1-2 years include Microalbumin 4 (2; 5) 4 (1; 5) 4 (2; 5) 0.86 53% No consensus

90 testing once per year Statin therapy No consensus 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.22

Use of ***ACE Consensus to 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.86 73% inhibitors or ARBs include Antiplatelet therapy No consensus 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.00

Overall value of Importance Modifiability Outcome inclusion Consensus Median (%) Indicator Median Median (min; MADM (min; max) (min; max) max) Hospital admission rate for diabetes 4 (3; 5) 3 (1; 5) 4 (1; 5) 0.67 53% No consensus long-term complications Hospital admission rate for diabetes 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.73 67% No consensus short-term complications Lower-extremity Consensus to 4.5 (3; 5) 4 (1; 5) 4 (1; 5) 0.72 73% amputation rate include Cardiovascular 4 (3; 5) 4 (2; 5) 4 (2; 5) 0.83 67% No consensus mortality rate *HbA1c testing=glycated hemoglobin testing *LDL-cholesterol=low-density lipoprotein cholesterol **ACE inhibitors= angiotensin converting enzyme (ACE) inhibitors; ARBs= angiotensin receptor blockers

3.3.1.2 Delphi Round II

Adjustments were made prior to the Delphi Round II, taking into account the suggestions and

comments of the panelists. The indicators that reached a first-round consensus were not

included in the second-round questionnaire. The remaining indicators were included in the

Round II survey together with new indicators that were suggested by the panel and could be

measured using Ontario administrative data.

The suggested new process indicators included serum creatinine test (with eGFR), use of

hypoglycemic drugs, baseline electrocardiography, and MRI of head/heart (to rule out

stroke/ischemic heart disease for prescription of statins or antiplatelet agents). The suggested

new outcome indicators included all-cause mortality, ocular complications due to diabetes,

and urinary/skin/soft tissue infections.

In Round II, 8 process and 7 outcome indicators were evaluated using the same assessment

criteria as in Round I. Levels of consensus and median values for each indicator for Round I

are summarized in Table 3. Of the 15 indicators, 3 process and 2 outcome indicators reached

the high level of consensus (at least 11 of 15 panelists (73%) rated an indicator a 4 or 5, or 1

or 2, and MADM was less or equal to 1.03), while 1 process and 2 outcome indicators

reached the moderate level of consensus (9 or 10 of 15 panelists (60-67%) rated an indicator

as 4 or 5, and MADM was less or equal to 1.03).

Table 3.3 Round II – Indicators for care for older adults with diabetes with comorbid hypertension

Potential for Overall value of Meaningfulness improvements in Process inclusion clinical practices Consensus Indicator Median Median Median MADM (%) (min; max) (min; max) (min; max) *LDL- cholesterol 3 (1: 5) 4 (1; 5) 4 (1; 5) 1.08 53% No consensus testing once per year Microalbumin Consensus to testing once per 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.73 60% include year Statin therapy 3 (2; 5) 4 (2; 5) 3 (2; 5) 1.12 No consensus

Antiplatelet therapy 3 (1; 5) 3 (1; 5) 2 (1; 5) 1.08 53% No consensus

New indicators

Serum creatinine testing (with Consensus to 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.47 87% eGFR) include

Use of oral hypoglycemic Consensus to 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.60 73% drugs include

Baseline electrocardiogra 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.27 No consensus phy

*MRI of Consensus to 1 (1, 4) 1 (1, 4) 1 (1, 4) 0.53 87% head/heart reject

Overall value of Importance Modifiability Outcome inclusion Consensus Median (%) Indicator Median Median (min; max) MADM (min; max) (min; max)

Hospital admission rate Consensus to for diabetes 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.53 67% include long-term complications Hospital admission rate Consensus to for diabetes 4 (1; 5) 4 (3; 5) 4 (1; 5) 0.73 73% include short-term complications Consensus to Cardiovascular 4 (3; 5) 4 (3; 5) 4 (3; 5) 0.60 67% include mortality rate

New indicators

All-cause mortality 4 (2; 5) 3 (1; 5) 3 (1; 5) 0.86 No consensus

Ocular complications Consensus to 4 (3; 5) 4 (3; 5) 4 (3; 5) 0.47 73% due to diabetes include

Urinary/skin/sof 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.13 No consensus t tissue

93 infections

*LDL-cholesterol=low-density lipoprotein cholesterol **MRI= magnetic resonance imaging

Thus, 12 indicators in total reached second-round consensus, including 9 high-consensus and

3 moderate-consensus indicators. In particular, 11 quality indicators were rated as “should

include” or “must include” in the final list of indicators, and 1 indicator - MRI of head/heart,

was rated as “do not include” and was not included in the final list of quality indicators. The

final list of selected process/outcome indicators was categorized into groups according to the

categories of the Conceptual Model of Performance Measurement for People with Multiple

Chronic Conditions (70) (Table 4).

Table 3.4 High and moderate consensus indicators for care for older adults with diabetes with comorbid hypertension

Process indicators

Median Category Indicator Consensus level score

*HbA testing every 6 months 1c 4 High consensus

Eye examination every 1-2 years 5 On-going High consensus monitoring Microalbumin testing once per 4 Moderate consensus year Serum creatinine test (with 4 High consensus

eGFR) Use of hypoglycemic drugs 4 High consensus Pharmacological Treatment Use of **ACE inhibitors or ARBs 4 High consensus

Outcome Indicators

Median Category Indicator Consensus level score Hospital admission rate for 4 diabetes long-term complications Moderate consensus

Hospital admission rate for 4 High consensus diabetes short-term complications Healthcare utilization/ High consensus Lower-extremity amputation rate 4 Clinical effectiveness Cardiovascular mortality rate 4 Moderate consensus

Ocular complications due to 4 High consensus diabetes *HbA1c testing=glycated hemoglobin testing ** ACE inhibitors = angiotensin converting enzyme (ACE) inhibitors; ARBs= angiotensin receptor blockers

3.3.2 Selection of quality indicators for older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

3.3.2.1 Delphi Round I

Of the 11 potential quality indicators for ambulatory care for older adults with diabetes with

comorbid hypertension, 4 process and 1 outcome indicators reached the first level of

consensus (at least 11 of 15 panelists (73%) rated an indicator a 4 or 5, and MADM was less

or equal to 1.03). Three process indicators did not reach the consensus: 3 because less than

11 of 15 panelists (73%) of rated it as 4 or MADM was more than 1.03. The levels of

consensus and median values for each indicator for Round I are summarized in Table 5.

Table 3.5 Round I – Quality indicators for care for older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

Potential for Overall value of Meaningfulness improvements in Process inclusion clinical practices Consensus Indicators Median Median Median MADM (%) (min; max) (min; max) (min; max) *HbA1c testing Consensus 4 (3; 5) 4 (2; 5) 4 (3; 5) 0.36 80% every 6 month to include **LDL- cholesterol No 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.14 60% testing once consensus per year Eye Consensus examination 5 (1; 5) 5 (1; 5) 5 (1; 5) 0.72 80% to include every 1-2 years Microalbumin No testing once 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.79 60% consensus per year Statin therapy Consensus 4 (1.5; 5) 4 (1.5; 5) 4 (1; 5) 0.93 80% to include Use of ***ACE inhibitors or Consensus 5 (2; 5) 5 (2; 5) 5 (2; 5) 0.50 87% ARBs to include

Beta-blockers No therapy 3.5 (2; 5) 4 (2; 5) 4 (2; 5) 1.00 60% consensus

Modifiability Overall value of Importance Outcome inclusion Consensus Indicators Median Median Median (%) MADM (min; max) (min; max) (min; max) Hospital admission rate No for diabetes 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.79 67% consens long-term us complications Hospital admission rate No for diabetes 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.00 60% consens short-term us complications

Lower- No extremity 4 (1; 5) 3.5 (1; 5) 4 (1; 5) 0.72 67% consens amputation rate us Consens Cardiovascular 4 (3; 5) 4 (2; 5) 4 (2; 5) 0.57 80% us to mortality rate include *HbA1c testing=glycated hemoglobin testing **LDL-cholesterol=low-density lipoprotein cholesterol ** ACE inhibitors = angiotensin converting enzyme inhibitors; ARBs= angiotensin receptor blockers

3.3.2.2 Delphi Round II

Adjustments were made prior to the Delphi Round II, taking into account the suggestions and

comments of the panelists. The indicators that reached a first-round consensus were not

included in the second-round questionnaire. The remaining indicators were included in the

Round II survey together with new indicators that were suggested by the panel. In this

second round, we included ratings for all indicators suggested by the panel that can be

measured using Ontario health administrative data.

In particular, the suggested new process indicator was antiplatelet therapy, while new

outcome indicators included all-cause mortality, hospital admissions for heart failure, ED

visits for diabetes-short-term complications, and bariatric surgery.

In Round II, 4 process and 7 outcome indicators were evaluated using the same assessment

criteria as in Round I. Levels of consensus and median values for each indicator for Round I

are summarized in Table 6. Of the 11 indicators, 1 process and 1 outcome indicators reached

the high level of consensus (at least 11 of 15 panelists (73%) rated an indicator as 4 or 5, and

MADM was less or equal to 1.03), while 2 process and 5 outcome indicators reached the

moderate level of consensus (9 or 10 of 15 panelists (60-67%) rated an indicator as 4 or 5,

and 1 or 2, and MADM was less or equal to 1.03).

Table 3.6 Round II – Quality indicators for care for older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

Potential for Overall value of Meaningfulness improvements in Process inclusion clinical practices Consensus Indicators Median Median Median MADM (%) (min; max) (min; max) (min; max) *LDL- cholesterol No 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.94 53% testing once per consensus year Microalbumin Consensus to testing once per 4 (1.; 5) 4 (1.; 5) 4 (1.; 5) 0.60 73% include year Beta-blocker No therapy 4 (1; 5) 4 (2; 5) 4 (1; 5) 1.00 53% consensus

New indicators

Consensus Antiplatelet 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.80 67% to include therapy

Modifiability Overall value of Importance Outcome inclusion Consensus Indicators Median Median Median (%) MADM (min; max) (min; max) (min; max) Hospital Consensus admission rate to include for diabetes 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.67 67% long-term complications Hospital Consensus admission rate to include for diabetes 4 (3; 5) 4 (3; 5) 4 (3; 5) 0.53 73% short-term complications Lower-extremity Consensus 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.67 67% amputation rate to include

New indicators

Hospital admission for Consensus 4 (3; 5) 4 (2; 5) 4 (2; 5) 0.71 67% heart failure to include

All-cause No mortality rate 4 (2; 5) 3 (1; 5) 3 (1; 5) 0.93 consensus

ED visits for diabetes short- Consensus 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.87 60% term to include complications

Bariatric surgery Consensus rate 2 (1; 5) 2 (1; 5) 2 (1; 5) 1.27 67% to reject

*LDL-cholesterol=low-density lipoprotein cholesterol

Thus, consensus was reached for 13 indicators in total, including 5 high-consensus and 8

moderate-consensus indicators. In particular, 11 indicators were rated as “should include” or

“must include” in the final list of indicators, and 1 indicator - bariatric surgery was rated as

“little reason to include” and was not included in the final list of quality indicators. The final

list of selected process/outcome indicators was categorized into groups according to the

categories of the Conceptual Model of Performance Measurement for People with Multiple

Chronic Conditions (70) (Table 7).

Table 3.7 High and moderate consensus indicators for care for older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

Process indicators

Median Category Indicator Consensus level score *HbA testing every 6 months 1c 4 High consensus

On-going Eye examination every 1-2 years 5 High consensus monitoring Microalbumin testing once per 4 Moderate consensus year Antiplatelet therapy 4 Moderate consensus

**Use of ACE inhibitors or Pharmacological 4 High consensus ARBs therapy Treatment Statin therapy 4 High consensus

Outcome Indicators

Median Category Indicator Consensus level score Hospital admission rate for 4 diabetes long-term complications Moderate consensus Hospital admission rate for 4 diabetes short-term complications High consensus Lower-extremity amputation rate 4 Moderate consensus Healthcare Cardiovascular mortality rate utilization/ 4 High consensus

Clinical effectiveness Ocular complications due to 4 High consensus diabetes Hospital admission for heart 4 Moderate consensus failure ED visits for diabetes short-term 4 Moderate consensus complications *HbA1c testing=glycated hemoglobin testing **ACE inhibitors = angiotensin converting enzyme inhibitors

3.3.3 Selection of quality indicators for older adults with diabetes with comorbid osteoarthritis

100

3.3.3.1 Delphi Round I

Of the 12 potential quality indicators for ambulatory care for older adults with diabetes with osteoarthritis (with moderate pain and moderate functional disability), only 1 process and 1 outcome indicator reached the first level of consensus (at least 11 of 15 panelists (73%) rated an indicator a 4 or 5, and MADM was less or equal to 1.03). The remaining 10 indicators did not reach the consensus: 5 indicators because of the median rating was 3, and 5 because less than 11 of 15 panelists (73%) rated it a 4 or 5, or MADM was more than 1.03. The levels of consensus and median values for each indicator for Round I are summarized in Table 8.

Table 3.8 Round I – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis

**LDL- cholesterol 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.13 No consensus testing once per year Eye examination Consensus to 5 (1; 5) 4 (1; 5) 4 (1; 5) 0.72 87% every 1-2 years include Microalbumin testing once per 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.67 60% No consensus year Acetaminophen as first-line 3 (2; 5) 4 (2; 5) 3 (2; 5) 1.07 No consensus therapy

Non-selective ***NSAIDs 3 (1; 5) 3 (1; 5) 3 (1; 4) 0.93 No consensus therapy Cox-selective 3 (1; 5) 3 (1; 4) 2 (1; 4) 1.00 No consensus

101

NSAIDs therapy 53%

Non-selective NSAIDs in combination with 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.93 No consensus misoprostol or proton pump inhibitors Overall value of Importance Modifiability inclusion Consensus Median Outcome Median Median (min; max) MADM (%) Indicators (min; max) (min; max)

Hospital admission rate for diabetes long- 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.93 No consensus term complications Hospital admission rate for diabetes short- 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.00 60% No consensus term complications

Lower-extremity Consensus to amputation rate 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.93 73% include

Cardiovascular mortality rate 4 (2; 5) 4 (1; 5) 4 (1; 5) 0.87 53% No consensus

*HbA1c testing=glycated hemoglobin testing **LDL-cholesterol=low-density lipoprotein cholesterol ***NSAIDs=non-steroidal anti-inflammatory drugs

3.3.3.2 Delphi Round II

Adjustments were made prior to the Delphi Round II, taking into account the suggestions and comments of the panelists. The indicators that reached a first-round consensus were not included in the second-round questionnaire. The remaining indicators were included in the

102

Round II survey together with new indicators that were suggested by the panel that could be

measured using health administrative data.

In particular, the suggested new process indicators were statin therapy, use of opioids, ACE

(angiotensin converting enzyme) inhibitors therapy, referral for home care, while new

outcome indicators included all-cause mortality, joint replacement therapy, and ED/hospital

admissions for fall. Two indicators were modified as “negative indicator” based on panelists’

comments and suggestions: use of non-selective NSAIDs (non-steroidal anti-inflammatory

drugs), and use of cox-selective NSAIDs.

In Round II, 12 process and 6 outcome indicators were evaluated using the same assessment

criteria as in Round I. Levels of consensus and median values for each indicator for Round I

are summarized in Table 9. Of the 18 indicators, only 1 outcome indicators reached the high

level of consensus (at least 11 of 15 panelists (73%) rated an indicator as 4 or 5, and MADM

was less or equal to 1.03), while 3 process and 1 outcome indicators reached the moderate

level of consensus (9 or 10 of 15 panelists (60-67%) rated an indicator as 4 or 5, and MADM

was less or equal to 1.03).

Table 3.9 Round II – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis

Potential for Meaningful Overall value of improvements in ness inclusion Process Indicators clinical practices Consensus Median Median Median (%) MADM (min; max) (min; max) (min; max) *HbA1c testing every 3 (1; 5) 4 (1; 5) 4 (1; 5) 0.87 67% Consensus to

103

6 month include

**LDL-cholesterol No 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.06 testing once per year consensus Microalbumin testing Consensus to 4 (1; 5) 4 (1; 5) 4 (2; 5) 0.73 60% once per year include Use of acetaminophen as No 3 (2; 5) 4 (2; 5) 3 (2; 5) 0.87 first-line therapy consensus

Non-selective ***NSAIDs in No combination with 3 (1; 5) 3 (1; 4) 3 (1; 4) 0.93 consensus misoprostol or proton pump inhibitors Modified indicators

Non-selective NSAID therapy Consensus to 4 (1; 5) 3 (1; 5) 4 (1; 4) 0.86 60% “negative include indicator” Cox-selective NSAID therapy No “negative 3 (1; 5) 3 (1; 4) 3 (1; 4) 0.93 53% consensus indicator” New indicators

Use of topical No 3 (1; 4) 3 (1; 5) 3 (1; 4) 0.87 NSAIDs consensus Statin therapy No 2 (1, 5) 3 (1, 5) 2 (1, 5) 0.93 53% consensus Use of opioids No 4 (1; 5) 3 (1; 5) 3 (1; 5) 1.00 consensus ****Use of ACEs No inhibitors 4 (1; 5) 3 (1; 5) 3 (1; 5) 1.20 consensus

Referral for home No care 3 (1; 4) 3 (1; 5) 3 (1; 4) 1.08 consensus

Overall value of

Importance Modifiability inclusion Consensus

Outcome Median Median Median (%) Indicators MADM (min; max) (min; max) (min; max) Hospital admission rate for No 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.71 diabetes long- consensus term

104 complications Hospital admission rate for Consensus diabetes short- 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.71 73% to include term complications Cardiovascular Consensus mortality rate 4 (2; 5) 4 (1; 5) 4 (1; 5) 0.67 60% to include

All-cause No mortality 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.20 47% consensus

Joint replacement No therapy 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.09 consensus

ED visits/hospital No admissions for 3 (2; 4) 3 (2; 4) 3 (2; 4) 0.73 consensus fall *HbA1c testing=glycated hemoglobin testing **LDL-cholesterol=low-density lipoprotein cholesterol ***NSAIDs=non-steroidal anti-inflammatory drugs **** ACE inhibitors = angiotensin converting enzyme inhibitors

Thus, consensus was reached for 7 indicators in total, including 3 high-consensus and 4

moderate-consensus indicators. The “negative” indicator “non-selective NSAIDs therapy”

was included in the final list as an indicator of inappropriate care or poor performance. The

final list of selected process/outcome indicators was categorized into groups according to the

categories of the Conceptual Model of Performance Measurement for People with Multiple

Chronic Conditions (70) (Table 10).

Table 3.10 High and moderate consensus indicators for care for older adults with diabetes with comorbid osteoarthritis

Process indicators

105

Median Category Indicator Consensus level score *HbA testing every 6 months 1c 4 Moderate consensus

On-going Eye examination every 1-2 years 5 High consensus monitoring Microalbumin testing once per 4 Moderate consensus year **Non-selective NSAID therapy Patient safety “negative indicator” 4 Moderate consensus

Outcome Indicators

Median Category Indicator score Consensus level

Hospital admission rate for 4 High consensus diabetes short-term complications Healthcare utilization/ Lower-extremity amputation rate 4 High consensus Clinical effectiveness Cardiovascular mortality rate 4 Moderate consensus

*HbA1c testing=glycated hemoglobin testing **NSAIDs=non-steroidal anti-inflammatory drugs

3.3.4 Quality indicators for older adults with diabetes with comorbid osteoarthritis and depression

3.3.4.1 Delphi Round I

Of the 16 potential quality indicators for ambulatory care for older adults with diabetes with osteoarthritis and depression, only 2 process indicators reached the first level of consensus

106

(at least 11 of 15 panelists (73%) rated an indicator a 4 or 5, and MADM was less or equal to

1.03). The panelists did not reach consensus on the remaining 14 indicators: 6 indicators

because of the median rating was 3, and 8 because less than 11 of 15 panelists (73%) rated it

a 4 or 5, and 1 or 2, and MADM was more than 1.03. Levels of consensus and median values

for each indicator for Round I are summarized in Table 11.

Table 3.11 Round I – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and depression

Potential for Overall value of Meaningfulness improvements in inclusion Process Indicators clinical practices Consensus Median Median Median (%) MADM (min; max) (min; max) (min; max) *HbA1c testing No 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.86 67% every 6 month consensus **LDL- cholesterol testing once per No 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.13 year consensus

Eye examination Consensus to every 1-2 years 5 (1; 5) 4 (1; 5) 4 (1; 5) 0.72 87% include

Microalbumin testing once per No 3 (2; 5) 4 (2; 5) 4 (2; 5) 1.00 53% year consensus

Use of No acetaminophen as 3 (1; 5) 3 (2; 5) 3 (2; 5) 1.00 consensus first-line therapy ***Non-selective No NSAIDs therapy 3 (1; 5) 2 (1; 5) 2 (1; 4) 0.86 67% consensus

Cox-selective No NSAIDs therapy 2 (1; 5) 2 (1; 4) 2 (1; 4) 0.80 60% consensus

Non-selective No 2 (1; 5) 2 (1; 5) 2 (1; 4) 0.86 60% NSAIDs therapy in consensus

107 combination with misoprostol or proton pump inhibitors Use of tri/tetracyclic antidepressant, No 3 (1; 5) 2 (1; 5) 2 (1; 5) 1.12 53% benzodiazepine, or consensus monoamine oxidase inhibitors Interval between ****SSRIs and Consensus to 4 (2; 4) 4 (2; 5) 4 (2; 5) 0.72 87% monoamine oxidase include inhibitors therapy At least 3 months antidepressant No 3 (1; 5) 3 (1; 5) 3 (1; 4) 0.93 treatment consensus (acute phase) At least 6 months antidepressant No 3 (1; 4) 3 (1; 5) 3 (1; 5) 1.13 treatment consensus (continuation phase) Overall value of Importance Modifiability Outcome inclusion Consensus Indicators Median Median Median MADM (%) (min; max) (min; max) (min; max) Hospital admission rate for diabetes No long-term 3 (1; 5) 4 (1; 5) 3 (1; 5) 0.93 consensus complications

Hospital admission rate for diabetes short-term No 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.86 60% complications consensus (hypo- or hyperglycemia) Lower-extremity No amputation rate 4 (1; 5) 3.5 (1; 5) 3 (1; 5) 1.07 consensus

Cardiovascular No mortality rate 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.67 67% consensus

*HbA1c testing=glycated hemoglobin testing **LDL-cholesterol=low-density lipoprotein cholesterol ***NSAIDs=non-steroidal anti-inflammatory drugs ****SSRIs= selective serotonin re-uptake inhibitors

108

3.3.4.2 Delphi Round II

Round II survey together with new indicators that were suggested by the panel. In this second round, we included ratings for all indicators suggested by the panel that could be measured using health administrative data.

In particular, the suggested new process indicators were use of selective serotonin reuptake inhibitor(SSRI)/serotonin norepinephrine reuptake Inhibitor (SNRI), use of gaba receptor agonists, use of topical NSAIDs, use of opioids, referral for home care, while new outcome indicators included joint replacement therapy, ED/hospital admissions for fall, hospital admissions for depression, and all-cause ED visits.

Three indicators were modified as “negative indicator” based on panelists’ comments and suggestions as indicators of inappropriate use of poor performance: use of non-selective

NSAIDs, use of cox-selective NSAIDs, as well as use of tetracyclic antidepressant, benzodiazepines, gaba receptor agonists, or monoamine oxidase inhibitors. The indicator

“use of tricyclic antidepressants” was excluded from the group of “negative” indicators and presented as a positive indicator in the Round II based on panelists’ suggestions.

109

In Round II, 15 process and 8 outcome indicators were evaluated using the same assessment

criteria as in Round I. Levels of consensus and median values for each indicator for Round I

are summarized in Table 12. Of the 23 indicators, only 1 outcome indicator reached a high

level of consensus (at least 11 of 15 panelists (73%) rated an indicator a 4, or 1 and 2, and

MADM was less or equal to 1.03), while 6 process and 1 outcome indicator reached the

moderate level of consensus (9 or 10 of 15 panelists (60-67%) rated an indicator a 4, and

MADM was less or equal to 1.03).

Table 3.12 Round II – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and depression

**LDL- cholesterol testing once per 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.93 No consensus year

Microalbumin testing once per Consensus to 4 (2; 5) 4 (2; 5) 4 (1; 5) 0.73 60% year include

Acetaminophen as first-line 3 (1; 5) 3 (2; 5) 3 (2; 5) 1.00 No consensus therapy

At least 3 months antidepressant 3 (1; 5) 3 (1; 5) 3 (1; 4) 0.80 No consensus treatment (acute phase) At least 6 3 (1; 4) 3 (1; 5) 3 (1; 5) 1.13 No consensus

110 months antidepressant treatment (continuation phase) ***Non-selective NSAID in combination with 2 (1; 5) 3 (1; 5) 3 (1; 4) 0.93 No consensus misoprostol or proton pump inhibitors Modified indicators

Non-selective NSAID therapy Consensus to “negative 4 (1; 5) 3 (1; 5) 4 (1; 4) 0.86 67% include indicator”

Cox-selective NSAID therapy “negative 3 (1; 5) 3 (1; 4) 3 (1; 4) 0.87 No consensus indicator”

Use of tetracyclic antidepressants, benzodiazepines, gaba receptor agonists, or Consensus to 3 (1; 5) 4(1; 5) 4 (1; 5) 0.87 60% monoamine include oxidase inhibitors “negative indicator”

New indicators

****Use of SSRI Consensus to or SNRI 2 (1; 5) 2 (1; 5) 2 (1; 5) 1.01 60% reject

Use of tricyclic antidepressants 3 (1; 5) 4 (1; 4) 4 (1; 4) 0.79 53% No consensus

Use of topical 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.08 No consensus NSAIDs

Use of opioids 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.13 53% No consensus

Referral for home 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.89 No consensus care

111

Importance Modifiability Outcome Overall value of Consensus Indicators inclusion Median Median Median MAD (%) (min; max) (min; max) (min; max) M Hospital admission rate for diabetes long- Consensus to 3 (1; 5) 4 (1; 5) 4 (2; 5) 0.67 67% term include complications

Hospital admission rate for Consensus to diabetes short- 4 (2; 5) 4 (1; 5) 4 (2; 5) 0.67 67% include term complications Lower-extremity amputation rate 4 (1; 5) 3 (1; 5) 3 (1; 5) 0.87 No consensus

Cardiovascular Consensus to mortality rate 4 (1; 4) 4 (1; 5) 4 (1; 5) 0.40 73% include

New indicators

ED visits/hospital admissions for 3 (2; 5) 4 (1; 5) 4 (2; 5) 0.67 53% No consensus falls

Hospital admission for 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.20 No consensus depression

All-cause ED visits 4 (1; 5) 3 (1; 5) 3 (1; 5) 0.93 No consensus

Joint replacement rate 3 (1; 5) 3 (1; 4) 3 (1; 5) 0.10 No consensus

*HbA1c testing=glycated hemoglobin testing **LDL-cholesterol=low-density lipoprotein cholesterol ***NSAIDs=non-steroidal anti-inflammatory drugs ****SSRIs= selective serotonin re-uptake inhibitors; SNRIs= serotonin–norepinephrine reuptake inhibitors

Thus, consensus was reached for 8 indicators in total, including 3 high-consensus and 6

moderate-consensus indicators. The “negative” indicators including use of non-selective

112

NSAIDs therapy, use of tetracyclic antidepressants, gaba receptor agonists, benzodiazepines, use of monoamine oxidase inhibitors were included as an indicator of inappropriate care or poor performance. One process indicator – use of selective serotonin re-uptake inhibitors

(SSRIs)/ serotonin–norepinephrine reuptake inhibitors (SNRIs), was rated with a median score of 2 and was not included in the final list of quality indicators. The final list of selected process/outcome indicators was categorized into groups according to the categories of the

Conceptual Model of Performance Measurement for People with Multiple Chronic

Conditions (70) (Table 13).

Table 3.13 High and moderate consensus indicators for care for older adults with diabetes with comorbid osteoarthritis and depression

Process indicators

Median Category Indicator Consensus level score

*HbA testing every 6 months 1c 4 Moderate consensus

On-going Eye examination every 1-2 years 5 High consensus monitoring Microalbumin testing once per 4 Moderate consensus year Interval between SSRIs and 4 High consensus monoamine oxidase therapy **Non-selective NSAIDs therapy - 4 Moderate consensus “negative indicator”

Use of tetracyclic antidepressants, Patient safety benzodiazepines, gaba receptor 4 Moderate consensus agonists, or monoamine oxidase inhibitors - “negative indicator”

Outcome Indicators

Category Indicator Median

113

score Consensus level

Hospital admission rate for 4 Healthcare diabetes long-term complications Moderate consensus Hospital admission rate for utilization/ 4 diabetes short-term complications Moderate consensus Clinical effectiveness Cardiovascular mortality rate 4 High consensus

*HbA1c testing=glycated hemoglobin testing **NSAIDs=non-steroidal anti-inflammatory drugs

3.3.5 Quality indicators for primary care for older adults with diabetes with comorbid osteoarthritis and hypertension

3.3.5.1 Delphi Round I

Of the 16 potential quality indicators for ambulatory care for older adults with diabetes with osteoarthritis and moderate depression, only 1 process and 1 outcome indicators reached the first level of consensus (at least 11 of 15 panelists (73%) rated an indicator as 4 or 5, and

MADM was less or equal to 1.03). The panelists did not reach consensus on the remaining

14 indicators: 6 indicators because of the median rating was 3, and 8 because less than 11 of

15 panelists (73%) rated it a 4 or 5, and 1 or 2, and MADM was more than 1.03. Levels of consensus and median values for each indicator for Round I are summarized in Table 14.

Table 3.14 Round I – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and hypertension

114

Potential for Overall value of Meaningfulness improvements in Process inclusion clinical practices Consensus Indicators Median Median Median (%) MADM (min; max) (min; max) (min; max) *HbA1c testing Consensus to every 6 month 4 (2; 5) 4 (2; 5) 4 (2; 5) 0.67 60% include

**LDL- cholesterol testing once per 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.08 No consensus year

Eye examination Consensus to 5 (1; 5) 4 (1; 5) 5 (1; 5) 0.79 80% every 1-2 years include

Microalbumin testing once per 3 (2; 5) 3 (2; 5) 3 (1; 5) 0.87 No consensus year

Statin therapy 3 (1; 5) 4 (1; 5) 3 (1; 5) 0.89 No consensus

***Use of ACE inhibitors or Consensus to 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.13 67% ARBs therapy include

Beta-blocker 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.07 No consensus therapy

Antiplatelet 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.79 No consensus therapy

Acetaminophen as first-line 3.5 (2; 5) 3.5 (2; 5) 3 (1; 5) 1.07 No consensus therapy

Non-selective ****NSAIDs 3 (1; 5) 3 (1; 5) 2 (1; 5) 1.00 53% No consensus therapy

Cox-selective NSAID therapy 3 (1; 5) 2 (1; 5) 2 (1; 5) 1.14 53% No consensus

Non-selective 2 (1; 5) 2 (1; 5) 2 (1; 5) 1.22 53% No consensus

115

NSAID in combination with misoprostol or proton pump inhibitors

Overall value of Importance Modifiability Outcome inclusion Consensus Indicators Median Median Median MADM (%) (min; max) (min; max) (min; max) Hospital admission rate for diabetes 4 (1; 5) 3 (1; 5) 4 (1; 5) 0.93 60% No consensus long-term complications Hospital admission rate for diabetes 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.00 60% No consensus short-term complications Lower- extremity 4 (1; 5) 4 (1; 5) 4 (1; 5) 1.00 53% No consensus amputation rate

Cardiovascular Consensus to mortality rate 4.5 (1; 5) 4 (1; 5) 4 (1; 5) 0.53 87% include

*HbA1c testing=glycated hemoglobin testing **LDL-cholesterol=low-density lipoprotein cholesterol ***ACEIs=angiotensin-converting enzyme inhibitors; ARBs=angiotensin II receptor blockers ****NSAIDs=non-steroidal anti-inflammatory drugs

3.3.5.2 Delphi Round II

Adjustments were made prior to the Delphi Round II, taking into account the suggestions and

comments of the panelists. The indicators that reached a first-round consensus were not

included in the second-round questionnaire. The remaining indicators were included in

Round II survey together with new indicators that were suggested by the panel that could be

measured using health administrative data.

116

In particular, the suggested new process indicators were use of topical NSAIDs, and referral

for home care, while new outcome indicators included joint replacement therapy, all-cause

ED mortality, and end-stage renal disease. Two indicators were modified as “negative

indicators” based on panelists’ comments and suggestions: use of non-selective NSAIDs, and

use of cox-selective NSAIDs.

In Round II, 11 process and 6 outcome indicators were evaluated using the same assessment

criteria as in Round I. Levels of consensus and median values for each indicator for Round I

are summarized in Table 15. Of the 17 indicators, only 2 process and 1 outcome indicator

reached a high level of consensus (at least 11 of 15 panelists (73%) rated an indicator a 4 and

5, and MADM was less or equal to 1.03), while 3 process and 1 outcome indicator reached

the moderate level of consensus (9 or 10 of 15 panelists (60-67%) rated an indicator a 4 or 5,

and MADM was less or equal to 1.03).

Table 3.15 Round II – Quality indicators for care for older adults with diabetes with comorbid osteoarthritis and hypertension

**LDL- 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.00 No consensus

117 cholesterol testing once per year

Microalbumin testing once per Consensus to 3 (2; 5) 4 (2; 5) 4 (1; 5) 0.80 60% year include

Statin therapy 3 (1; 5) 4 (1; 5) 3 (1; 5) 0.89 No consensus

***Use of ACE inhibitors or Consensus to 4 (1; 5) 4 (1; 5) 4 (1; 5) 0.67 73% ARBs include

Beta-blocker therapy 2 (1; 5) 2 (1; 4) 2 (1; 4) 1.00 67% Consensus to reject

Antiplatelet therapy 3 (1; 5) 3 (1; 5) 3 (1; 5) 1.01 No consensus

Acetaminophen as first-line 3 (1; 5) 3 (2; 5) 3 (2; 5) 1.06 No consensus therapy

Non-selective NSAID in combination with 3 (1; 5) 2 (1; 5) 3 (1; 4) 1.00 No consensus misoprostol or proton pump inhibitors Modified indicators

Non-selective ****NSAID therapy Consensus to 4 (1; 5) 4 (1; 5) 4 (1; 4) 0.93 60% “negative include indicator”

Cox-selective NSAID therapy “negative 3 (1; 5) 2 (1; 5) 3 (1; 4) 1.02 No consensus indicator”

New indicators

Use of topical NSAIDs 3 (1; 5) 3 (1; 5) 3 (1; 5) 0.93 No consensus

118

Referral for home care 3 (1; 4) 3 (1; 5) 3 (1; 5) 0.87 No consensus

Overall value of Importance Modifiability Outcome inclusion Consensus Indicators Median Median Median MADM (%) (min; max) (min; max) (min; max) Hospital admission rate for Consensus to diabetes long- 4 (2; 5) 4 (1; 5) 4 (3; 5) 0.64 60% include term complications Hospital admission rate for Consensus to diabetes short- 4 (1; 5) 4 (2; 5) 4 (2; 5) 0.60 73% include term complications Lower-extremity amputation rate 4 (1; 5) 3 (1; 5) 3 (1; 5) 0.73 No consensus

New indicators

All-cause mortality 4 (1; 5) 3 (1; 5) 3 (1; 5) 1.07 No consensus

Joint replacement therapy 3 (1; 4) 3 (1; 4) 3 (1; 4) 0.93 No consensus

End-stage renal disease 4 (2; 5) 3 (2; 5) 4 (2; 5) 0.80 53% No consensus

Thus, consensus was reached for 8 indicators in total, including 3 high-consensus and 5

moderate-consensus indicators. The “negative” indicator “use of non-selective NSAIDs

therapy” was included as an indicator of inappropriate care or poor performance. One

indicator – use of beta blockers, was rated as “little reason to include” and was not included

in the final list of quality indicators. The final list of selected process/outcome indicators was

119 categorized into groups according to the categories of the Conceptual Model of Performance

Measurement for People with Multiple Chronic Conditions (70) (Table 16).

Table 3.16 High and moderate consensus indicators for care for older adults with diabetes with comorbid osteoarthritis and hypertension

Process indicators

Median Category Indicator Consensus level score

*HbA testing every 6 1c 4 Moderate consensus

On-going Eye examination every 1-2 years 5 High consensus monitoring Microalbumin testing once per 4 Moderate consensus year ** Pharmacological Use of ACE inhibitors or ARBs therapy 4 High consensus Treatment

Non-selective ***NSAID therapy Patient safety 4 Moderate consensus “negative indicator”

Outcome Indicators

Median Category Indicator Consensus level score

Hospital admission rate for 4 Healthcare diabetes short-term complications Moderate consensus Hospital admission rate for utilization/ 4 Moderate consensus diabetes long-term complications Clinical effectiveness Cardiovascular mortality rate 4 High consensus

*HbA1c testing=glycated hemoglobin testing **ACEIs=angiotensin-converting enzyme inhibitors; ARBs=angiotensin II receptor blockers ***NSAIDs=non-steroidal anti-inflammatory drugs

120

3.4 Discussion

This two-round Delphi study identified a set of quality indicators for ambulatory care for older adults with diabetes with different types of comorbid conditions, including:

Diabetes with concordant comorbid conditions:

 Diabetes with comorbid hypertension

 Diabetes with comorbid hypertension and chronic ischemic heart disease

Diabetes with discordant comorbid conditions:

 Diabetes with comorbid osteoarthritis

 Diabetes with comorbid osteoarthritis and major depression

Diabetes with both types of comorbid conditions:

 Diabetes with comorbid hypertension and osteoarthritis.

To our knowledge this study is unique in its focus on developing quality indicators for care for older adults with diabetes with comorbid conditions and listing potentially harmful indicators.

The initial list of quality indicators was identified from literature using a systematic approach. As previously stated, there was scarce information on quality indicators for care for patients with diabetes with comorbidities. Our study focused on clinical process and outcome indicators that were categorized into domains, including screening, on-going monitoring, pharmacological treatment, patient safety, and healthcare utilization/ clinical

121 effectiveness, which are accessible through Ontario administrative data. The feasibility of measuring each of these indicators using administrative data was assessed prior to conducting the Delphi study.

A set of 48 quality indicators was developed by means of a Delphi study of care for older adults with diabetes with comorbid conditions in ambulatory care settings.

The study identified:

 8 high-consensus and 3 moderate-consensus indicators for care for older adults

with diabetes with comorbid hypertension;

 7 high-consensus and 6 moderate-consensus indicators for care for older adults

with diabetes with comorbid hypertension and chronic ischemic heart disease;

 3 high-consensus and 4 moderate-consensus indicators for care for older adults

with diabetes with comorbid osteoarthritis;

 3 high-consensus and 6 moderate-consensus indicators for care for older adults

with diabetes with comorbid osteoarthritis and major depression;

 3 high-consensuses and 5 moderate-consensuses for care for older adults with

diabetes with comorbid hypertension and osteoarthritis.

Overall, the panelists did not reach consensus on 44 quality indicators for care for older adults with five selected disease combinations, 4 indicators were rated as “do not include” or

“little reason to include” and were excluded from the final list of quality indicators, and 10 indicators suggested by the panelists were not included in this Delphi study because they are not amenable to measurement using health administrative data. (Appendix 12)

122

Panelists reached rapid consensus on quality indicators for care for older adults with diabetes with concordant conditions, but it was more complicated for care for diabetes patients with discordant or both types of comorbid conditions. This finding may reflect the fact that diabetes clinical guidelines do provide a few recommendations for care of patients with diabetes with comorbid concordant chronic conditions, including hypertension and ischemic heart disease, such as use of antihypertensive drugs, antiplatelet therapy, or use of statins

(156). Conversely, there are no guideline recommendations regarding the care of patients with diabetes with comorbid discordant chronic conditions, including depression and osteoarthritis.

The possible explanation can be that diabetes-concordant conditions are related to diabetes in terms of sharing the underlying predisposing factors and a common management plan. In contrast, diabetes-discordant conditions are not directly related to diabetes in their pathogenesis and are more likely to add to the complexity of clinical decision-making in the management of diabetes (22). A recent population-based study conducted in Ontario demonstrated that 20.7% of people with diabetes had diabetes-concordant conditions, 15.6% had diabetes-discordant conditions and 49.8% had both concordant and discordant conditions

(157).

In general, current clinical guidelines rarely consider the cumulative impact of clinical recommendations, including screening, monitoring and treatment, on individuals with multiple chronic conditions (158). Thus, there is a lack of scientific evidence on which to

123 build quality indicators for care for people with multiple chronic conditions. Moreover, prior research results demonstrated that over 65% of clinical trials excluded individuals over the age of 70 years, that makes it difficult to understand the impact of medical interventions on these populations (159).

We observed several inconsistencies between guideline recommendations and the Expert

Panel members’ opinions on testing and its frequency, as well as some pharmacological treatment recommendations in older adults with selected five disease combinations. Several panelists recommended less frequent HbA1c testing among older adults with controlled diabetes instead of testing every six months as per diabetes guideline recommendations

(156).

The annual LDL-cholesterol testing did not reach consensus and wasn’t included in the list of indicators for care for older adults with any of the selected disease combinations. The panellists mentioned that “LDL-C testing becomes less important in older people, especially without a measure of frailty”. Nevertheless, the literature suggests regular LDL-cholesterol testing and control for vascular protection and lowering the risk of development and progression of serious vascular complications for individuals with diabetes (156).

Although acetaminophen is recommended as a first choice therapy for osteoarthritis treatment because of its effectiveness and lower toxicity than other oral agents (136, 137,

160), this indicator did not reach the consensus and was not included in the final list of indicators for care for older adults with diabetes with comorbid osteoarthritis. The panelists

124 mentioned that “effectiveness of acetaminophen has been questioned”, or “it may be difficult to measure use of acetaminophen using administrative data and this could lead to underestimation of this drug”.

The Expert panel members considered the use of NSAIDs as a “negative” indicator for care for older adults with diabetes with comorbid osteoarthritis, despite the fact that the literature recommends NSAIDs to treat osteoarthritis if a patient does not have a satisfactory clinical response to full-dose acetaminophen (136, 160). Prior studies showed use of NSAIDs in the elderly is effective for the treatment of osteoarthritis, although the risk of serious adverse events is also clearly increased, including upper gastrointestinal bleeding and cardiovascular events, such as myocardial infarction and stroke (160-162).

The literature suggests avoidance of anticholinergic antidepressant drugs, including tricyclic or tetracyclic antidepressants, and MAO inhibitors, among older adults due to side effects, including delirium (132, 163). However, the majority of panelists suggested excluding use of tricyclic antidepressants from the list of “negative” indicators for care for older adults with diabetes comorbid with depression and osteoarthritis, because “risk of tricyclics is over emphasized”, and “recent reports of fewer falls with tricyclics”, or “they can be used appropriately in low doses for diabetic neuropathy and may therefore not be a negative indicator”; thus, this indicator was presented as a positive indicator in Round II.

Overall, reasons for the small number of defined outcome indicators may be the limited scientific evidence linking structure and process to outcomes of care for people with

125 multimorbidity, data limitation issues, or challenges to the accurate measurement because multiple factors contribute to a patient’s health outcomes. Future research is required to develop evidence-based outcome indicators that are sensitive to ambulatory care for older adults with multiple chronic conditions.

The defined quality indicators form the foundation for future development of quality indicators in the context of various disease combinations. The feedback from the panelists emphasized the importance of developing indicators related to such aspects of care as self- management, patient education, patient-physician relationships, patient's preferences and goals, as well as patient adherence to medication, diet, etc.

The panelists also highlighted the role of frailty level in developing quality indicators for care for older adults with multiple chronic conditions. In particular, they mentioned the importance of considering frailty level for discussing pharmacological treatment or setting a testing target, such as HbA1c testing, in older adults with diabetes. The literature suggests that older adults with diabetes with moderate or advanced frailty have a reduced life expectancy and high level of functional dependency (156). Previous research revealed that stringent glycemic control in these patients results in fewer episodes of significant hyperglycemia but also more episodes of severe hypoglycemia (164).

The main advantages of the present Delphi study are that there was no need for participants to meet and, hence it served as a relatively inexpensive method; it allowed the involvement of participants from disparate geographical areas and has been used in international health

126 research. Delphi participants were purposefully selected to apply their knowledge and experience to appraise and develop a list of indicators in the context of assessing care of older adults with multiple chronic conditions.

Within the Delphi study, panelists do not meet face to face and this enables a relatively large group of 15 experts from 3 different provinces of Canada to be consulted for the development of a set of indicators. In our panel, most of the 15 panelists were in general practice or geriatrics, and the representativeness of the panel was ensured by including clinical pharmacists and a senior methodologist from a quality measurement organization from across Canada. All experts had been involved in multimorbid patient treatment decisions and/or in a number of research studies or quality development activities focusing on patients with multiple chronic conditions.

Feedback of responses to panelists and the opportunity to revise earlier ratings of the indicators are the main features of the Delphi technique; and this usually requires at least two rounds (165). The anonymity of panelists and their responses removes the effects of dominant personalities (166). We used median scores, the range of panel scores and mean absolute deviation from median to give a full picture of results.

3.4.1 Limitations

The main limitations of the Delphi study include purposeful selection of the panelists, attrition rate and non-response bias (155, 167). The two-phase Delphi study and

127 incorporation of reminder letters helped reduce the impact of attrition. Fifteen panelists participated in both rounds of this study. Use of consensus methods in health services research has been criticized in relation to validity (78). However, evidence suggests that if the expert panel is representative of the area of knowledge, then content validity can be assumed (168).

The Delphi technique has been criticised for having no evidence of reproducibility and for being only reflective of the opinion specifically of the invited experts (169). Many factors influence ratings in a consensus method and the final selection of quality indicators is obviously sensitive to the panel composition. Despite some limitations, a Delphi technique is an appropriate first step to critically appraise and develop quality indicators for care for older adults with multiple chronic conditions and to inform future research using a well-established and structured process.

The study findings demonstrate that there was limited research on quality indicators for care for patients with diabetes with comorbid conditions.

There was no in-person meeting of the Panel members during the present Delphi study due to time constraints and diverse geographical representation among the panel members.

Evidence shows that having an in-person meeting might result in the situation when some panel members may dominate the whole consensus process, contradicting one of basic rules of the Delphi technique (78, 139). Nevertheless, an in-person meeting at the end of the

Delphi study may be useful in clarification of reasons for disagreement though face-to-face exchange of the information when reaching a consensus is difficult (139).

128

This present study focused only on clinical process and outcome indicators that are amenable to measurement using administrative data. Consequently, patient-reported data, including patients/ caregivers’ preferences and goals for care, were not captured in this study.

Meanwhile, one of the six Institute of Medicine (IOM) specified quality of care aims include patient-centeredness (170). The IOM defines patient-centered care as “care that is respectful of and responsive to individual patient preferences and needs and that is guided by patient values” (170).

Future research is required to develop quality indicators that would reflect various aspects of care provided to older adults with selected comorbid conditions, including clinical measures along with the patients’ preferences and goals for care, self-management, patient education, and patient compliance with treatment and diet, as well as quality of life measures and efficiency of care. There is a need to develop specific outcome measures for older people with diabetes with comorbidities to reflect what matters most to patients, which is the effect of all their conditions on their health status.

We did not include in our study indicators related to laboratory values such as HbA1c, microalbumin or LDL-cholesterol level. We also were not able to control for severity of chronic conditions due to limitations to clinical sensitivity of administrative data. We asked our panel members to consider all conditions as if they were at moderate severity.

Meanwhile, the frequency of testing and use of particular mediations may depend on the severity of particular illnesses or their combinations. Thus, the developed set of indicators

129 provides a starting point for further investigations that might explore selection of quality indicators for care for older adults with multiple chronic conditions considering severity of illnesses.

3.5 Conclusions

Quality indicators are important tools used for both quality assessment and quality improvement in healthcare systems. The defined indicators are not intended to provide a comprehensive tool set for measuring quality of care for older adults with diabetes with comorbidities. Rather, they address clinical aspects of care and can be used as a starting point for further development of quality indicators and use in ambulatory care settings. The developed indicators are useful for health care providers, managers and decision makers and can be used to evaluate the quality of overall care for older adults with selected disease combinations in ambulatory care settings.

The developed indicators are useful for health care providers, managers and policy makers and can be used to evaluate the quality of care for older adults with selected disease combinations in ambulatory care settings. In particular, they can allow health care providers to initiate local quality improvement initiative, systems managers to identify and correct system-wide problems, and policy makers to plan for future systems of care for older adults with selected disease combinations.

130

As the selected disease combinations burden is high among older adults, and much of it is presented clinically to general practitioners, incorporation of these indicators to routine ambulatory care practice is recommended. Despite the indicators were developed to assess the quality of ambulatory care for older diabetes patients in Canada, they can be useful to others undertaking evaluation or research on ambulatory care for multimorbid diabetes patients.

Quality indicators should be valid and sensitive to the changes they are intended to detect, and should be linked to improving patient outcomes (75). Future research is required to implement the developed set of quality indicators in this study, and to examine the association between defined processes and outcomes of care for older adults with multiple chronic conditions.

The study findings suggest lack of agreement between Expert panel members with respect to developing indicators for older diabetes patients with discordant comorbid conditions.

We recommend developing quality indicators for the larger diabetes population with various comorbid conditions comorbid conditions so the quality of care can be assessed within these

131 disease combinations. This can facilitate targeting the interventions towards improving the ambulatory care practices for care for specific groups among older adults with diabetes.

132

CHAPTER 4

Evaluating quality of overall care among older adults with diabetes with comorbid chronic conditions: a retrospective cohort study

Abstract

Background: With an increasing number and complexity of comorbid chronic conditions among older adults with diabetes, current disease-specific guidelines fail to address multiple treatment needs in this group of patients. The literature suggests that an optimal approach to treat patients with any combination of co-existing diseases is not the same as the sum of treatments for the separate diseases. However, a single condition focus in both clinical care and research remains and limits the assessment of care for the whole person with multiple chronic conditions.

Purpose: 1) to examine the quality of care among adults aged 65 and over with a diagnosis of diabetes with selected comorbid conditions in Ontario, in the time period from 2010 to 2014, 2) to examine the difference in the quality of care between patients with 2 vs. 1 selected concordant vs. discordant comorbid conditions, 3) to examine associations between the quality of care and hospitalizations, among older adults with diabetes with selected comorbid conditions.

Methods: This retrospective cohort study used linked provincial health administrative databases housed at the Institute for Clinical Evaluative Sciences in Ontario. All people 65 years of age and older who had diabetes comorbid with hypertension, chronic ischemic heart disease, osteoarthritis or major depression between April 1, 2010 and March 31, were included in this study. The cohort was stratified into four disease combinations, including concordant conditions: 1) diabetes with comorbid hypertension, 2) diabetes with comorbid hypertension and chronic ischemic heart disease, and discordant conditions: 1) diabetes comorbid with osteoarthritis, 2) diabetes with comorbid osteoarthritis and major depression. A specific set of process and outcome indicators developed by means of a Delphi panel was used for assessing the quality of care among older diabetes patients with comorbidities (Chapter 3). A generalized estimating equations approach was used to examine associations between the quality of care and the likelihood of hospitalizations among older diabetes patients with selected comorbid conditions.

Findings: The study findings suggest that patients are at risk of suboptimal care with additional comorbid condition, especially those with discordant comorbid conditions. The incidence of hospitalizations markedly increased in older adults with diabetes with 2 vs. 1 selected comorbid condition, especially in those with discordant conditions. The median score of continuity of care was higher among older adults with concordant comorbid conditions compared to those with discordant conditions; and it declined with additional selected comorbid conditions among older diabetes patients, especially in those with

133 discordant comorbid conditions. The study findings suggest that greater continuity of care was associated with lower hospital utilization for older diabetes patients with both concordant and discordant conditions. The study findings also suggest that the likelihood of hospitalizations increases with the number of prescribed drugs among older adults with comorbid concordant or discordant conditions.

Conclusions: Older adults with diabetes are a diverse and heterogeneous group; therefore, there is a need for a holistic approach in education and clinical care of patients taking into account concomitant conditions that affect patient’s overall health status. Older diabetes patients with discordant conditions- comorbid osteoarthritis with or without major depression need more targeted interventions and collaboration between healthcare providers to improve quality of care and reduce hospitalizations. Chronic condition management programs among older adults with diabetes must incorporate levers to promote continuity, especially for patients with discordant conditions who are more likely to see multiple providers compared to those with concordant conditions.

134

4.1 Introduction

Multimorbidity, defined as the co-existence of two or more conditions in an individual without defining an index disease, is commonly experienced in older adults (5, 74, 171, 172).

Evidence suggests that multimorbidity will become more prevalent as populations ages (172,

173). Data from the Canadian Community Health Survey shows that 71% of Canadians of 65 to 79 years old and 82% of adults aged 80 and older have two or more chronic conditions (4,

10).

Data from several community surveys across Canada indicate that multimorbidity is affecting the most vulnerable groups in society, including those who are less educated, have low incomes, and are living in rural communities (10-12).

Primary care is well placed to have an important impact on outcomes of care for people with multiple chronic conditions (58, 67). In Ontario, primary care includes many different practice models, ranging from small solo-physician practices to large, regionally organized interdisciplinary teams (60). These models differ mainly by physician compensation methods

(e.g. fee-for-service, capitation) and several organizational elements, including patient enrollment, quality incentives, provision of after-hour services, and inter-disciplinary teams

(61, 62).

The literature suggests that different primary care models serve different populations of patients and are associated with different outcomes, including rates for hospital admissions and emergency department visits (61, 74). Several studies found that patients at Community

Health Centres (CHC) had the highest overall performance of chronic disease management

135

(60, 62). Recent research revealed that patients enrolled in blended capitation models received higher quality of diabetes care than those enrolled in blended fee-for-service models, while the biggest gap in diabetes care quality occurred for patients not enrolled in any primary care model (64).

People with multiple chronic conditions are likely to have frequent consultations with various health providers with cross-referral related to the co-existing conditions (67).

Therefore, high continuity of care may mitigate some challenges in managing care for people with diabetes with concurrent comorbidities. In the last two decades, over a dozen indices have been developed to assess how care is concentrated among the different providers that a patient sees. The most widely used measures of continuity of care include Usual Provider of

Care (UPC) and the Continuity of Care (COC) indices (68). The UPC index measures the proportion of visits with a usual provider over a given period of time.

The continuity of care (COC) index measures both the dispersion and concentration of care among all providers seen, and can be adapted to capture aspects of the coordination of care by attributing referral visits back to the referring provider (68, 72). Advantages of the COC index include: 1) it is a widely used measure, permitting comparison between studies, 2) it may be adapted to measure concentration of care at a care site or with a group of providers,

3) it accounts for number of different providers seen, and 4) it is sensitive to shifts in the distribution of visits among providers (68, 72).

136

The literature provides mixed results on the level of continuity of care among people with multiple chronic conditions. Previous study results demonstrate that levels of continuity of care, measured using continuity of care (COC) indexes, were lower among people with multiple chronic conditions, compared to those with single conditions (67, 72). But results of a population-based study conducted in Ontario demonstrate that continuity of care was relatively low among people with one or more concomitant chronic conditions (74). This study also found that the risk of hospitalizations was lower among people with greater continuity of care (74).

A growing body of evidence suggests that individuals with multiple chronic conditions have different care needs as compared to those with single chronic conditions; therefore, processes of care addressing patients’ single complains are not ideal for those with multiple chronic conditions (32, 143-145). People with multiple chronic conditions who are treated with clinical guidelines for single diseases may be susceptible to the drug-drug interactions due to uncertainty about the benefits and harms of simultaneous treatment, as well as drug-disease interactions due to a potential risk of worsening one condition by treating a co-existing one

(34, 46). Thus, there is a need to identify new processes of care for people with multiple chronic conditions oriented toward patients’ overall needs taking into account the multimorbidity patterns, continuity of care and integration of patients’ needs (143).

A growing body of evidence shows that people with multiple chronic conditions, compared to those with singe diseases, are more likely to experience negative health outcomes,

137 including increased hospital admissions and emergency department visits, high use of ambulatory services, poor quality of life and increased care costs (11, 32, 33, 174).

The literature suggests that adults with multiple chronic conditions, in particular those aged

65 and over, are significant users of healthcare services, and account for more than two-thirds of healthcare spending (31). Prior research found that Ontarians with three or more diagnoses had 56% more primary care visits, 76% more specialist visits, 256% more inpatient hospital stays, 11% more emergency department visits, and 68% more prescriptions, as compared to those with a single condition (35, 36). Evidence shows that diabetes mostly occurs in conjunction with other chronic conditions (30, 157). Recent study results, conducted in

Ontario, demonstrate that more than 90% of community-dwelling older adults with diabetes had at least one comorbid condition; and more than 40% of the cohort had 5 or more comorbidities (24).

Comorbid conditions in people with diabetes can be categorized into diabetes-concordant and diabetes-discordant conditions (22). Diabetes-concordant conditions are related to diabetes in terms of sharing the underlying predisposing factors and a common management plan with diabetes. In contrast, diabetes-discordant conditions are not directly related to diabetes in their pathogenesis and are more likely to add to the complexity of clinical decision-making in the management of diabetes (22).

The presence of comorbid conditions among older adults with diabetes, especially discordant or both types of conditions, has been shown to be associated with higher rates of hospital

138 admissions (25, 157). A recent study indicated that there is a trend of increasing use of healthcare services, including hospitalizations, emergency department visits and physician visits, with increase in the number of comorbid conditions among older adults with diabetes

(24).

The results of a population-based study conducted in Ontario demonstrate that diabetes patients with comorbidities were more likely to be hospitalized for diabetes-related short- term complications, including hyper – or hypoglycemia, compared to those without comorbidities (157). Uncontrolled diabetes can contribute to long-term microvascular complications affecting the kidneys, eyes, and nerves, as well as an increased risk for diabetic foot complications and cardiovascular disease (156).

Previous study results showed that hospitalisation rates appeared to be associated with primary health care visits in a nonlinear fashion among people with chronic conditions, including diabetes, hypertension, and ischemic heart disease (175). In particular, as compared to patients who had less than 5 primary care visits per year, having 5 to 12 visits per year was associated with decreased risk of hospitalization, while having more than 12 visits per year was associated with increased risk of hospitalization (175).

4.1.1 Rationale for the study

Primary care physicians face difficulties in addressing the complex multifaceted needs of older adults with multiple chronic conditions. Actually, treatment of people with multiple

139 chronic conditions often requires “trade-off” decisions, because current clinical guidelines may be impractical in the presence of multiple chronic conditions (65).

Treating one condition in older diabetes patients with comorbid conditions may cause undesirable consequences with regard to their other conditions. For instance, nonsteroidal anti-inflammatory medications for pain relief from osteoarthritis would aggravate hypertension and renal disease in diabetes patients (162, 176); or beta-blockers for treating diabetes and heart failure will lead to bronchospasm in asthmatic patients (177).

The literature suggests that an optimal approach to treat patients with any combination of co- existing diseases is not the same as the sum of treatments for the separate diseases (178).

However, a single condition focus in both clinical care and research remains and limits the assessment of care for the whole person with multiple chronic conditions. This study aimed to explore whether the quality of care for older people with diabetes is differentially affected by types and number of comorbid chronic conditions; as well as to examine the association between processes of care measures and the likelihood of hospitalizations among older adults with diabetes with selected comorbid conditions.

This study is focused on older adults because they are more likely than younger individuals to have comorbid chronic conditions that can be complex and difficult to manage (5, 41).

Diabetes has been chosen as a condition of interest due to the high burden of co-existing chronic conditions in this group of patients (5, 24). Four chronic conditions among diabetes patients, including hypertension, chronic ischemic heart disease, osteoarthritis, and major

140 depression were selected for the purpose of this study, since they have been identified as the most common conditions among people diagnosed with diabetes (23, 24).

These chronic conditions were grouped into four disease combinations that represented the most prevalent clusters of three co-existing conditions across multimorbidity groupings based on previous research results (5, 24). Finally, the defined four disease combinations were categorized into two groups by comorbidity type (22), including:

 Diabetes with concordant comorbid conditions:

o Diabetes with comorbid hypertension

o Diabetes with comorbid hypertension and chronic ischemic heart disease

 Diabetes with discordant comorbid conditions:

o Diabetes with comorbid osteoarthritis

o Diabetes with comorbid osteoarthritis and major depression

4.1.2 Research questions

1. What is the quality of care among adults aged 65 and over with a diagnosis of

diabetes with selected concordant vs. discordant comorbid conditions in Ontario, in

the time period from 2010 to 2014?

2. What is the difference in the quality of care between patients with 2 vs. 1 selected

concordant vs. discordant comorbid conditions?

3. Is there an association between the quality of care and the likelihood of

hospitalizations among older adults with diabetes with selected comorbid conditions?

141

Hypothesis #1: Older adults with diabetes with concordant comorbid conditions will have better quality of care and less hospitalizations, compared to those with discordant comorbid conditions.

Hypothesis #2: Each additional selected condition among older adults with diabetes will result in worse quality of care and increase in hospitalizations, regardless of comorbidity type.

4.2 Methods

4.2.1 Study design and study participants

This is a retrospective cohort study conducted in Ontario, Canada using linked provincial health administrative databases. The study observation period extended from April 1, 2010 to

March 31, 2014. We identified a cohort of people 65 years of age and older who had diabetes as of April 1, 2010, using Ontario Diabetes Database (ODD) (179, 180). The ODD is a validated database that identifies all adults aged 20 years and older with diabetes in Ontario from April 1, 1991 (179, 180). The use of ODD provided high level sensitivity (86%) and specificity (97%) in identifying individuals in whom diabetes was reported in primary care charts (64, 179).

Patients were considered to have diabetes if diagnosed with the International Classification of Diseases (ICD 10) code of E10 to E14. Individuals were included if they had either one hospital admission or two Ontario Health Insurance Plan (OHIP) claims with a diagnosis of

142 diabetes within 2 years. The ODD does not distinguish between diabetes type 1 and 2 and excludes gestational diabetes, although the large majority of people identified would be expected to have type 2 diabetes (179).

Four other chronic conditions among patients with diabetes were selected as the most prevalent conditions, including osteoarthritis, hypertension, chronic ischemic heart disease and major depression/dysthymia (5, 24). Patients were considered to have hypertension if diagnosed with the International Classification of Diseases (ICD 10) code of I10 to I15, or

ICD 9 code of 401 to 405. Patients were considered to have chronic ischemic heart disease if diagnosed with the International Classification of Diseases (ICD 10) code of I25, or ICD 9 code of 412 and 414.

Patients were considered to have osteoarthritis if diagnosed with the ICD 10 code of M15 to

M19, or ICD 9 code of 715. Depression in this study connotes major depression and dysthymia, since most clinical practice guidelines only address treatment of major depression

(84). Previous studies demonstrate that treatments for major depression also apply to dysthymic disorders (113-115). Patients were considered to have major depression/dysthymia if diagnosed with the ICD 10 code F32.0- F32.3, F32.9- F33.4, F33.9,

F34.1, or ICD 9 code of 296.

The same algorithm was used that requires one diagnosis recorded in acute care (DAD), or two diagnoses recorded in ambulatory care (physician) records within a two-year period prior to the index date (April 1, 2010) to define the four other conditions. Validation studies of

143 algorithms used to create ICES registries found sensitivity values for hypertension to be 72%

(5). Validation studies of algorithms used for detecting major depression in administrative data found a suboptimal level of sensitivity (30%) and a high level specificity (98%) (181).

Validation studies of algorithms used for detecting patients with osteoarthritis and ischemic heart disease in administrative data found sensitivity values to be 50% and 69%, respectively

(182).

The resulting database is cumulative, such that older adults with five selected disease combinations remain in the database once identified as of April 1, 2010 up to March 31,

2014 (Appendix 13). Patients were excluded if they fell under the following criteria: had an invalid health card number, were younger than 65 or older than 105 years old, died before the index date (April 1, 2010), or had no contact with the health care system in the last 5 years before the index date.

Data sources for this study included: the Canadian Institute for Health Information (CIHI)

Discharge Abstract Database (DAD) which consists of data on all hospital discharges in

Ontario; the OHIP database which contains information on patient contact with physicians in both ambulatory and hospital settings; the Registered Persons Database (RPDB) which contains information regarding the demographics of persons eligible for health care coverage in Ontario; the Client Agency Program Enrolment (CAPE) database which identifies patients belonging to the primary care models; and the Ontario Drug Benefit (ODB) claims database which contains comprehensive records of prescription medications dispensed in outpatient pharmacies to Ontario residents eligible for public drug coverage, specifically those aged 65

144 and over. Canada census data were also used to derive population estimates by age and sex in each year. All databases were linked using unique, encoded identifiers and analyzed at the

Institute of Clinical Evaluative Sciences (ICES) in Toronto, Ontario.

4.2.2 Measures

A specific set of process and outcome measures was developed by means of a Delphi panel

(Chapter 3) for assessing the quality of care for older adults with each particular disease combination in ambulatory care settings (Table 4.1).

Table 4.1 Process and outcome measures for ambulatory care for older adults with four selected disease combinations

Concordant conditions Discordant conditions

Measure Diabetes with Diabetes with Diabetes with Diabetes with comorbid comorbid comorbid comorbid hypertension hypertension and osteoarthritis osteoarthritis and chronic ischemic heart major depression disease Process measures

*HbA1c testing X X X X

Eye examination X X X X

Use of oral X X X X hypoglycemic drugs Use of angiotensin- converting enzyme X X (ACE) inhibitors Use of angiotensin X X

145

II receptor blockers (ARBs) Us of antiplatelet X drugs Use of statins X

Use of *NSAIDs- X X “negative” indicator Use of tetracyclic antidepressant – X “negative indicator” Use of monoamine oxidase inhibitors X (MAO) – “negative indicator” Use of benzodiazepines – X “negative indicator” Use of gaba receptor agonists – “negative X indicator” Outcome measures

All-cause X X X X hospitalizations Hospitalizations for diabetes long-term X X X X complications Hospitalizations for diabetes short-term X X X X complications *HbA1c=glycated hemoglobin **NSAIDs=non-steroidal anti-inflammatory drugs

4.2.2.1 Dependent variables

All outcome indicators were measured in each year, from April 1, 2010 to March 31, 2014 to see their fluctuations over 4-year period compared to the baseline data of 2010.

146

The primary outcome in this study was a likelihood of having at least one hospital admission in each year, during the period of time from April 1, 2010 to March 3, 2014. This outcome measure had a value 1 (yes) if an older adult with any of three selected disease combinations had at least one all-cause hospitalization in each year, and 0 (no) if did not.

The secondary outcomes in this study were a likelihood of having at least one hospitalization for diabetes-related short-term or long-term complications in each year, during the period of time from April 1, 2010 to March 3, 2014. Hospital admissions for diabetes-related complications were identified using the most responsible diagnosis or principal diagnosis code of diabetes-related short-term or long-term complications as defined by the OECD

Health Care Quality Indicator Project (183, 184).

Diabetes-related short-term complications consisted of diabetic ketoacidosis, hyperglycemic hyperosmolar coma, mixed ketoacidosis, or hypoglycemic or insulin coma (ICD-10 Codes:

E10.0, E10.1, E10.11, E10.12, E11.0, E11.1, E11.11, E11.12, E13.0, E13.1,E13.11, E13.12,

E14.0, E14.1, E14.11, E14.12) (183, 184).

Diabetes-related long-term complications included micro- or macrovascular complications, including ophthalmic, renal, neurological, circulatory complications or multiple complications (ICD-10 Codes: E10.2, E10.3, E10.4, E10.5, E10.6, E10.7, E11.2, E11.3,

E13.4, E11.5, E11.6, E11.7, E13.2, E13.3, E13.4, E13.5, E13.6, E13.7, E14.2, E14.3, E14.4,

E14.5, E14.6, E14.7) (183, 184).

147

Each outcome measure had a value 1 (yes) if an older adults with any of three selected disease combinations had at least one hospitalization for diabetes-related short-term or long- term complications in each year, respectively, and 0 (no) if did not, over the 4-year study period.

4.2.2.2 Independent variables/ Process measures

A specific set of processes indicators was developed by means of a Delphi study for evaluating the quality of care for older adults with four selected disease combinations in ambulatory care settings, respectively.

The cohort of older adults with diabetes was stratified into mutually exclusive concordant and discordant disease combinations with 2 vs. 3 chronic conditions in each.

Concordant disease combinations included:

 Diabetes with comorbid hypertension and without chronic ischemic heart disease

 Diabetes with comorbid hypertension and chronic ischemic heart disease

Discordant disease combinations included:

 Diabetes with comorbid osteoarthritis and without major depression;

 Diabetes with comorbid osteoarthritis and major depression

All process indicators were measured if occurred in each year and were deemed to meet the recommended quality standards, during the period of time from April 1, 2010 to March 31,

2014 to see their fluctuations over 4-year period compared to the baseline data of 2010.

148

1. Processes of care measures for older adults with diabetes comorbid with hypertension

a. Having 1 or 2 glycated hemoglobin (HbA1c) tests per year - (yes =1, no =0),

b. Having 3 or more HbA1c tests per year - (yes =1, no =0),

c. Annual eye examination - (yes =1, no =0),

d. Use of oral hypoglycemic drugs in each year - (yes =1, no =0),

e. Use of angiotensin-converting-enzyme (ACE) inhibitors in each year - (yes

=1, no =0),

f. Use of angiotensin II receptor blockers (ARBs) in each year - (yes =1, no =0),

g. *Continuity of care (COC) index in each year – (>0.57=1 – greater continuity

or concentration of care, ≤0.57 = 0 –less continuity of care),

h. Number of prescribed drugs in each year – (<5, 6-10, ≥11 drugs) (49, 52)

*Continuity of care was measured use Bice’s COC index that measures both the dispersion and concentration of care among all providers seen, and can be adapted to capture aspects of the coordination of care by attributing referral visits back to the referring provider (68, 72). We calculated the COC using billing codes for any outpatient physician visit, including all primary and specialty care visits. The possible index values range from one, if all visits made to the same physician, and just greater than zero if visits are made to different physicians (72). We categorized COC index as having a high vs. low continuity or concentration of care using the median COC score for each selected disease combination, respectively.

149

2. Processes of care measures for older adults with diabetes comorbid with hypertension

and chronic ischemic heart disease

a. Having 1 or 2 glycated hemoglobin (HbA1c) tests per year - (yes =1, no =0),

b. Having 3 or more HbA1c tests per year - (yes =1, no =0),

c. Annual eye examination - (yes =1, no =0),

d. Use of oral hypoglycemic drugs in each year - (yes =1, no =0),

e. Use of angiotensin-converting-enzyme (ACE) inhibitors in in each year - (yes

=1, no =0),

f. Use of angiotensin II receptor blockers (ARBs) in each year - (yes =1, no =0),

g. Use of antiplatelet drugs in each year - (yes =1, no =0),

h. Use of statins in each year - (yes =1, no =0),

i. Continuity of care (COC) index in each year – (>0.49=1 – greater continuity

or concentration of care, ≤0.49 = 0 –less continuity of care).

j. Number of prescribed drugs in each year – (<5, 6-10, ≥11 drugs)

3. Processes of care measures for older adults with diabetes comorbid with osteoarthritis

a. Having 1 or 2 HbA1c tests per year - (yes =1, no =0),

b. Having 3 or more HbA1c tests per year - (yes =1, no =0),

c. Annual eye examination - (yes =1, no =0),

d. Use of oral hypoglycemic drugs in each year - (yes =1, no =0),

150

e. Use of non-steroidal anti-inflammatory drugs (NSAIDs) in each year –

“negative” indicator - (yes =1, no =0),

f. Continuity of care (COC) index in each year – (≥0.53=1 – greater continuity

or concentration of care, ≤0.53 = 0 –less continuity of care)

g. Number of prescribed drugs in each year – (<5, 6-10, ≥11 drugs)

4. Processes of care measures for older adults with diabetes comorbid with osteoarthritis

and major depression

a. Having 1 or 2 HbA1c tests per year - (yes =1, no =0),

b. Having 3 or more HbA1c tests per year - (yes =1, no =0),

c. Annual eye examination - (yes =1, no =0),

d. Use of oral hypoglycemic drugs in each year - (yes =1, no =0),

e. Use of non-steroidal anti-inflammatory drugs (NSAIDs) in each year –

“negative” indicator (yes =1, no =0),

f. Use of tetracyclic antidepressants in each year – “negative” indicator (yes =1,

no =0),

g. Use of monoamine oxidase (MAO) inhibitors in each year – “negative”

indicator (yes =1, no =0),

h. Use of gaba receptor agonists in each year – “negative” indicator (yes =1, no

=0),

i. Use of benzodiazepines in each year – “negative” indicator (yes =1, no =0),

151

j. Continuity of care (COC) index in each year – (>0.36=1 – greater continuity

or concentration of care, ≤0.36 = 0 – less continuity of care),

k. Number of prescribed drugs in each year – (<5, 6-10, ≥11 drugs)

4.2.2.3 Covariates

We also controlled for other factors that could confound the relationship between processes of care measures and hospitalizations. These additional factors included age (coded as 65-74;

75-84; 85-94; 95 and over); sex (coded as 0 = male, 1 = female); geographic location measured by the Rurality Index for Ontario (RIO) (≤40 = non-rural and >40 = rural) (185); income quintile (ranging from Q1 = lowest income to Q5=highest income) (186); primary care models categorized into: 1) non-capitated, including the comprehensive care model, family health groups, and all other non-rostered models where physicians largely operate on a fee-for-service basis; 2) capitated, including family health networks and family health organizations; and 3) capitated+, including family health teams and other rostered models with additional incentives for interdisciplinary care (31, 187); level of multimorbidity (i.e., chronic disease burden) as the number of prevalent chronic conditions in addition to five selected chronic conditions (5, 24), including heart failure, acute myocardial infarction, cardiac arrhythmia, stroke, COPD, asthma, cancer, renal disease, other mood disorders, dementia, psychiatric diseases other than mood disorders and dementia, rheumatoid arthritis, or osteoporosis (Appendix 14) - this was coded as zero, one, two, three, four, or five-plus; number of primary care visits, including office-based visit with a general practitioner or

152 family physician; duration of each condition of interest in the particular disease combinations, including diabetes, hypertension, chronic ischemic heart disease, major depression or osteoarthritis (in years).

4.2.2.4 Data linkage and data collection

All data were housed and analyzed at the Institute for Clinical Evaluation Sciences (ICES) in

Toronto, Ontario. Unique encrypted identifiers were used for linkage across the databases.

Chronic conditions cohorts were linked to primary care services, hospital in-patient care and prescription drug use data. The data linkage was performed at the level of individual patients by year.

The study received approval from the Sunnybrook Health Sciences Research Ethics Board and the University of Toronto.

4.2.3 Statistical analysis

The cohort was stratified by four mutually exclusive disease combinations in older adults as follows:

Concordant disease combinations, including:

 Older adults with diabetes with comorbid hypertension and without chronic

ischemic heart disease

153

 Older adults with diabetes with comorbid hypertension and chronic ischemic

heart disease

Discordant disease combinations, including:

 Older adults with diabetes with comorbid osteoarthritis and without major

depression

 Older adults with diabetes with comorbid osteoarthritis and major depression.

Participant characteristics were described using proportions, means (standard deviation

(SD)), and medians (inter-quartile range (IQR)) where appropriate. To address research questions #1 and #2, univariate analysis was performed to estimate the processes of care measures and hospitalizations, in each fiscal year, from April 1, 2010 to March 31, 2014.

Marginal logistic models using generalized estimating equations (PROC GENMOD procedure with invoking a REPREATED statement in SAS) were performed to address research question #3. Marginal logistic regression models, binary and multivariate, were performed to examine associations between the processes of care measures, from April 1,

2010 to March 31, 2013, and the likelihood of hospitalisations during the follow-up period, from 2011-2014, among older adults with each particular disease combination, respectively.

The choice of generalized estimating equations approach was based on the nature of the longitudinal data and the study objectives. A straightforward application of the standard logistic models to longitudinal data is not appropriate, owing to the lack of independence among repeated measures obtained from the same individual that may result in invalid

154 statistical inferences (188). The generalized estimating equations are used to make inferences about the mean response in the population, to make inference about differences in quality of care between two group of patients, to account for within-subject correlation among the repeated responses, to deal with different numbers of observations per patient, and to estimate model parameters, using the available information (188).

4.2.3.1 Regression models

Multiple regressions were performed to examine a temporal relationship between the quality of ambulatory care for older diabetes patients with comorbidities, from 2010 to 2013, and the likelihood of hospitalizations during the follow-up period, from 2011-2014, using marginal logistic models with logit link function and binominal distribution (188):

Logit (µij /1-µi) =β0+β1timeij+ β2Xij2 + β3Xij3 + · · · + βpXijp.

Where,

E (Yij/Xij)= µij

th th Yij denote response variable for i subject on j occasion (year), and

th th Xij =1 if i subject had a test on j occasion (year), and Xij =0 if did not have;

Time = persons-years of follow-up – total duration of observed follow-up, from 2010

to 2014.

155

 An unstructured working correlation matrix was used for the purpose of this study

specified in terms of log odds ratios for all pairs of responses that were selected

(188):

 The correlation of repeated measures over time was specified using REPEATED

SUBJECT statement (188).

Multicollinearity was tested by examining the variance inflation factors (VIF). The results of variance inflation factor for all of the models were between 1.005 and 1.3025, which is less than the upper limit of 10 that indicates multicollinearity. This means that multicollinearity was not a problem for this analysis.

Risk estimates were presented as adjusted odds ratios (AORs) and corresponding 95 %

Confidence Intervals (CIs). All data analyses were performed with SAS package version 9.3

(SAS Institute, Cary, 145 North Carolina). The level of statistical significance was considered p less than 0.05.

4.3 Results

4.3.1. Quality of overall care and hospitalizations among patients with concordant disease combinations

156

4.3.1.1 Quality of overall care and hospitalizations among older adults with diabetes with comorbid hypertension

The cohort of older adults with diabetes with comorbid hypertension and without chronic ischemic heart disease included 273,592 patients. The baseline characteristics of patients are presented in Table 4.2. About 85% of diabetes patients with comorbid hypertension were between 65 and 84 years. Slightly over half of the people were female (56.5%). Nearly all patients (78.4%) resided in non-rural areas and the majority were enrolled in non-capitated primary care models (68.3%).

The median score of the continuity of care (COC) index was 0.57 (IQR=0.46). The mean duration of diabetes and hypertension were 9.9 (SD=5.80) years and 13.1 (SD=5.65) years, respectively. The mean number of drugs prescribed in older adults with diabetes with comorbid hypertension was 10.6 (SD=5.89). Nearly half of the people (44.7%) were prescribed 11 or more medications. About 56% of older adults with diabetes with comorbid hypertension had 1 or 2 other comorbid chronic conditions, while about 22.4% of patients had 3 or more comorbid conditions.

Table 4.2 Baseline characteristics of older diabetes patients with comorbid hypertension

Diabetes with comorbid Characteristic hypertension and without chronic ischemic heart disease Number of individuals 273,592 Age in years, mean (SD) 76.2 (7.18)

157

Age in groups, n (%) 65 – 74 127,469 (46.6) 75 – 84 106,336 (38.9) 85 – 94 37,194 (13.6) 95+ 2,593 (0.9) Sex, n (%) Female 154,565 (56.5) Male 119,027 (43.5) *Continuity of care (COC) index Mean, (SD) 0.59 (0.28) Median, (IQR) 0.57 (0.36-0.82) Number of drugs, mean (SD) 10.6 (5.89) Number of drugs, n (%) ≤5 drugs 48,210 (17.6%) 6-10 drugs 103,032 (37.7%) ≥11 drugs 122,350 (44.7%) Income quintiles, n (%) Q1 lowest income 57,053 (21.7) Q2 58,237 (22.1) Q3 52,967 (20.1) Q4 50,668 (19.2) Q5 highest income 44,653 (16.9) **RIO index, n (%) ≤40 214,443 (78.4) >40 59,149 (21.6) ***Primary care models, n (%) Fee-for-service 113,465 (68.3) Capitated+ 3,070 (1.8) Capitated 49,661 (29.9) Comorbidities, n (%) 0 CC 59,149 (21.6) 1 CC 88,411 (32.3) 2 CC 64,965 (23.7) 3 CC 34,914 (12.8) 4 CC 16,382 (6.0) 5 or more CC 9,771 (3.6) Number of primary care visits, mean (SD) 6.1 (5.77) Duration of diabetes in years, mean (SD) 9.90 (5.80) Duration of hypertension in years, mean (SD) 13.1 (5.65)

* Calculated using the Bice index ** Geographic location (≤40=non-rural; >40=rural). *** Noncapitated models include nonrostered models and those that operate on a fee-for-service basis; capitated models include family health networks and family health organizations operating on a capitation funding scheme; and the capitated+ models include family health teams and other rostered models operating on a capitated funding scheme with additional incentives for interdisciplinary care. 158

Table 4.3 presents the distribution of process of care measures and hospitalizations among

older adults with diabetes with comorbid hypertension. The proportion of patients who

received 1 or 2 HbA1c tests per year and annual eye examinations was relatively stable over

4-year time period (~45% and ~65%, respectively). The proportion of patients who received

3 or more HbA1c tests per year slightly increased form 28.5% in 2010 to 29.7% in 2014.

The proportion of patients who were prescribed ACE inhibitor therapy gradually declined

from 40.4% in 2010 to 37.7% in 2014. The proportion of patients who were prescribed oral

hypoglycemic drugs and ARB therapy was 54.2% and 22.7% in 2010, respectively, and it

remained stable over time. The proportion of patients who had high continuity of care

slightly declined from 51.0% in 2010 to 44.9% in 2014.

There was a gradual increase in the incidence of all-cause hospitalization and hospitalizations

for diabetes-related long-term complications from 15.6% and 6.3% in 2010 to 16.8% and

7.7% in 2014, respectively. There was a higher incidence of hospitalizations for diabetes-

related long-term complications (6.3%) than short-term complications (0.2%) among older

adults with diabetes with comorbid hypertension.

Table 4.3 Distribution of process measures and hospitalizations among adults with diabetes with comorbid hypertension

Year 1 Year 2 Year 3 Year 4 Measure, n (%) (n=273,592) (n=261,679) (n=250,213) (n=238,408)

Process measures, n (%)

159

Having 1 or 2 *HbA1c 124,336 (45.4) 179,651 (45.1) 114,321 (45.6) 107,929 (45.3) tests per year Having 3 or more 77,942 (28.5) 76,687 (29.3) 68,510 (37.4) 70,813 (29.7) HbA1c tests per year Annual eye 177,080 (64.7) 171,468 (65.5) 160,468 (64.1) 155,180 (65.1) examination Use of oral 148,344 (54.2) 142,181 (54.3) 135,981 (54.1) 128,216 (53.8) hypoglycemic drugs Use of **ACE 110,641 (40.4) 104,068 (39.8) 96,979 (38.8) 89,912 (37.7) inhibitors Use of ***ARBs 62,169 (22.7) 58,478 (22.4) 54,506 (21.8) 51,639 (21.7)

****COC index>0.57 139,395 (51.0) 127,785 (48.9) 118,082 (47.2) 107,072 (44.9)

Outcome measures, n (%)

All-cause 45,520 (15.6) 41,736 (16.0) 41,742 (16.7) 40.052 (16.8) hospitalizations Hospitalizations for diabetes long-term 17,311 (6.3) 17,835 (6.8) 18,555 (7.4) 18,348 (7.7) complications Hospitalizations for diabetes short-term 406 (0.2) 385 (0.2) 414 (0.2) 401 (0.2) complications *HbA1c- glycated hemoglobin ** ACE inhibitors – angiotensin-converting enzyme inhibitors ***ARBs- angiotensin II receptor blockers **** Calculated using the Bice index

The results of bivariate regression analysis (Table 4.4) showed that all process measures were

significantly associated with the likelihood of hospitalizations among older diabetes patient

with comorbid hypertension and were included in the multivariate marginal logistic models.

Table 4.4 Bivariate associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension

160

Model 2 Model 3 Model 1 Hospitalizations for Hospitalizations for All-cause diabetes short-term diabetes long-term Measure, n (%) hospitalizations complications complications Unadjusted Unadjusted Unadjusted OR (95%CI) OR (95% CI) OR (95% CI) Having *HbA1c tests No Ref. Ref. Ref. 1 or 2 HbA1c tests 0.79 (0.76-0.81) 1.05 (1.03-1.07) 1.03 (1.01-1.05) 3 or more HbA1c tests 0.75 (0.70-0.78) 1.03 (1.01-1.06) 1.08 (1.03-1.10) Eye examination No Ref. Ref. Ref. Yes 0.66 (0.65-0.67) 0.47 (0.43-0.53) 0.63 (0.62-0.65) Use of oral hypoglycemic drugs No Ref. Ref. Ref. Yes 0.78 (0.76-0.80) 1.33 (1.20-1.48) 1.36 (1.33-1.38) Use of **ACE-inhibitors No Ref. Ref. Ref. Yes 1.29 (1.28-1.31) 1.60 (1.45-1.77) 1.54 (1.51-1.57) Use of ***ARBs No Ref. Ref. Ref. Yes 1.05 (1.04-1.07) 0.88 (0.77-0.99) 1.17 (1.15-1.20) ****Continuity of care index COC≤0.49 Ref. Ref. Ref. COC>0.49 0.61 (0.58-0.63) 0.55 (0.53-0.57) 0.59 (0.57-0.61) *HbA1c- glycated hemoglobin ** ACE inhibitors – angiotensin-converting enzyme inhibitors ***ARBs- angiotensin II receptor blockers **** Calculated using the Bice index

Table 4.5 presents the results of multivariate marginal logistic models that aimed to examine

the association between process measures for care for older adults with diabetes with

comorbid hypertension and the likelihood of hospitalizations. After adjusting for other

covariates, patients who received 1 or 2, or 3 or more HbA1c tests were significantly less

likely to have all-cause hospitalizations (AOR=0.90, 95% CI 0.88-0.92, and AOR=0.84,

95% CI 0.82-0.86, respectively), compared to those who did not complete the recommended

161

HbA1c testing. After adjusting for other covariates, patients who received 1 or 2, or 3 or more HbA1c tests were significantly more likely to have hospitalizations for diabetes-related long-term complications (AOR=1.07, 95% CI 1.03-1.11, and AOR=1.13, 95% CI 1.08-1.19, respectively), compared to those who did not complete the recommended HbA1c testing.

There was no significant relationship found between the receipt of recommended frequency of HbA1c testing and the likelihood of hospitalizations for diabetes-related short-term complications.

After adjusting for other covariates, there was observed a strong protective association between having annual eye examination and the likelihood of all-cause hospitalizations, including those for diabetes-related short-term and long-term complications among older diabetes patients with comorbid hypertension (AOR=0.85, 95% CI 0.83-0.87, AOR=0.61,

95% CI 0.52-0.71, and AOR=0.80, 95% CI 0.78-0.82, respectively).

After adjusting for other covariates, patients who were prescribed ACE inhibitors were significantly more likely to have an episode of all-cause hospitalizations as well as hospitalizations for diabetes-related short-term and long-term complications (AOR=1.04,

95% CI 1.02-1.06, AOR=1.21, 95% CI 1.02-1.44, and AOR=1.13, 95% CI 1.11-1.17, respectively), compared to patients who were not prescribed.

After adjusting for other covariates, patients who were prescribed ARB therapy were significantly less likely to be hospitalized for all-cause hospitalizations (AOR=0.93, 95% CI

162

0.92-0.95), compared to patients who did not use ARBs. There was no association found between ARB therapy and hospitalizations for diabetes-related complications.

After adjusting for other covariates, there was observed a protective association between having high continuity of care (COC>0.57) and the likelihood of all-cause hospitalizations, hospitalizations for diabetes-related short-term and long-term complications (AOR=0.70,

95% CI 0.69-0.72, AOR=0.62, 95% CI 0.58-0.67, and AOR=0.69, 95% CI 0.67-0.72, respectively). After adjusting for other variables, increase in number of prescribed drugs was associated with increased odds of both all-cause hospitalizations, and hospitalizations for diabetes-related short-term and long-term complications (AOR=1.06, 95% CI 1.04-1.07,

AOR=1.04, 95% CI 1.02-1.05, and AOR=1.08, 95% CI 1.06-1.11, respectively).

Table 4.5 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension, adjusted for other covariates

Characteristic Model 1 Model 2 Model 3 All-cause Hospitalizations for Hospitalizations for hospitalisations diabetes short-term diabetes long-term AOR (95% CI) complications complications AOR (95% CI) AOR (95% CI) Having *HbA1c tests No Ref. Ref. Ref. 1 or 2 HbA1c tests 0.90 (0.88-0.92) 1.07 (0.86-1.17) 1.07 (1.03-1.11) 3 or more HbA1c tests 0.84 (0.82-0.86) 1.04 (0.82-1.18) 1.13 (1.08-1.19) Eye examination

163

No Ref. Ref. Ref. Yes 0.85 (0.84-0.87) 0.61 (0.52-0.71) 0.80 (0.78-0.82) Use of oral hypoglycemic drugs No Ref. Ref. Ref. Yes 0.88 (0.86-0.90) 1.11 (0.92-1.34) 1.30 (1.27-1.34) Use of **ACE-inhibitors No Ref. Ref. Ref. Yes 1.04 (1.02-1.06) 1.21 (1.02-1.44) 1.13 (1.11-1.17) Use of ***ARBs No Ref. Ref. Ref. Yes 0.93 (0.92-0.95) 1.01 (0.83-1.12) 1.03 (0.98-1.06) ****COC index COC≤0.57 Ref. Ref. Ref. COC>0.57 0.70 (0.69-0.72) 0.62 (0.58-0.67) 0.69 (0.67-0.72) Number of drugs 1.06 (1.04-1.07) 1.04 (1.02-1.05) 1.08 (1.06-1.11) Age 1.04 (1.03-1.05) 1.00 (0.99-1.02) 1.04 (1.03-1.05) Sex Female Ref. Ref. Ref. Male 1.40 (1.36-1.44) 0.88 (0.74-1.03) 1.40 (1.36-1.44) Income quintiles Q1 lowest income Ref. Ref. Ref. Q2 0.93 (0.90-0.97) 0.72 (0.57-0.90) 0.92 (0.90-0.96) Q3 0.95 (0.90-0.99) 0.95 (0.76-1.19) 0.92 (0.88-0.99) Q4 0.89 (0.83-0.93) 0.80 (0.61-1.14) 0.87 (0.83-0.90) Q5 highest income 0.87 (0.82-0.92) 0.62 (0.47-0.82) 0.86 (0.82-0.94) *****RIO index ≤40 Ref. Ref. Ref. >40 1.14 (1.09-1.19) 0.88 (0.67-1.12) 1.14 (1.08-1.19) Duration of diabetes 1.03 (1.01-1.05) 1.12 (1.09-1.13) 1.07 (1.06-1.08) Duration of hypertension 1.02 (1.01-1.03) 0.98 (0.97-0.99) 1.00 (0.98-1.01) Number of primary care 1.02 (1.0-1.04) 0.98 (0.97-1.00) 0.99 (0.98-1.00) visits ******Primary care models Capitated+ Ref. Ref. Ref. Fee-for-service 0.77 (0.76-0.79) 1.05 (0.88-1.26) 0.75 (0.73-0.76) Capitated 1.09 (1.02-1.13) 0.79 (0.51-1.28) 1.03 (0.92-1.14) Comorbidities 0 CC Ref. Ref. Ref. 1 CC 1.17 (1.13-1.22) 1.08 (0.82-0.36) 1.17 (1.12-1.22) 2 CC 1.37 (1.33-1.40) 0.84 (0.65-1.09) 1.36 (1.31-1.43) 3 CC 1.65 (1.58-1.70) 1.10 (0.82-1.32) 1.64 (1.56-1.72) 4 CC 2.00 (1.89-2.12) 1.59 (1.22-1.85) 2.01 (1.90-2.13) 5 or more CC 2.32 (2.16-2.44) 1.13 (0.71-1.32) 2.31 (2.15-2.40) *HbA1c- glycated hemoglobin **ACE inhibitors – angiotensin-converting enzyme inhibitors

164

***ARBs- angiotensin II receptor blockers **** Calculated using the Bice index ***** Geographic location (≤40=non-rural; >40=rural). ****** Noncapitated models include nonrostered models and those that operate on a fee-for-service basis; capitated models include family health networks and family health organizations operating on a capitation funding scheme; and the capitated+ models include family health teams and other rostered models operating on a capitated funding scheme with additional incentives for interdisciplinary care.

4.3.1.2 Quality of overall care and hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

The cohort of older adults with diabetes with comorbid hypertension and chronic ischemic heart disease contained 141,947 patients. The baseline characteristics of patients are presented in Table 4.6. About 85% of people with comorbid hypertension were aged between

65 and 84 years. Over half were female (57.8%). Nearly all people (92.3%) resided in non- rural areas and the majority were enrolled in non-capitated primary care models (63.7%).

The median value of the continuity of care (COC) index was 0.49 (IQR=0.44). The mean number of drugs prescribed in older adults with diabetes with comorbid hypertension and chronic ischemic heart disease was 13.4 (SD=6.52). About 64.4% of patients were prescribed

11 or more medications. About 47.5% of older adults with diabetes with comorbid hypertension and chronic ischemic heart disease had 1 or 2 other comorbid conditions, while about 41% had 3 or more conditions.

165

Table 4.6 Baseline characteristics of older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

Diabetes with comorbid Characteristic hypertension and chronic ischemic heart disease Number of individuals 141,947 Age in years, mean (SD) 77.4 (7.12) Age in groups, n (%) 65 – 74 54,593 (38.4) 75 – 84 61.883 (43.6) 85 – 94 23,950 (16.9) 95+ 1,521 (1.1) Sex, n (%) Female 81,987 (57.8) Male 59,960 (42.2) *Continuity of care (COC) index Mean, (SD) 0.51 (0.27) Median, (IQR) 0.49 (0.29-0.73) Number of drugs, mean (SD) 13.4 (6.52) Number of drugs, n (%) ≤5 drugs 10,924 (7.7%) 6-10 drugs 39,583 (27.9%) ≥11 drugs 91,440 (64.4%) Income quintiles, n (%) Q1 lowest income 29,478 (22.0) Q2 29,496 (22.0) Q3 26,765 (20.0) Q4 25,649 (19.10) Q5 highest income 22,657 (16.90) **RIO index, n (%) ≤40 131,065 (92.3) >40 10,882 (7.7) ***Primary care models, n (%) Fee-for-service 67,557 (63.7) Capitated+ 1,638 (1.8) Capitated 36,608 (34.5) Comorbidities, n (%) 0 CC 15,859 (11.2) 1 CC 33,105 (23.3)

166

2 CC 34,350 (24.2) 3 CC 26,547 (18.7) 4 CC 16,972 (12.0) 5 or more CC 15,114 (10.7) Number of primary care visits, mean (SD) 7.6 (6.99)

Duration of diabetes, mean (SD) 10.7 (6.02) Duration of hypertension, mean (SD) 13.8 (5.44) Duration of chronic ischemic heart disease, mean (SD) 7.13 (2.68)

* Calculated using the Bice index ** Geographic location (≤40=non-rural; >40=rural). *** Noncapitated models include nonrostered models and those that operate on a fee-for-service basis; capitated models include family health networks and family health organizations operating on a capitation funding scheme; and the cap itated+ models include family health teams and other rostered models operating on a capitated funding scheme with additional incentives for interdisciplinary care.

Table 4.7 presents the distribution of process measures and hospitalizations among older

adults with diabetes with comorbid hypertension and chronic ischemic heart disease.

Proportion of patients who received 1 or 2 HbA1c tests per year and annual eye examination

was relatively stable over 4-year time period (~43% and 65%, respectively). The proportion

of patients who were prescribed ACE inhibitor and ARB therapy gradually declined from

48.8% and 23.3% in 2010 to 45.0% and 21.8% in 2014, respectively.

The proportion of patients who were prescribed oral hypoglycemic drugs and statins was

~51.0% and ~79.0%, respectively, over the 4-year study period. The proportion of patients

who were prescribed antiplatelet therapy gradually declined from 24.6% in 2010 to 22.2% in

2014, respectively. The proportion of patients who had greater continuity of care gradually

declined from 49.1% in 2010 to 45.2% in 2014.

167

The incidence of hospitalizations among older adults with diabetes with comorbid

hypertension and chronic ischemic heart disease was relatively stable over the 4-year study

period. In particular, the incidence of all-cause hospitalizations was ~25.0%, while incidence

of hospitalizations for diabetes-related long-term and short-term complications was ~15.0%

and 0.2%, respectively.

Table 4.7 Distribution of process of care and outcome measures among older adults with diabetes comorbid with hypertension and chronic ischemic heart disease

Year 1 Year 2 Year 3 Year 4 Measure (n=141,947) (n=132,096) (n=122,980) (n=113,947)

Process measures, n (%)

Having 1 or 2 *HbA1c 61,505 (43.3) 57,278 (43.4) 54,260 (44.1) 49,768 (43.7) tests per year Having 3 or more 42,194 (29.7) 39,926 (30.2) 34,720 (28.2) 35,102 (30.8) HbA1c tests per year Annual eye 92,623 (65.3) 86,919 (65.8) 78,767 (64.1) 73,850 (64.8) examination Use of oral 72,686 (51.2) 67,348 (51.0) 62,156 (50.5) 56,992 (50.0) hypoglycemic drugs Use of **ACE-inhibitors 69,296 (48.8) 63,101 (47.8) 56,988 (46.3) 51,256 (45.0)

Use of ***ARBs 32,997 (23.3) 30,053 (22.8) 27,152 (22.1) 24,872 (21.8)

Use of antiplatelet 34,868 (24.6) 31,574 (23.9) 28,411 (23.1) 25,267 (22.2) drugs Use of statins 12,845 (79.5) 104,650 (79.2) 97.253 (79.1) 89,927 (78.9)

****COC index >0.49 69,660 (49.1) 64,230 (48.6) 58,284 (47.4) 51,496 (45.2)

Outcome measures, n (%)

All-cause 35,157 (24.8) 32,677 (24.7) 32,677 (25.0) 27,999 (24.6)

168 hospitalizations Hospitalizations for diabetes long-term 20,629 (14.5) 19,152 (14.5) 19,152 (15.0) 17,438 (15.3) complications Hospitalizations for diabetes short-term 263 (0.2) 237 (0.2) 218 (0.2) 234 (0.2) complications *HbA1c- glycated hemoglobin ** ACE inhibitors – angiotensin-converting enzyme inhibitors ***ARBs- angiotensin II receptor blockers **** Calculated using the Bice index

The results of bivariate regression analysis (Table 4.8) showed that all process measures were

significantly associated with the likelihood of hospitalizations and were included in the

multivariate marginal logistic models.

Table 4.8 Bivariate associations between process of care measures and hospitalizations, among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

Hospitalizations for Hospitalizations for All-cause diabetes short-term diabetes long-term hospitalizations Measure, n (%) complications complications Unadjusted Unadjusted Unadjusted OR (95%CI) OR (95% CI) OR (95% CI) Having *HbA1c tests No Ref. Ref. Ref. 1 or 2 HbA1c tests 0.77 (0.75-0.78) 0.89 (0.75-0.95) 1.04 (1.02-1.07) 3 or more HbA1c tests 0.71 (0.69-0.72) 0.92 (0.81-1.02) 1.06 (1.04-1.08) Eye examination No Ref. Ref. Ref. Yes 0.70 (0.69-0.71) 0.55 (0.48-0.63) 0.73 (0.71- 0.74) Use of oral hypoglycemic drugs

169

No Ref. Ref. Ref. Yes 0.82 (0.81-0.84) 1.13 (1.03-1.29) 1.34 (1.32-1.37) Use of **ACE-inhibitors No Ref. Ref. Ref. Yes 1.11 (1.08-1.15) 1.20 (1.05-1.37) 1.13 (1.10-1.16) Use of ***ARBs No Ref. Ref. Ref. Yes 1.09 (1.07-1.11) 1.06 (0.90-1.23) 1.17 (1.14-1.19) Use of antiplatelet drugs No Ref. Ref. Ref. Yes 1.18 (1.16-1.20) 1.41 (1.22-1.62) 2.11 (2.07-2.15) Use of statins No Ref. Ref. Ref. Yes 0.94 (0.92-0.96) 1.00 (0.99-1.02) 1.03 (1.01-1.06) ****COC index COC≤0.49 Ref. Ref. Ref. COC>0.49 0.65 (0.62-0.68) 0.60 (0.57-0.63) 0.66 (0.62-0.69) *HbA1c- glycated hemoglobin ** ACE inhibitors – angiotensin-converting enzyme inhibitors ***ARBs- angiotensin II receptor blockers **** Calculated using the Bice index

Table 4.9 presents the results of multivariate marginal logistic models that aimed to examine

the association between process measures and the likelihood of hospitalizations among older

adults with diabetes with comorbid hypertension and chronic ischemic heart disease. After

adjusting for other covariates, patients who received 1 or 2, or 3 or more HbA1ctests were

significantly less likely to have all-cause hospitalizations (AOR=0.88, 95% CI 0.85-0.91, and

AOR=0.86, 95% CI 0.83-0.88, respectively), but were significantly more likely to be

hospitalized for diabetes-related long-term complications (AOR=1.10, 95% CI 1.06-1.14,

and AOR=1.21, 95% CI 1.16-1.26, respectively), compared to those who did not complete

the recommended HbA1c testing. There was no association between the receipt of

recommended frequency of HbA1c testing and the likelihood of hospitalizations for diabetes-

related short-term complications.

170

After adjusting for other covariates, a strong protective association was observed between having annual eye examination and the likelihood of all-cause hospitalizations, as well as hospitalizations for diabetes-related short-term and long-term complications (AOR=0.90,

95% CI 0.88-0.92, AOR=0.55, 95% CI 0.45-0.68, and AOR=0.86, 95% CI 0.82-0.93, respectively).

Patients who were prescribed oral hypoglycemic drugs had significantly lower odds of all- cause hospitalizations (AOR=0.88, 95% CI 0.86-0.90) and significantly higher odds of hospitalizations for diabetes-related long-term complications (AOR=1.40, 95% CI 1.35-

1.43), compared to those who were not.

After adjusting for other covariates, the odds of all-cause hospitalizations and hospitalizations for diabetes-related long-term complications was significantly higher in patients who were prescribed ACE inhibitors (AOR=1.03, 95% CI 1.01-1.05 and

AOR=1.03, 95% CI 1.02-1.07, respectively), compared to those who were not. There was no association between ARB therapy and the likelihood of diabetes-related hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease. After adjusting for other covariates, the odds of all-cause hospitalizations and hospitalizations for diabetes-related long-term complications was significantly higher in patients who were prescribed antiplatelet drugs (AOR=1.08, 95% CI 1.06-1.1 and

AOR=1.14, 95% CI 1.10-1.18, respectively), compared to those who were not.

171

After adjusting for other covariates, there was observed a strong protective association between having high continuity of care (COC>0.49) and the likelihood of all-cause hospitalizations, hospitalizations for diabetes short-term and long-term complications

(AOR=0.74, 95% CI 0.72-0.77, AOR=0.70, 95% CI 0.57-0.87, and AOR=0.76, 95% CI

0.73-0.78, respectively). After adjusting for other variables, the odds of all-cause hospitalizations, hospitalizations for diabetes-related short-term and long-term complications increased with each unit increase in number of prescribed drugs (AOR=1.05, 95% CI 1.02-

1.07, AOR=1.07, 95% CI 1.04-1.08, and AOR=1.06, 95% CI 1.05-1.07, respectively).

Table 4.9 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease, adjusted for other covariates

Model 3 Model 1 Model 2 Hospitalizations for All-cause Hospitalizations for Characteristic diabetes long-term hospitalisations diabetes short-term complications AOR (95% CI) complications AOR (95% CI) AOR (95% CI)

Having *HbA1c tests No Ref. Ref. Ref. 1 or 2 HbA1c tests 0.88 (0.85-0.91) 1.03 (0.78-1.24) 1.10 (1.06-1.14) 3 or more HbA1c tests 0.86 (0.83-0.88) 1.06 (0.78-1.26) 1.21 (1.16-1.26) Eye examination No Ref. Ref. Ref. Yes 0.90 (0.88-0.92) 0.55 (0.45-0.68) 0.86 (0.82-0.93) Use of oral hypoglycemic drugs No Ref. Ref. Ref. Yes 0.88 (0.86-0.90) 1.14 (0.90-1.32) 1.40 (1.35-1.43) Use of **ACE inhibitors No Ref. Ref. Ref. Yes 1.03 (1.01-1.05) 1.14 (0.91-1.32) 1.03 (1.02-1.07) Use of ***ARBs No Ref. Ref. Ref. Yes 0.98 (0.96-1.01) 0.98 (0.76-1.24) 1.02 (0.99-1.06)

172

Use of antiplatelet drugs No Ref. Ref. Ref. Yes 1.08 (1.06-1.11) 0.92 (0.73-1.18) 1.14 (1.10-1.18) Use of statins No Ref. Ref. Ref. Yes 0.89 (0.86-0.92) 0.82 (0.63-1.08) 0.98 (0.96-0.99) ****COC index COC>0.49 Ref. Ref. Ref. COC≤0.49 0.74 (0.72-0.77) 0.70 (0.57-0.87) 0.76 (0.73-0.78) Number of drugs 1.05 (1.02-1.07) 1.07 (1.04-1.08) 1.06 (1.05-1.07) Age 1.03 (1.02-1.04) 0.99 (0.97-1.01) 1.02 (1.01-1.03) Sex Female Ref. Ref. Ref. Male 1.15 (1.12-1.18) 0.90 (0.72-1.11) 1.22 (1.19-1.26) Income quintiles Q1 lowest income Ref. Ref. Ref. Q2 0.99 (0.97-1.03) 1.07 (0.80-1.35) 0.94 (0.90-0.98) Q3 1.03 (0.99-1.07) 0.98 (0.77-1.33) 0.92 (0.89-0.98) Q4 1.05 (0.98-1.09) 0.88 (0.64-1.20) 0.91 (0.87-0.95) Q5 highest income 1.04 (0.95-1.07) 0.85 (0.60-1.20) 0.90 (0.85-0.94) *****RIO index ≤40 Ref. Ref. Ref. >40 1.16 (1.12-1.20) 0.91 (0.59-1.42) 1.10 (1.04-1.15) Duration of diabetes 1.02 (1.01-1.03) 1.10 (1.08-1.12) 1.06 (1.04-1.09) Duration of hypertension 1.01 (1.00-1.03) 0.98 (0.96-0.99) 1.01 (1.00-1.03) Duration of ischemic 1.01 (1.00-1.02) 0.96 (0.92-0.99) 1.02 (1.00-1.04) heart disease Number of primary care 1.01 (1.00-1.03) 0.99 (0.98-1.01) 1.00 (0.99-1.02) visits ******Primary care models Capitated+ Ref. Ref. Ref. Fee-for-service 0.78 (0.76-0.80) 1.09 (0.86-1.35) 0.75 (0.72-0.77) Capitated 1.08 (0.99-1.13) 0.41 (0.11-0.66) 1.08 (0.97-1.21) Comorbidities 0 CC Ref. Ref. Ref. 1 CC 1.21 (1.16-1.27) 1.09 (0.70-1.38) 1.15 (1.09-1.22) 2 CC 1.43 (1.37-1.48) 1.23 (0.80-1.49) 1.41 (1.34-1.49) 3 CC 1.69 (1.61-1.75) 1.32 (0.85-1.58) 1.69 (1.59-1.79) 4 CC 1.98 (1.89-2.09) 1.40 (0.99-1.64) 2.09 (1.96-2.22) 5 or more CC 2.27 (2.15-2.35) 1.23 (0.93-1.44) 2.30 (2.15-2.40)

*HbA1c- glycated hemoglobin **ACE inhibitors – angiotensin-converting enzyme inhibitors ***ARBs- angiotensin II receptor blockers **** Calculated using the Bice index ***** Geographic location (≤40=non-rural; >40=rural).

173

****** Noncapitated models include nonrostered models and those that operate on a fee-for-service basis; capitated models include family health networks and family health organizations operating on a capitation funding scheme; and the capitated+ models include family health teams and other rostered models operating on a capitated funding scheme with additional incentives for interdisciplinary care.

4.3.2 Quality of overall care and hospitalizations among older adults with diabetes with discordant comorbid conditions

4.3.2.1 Quality of overall care and hospitalizations among older adults with diabetes with comorbid osteoarthritis and without major depression

The cohort of older adults with diabetes with comorbid osteoarthritis included 255,214 patients. The baseline characteristics of patients are presented in Table 4.10. About 84% of people with comorbid osteoarthritis were aged between 65 and 84 years. Slightly over half of the people were female (54.8%). Nearly all people resided in non-rural areas (93%) and the majority were enrolled in non-capitated primary care models (69.2%).

The median value of the continuity of care (COC) index among older adults with diabetes with comorbid osteoarthritis was 0.53 (IQR=0.38). The mean duration of diabetes and osteoarthritis were 10.0 (SD=5.88) years and 7.17 (SD=2.57) years, respectively. The mean number of drugs prescribed in older adults with diabetes with comorbid osteoarthritis was

12.1 (SD=6.42). About 55.2% of people were prescribed 11 or more medications. About

49.4% of older diabetes patient with comorbid osteoarthritis had 1 or 2 other comorbid chronic conditions, while about 45.9% of patients had 3 or more comorbid conditions.

174

Table 4.10 Baseline characteristics of older adults with diabetes with comorbid osteoarthritis

Diabetes with comorbid Characteristic osteoarthritis and without major depression Number of individuals 255,214 Age in years, mean (SD) 76.6 (7.24) Age in groups, n (%) 65 – 74 112,046 (43.9) 75 – 84 102,717 (40.2) 85 – 94 37,900 (14.9) 95+ 2,551 (1.0) Sex, n (%) Female 139,951 (54.8) Male 115,263 (45.2) *Continuity of care (COC) index Mean, (SD) 0.55 (0.26) Median, (IQR) 0.53 (0.32-0.77) Number of drugs, mean (SD) 12.1 (6.42) Number of drugs, n (%) ≤5 drugs 33,768 (13.2%) 6-10 drugs 80,695 (31.6%) ≥11 drugs 140,751 (55.2%) Income quintiles, n (%) Q1 lowest income 53,174 (21.6) Q2 53,884 (22.0) Q3 48,922 (20.0) Q4 47,143 (19.3) Q5 highest income 41,855 (17.1) **RIO index, n (%) ≤40 237,312 (93.0) >40 17,.902 (7.0) ***Primary care models, n (%) Fee-for-service 108,522 (69.2) Capitated+ 45,000 (29.0) Capitated 2,818 (1.8) Comorbidities, n % 0 CC 12,061 (4.7%) 1 CC 58,547 (22.9%) 2 CC 67,635 (26.5%) 3 CC 50,641 (19.8%) 4 CC 32,778 (12.8%) 5 or more CC 33,552 (13.3%)

175

Number of primary care visits, mean (SD) 7.34 (6.60) Duration of diabetes in years, mean (SD) 10.0 (5.88) Duration of osteoarthritis in years, mean (SD) 7.17 (2.57)

incentives for interdisciplinary care.

Table 4.11 presents the distribution of process measures and hospitalizations among older adults with diabetes with comorbid osteoarthritis. The proportion of patients who received 1 or 2 HbA1c tests per year and annual eye examination was relatively stable over 4-year time period (~45% and 67%, respectively). The proportion of patients who were prescribed

NSAIDs gradually declined from 20.8% in 2010 to 17.9% % in 2014. The proportion of patients who were prescribed oral hypoglycemic drugs was ~51.0%, over time 4-year study period. The proportion of patients who had high continuity of care (COC>0.53) gradually declined from 50.0% in 2010 to 44.8% in 2014.

The incidence of all-cause hospitalizations and hospitalisations for diabetes short-term complications was relatively stable (~20.0% and 0.2%, respectively), while the incidence of hospitalizations for diabetes-related long-term complications gradually increased from 8.8% in 2010 to 9.9% in 2014.

176

Table 4.11 Distribution of process of care and outcome measures among older adults with diabetes with comorbid osteoarthritis

Year 1 Year 2 Year 3 Year 4 Measure (n=255,214) (n=242,750) (n=230,800) (n=185,979)

Process measure, n (%)

Having 1 or 2 *HbA1c tests 114,746 (45.0) 109,027 (44.9) 105,108 (45.5) 98,970 (45.3) per year Having 3 or more HbA1c 72,469 (28.4) 70,180 (28.9) 62,525 (27.1) 64,450 (29.5) tests Eye examination annually 171,803 (67.3) 164,860 (67.9) 153,184 (66.4) 146,863 (67.2)

Use of oral hypoglycemic 130,599 (51.2) 124,474 (51.3) 180,601 (51.2) 111,234 (50.8) drugs Use of **NSAIDs– “negative” 52,952 (20.8) 47,0.22 (19.4) 43,136 (18.7) 39,288 (17.9)

***COC index>0.53 127,617 (50.0) 117,605 (48.5) 108,292 (46.9) 97,814 (44.8)

Outcome measure, n (%)

All-cause hospitalizations 49,873 (19.5) 48,023 (19.8) 47,109 (20.4) 43,944 (20.1)

Hospitalizations for diabetes 363 (0.1) 333 (0.1) 345 (0.2) 345 (0.2) short-term complications Hospitalizations for diabetes 22,454 (8.8) 22,178 (9.2) 22,379 (9.7) 21,703 (9.9) long-term complications *HbA1c- glycated hemoglobin ** NSAID- non-steroidal anti-inflammatory drugs *** Calculated using the Bice index

The results of bivariate regression analysis (Table 4.12) showed that all process measures

were significantly associated with the likelihood of hospitalizations among older adults with

diabetes with comorbid osteoarthritis and were included in the multivariable marginal

logistic models.

177

Table 4.12 Bivariate associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid osteoarthritis

Model 2 Model 2 Model 1 Hospitalizations for Hospitalizations for All-cause diabetes long-term diabetes short-term Measure hospitalisations complications complications Unadjusted Unadjusted Unadjusted OR (95% CI) OR (95% CI) OR (95% CI) *HbA1c testing annually No Ref. Ref. Ref. 1 or 2 HbA1c tests 0.77 (0.76-0.78) 1.05 (1.02-1.08) 0.85 (0.74-0.96) 3 or more HbA1c tests 0.72 (0.71-0.73) 1.10 (1.07-1.12) 0.98 (0.85-1.22) Annual eye examination No Ref. Ref. Ref. Yes 0.68 (0.67-0.69) 0.69 (0.68-0.70) 0.50 (0.45-0.56) Use of oral hypoglycemic drugs No Ref. Ref. Ref. Yes 0.87 (0.85-0.88) 1.41 (1.39-1.43) 1.38 (1.23-1.54) Use of **NSAIDs No Ref. Ref. Ref. Yes 0.99 (0.97-0.99) 0.90 (0.88-0.92) 0.72 (0.62-0.84) ***COC index COC≤0.53 Ref. Ref. Ref. COC>0.53 0.61 (0.59-0.64) 0.62 (0.58-0.64) 0.58 (0.55-0.61) *HbA1c- glycated hemoglobin ** NSAID- non-steroidal anti-inflammatory drugs *** Calculated using the Bice index

Table 4.13 presents the results of multivariate marginal logistic models that aimed to

examine the association between process measures for care for older adults with diabetes

with comorbid osteoarthritis and the likelihood of hospitalizations. After adjusting for other

covariates, patients who received 1 or 2, or 3 or more HbA1c tests were significantly less

likely to have an episode of all-cause hospitalization (AOR=0.88, 95% CI 0.86-0.90, and

AOR=0.83, 95% CI 0.81-0.85, respectively), but were significantly more likely to be

178 hospitalized for diabetes-related long-term complications (AOR=1.12, 95% CI 1.09-1.15, and AOR=1.19, 95% CI 1.15-1.23, respectively), compared to those who did not complete the recommended HbA1c testing. There was no association between the receipt of recommended frequency of HbA1c testing and the likelihood of hospitalizations for diabetes- related short-term complications among older diabetes patients with comorbid osteoarthritis.

After adjusting for other covariates, there was observed a strong protective association between having annual eye examination and the likelihood of all-cause hospitalizations, as well as hospitalizations for diabetes-related short-term and long-term complications

(AOR=0.89, 95% CI 0.86-0.92, AOR=0.57, 95% CI 0.48-0.64, and AOR=0.88, 95% CI

0.86-0.90, respectively).

After adjusting for other covariates, patients who were prescribed oral hypoglycemic drugs were significantly less likely to be hospitalized for all-cause hospitalizations (AOR=0.89,

95% CI 0.87-0.91), but significantly more likely to be hospitalized for diabetes-related long- term complications (AOR=1.47, 95% CI 1.42-1.51), compared to patients who were not.

There was no association between use of oral hypoglycemic drugs and hospitalizations for diabetes-related short-term complications.

After adjusting for other covariates, there was observed a strong protective association between having high continuity of care (COC>0.53) and the likelihood of all-cause hospitalizations, hospitalization for diabetes-related short-term and long-term complications

(AOR=0.73, 95% CI 0.72-0.74, AOR=0.71, 95% CI 0.60-0.84, and AOR=0.74, 95% CI

179

0.72-0.76, respectively). After adjusting for other variables, an increase in each unit of

prescribed drugs was associated with increased odds of both all-cause hospitalizations, and

hospitalisations for diabetes-related short-term and long-term complications (AOR=1.06,

95% CI 1.04-1.08, AOR=1.05, 95% CI 1.03-1.07, and AOR=1.07, 95% CI 1.06-1.08,

respectively).

Table 4.13 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid osteoarthritis, adjusted for other covariates

Model 1 Model 2 Model 2 All-cause Hospitalizations for Hospitalizations for Measure hospitalisations diabetes long-term diabetes short-term AOR (95% CI) hospitalisations complications AOR (95% CI) AOR (95% CI) *HbA1c testing annually No Ref. Ref. Ref. 1 or 2 HbA1c tests 0.88 (0.86-0.90) 1.12 (1.09-1.15) 1.07 (0.84-1.20) 3 or more HbA1c tests 0.83 (0.81-0.85) 1.19 (1.15-1.23) 1.04 (0.80-1.26) Annual eye examination No Ref. Ref. Ref. Yes 0.89 (0.86-0.92) 0.88 (0.86-0.90) 0.57 (0.48-0.67) Use of oral hypoglycemic drugs No Ref. Ref. Ref. Yes 0.89 (0.87-0.91) 1.47 (1.42-1.51) 1.21 (0.95-1.39) Use of **NSAIDs No Ref. Ref. Ref. Yes 0.92 (0.89-0.93) 0.82 (0.79-0.84) 0.65 (0.52-0.82) ***COC index COC≤0.53 Ref. Ref. Ref. COC>0.53 0.73 (0.72-0.74) 0.74 (0.72-0.76) 0.71 (0.60-0.84) Number of prescribed drugs 1.06 (1.04-1.08) 1.07 (1.06-1.08) 1.05 (1.03-1.07) Age 1.03 (1.02-1.04) 1.03 (1.02-1.04) 0.99 (0.98-1.01) Sex Female Ref. Ref. Ref. Male 1.22 (1.20-1.24) 1.36 (1.33-1.39) 0.88 (0.74-1.04)

180

Income quintiles Q1 lowest income Ref. Ref. Ref. Q2 1.02 (0.96-1.05) 0.94 (0.90-0.99) 0.95 (0.74-1.21) Q3 0.97 (0.94-0.99) 0.98 (0.94-1.02) 1.10 (0.86-1.40) Q4 0.97 (0.94-0.99) 0.94 (0.90-0.98) 1.06 (0.82-1.37) Q5 highest income 1.48 (1.40-1.56) 0.95 (0.92-0.99) 0.70 (0.53-0.93) ****RIO index ≤40 Ref. Ref. Ref. >40 1.19 (1.16-1.24) 1.15 (1.09-1.20) 0.83 (0.57-1.28) Duration of diabetes 1.02 (1.01-1.03) 1.06 (1.05-1.07) 1.10 (1.08-1.12) Duration of osteoarthritis 0.99 (0.97-1.01) 0.98 (0.96-0.99) 0.98 (0.95-1.01) Number of primary care visits 1.02 (1.01-1.03) 1.01 (0.99-1.02) 0.99 (0.98-1.01) *****Primary care models Capitated+ Ref. Ref. Ref. Fee-for-service 0.77 (0.76-0.78) 0.73 (0.71-0.75) 1.01 (0.84-1.23) Capitated 1.04 (1.02-1.06) 0.92 (0.79-1.27) 0.87 (0.51-1.14) Comorbidities 0 CC Ref. Ref. Ref. 1 CC 1.10 (1.04-1.15) 1.30 (1.18-1.42) 1.08 (0.85-1.40) 2 CC 1.26 (1.19-1.32) 1.70 (1.55-1.87) 0.95 (0.57-1.45) 3 CC 1.48 (1.40-1.56) 2.27 (2.10-2.38) 1.19 (0.77-1.51) 4 CC 1.77 (1.68-1.86) 2.90 (2.69-3.10) 1.32 (0.99-1.77) 5 or more CC 2.12 (1.60-1.46) 3.91 (3.60-4.30) 1.30 (0.91-0.75) *HbA1c- glycated hemoglobin ** NSAID- non-steroidal anti-inflammatory drugs *** Calculated using the Bice index **** Geographic location (≤40=non-rural; >40=rural). ***** Noncapitated models include nonrostered models and those that operate on a fee-for-service basis; capitated models include family health networks and family health organizations operating on a capitation funding scheme; and the capitated+ models include family health teams and other rostered models operating on a capitated funding scheme with additional incentives for interdisciplinary care.

4.3.2.2 Quality of overall care and hospitalizations among older adults with diabetes with comorbid osteoarthritis and major depression

The cohort of older adults with diabetes with comorbid osteoarthritis and major depression

contained 2,444 individuals. The baseline characteristics of patients are presented in Table

4.14. About 86% of older adults with diabetes with comorbid osteoarthritis and major

181 depression were aged between 65 and 84 years. Slightly over half of the people were female

(63.2%).

Nearly all people with diabetes (93.8%) resided in non-rural areas and the majority were enrolled in non-capitated primary care models (67.8%). The median value of the continuity of care (COC) index was 0.36 (IQR=0.38). The mean number of drugs prescribed in older adults with diabetes with comorbid osteoarthritis and major depression was 17.1 (SD=7.6).

About 76.6% of patients were prescribed 11 or more medications. About 34% of older diabetes patients with comorbid osteoarthritis and major depression had 1 or 2 other comorbid chronic conditions, while about 62.9% of patients had 3 or more comorbid conditions.

Table 4.14 Baseline characteristics of older diabetes patients with comorbid osteoarthritis and major depression

Diabetes with comorbid Characteristic osteoarthritis and major depression Number of individuals 2,444 Age in years, mean (SD) 75.7 (7.12) Age in groups, n (%) 65 – 74 1,194 (48.9) 75 – 84 906 (37.1) 85 – 94 333 (13.6) 95+ 11 (0.4) Sex, n (%) Female 1,545 (63.2) Male 899 (36.8) *Continuity of care (COC) index Mean, (SD) 0.42 (0.26)

182

Median, (IQR) 0.36 (0.21-0.59) Number of drugs, mean (SD) 17.1 (7.6)

Number of drugs, n (%) ≤5 drugs 136 (5.7%) 6-10 drugs 433 (17.7%)

≥11 drugs 1,875 (76.6%) Income quintiles, n (%)

Q1 lowest income 589 (26.1) Q2 504 (22.3) Q3 414 (18.4)

Q4 360 (15.0) Q5 highest income 388 (17.2) ** RIO index, n (%) ≤40 2,293 (93.8)

>40 151 (6.2) *** Primary care models, n (%) Fee-for-service 982 (67.8)

Capitated+ 45 (3.1)

Capitated 421 (29.1) Comorbidities, n % 0 CC 77 (3.1%)

1 CC 335 (13.7%) 2 CC 495 (20.3%)

3 CC 490 (20.1%) 4 CC 428 (17.5%) 5 or more CC 619 (25.3%)

Number of primary care visits, mean (SD) 7.8 (7.4) Duration of diabetes, mean (SD) 10.3 (6.01)

Duration of major depression, mean (SD) 3.3 (1.62) Duration of osteoarthritis, mean (SD) 7.4 (2.61)

scheme with additional incentives for interdisciplinary care.

183

Table 4.15 presents the distribution of process measures and hospitalizations among older adults with diabetes with comorbid osteoarthritis and major depression. Proportion of patients who received 1 or 2 HbA1c tests per year was relatively stable over 4-year time period (~40%). Proportion of patients who had annual eye examination gradually declined from 56.7% in 2010 to 54.6% in 2014. The proportion of patients who were prescribed

NSAIDs or benzodiazepines gradually declined from 18.5% and 41.4% in 2010 to 16.5% and

36.6% in 2014, respectively. The proportion of patients who were prescribed oral hypoglycemic drugs was relatively stable over 4-year time period (~45%).

The proportion of patients who had greater continuity of care (COC>0.36) gradually increased from 50.0% in 2010 to 55.4% in 2014. The proportion of patients who were prescribed SSRI or tetracyclic antidepressants gradually declined from 45.2% and 14.2% in

2010 to 38.1% and 12.7% in 2014, respectively. The proportion of patients who were prescribed SNRIs was relatively stable over 4-year time period (~22%). Only a few older diabetes patients with comorbid osteoarthritis and major depression were prescribed gaba receptor agonists or MAO inhibitors.

The incidence of hospitalizations among older adults with diabetes with comorbid osteoarthritis and major depression was relatively stable over 4-year study period (~30%).

The incidence of hospitalizations for diabetes-related long-term complications gradually declined from 15.2% to 14.0%. There were observed only a few episodes of hospitalizations for diabetes-related short-term complications among older adults with diabetes with comorbid osteoarthritis and major depression.

184

Table 4.15 Distribution of process of care and outcome measures among older adults with diabetes with comorbid osteoarthritis and major depression

Year 1 Year 2 Year 3 Year 4 Measure (n=2,444) (n=2,214) (n=2,018) (n=1,905) Process measure, n (%)

Having 1 or 2 *HbA1c tests 964 (39.4) 877 (39.6) 847 (42.0) 770 (41.6) per year Having 3 or more HbA1c 669 (27.9) 639 (26.6) 540 (22.5) 553 (23.0) tests Eye examination annually 1,386 (56.7) 1,239 (56.0) 1,094 (54.2) 1,010 (54.6)

Use of oral hypoglycemic 1,102 (45.1) 1,011 (45.7) 922 (45.7) 846 (45.7) drugs Use of tetracyclic 348 (14.2) 323 (14.6) 283 (14.0) 248 (12.7) antidepressants–“negative” Use of benzodiazepines– 1,011 (41.4) 861 (38.9) 779 (38.6) 678 (36.6) “negative” Use of gaba receptor agonist–“negative” <6 (0.2) <6 (0.1) <6 (0.1) <6 (0.1)

Use of **MAOIs–“negative” 9 (0.4) 8 (0.4) 7 (0.4) 6 (0.3)

Use of ***NSAIDs– “negative” 452 (18.5) 375 (16.9) 322 (16.0) 306 (16.5)

****COC index>0.36 1,221 (50.0) 1,223 (55.2) 1,137 (56.3) 1,025 (55.4)

Outcome measure, n (%)

All-cause hospitalizations 761 (31.1) 636 (28.7) 608 (30.1) 536 (29.0)

Hospitalizations for diabetes 8 (0.3) 6 (0.2) 8 (0.4) 6 (0.3) short-term complications Hospitalizations for diabetes 371 (15.2) 321 (14.5) 295 (14.6) 259 (14.0) long-term complications *HbA1c- glycated hemoglobin **MAO inhibitors - monoamine oxidase inhibitors *** NSAID- non-steroidal anti-inflammatory drugs *** Calculated using the Bice index

185

The results of bivariate regression analysis (Table 4.16) showed that all process measures

were significantly associated with the likelihood of hospitalizations and were included in the

multivariate marginal logistic models, except use of gaba receptor antagonist and MAO

inhibitors. The hospitalization for diabetes-related short-term complications was not

considered as an outcome variable due to rare episodes among older diabetes patients with

osteoarthritis and major depression (from 6 to 8 episodes per year).

Table 4.16 Bivariate associations between process measures and the likelihood of hospitalizations among older adults with diabetes with osteoarthritis and major depression

Model 1 Model 2 All-cause Hospitalizations for diabetes Measure hospitalisations long-term complications Unadjusted Unadjusted OR (95% CI) OR (95% CI) *HbA1c testing annually No Ref. Ref. 1 or 2 HbA1c tests 0.87 (0.77-0.97) 1.07 (1.03-1.23) 3 or more HbA1c tests 0.83 (0.73-0.95) 1.21 (1.02-1.47) Annual eye examination No Ref. Ref. Yes 0.70 (0.64- 0.77) 0.72 (0.64-0.81) Use of oral hypoglycemic drugs No Ref. Ref. Yes 0.96 (0.86-1.07) 1.45 (1.26-1.68) Use of **NSAIDs No Ref. Ref. Yes 0.99 (0.88-1.12) 0.94 (0.80-1.10) Use of tetracyclic antidepressants No Ref. Ref. Yes 1.36 (1.17-1.58) 1.51 (1.26 -1.81) Use of benzodiazepines No Ref. Ref. Yes 1.33 (1.20-1.48) 1.7 (1.02- 1.34)

186

Use of ***MAOIs No Ref. Ref. Yes 1.36 (0.48-3.86) 0.80 (0.22-2.92) Use of gaba receptor agonists No Ref. Ref. Yes 1.92 (0.32-3.58) 1.68 (0.39-7.23) ****COC index COC≤0.36 Ref. Ref. COC>0.36 0.73 (0.70- 0.76) 0.75 (0.72- 0.78) *HbA1c- glycated hemoglobin **MAO inhibitors - monoamine oxidase inhibitors *** NSAID- non-steroidal anti-inflammatory drugs *** Calculated using the Bice index

Table 4.17 presents the results of multivariate marginal logistic models that aimed to

examine the association between process measures for care for older adults with diabetes

with comorbid hypertension and the likelihood of hospitalizations. There was no association

between the receipt of recommended frequency of HbA1c testing and the likelihood of

hospitalizations among older diabetes patients with comorbid osteoarthritis and major

depression.

After adjusting for other covariates, there was observed a strong protective association

between having annual eye examination and the likelihood of all-cause hospitalizations, as

well as hospitalization for diabetes-related long-term complications (AOR=0.85, 95% CI

0.75-0.97, and AOR=0.88, 95% CI 0.61-0.99, respectively). After adjusting for other

covariates, the odds of hospitalizations for diabetes-related long-term complications was

higher among patients who were prescribed oral hypoglycemic drugs, compared to those who

were not (AOR=1.56, 95% CI 1.24-1.89), after adjusting for other covariates.

187

There was no association between use of NSAIDs, tetracyclic antidepressants, and

benzodiazepines and the likelihood of hospitalizations among older diabetes patients with

comorbid osteoarthritis and major depression. After adjusting for other covariates, a

protective association was observed between having greater continuity of care (COC>0.36)

and the likelihood of all-cause hospitalizations and hospitalizations for diabetes-related long-

term complications (AOR=0.84, 95% CI 0.72-0.93, and AOR=0.72, 95% CI 0.57-0.89,

respectively). After adjusting for other variables, an increase in each unit of prescribed drugs

was associated with increased odds of all-cause hospitalisations and hospitalizations for

diabetes-related long-term complications (AOR=1.06, 95% CI 1.05-1.07, and AOR=1.05,

95% CI 1.02-1.06, respectively).

Table 4.17 Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with osteoarthritis and major depression, adjusted for other covariates

Model 1 Model 2 All-cause Hospitalizations for Measure hospitalisations diabetes long-term AOR (95% CI) complications AOR (95% CI) Having *HbA1c testing No Ref. Ref. 1 or 2 HbA1c tests 0.93 (0.76-1.13) 1.27 (0.96-1.44) 3 or more HbA1c tests 0.82 (0.69-1.03) 1.34 (0.99-1.77) Annual eye examination No Ref. Ref. Yes 0.85 (0.75-0.97) 0.88 (0.61-0.99) Use of oral hypoglycemic drugs No Ref. Ref. Yes 0.93 (0.78-1.10) 1.56 (1.24-1.89) Use of **NSAIDs No Ref. Ref.

188

Yes 0.97 (0.80-1.17) 0.84 (0.64-1.10) Use of tetracyclic antidepressants No Ref. Ref. Yes 0.98 (0.78-1.20) 1.14 (0.86-1.32) Use of benzodiazepines No Ref. Ref. Yes 0.90 (0.76-1.07) 0.71 (0.60-1.04) ***COC index COC≤0.36 Ref. Ref. COC>0.36 0.84 (0.72-0.93) 0.72 (0.57-0.89) Number of prescribed drugs 1.06 (1.05-1.07) 1.05 (1.02-1.06) Age 1.02 (1.01-1.04) 1.02 (1.01-1.04) Sex Female Ref. Ref. Male 1.15 (0.97-1.23) 1.32 (1.11-1.61) Income quintiles Q1 lowest income Ref. Ref. Q2 1.02 (0.79-1.3) 1.04 (0.74-1.46) Q3 0.99 (0.78-1.28) 0.90 (0.63-1.26) Q4 1.03 (0.79-1.34) 1.05 (0.73-1.51) Q5 highest income 1.05 (0.82-1.35) 0.91 (0.63-1.29) ****RIO index ≤40 Ref. Ref. >40 1.27 (0.95-1.57) 1.26 (0.87-1.83) Duration of diabetes 1.01 (0.99-1.02) 1.07 (1.05-1.10) Duration of osteoarthritis 0.92 (0.97-1.03) 1.01 (0.96-1.02) Duration of depression 0.95 (0.89-1.01) 0.94 (0.86-1.02) Number of primary care visits 1.02 (1.01-1.03) 1.00 (0.98-1.03) *****Primary care models Capitated+ Ref. Ref. Fee-for-service 0.83 (0.68-1.02) 0.78 (0.60-1.00) Capitated 0.97 (0.51-1.89) 1.38 (0.60-2.99) Comorbidities 0 CC Ref. Ref. 1 CC 0.81 (0.62-1.02) 0.41 (0.18-0.82) 2 CC 1.05 (0.68-1.21) 0.87 (0.41-1.45) 3 CC 1.27 (0.71-1.81) 1.14 (0.53-1.68) 4 CC 1.39 (0.82-1.98) 1.31 (0.61-1.78) 5 or more CC 1.55 (0.97-2.23) 1.68 (0.98-2.01) *HbA1c- glycated hemoglobin ** NSAID- non-steroidal anti-inflammatory drugs ***Calculated using the Bice index *** Geographic location (≤40=non-rural; >40=rural). **** Noncapitated models include nonrostered models and those that operate on a fee-for-service basis; capitated models include family health networks and family health organizations operating on a capitation funding scheme; and the capitated+ models include family health teams and other rostered models operating on a capitated funding scheme with additional incentives for interdisciplinary care.

189

4.3.3 Sensitivity analysis results

The study findings were tested by performing sensitivity analyses considering older adults with diabetes with comorbid hypertension with and without chronic ischemic heart disease, as well as older diabetes patients with comorbid osteoarthritis after excluding all other comorbidities from the multivariate models. Unfortunately, it was not possible to run a sensitivity analysis for a cohort of older adults with diabetes comorbid with osteoarthritis and major depression because over 97% of those patients had at least one comorbid condition.

In sensitivity analyses (Appendix 15), the findings observed for the associations between the process of care measures and the likelihood of hospitalizations for three cohorts of older adults with diabetes with comorbid hypertension, with comorbid hypertension and chronic ischemic heart disease, and with comorbid osteoarthritis were essentially unchanged following the removal of all other comorbid conditions identified for the purpose of this study from the models.

4.4 Discussion

The study findings demonstrate that the quality of care declined in older adults with each additional selected comorbid condition, especially in those with discordant conditions – comorbid osteoarthritis with and without major depression. Diabetes care quality, defined as meeting the recommended frequency of diabetes monitoring tests and the use of oral

190 hypoglycemic therapy, was slightly lower among diabetes patients with 2 rather than

1concordant conditions.

However, there was an increase in use of antihypertensive drugs in patients with hypertension and chronic ischemic heart disease compared to those with comorbid hypertension. Conversely, the use of NSAIDs, defined as a negative indicator (inappropriate care) was lower in older diabetes patients with osteoarthritis and major depression compared to those with comorbid osteoarthritis.

The study findings support the underlying premise of the framework of Concordance and

Discordance proposed by Piette and Kerr that hypothesizes that the effects of comorbidity on patients with diabetes differ depending on the nature of comorbid conditions (22). It is known that diabetes-concordant conditions among diabetes patients share underlying predisposing factors and management goals with diabetes that results in better care (27, 64).

While diabetes-discordant conditions that do not share management goals with diabetes, such as depression or arthritis, and can impair patients’ functioning and may pose a significant barrier to management plan adherence resulting in worse health outcomes (22, 189).

The literature suggests that physicians may prioritize treatment of concordant conditions over discordant conditions, because a single treatment plan can improve the status of more than one condition (190). Blood pressure and cholesterol targets, increased physical activity, as well as the use of acetylsalicylic acid (ASA) or antihypertensive therapy are identical for patients with diabetes and cardiovascular conditions, including hypertension and ischemic

191 heart disease (191). Thus, for the majority of patients, management of cardiovascular conditions enhances the management of diabetes.

The incidence of hospitalizations markedly increased in older adults with diabetes with 2 vs.

1 selected comorbid condition, especially in those with discordant conditions. This study finding is consistent with previous research that found a higher rate of hospital admission among people with diabetes with discordant than concordant comorbid conditions, especially in those with mental conditions (192). A recent study indicated that there is a trend of increasing use of healthcare services, including hospitalizations, emergency department visits and physician visits, with increase in number of comorbid conditions among older adults with diabetes (24). There was a higher incidence of hospitalizations for diabetes-related long- term complications than short-term complications among older diabetes patients with both concordant and discordant conditions.

Meeting HbA1c testing goal, oral hypoglycemic drug therapy and hospitalizations among older adults with diabetes with 2 vs. 1 comorbid concordant/discordant conditions

Overall, the proportion of diabetes patients meeting recommended frequency of HbA1c testing goals was low (65-73%) for those with both concordant and discordant conditions.

The proportion of patients who met the recommended HbA1c testing goal or were prescribed oral hypoglycemic drugs was lower in older diabetes patients with 2 comorbid conditions compared to those with 1 condition; this decline was more significant in patients with

192 discordant conditions. However, patients who were not prescribed oral hypoglycemic drugs might be receiving insulin. Use of insulin was not assessed in this cohort.

This finding appears to contradict the results of previous research that found that people with diabetes with 2 or more comorbid conditions were more likely to achieve the target HbA1c testing frequency compared to those with no or one comorbid condition (30). However, the authors assessed the role of number of concordant and discordant conditions on the achievement of diabetes testing goals without specification of individual concordant and discordant conditions, despite the fact that certain conditions may have a greater impact on diabetes care than other conditions (30).

The current study findings suggest an association between meeting HbA1c testing frequency goals or oral hypoglycemic drug therapy and reduction in the likelihood of all-cause hospitalizations in older people with diabetes comorbid with both concordant and discordant conditions. Evidence shows that diabetes in older adults was associated with an increased odds of having a history of dizziness, falls, risk of cardiovascular events, or functional disability that may result in hospitalizations (193). The goal of both completing the recommended frequency of Hb1Ac tests and oral hypoglycemic drug therapy is to ensure optimal glycemic control (191) that can prevent development of cardiovascular events as well as fall-related events due to hypoglycemia and subsequent hospitalizations (194, 195).

Conversely, frequent HbA1c testing may be a marker for poor or inadequate glycemic control. Some patients who have more frequent HbA1c testing may be at increased risk for hospitalizations.

193

However, the study results demonstrate that completing the recommended frequency of

HbA1c testing or oral hypoglycemic drug therapy alone does not necessarily have a protective effect on hospitalizations for diabetes-related long-term complications in older diabetes patients with comorbid conditions. The possible explanation might be that poor or inadequate glycemic control among older adults with diabetes might results in more frequent

HbA1c testing. This finding is consistent with previous research results indicating that meeting recommended frequency of HbA1c tests among diabetes patients with comorbidities was associated with high odds of hospitalizations for diabetes long-term complications, while receipt of all three diabetes recommended tests, including HbA1c and LDL-C tests, and eye examinations had a strong protective effect on the likelihood of hospitalizations for diabetes- related long-term complications among diabetes patients with comorbidities (157).

Moreover, several other factors, including self-monitoring of glucose level, blood glucose level control, life style change, adherence to diet and medication regimens, and patient education may play more important roles in preventing diabetes-related complications and subsequent hospitalizations than meeting HbA1c testing goal or oral hypoglycemic drug therapy alone (156).

Annual eye examination and hospitalisations among older adults with diabetes with 2 vs. 1 comorbid concordant/discordant conditions

194

Diabetic retinopathy is one of the most prevalent complications among older adults with diabetes (156). Loss or impairment of vision is associated with significant morbidity, including increased falls, hip fracture and a 4-fold increase in mortality (196). Annual retinal examinations are recommended for early detection and treatment of diabetes-related eye complications (156).

Overall, proportion of patients having an annual eye examination performed was low (56-

67%) in older adults with both concordant and discordant conditions. The proportion of patients who had an annual eye exam declined in this patients with 2 vs. I comorbid condition; this decline was significant in those with discordant comorbid conditions (with comorbid osteoarthritis and major depression). Eye examinations may place a greater burden on patients in terms of scheduling appointments and transportation and require more cooperation from patients and caregivers in completing the examination, especially in those with comorbid osteoarthritis and major depression.

Having an annual eye examination performed among older adults with both concordant and discordant comorbid conditions had a positive association on the reduction of all-cause hospitalizations, including those for diabetes-related complications. This finding is consistent with previous research results (157). Eye examinations may prevent visual impairment which may in turn be associated with a lower risk for hospitalization. Other confounding factors may also contribute to this association. Patients who get eye examinations may be less severely ill and more proactive in managing their health, and have a lower risk for hospitalization.

195

Continuity of care and hospitalizations among older adults with diabetes with 2 vs. 1 comorbid concordant/discordant conditions

The median score of continuity of care was greater in older diabetes patients with concordant rather than discordant comorbid conditions; however, it declined with additional comorbid conditions, especially in those with discordant conditions. A possible explanation for the decline in the continuity of care score might be that older adults with a higher number of comorbidities are more likely to see multiple providers, especially those with discordant conditions. The proportion of other comorbidities in older diabetes patients with 2 of the selected conditions was higher than that in those with 1 selected condition, especially in patients with discordant conditions.

The study findings suggest an association between greater continuity of care and reduction in all-cause hospitalizations, including those for diabetes-related complications, in older people with diabetes with comorbid concordant and discordant conditions. This finding is consistent with other study results (197-199). Grunier and colleagues (74) found that the risk of hospitalizations was reduced in people with one or more chronic conditions, when visits are concentrated with a single physician.

Number of prescribed drugs and hospitalizations among older adults with diabetes with 2 vs. 1 comorbid concordant/discordant conditions

196

The study results demonstrate that the mean number of prescribed drugs increased in older diabetes patients with 2 vs. 1 comorbid condition, especially in those with discordant conditions (17 vs. 12 prescriptions). People with multiple chronic conditions are more likely to see multiple providers. This has been shown to carry a risk associated with failure of communication among professionals that is essential for evaluating and monitoring the patient’s therapeutic regimen, especially in those with discordant conditions (52).

The study findings suggest that the likelihood of all-cause hospitalizations, including those for diabetes –related complications, increases with the number of prescribed drugs among older adults with comorbid concordant or discordant conditions. This study finding is consistent with previous research results (200, 201). The literature suggests that the consequences of polypharmacy include adverse drug reactions, drug-drug interactions, drug- disease interactions, improper drug election, nonadherence/failure to receive drug and increased risk of hospitalizations and mortality (49, 202, 203).

Therefore, polypharmacy can be considered a marker of disease severity and clinical complexity for older patients with diabetes with comorbidities. Further research is warranted to disclose the relationship between polypharmacy and high risk of hospitalizations among older adults with diabetes taking into account possible confounders, including multimorbidity, disease severity, clinical complexity and frailty.

ACE inhibitor or ARB therapy and hospitalizations among older adults with diabetes with comorbid hypertension with and without chronic ischemic heart disease

197

Use of ACE inhibitors or ARBs in older adults with diabetes with comorbid hypertension is strongly recommended (156). The study results demonstrate that the proportion of patients who were prescribed ACE inhibitors was higher in older adults with comorbid hypertension and chronic ischemic heart disease compared to those without ischemic heart disease. The proportion of patients who were prescribed ARBs was almost similar in both groups of patients.

The study findings suggest an association between ACE inhibitor therapy and increase in all- cause hospitalizations, including those for diabetes-related complications, among patients with hypertension with or without comorbid ischemic heart disease, while there was no association between use of ARB therapy and the likelihood of hospitalizations. The information regarding the benefit of ACE inhibitors or ARBs on vascular protection among older adults with diabetes remains controversial.

Several studies found that ACE inhibitors could reduce risk of major vascular events in diabetes patients with hypertension (204, 205), while other studies failed to find such a beneficial effect of ACE inhibitors or ARBs among people with diabetes with comorbid hypertension (206, 207). Another study found that both ACE inhibitors and ARBs were not associated with a decrease in the risk for stroke in patients with diabetes (208), while a meta- analysis of randomized clinical trials found that the use of ACE inhibitors was associated with reduction in all-cause and cardiovascular mortality among patients with diabetes (209).

198

Little research has been undertaken to assess the benefits of ACE inhibitors or ARBs, specifically in older populations with diabetes. Therefore, further research is needed to evaluate the beneficial effect of various ACE inhibitors or ARBs in diabetes patients aged 65 years and older with hypertension with and without comorbid ischemic heart disease after taking into account disease severity and clinical complexity.

Antiplatelet and statin therapy and hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease

The proportion of patients who were prescribed antiplatelet therapy was low - about 22%. A possible explanation might be that acetylsalicylic acid (ASA) use is not reliably captured in

Ontario Drug Benefit formulary because of significant over-the-counter use of this medication. Another explanation might be that they had been prescribed other medications such as Plavix or warfarin instead of ASA due to other comorbid conditions such as atrial fibrillation or stroke. The study findings suggest an association between antiplatelet therapy and increase in all-cause hospitalizations, including those for diabetes-related long-term complications, among older adults with comorbid hypertension and chronic ischemic heart disease.

The Canadian Cardiovascular Society Guidelines support the use of low dose of ASA in diabetes patients with comorbid chronic ischemic heart disease for secondary prevention of vascular events (210). However, there is no large randomized clinical trial that has specifically examined the use of antiplatelet therapy for secondary prevention exclusively in older adults with diabetes. The results of a meta-analysis of trials demonstrate that

199 antiplatelet therapy in patients with diabetes failed to show a significant reduction in the risk of serious vascular events (211), while another study results showed oral antiplatelet agents, mainly aspirin, to be protective against vascular events in patients with diabetes (212). This is an area for further exploration.

About 80% of older adults with diabetes with comorbid hypertension and chronic ischemic heart disease were prescribed statins. A strong protective association was observed between statin therapy and all-cause hospitalizations, including those for diabetes-related long-term complications. This study finding is consistent with previous research results (213-215).

NSAIDs therapy as a “negative” indicator among older adults with diabetes with comorbid osteoarthritis

About 20% of older diabetes patients with comorbid osteoarthritis with and without major depression were prescribed NSAID therapy. One possible explanation for such a low rate might be that NSAIDs use is not reliably captured in the Ontario Drug Benefit formulary because of over-the-counter use of this medication. The study findings suggest an association between NSAID therapy and reduction in all-cause hospitalizations, including those for diabetes-related complications in older diabetes patients with comorbid osteoarthritis.

The recent review of evidence from the Osteoarthritis Research Society International

(OARSI) suggests that use of NSAID therapy for osteoarthritis management provides better efficacy than acetaminophen for relief of chronic inflammatory pain (216). However, the

200 literature suggests that all older adults who were prescribed an NSAID should be monitored for gastro-intestinal, cardiovascular and renal adverse events (217, 218).

“Negative” indicators among older adults with diabetes with comorbid osteoarthritis with and major depression

A few older diabetes patients with comorbid osteoarthritis and major depression were prescribed gaba receptor agonists or MAO inhibitors. About 40% of patients were prescribed benzodiazepines, and about 14% were prescribed tetracyclic antidepressants. There was no association found between use of benzodiazepines or tetracyclic antidepressants and all-cause hospitalizations, including those for diabetes-related complications. All these drugs are considered inadequate as first-or-second line depression treatments among older adults, based on lack of efficacy and/or their side-effect profiles (131, 219).

The cohort of older adults with diabetes with comorbid osteoarthritis and major depression was very small (2,444 patients). Previous research indicated mood disorders, including bipolar disorders, depression, manic episodes, and anxiety disorders, as one of the 10 top conditions among older adults with diabetes. This study was focused on quality of care of major depression/dysthymia among older adults, because the indicators that were developed using a Delphi study were applicable to only major depression/dysthymia.

201

4.4.1 Strengths and limitations

Our study sheds light on limited research evidence regarding the assessment of the quality of care among older adults with diabetes comorbid with 2 vs. 1 selected concordant/ discordant comorbid conditions. The study cohort came from validated administrative health databases that represent the entire Ontario population with a diagnosis of diabetes aged 65 and older.

Administrative data have the advantage of being population-based and are relatively inexpensive compared to the other potential sources of data for ambulatory care evaluation.

Four chronic conditions, including hypertension, chronic ischemic heart disease, osteoarthritis and major depression were selected for the purpose of this study as the most prevalent conditions among older adults with diabetes based on previous research (24).

The clinical process and outcome measures, as judged to be relevant by the Delphi Panel, were used for assessing clinical aspects of care and related outcomes among older adults with selected disease combinations. Quality indicators have been widely used in healthcare research projects. Mainz (68) stated that clinical indicators create the basis for quality improvement and prioritization in the health care system. Process measures are more appealing to healthcare providers for quality improvement purposes, because they represents aspects of care over which they have the most control; while consumers and payers of care may be more interested in outcome indicators, because they actually assess whether a patient’s status improves or not (75, 78).

202

The study findings suggest that older adults with diabetes are a diverse and heterogeneous group. In particular, the cohort of older adults with selected disease combinations had a high proportion of 15 other comorbid conditions, especially those with comorbid osteoarthritis with and without major depression. Therefore, there is a need to develop specific outcome measures for older people with diabetes with various types of comorbid conditions to reflect what matters most to patients, which is the effect of all their conditions on their health status.

Our process and outcome measures were limited to those available in Ontario administrative data. Consequently, we lacked data related to laboratory tests done in hospitals or paid for privately. Several quality measures, such as blood glucose level control, life style changes, self-monitoring of blood glucose (SMBG), and patients education, could reveal and explain important aspects of the associations between process of care measures and hospitalizations as reported here.

We also could not include patient-reported data, including patient perception on diabetes care, patient preferences and goals of care, self-management ability, healthcare access barriers, coordination of care and other factors that are known to have an impact on our study outcomes.

There is a potential for misclassifying people based on their comorbidity profiles. Reliance on the most responsible diagnosis field to identify diabetes-related hospitalizations in this study may underestimate hospitalization rates.

203

We were not able to account for severity of selected chronic conditions due to limitation of the administrative data that may lead to biased estimates. For instance, factors positively correlated with disease severity but unmeasured in the regression models may lead to observed expected (e.g. polypharmacy and hospitalisations) and unexpected (e.g. use of ACE inhibitors and the likelihood of hospitalisations) findings.

The common chronic co-existing conditions that were selected for this study may not reflect all existing comorbidities in patients with diabetes, however past work shows that 12 common conditions are likely sufficient for a study in multimorbdity (220). Other measures of comorbidity could be considered for risk-adjustment purposes, including Johns–Hopkins

Adjusted Clinical Group (ACG) Index (221), Charlson or Elixhauser comorbidity indices

(222).

4.5 Conclusions

Older adults with diabetes are a diverse and heterogeneous group; therefore, there is a need for a holistic approach in education and clinical care of patients taking into account concomitant conditions that affect a patient’s overall health status. The study findings suggest that patients are at increased risk of suboptimal care with additional selected comorbid condition, especially those with discordant comorbid conditions. Therefore, patients with diabetes-discordant conditions - comorbid osteoarthritis with or without major depression, need more targeted interventions and collaboration between healthcare providers

204 to improve quality of care and reduce hospitalization, while considering other co-existing conditions.

Any additional comorbid condition may affect the older adult to a greater or lesser magnitude at any one time, and may or may not be a dominant condition (223). Therefore, there is a need for discussion about how to prioritize diabetes treatment along with the care for other comorbid chronic conditions, as well as goals and preferences for treatment. This should be an iterative process and should be revisited. Care prioritization and shared decision making would be beneficial for both older adults and primary care providers, who are already challenged by the number of guideline recommendations (46, 224).

The study findings also support the importance of continuity of care for older diabetes patients with comorbid chronic conditions. The study demonstrated that continuity of care scores were higher among older adults with concordant comorbid conditions; and it declined with additional selected comorbid conditions among older diabetes patients, especially in those with discordant comorbid conditions. The study identified an association between increased continuity of care and lower hospital utilization for older diabetes patients with both concordant and discordant conditions.

205 review of therapeutic regimens should be performed among older people with diabetes to optimize and reduce and simplify their drug regimens where possible.

The study findings revealed a higher incidence of hospitalizations for diabetes-related long- term complications than short-term complications among older diabetes patients with both concordant and discordant conditions. Therefore, vascular protection strategies should be emphasized among older adults with diabetes, especially those with comorbid conditions.

206

CHAPTER 5

Synthesis

5.1 Study overview

The aim of this dissertation was to evaluate the quality of care for older adults with diabetes with comorbidities in ambulatory care settings, including:

 Diabetes with concordant comorbid conditions:

o Diabetes with comorbid hypertension

o Diabetes with comorbid hypertension and chronic ischemic heart disease

 Diabetes with discordant comorbid conditions:

o Diabetes with comorbid osteoarthritis

o Diabetes with comorbid osteoarthritis and major depression

This aim was achieved through three inter-related studies: 1) identifying a set of quality indicators for evaluating ambulatory care for older adults with selected chronic conditions, including diabetes, hypertension, chronic ischemic heart disease, major depression and osteoarthritis, 2) developing a set of quality indicators for evaluating the quality of overall care for older adults with diabetes with selected comorbid concordant/ discordant conditions by means of a Delphi method; and 3) assessing the quality of overall care for older adults with diabetes with selected comorbid concordant/discordant condition in ambulatory care settings using developed indicators, as well as examining associations between process of care measures and the likelihood of hospitalisations among older diabetes patients with selected comorbid conditions.

207

5.2 Main findings

The systematic review was conducted to identify evidence-based and valid quality indicators for care for older adults with chronic conditions, including diabetes, hypertension and chronic ischemic heart disease major depression, and osteoarthritis. All included studies used the most rigorous method of developing quality indicators, including combination of a systematic literature search with appraisal of candidate indicators using consensus techniques. The AIRE instrument was used to assess the methodological aspects of the quality indicator development process. All identified quality indicators assess clinical aspects of ambulatory care for older adults with selected chronic conditions. The identified indicators were categorized according to Donabedian’s structure- process-outcome framework.

Eight process and 4 outcome indicators were identified for diabetes care, 7 process indicators were identified for hypertension care, and 6 process indicators were identified for chronic ischemic heart disease, 6 process indicators were identified for major depression care, and 3 process indicators were identified for osteoarthritis care. Only indicators amenable to measurement using Ontario administrative data were extracted for the purpose of this study.

The study results indicate that relatively little research has been done to develop quality indicators to assess the quality of ambulatory care for older adults with selected chronic conditions. The study results also demonstrate that there were only a few published indicators for care for older adults with diabetes with comorbid hypertension or chronic

208 ischemic heart disease, and there were no defined indicators for care for diabetes with comorbid osteoarthritis or major depression.

The second study defined a set of quality indicators for ambulatory care for older adults with five selected disease combinations using a two-stage Delphi panel, using a set of candidate indicators identified after the 1st study as a starting point. The study focused on clinical process and outcome indicators that were categorized into domains, including on-going monitoring, pharmacological treatment, patient safety, and healthcare utilization/clinical effectiveness, which are accessible through administrative data.

A fifteen-member expert panel with broad geographical and clinical representation participated in this study, including primary care physicians, geriatricians, general internists and clinical pharmacists. Panelists were asked to rate process indicators for meaningfulness and potential for improvements in clinical practices, while outcome indicators were rated for importance and modifiability. Finally, the panelists were asked to rate each indicator for

“overall value for inclusion” in the context of each particular disease combination. Two levels of consensus, high and moderate, were defined for the final selection of quality indicators based on the value of the median, its frequency and value of the mean absolute deviation from the median.

Overall, twenty four high-consensus and twenty four medium-consensus indicators were selected for assessing care for older adults with five selected disease combinations. The developed indicators are useful for health care providers, managers and decision makers and

209 can be used to evaluate the quality of overall care for older adults with selected disease combinations in ambulatory care settings using administrative data.

Panelists reached rapid consensus on quality indicators for care for older adults with diabetes with concordant comorbid conditions, but it was more complicated for care for diabetes patients with discordant or both types of comorbid conditions. The study results indicate several inconsistencies between guideline recommendations and the Expert panel members’ opinion with respect to diabetes testing and its frequency, as well as pharmacological treatment for older adults with diabetes comorbid with osteoarthritis with or without major depression. Five indicators were defined as indicators of inappropriate care or poor performance – “negative”, including use of NSAIDs in older adults with diabetes and osteoarthritis, and use of tetracyclic antidepressants, MAO inhibitors, benzodiazepines, and gaba receptor agonists in older diabetes patients with comorbid major depression and osteoarthritis.

The third study examined the quality of care for older adults with selected disease combinations in ambulatory care settings, as well as associations between the process of care measures and hospitalizations. This was a retrospective cohort study conducted in Ontario,

Canada using linked provincial health administrative databases. The study observation period extended from April 1, 2010 to March 31, 2014.

The primary outcome in this study was a likelihood of having at least one all-cause hospitalization, while the secondary outcomes were a likelihood of at least one

210 hospitalization for diabetes-related short-term or long-term complications, during the period of time from April 1, 2010 to March 3, 2014. The quality of care was measured using process indicators developed through the aforementioned Delphi panel. Multiple regressions using generalized estimating equations approach were performed to examine associations between the quality of care for older diabetes patients with comorbidities, from 2010 to 2013, and the likelihood of hospitalizations during the follow-up period, from 2011-2014. The adjustments were made for patient and physician characteristics, duration of the selected conditions, and primary care visits.

The study findings indicate that the quality of care among older adults with diabetes was lower in those with 2 vs. 1 selected comorbid condition, especially in those with discordant conditions- osteoarthritis with vs. without major depression. The study findings also demonstrate that the incidence of hospitalizations markedly increased in those with 2 vs. 1 selected comorbid conditions, especially in those with comorbid osteoarthritis with vs. without major depression.

The median score of continuity of care was greater in older adults with selected concordant than discordant comorbid conditions; however, it declined in those with 2 vs. 1 selected comorbid conditions. The study findings suggest an association between greater continuity of care and reduction in hospitalizations among older adults with both concordant and discordant chronic conditions. The study findings indicate that the mean number of prescribed drugs increased in older adults with diabetes with comorbid 2 vs. 1 selected chronic condition, especially in those with comorbid osteoarthritis with vs. without major

211 depression. The study findings suggest that the likelihood of hospitalizations increases with increase in number of prescribed drugs among older adults with both concordant and discordant conditions.

5.3 Implications for practice

Quality indicators are important tools used for both quality assessment and quality improvement in healthcare systems. Despite the growing prevalence of multimorbidity among older adults, we still lack validated and evidence-based indicators for assessing quality of care among people with multimorbidity. The study findings suggest there were a few published quality indicators for care for patients with diabetes comorbid with hypertension or chronic ischemic heart disease, and there were no indicators for care for diabetes patients with comorbid osteoarthritis or major depression. The quality indicators for care for older adults with selected disease combinations that were developed using a Delphi study are useful for health care providers, managers and decision makers and provide an excellent starting point for further development and use in ambulatory care settings.

As the multimorbidity burden is high among older adults with diabetes, and much of it is presented clinically to general practitioners and family physicians, incorporation of these indicators to routine ambulatory care practice is recommended. Periodic assessment of the quality of care for older adults with diabetes can be useful to better understand the multimorbidity issue and support the decision-making process of general practitioners during

212 the management of older diabetes patients with comorbidities. Moreover, this would identify areas that need improvement for care for older adults with selected disease combinations and that could be addressed through quality improvement plans.

Therefore, patients with diabetes-discordant conditions - osteoarthritis with or without major depression, need more targeted interventions and collaboration between healthcare providers to improve quality of care and reduce hospitalization. These findings can help inform policy makers in developing a public health strategy for subpopulations of at-risk, older adults with diabetes.

Because of high prevalence of comorbidity among older adults with diabetes, it may not be possible to create clinical guidelines that could apply to all individuals with diabetes. In this situation, shared decision-making can be an important key element for developing an individualized treatment plan for each patient with particular types of concomitant chronic conditions (224). There is a necessity for discussion between older adults, their caregivers

(e.g., family members) and physicians about the risks of diabetes and other co-existing

213 chronic conditions, options for treatment, as well as patients’ goals and preferences for treatment especially in those with discordant conditions.

Care prioritization and shared decision making would be beneficial for both older adults and primary care providers, who are already challenged by the number of guideline recommendations (46, 224). The literature suggests that failure to discuss treatment goals with patients with multiple chronic conditions raises a risk that providers may focus on aspects of care that are not aligned with patient priorities. This can result in complex issues, including poor patient-physician relationship and treatment concordance, poor medication adherence, and poor health outcomes (225-227).

The study findings demonstrate that older diabetes patients with comorbidities, especially with discordant conditions, are likely to be prescribed a large number of drugs, and the more drugs they are prescribed the higher is the risk of hospitalizations. Therefore, the regular review of therapeutic regimens should be performed among older diabetes patients with comorbidities to reduce and simplify their drug regimens where possible. The patient perspective on medicine-taking also needs to be determined. In particular, compromise may

214 be appropriate between the patient’s choice and view of the prescriber. Future clinical trials focused on the older adults with specific disease combinations could also provide data about optimal treatment strategies.

5.4 Implications for policy

This study highlighted prevalent multimoribdity clusters among older adults with diabetes, including both concordant and discordant comorbidities. Explicit consideration of multimorbidity clusters among older adults with diabetes is important because appropriate management of individual diseases in isolation may not be optimal for patients with multimorbidity due to unique disease-disease or disease-treatment interactions. Furthermore, determining specific multimorbidity subgroups among patients with diabetes at increased risk of adverse health outcomes has important policy implications and provides targets for tailored prevention.

The study findings support the importance of continuity of care for older diabetes patients with comorbid chronic conditions. The study also found that the continuity of care score was higher among older adults with concordant comorbid conditions; and it declined with additional selected comorbid conditions among older diabetes patients, especially in those with discordant comorbid conditions. It is possible that lower COC scores in patients with discordant comorbid conditions reflected involvement of multiple specialists. These patients

215 may have been more complex or suffering from more severe disease, hence at higher risk of hospitalization.

The study findings also suggest that greater continuity of care is associated with lower hospital utilization for older diabetes patients with both concordant and discordant conditions. Therefore, policy makers and clinicians involved in the care of older adults should implement efforts to improve the continuity of care in older diabetes patients with comorbidities. In particular, disease management programs among older diabetes patients must incorporate levers to promote continuity, especially for patients with discordant conditions who are more likely to see multiple providers compared to those with concordant conditions.

Thus, there is a need to support service developments that provide integration of care for older diabetes patients with comorbidities across multiple disease domains and providers rather than focusing on integration between primary, secondary and tertiary care within single disease domains (228). For an older diabetes patient with comorbidities the challenge is to find a way to encourage health care providers to manage all chronic conditions collectively instead of focusing on a single disease treatment. The payment system, e.g. providing incentives, needs to foster interaction across multiple providers.

Previous research indicated that diabetes-discordant comorbidity increases the health care demand as much as diabetes-related comorbidity (25, 229). Current single-disease approach of diabetes care should be extended with additional care modules, which must include

216 different types of comorbid chronic condition to meet the complex health care demands of older adults with diabetes in the future. Perhaps in the long term, current diabetes care programs must be integrated with other chronic diseases care programs, especially among older adults with multimorbidity.

5.5 Implications for research

The quality indicators that were defined using a Delphi approach are not intended to provide a comprehensive tool set for measuring quality of care for older adults with diabetes with comorbidities. Rather, they address clinical aspects of care and can be used as a starting point for further development and use in ambulatory care settings.

There is a need to develop specific outcome measures for older people with diabetes with comorbidities to reflect what matters most to patients, which is the effect of all their conditions on their health status. This can be done through engaging clinicians, patients and

217 families in the identification of meaningful outcomes and then determining how they could be collected systematically.

The current study defined quality indicators for older diabetes patients with selected comorbid conditions by considering all conditions as at moderate severity. However, the care priority might change with respect to the severity of particular co-existing conditions.

Therefore, future research is needed for considering different scenarios for each disease combination, depending on disease severity and clinical complexity.

The results of this project point to further examinations of quality of overall care among older adults with various disease combinations. Given the differences observed in both quality of care and hospitalization rate across the different disease combinations, there is a need to better understand the relationship between the process of care measures and hospitalizations for older adults, including specific multimorbidity clusters.

Future research is needed for measuring the quality of care in the larger diabetes population and reporting by different stratification, including age, sex, and primary care models to see if there are any patterns in certain groups and target the interventions towards improving the practices for specific sub groups. Future research is recommended for assessing the association between the quality of care measures and hospitalizations after accounting for severity of conditions.

Most studies on the prescription and use of antidepressants have been conducted on younger populations, and when mixed-aged groups have been studied older adults have been

218 underrepresented. This limits the ability to generalize from these study findings when treating older people. Little research has been done to examine the benefits and harms of prescribing antidepressants in older adults with diabetes comorbid with major depression.

The literature suggests that selective serotonin reuptake inhibitors (SSRI) or serotonin- norepinephrine reuptake inhibitors (SNRIs) are relatively safe in the elderly, since they have lower anticholinergic effects than older antidepressants and are thus better tolerated by older adults (230). Surprisingly, the Delphi panelists reached consensus for exclusion of use of

SSRIs/SNRIs from the list of final indicators for assessing depression care quality. Future research is needed to evaluate the impact of antidepressants in this subpopulation taking into account severity of major depression as well as other comorbid conditions.

Future research is needed for assessing the quality of care for older diabetes with selected comorbid conditions using indicators that were not included in this study because they are not amenable to measurement using Ontario administrative data, including self-management, test values, BP measurement, microalbumin or creatinine testing, quality of life, coordination of care, patient satisfaction and involvement in decision making processes regarding the disease treatment.

Increasingly, patient-centered care is considered an essential element of quality within health systems, in particular for older adults with complex needs. Thus, it is crucial to understand which problems/conditions are important to whom and why, and how they change over time.

219

Future research is needed to examine the extent to which self-management priorities of older adults with diabetes align with primary care physicians’ treatment decision making.

There is a lack of information about main challenges to coordination of care among older diabetes patients with concordant vs. discordant comorbid conditions, and how primary care physicians might better coordinate care for those patients, both within and across practices and settings. There is a need to explore the main causes and consequences of poor coordination of care from the perspectives of older diabetes patients with comorbidities and their primary care physicians.

220

References

1. World Health Organization. Global status report on noncommunicable diseases 2010: description of the global burden of NCDs, their risk factors and determinants. Geneva: World Health Organization, 2011. Available from http://www.who.int/nmh/publications/ncd_report2010/en/ Accessed November 26, 2015.

2. Verma J, Samis S. Canada's population is aging. Healthc Pap 2011;11(1):3-5.

3. Public Health Agency of Canada. The Chief Public Health Officer's Report on the State of Public Health in Canada: Growing older - adding life to years. 2010.

4. Canadian Institute for Health Information. Seniors and the health care system: what is the impact of multiple chronic conditions. Ottawa, Canada: Canadian Institute for Health Information; 2011.

5. Kone Pefoyo AJ, Bronskill SE, Gruneir A, Calzavara A, Thavorn K, Petrosyan Y, et al. The increasing burden and complexity of multimorbidity. BMC Public Health 2015;15(1):415.

6. Boyd CM, Fortin M. Future of multimorbidity research: how should understanding of multimorbidity inform health system design? Public Health Reviews 2012;32(2):451-474.

7. Fortin M, Bravo G, Hudon C, Vanasse A, Lapointe L. Prevalence of multimorbidity among adults seen in family practice. Ann Fam Med 2005;3(3):223-8.

8. Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Arch Intern Med 2002;162(20):2269-76.

9. Ontario Medical Association. OMA Policy on Chronic Disease Management. Ont Med Rev 2010:20-37.

10. Health Council of Canada. Population patterns of chronic health conditions in Canada - a data supplement to why healthcare renewal matters: learning from Canadians with chronic health conditions. Toronto: Health Council of Canada; 2007.

11. Fortin M, Bravo G, Hudon C, Lapointe L, Almirall J, Dubois M-F, et al. Relationship between multimorbidity and health-related quality of life of patients in primary care. Quality of Life Research 2006;15(1 DO - 10.1007/s11136-005-8661-z):83-91 LA - English.

12. Health Council of Canada. How do sicker Canadians with chronic disease rate the health care system? Results from the 2011 Commonwealth Fund International Health Policy Survey of sicker adults. Toronto, Canada: Health Council of Canada; 2011.

13. Whiting DR, Guariguata L, Weil C, Shaw J. IDF diabetes atlas: global estimates of the prevalence of diabetes for 2011 and 2030. Diabetes Res Clin Pract 2011;94(3):311-21.

221

14. Canadian Diabetes Association. An economic tsunami, the cost of diabetes in Canada. Toronto, ON: Canadian Diabetes Association; 2009.

15. Canadian Diabetes Association. Keeping diabetes prevention and management as health priorities in Ontario. Toronto, ON: Brief to the Standing Committee on Finance and Economic Affairs, 2014.

16. Kirkman MS, Briscoe VJ, Clark N, Florez H, Haas LB, Halter JB, et al. Diabetes in older adults. Diabetes Care 2012;35(12):2650-64.

17. Druss BG, Marcus SC, Olfson M. The most expensive medical conditions in America. Health Aff (Millwood) 2002;21:105-111.

18. Currie CJ, Morgan CL, Dixon S, McEwan P, Marchant N, Bearne A, et al. The financial costs of hospital care for people with diabetes who have single and multiple macrovascular complications. Diabetes Res Clin Pract 2005;67(2):144-51.

19. Brandle M, Zhou H, Smith BR, Marriott D, Burke R, Tabaei BP, et al. The direct medical cost of type 2 diabetes. Diabetes Care 2003;26(8):2300-4.

20. Ohinmaa A, Jacobs P, Simpson S, Johnson JA. The projection of prevalence and cost of diabetes in Canada: 2000 to 2016. Canadian Journal of Diabetes 2004;28(2).

21. Rosella LC, Lebenbaum M, Fitzpatrick T, O'Reilly D, Wang J, Booth GL, et al. Impact of diabetes on healthcare costs in a population-based cohort: a cost analysis. Diabet Med 2015;33(3):395-403.

22. Piette JD, Kerr EA. The impact of comorbid chronic conditions on diabetes care. Diabetes Care 2006;29(3).

23. Luijks H, Schermer T, Bor H, van Weel C, Lagro-Janssen T, Biermans M, et al. Prevalence and incidence density rates of chronic comorbidity in type 2 diabetes patients: an exploratory cohort study. BMC Med 2012;10:128.

24. Gruneir A, Markle-Reid M, Fisher K, Reimer H, Ma X, Ploeg J. Comorbidity Burden and Health Services Use in Community-Living Older Adults with Diabetes Mellitus: A Retrospective Cohort Study. Can J Diabetes 2016;40(1):35-42.

25. Struijs JN, Baan CA, Schellevis FG, Westert GP, van den Bos GA. Comorbidity in patients with diabetes mellitus: impact on medical health care utilization. BMC Health Serv Res 2006;6:84.

26. Ciechanowski PS, Katon WJ, Russo JE. Depression and diabetes: impact of depressive symptoms on adherence, function, and costs. Arch Intern Med 2000;160(21):3278-85.

222

27. Woodard LD, Urech T, Landrum CR, Wang D, Petersen LA. Impact of comorbidity type on measures of quality for diabetes care. Med Care 2011;49(6):605-10.

28. Pentakota SR, Rajan M, Fincke BG, Tseng CL. Does diabetes care Differ by type of chronic comorbidity? Diabetes Care 2012;35.

29. Woodard LD, Landrum CR, Urech TH, Wang D, Virani SS, Petersen LA. Impact of clinical complexity on the quality of diabetes care. Am J Manag Care 2012;18(9):508-14.

30. Magnan EM, Palta M, Johnson HM, Bartels CM, Schumacher JR, Smith MA. The impact of a patient's concordant and discordant chronic conditions on diabetes care quality measures. J Diabetes Complications 2014;29(2):288-94.

31. Kiran T, Victor JC, Kopp A, Shan BR, Glazier RH. The relationship between financial incentives and quality of diabetes care in Ontario, Canada. Diabetes Care 2012;35:1038-1046.

32. Fortin M, Soubhi H, Hudon C, Bayliss EA, van den Akker M. Multimorbidity's many challenges. BMJ 2007;334(7602):1016-7.

33. Fortin M, Bravo G, Hudon C, Lapointe L, Dubois MF. Relationship between psychological distress and multimorbidity of patients in family practice. Ann Fam Med 2006;4:417-22.

34. Fortin M, Contant E, Savard C, Hudon C, Poitras ME, Almirall J. Canadian guidelines for clinical practice: an analysis of their quality and relevance to the care of adults with comorbidity. BMC Fam Pract 2011;12:74.

35. Glynn LG, Valderas JM, Healy P, Burke E, Newell J, Gillespie P, et al. The prevalence of multimorbidity in primary care and its effect on health care utilization and cost. Fam Pract 2011;28(5):516-23.

36. Iron K, Lu H, Manuel D, Henry D, Gershon A. Using linked health administrative data to assess the clinical and healthcare system impact of chronic diseases in Ontario. Healthc Q 2011;14(3):23-7.

37. Tinetti ME, Fried TR, Boyd CM. Designing health care for the most common chronic condition--multimorbidity. Jama 2012;307(23):2493-4.

38. Nicholson K, Terry AL, Fortin M, Williamson T, Thind A. Understanding multimorbidity in primary health care. Canadian Family Physician. 2015;61(10):918. Canadian Family Physician 2015;61(1):918.

39. van den Akker M, Buntinx F, Metsemakers JF, Roos S, Knottnerus JA. Multimorbidity in general practice: prevalence, incidence, and determinants of co-occurring chronic and recurrent diseases. J Clin Epidemiol 1998;51(5):367-75.

223

40. Carral F, Aguilar M, Olveira G, Mangas A, Domenech I, Torres I. Increased hospital expenditures in diabetic patients hospitalized for cardiovascular diseases. J Diabetes Complications 2003;17(6):331-6.

41. Norlund A, Apelqvist J, Bitzen PO, Nyberg P, Schersten B. Cost of illness of adult diabetes mellitus underestimated if comorbidity is not considered. J Intern Med 2001;250(1):57-65.

42. Gijsen R, Hoeymans N, Schellevis FG, Ruwaard D, Satariano WA, van den Bos GA. Causes and consequences of comorbidity: a review. J Clin Epidemiol 2001;54(7):661-74.

43. Kirkman MS, Briscoe VJ, Clark N, Florez H, Haas LB, Halter JB, et al. Diabetes in older adults: a consensus report. J Am Geriatr Soc 2012;60(12):2342-56.

44. Centers for Disease Control and Prevention. Diabetes Public Health Resource. Available from www.cdc.gov/diabetes. Accessed 27 September 2012.

45. Guiding principles for the care of older adults with multimorbidity: an approach for clinicians: American Geriatrics Society Expert Panel on the Care of Older Adults with Multimorbidity. J Am Geriatr Soc 2012;60(10):E1-E25.

46. Boyd CM, Darer J, Boult C, Fried LP, Boult L, Wu AW. Clinical practice guidelines and quality of care for older patients with multiple comorbid diseases: implications for pay for performance. JAMA 2005;294(6):716-24.

47. Mutasingwa DR, Ge H, Upshur REG. How applicable are clinical practice guidelines to elderly patients with multimorbidity. Can Fam Physician 2011;57:e253-e262.

48. Lugtenberg M, Burgers JS, Clancy C, Westert GP, Schneider EC. Current guidelines have limited applicability to patients with comorbid conditions: a systematic analysis of evidence-based guidelines. PLoS ONE 2011;6(10):e25987. doi:10.1371/journal.pone.0025987.

49. Nobili A, Garattini S, Mannucci PM. Multiple diseases and polypharmacy in the elderly: challenges for the internist of the third millennium. Journal of Comorbidity 2011;1(1):28-44.

50. Tobe SW, Stone JA, Brouwers M, Bhattacharyya O, Walker KM, Dawes M, et al. Harmonization of guidelines for the prevention and treatment of cardiovascular disease: the C-CHANGE Initiative. Cmaj 2011;183(15):E1135-50.

51. Hajjar ER, Cafiero AC, Hanlon JT. Polypharmacy in elderly patients. American Journal of Geriatric Pharmacotherapy 2007;5(4).

224

52. Calderon-Larranaga A, Poblador-Plou B, Gonzalez-Rubio F, Gimeno-Feliu LA, Abad-Diez JM, Prados-Torres A. Multimorbidity, polypharmacy, referrals, and adverse drug events: are we doing things well? Br J Gen Pract 2012;62(605):e821-6.

53. Linjakumpu T, Hartikainen S, Klaukka T. Use of medications and polypharmacy are increasing among the elderly. Clin Epidemiol 2002;55:809-817.

54. Jorgensen T, Johansson S, Kennerfalk A. Prescription drug use, diagnoses, and healthcare utilization among the elderly. Ann Pharmacother 2001;35:1004-1009.

55. Schoen C, Osborn R, Doty MM, Bishop M, Peugh J, Murukutla N. Toward higher- performance health systems: adults' health care experiences in seven countries, 2007. Health Aff (Millwood) 2007;26(6):w717-34.

56. Emslie-Smith A, Dowall J, Morris A. The problem of polypharmacy in type 2 diabetes. Br J Diabetes Vasc Dis 2003;3:54-56.

57. Campbell-Scherer D. Multimorbidity: a challenge for evidence-based medicine. Evid Based Med 2010;15(6):165-6.

58. Laux G, Kuehlein T, Rosemann T, Szecsenyi J. Co- and multimorbidity patterns in primary care based on episodes of care: results from the German CONTENT project. BMC Health Serv Res 2008;8:14.

59. Institute of Medicine (US). Committee on the Future of Primary Care: America’s Health in a New Era. Washington, DC: National Academy Press; 1996.

60. Russell GM, Dahrouge S, Hogg W, Geneau R, Muldoon L, Tuna M. Managing chronic disease in ontario primary care: the impact of organizational factors. Ann Fam Med 2009;7(4):309-18.

61. Glazier RH, Zagorski BM, Rayner J. Comparison of primary care models in Ontario by demographics, case mix and emergency department use, 2008/2009 to 2009/2010. ICES Investigate Report 2012.

62. Ettner SL, Thompson TJ, Stevens MR, Mangione CM, Kim C, Neil Steers W, et al. Are physician reimbursement strategies associated with processes of care and patient satisfaction for patients with diabetes in managed care? Health Serv Res 2006;41(4 Pt 1):1221-41.

63. Glazier RH, Klein-Geltink J, Kopp A, Sibley LM. Capitation and enhanced fee-for- service models for primary care reform: a population-based evaluation 10.1503/cmaj.081316. CMAJ 2009;180.

64. Kiran T, Victor JC, Kopp A, Shah BR, Glazier RH. The relationship between primary care models and processes of diabetes care in Ontario. Can J Diabetes 2014;38(3):172-8.

225

65. Lee L, Heckman G. Meeting the challenges of managing seniors with multiple complex conditions: the central role of primary care. CGS Journal pf CME 2012;2(2):23-27.

66. Noel PH, Frueh BC, Larme AC, Pugh JA. Collaborative care needs and preferences of primary care patients with multimorbidity. Health Expect 2005;8(1):54-63.

67. Salisbury C, Johnson L, Purdy S, Valderas JM, Montgomery AA. Epidemiology and impact of multimorbidity in primary care: a retrospective cohort study. Br J Gen Pract 2011;61(582):e12-21.

68. Reid R. Defusing the Confusion: Concepts and measures of continuity of healthcare. Ottawa: Canadian Health Services Research Foundation. 2002.

69. Waibel S, Henao D, Aller MB, Vargas I, Vazquez ML. What do we know about patients' perceptions of continuity of care? A meta-synthesis of qualitative studies. Int J Qual Health Care 2012;24(1):39-48.

70. Giovannetti ER, Dy S, Leff B, Weston C, Adams K, Valuck TB, et al. Performance Measurement for People With Multiple Chronic Conditions: Conceptual Model. Am J Manag Care 2013;19(10):e359-e366.

71. Carney PA, Conry CM, Mitchell KB, Ericson A, Dickinson WP, Martin JC, et al. The Importance of and the Complexities Associated With Measuring Continuity of Care During Resident Training: Possible Solutions Do Exist. Fam Med 2016;48(4):286-93.

72. Bice TW, Boxerman SB. A quantitative measure of continuity of care. Med Care 1977;15(4):347-9.

73. Haggerty JL, Freeman GK, Starfield BH, Adair CE, McKendy R. Continuity of care: a multidisciplinary review. BMJ 2003;327:1219-21.

74. Gruneir A, Bronskill SE, Maxwell CJ, Bai YQ, Kone AJ, Thavorn K, et al. The association between multimorbidity and hospitalization is modified by individual demographics and physician continuity of care: a retrospective cohort study 10.1186/s12913-016-1415-5. BMC Health Services Research 2016;16(1):1-9.

75. Mainz J. Defining and classifying clinical indicators for quality improvement. Int J Qual Health Care 2003;15(6):523-30.

76. Lohr K, Committee to Design a Strategy for Quality Review and Assurance in Medicare, eds. Medicare: a strategy for quality assurance, Vol. 1. Washington, DC: National Academy Press; 1990.

77. Canadian Council on Health Services Accreditation. A guide to the development and use of performance indicators. Ottawa: Canadian Council on Health Services Accreditation, 1996. Available from http://www.cchsa.ca Accessed October 6, 2015.

226

78. Campbell SM, Braspenning J, Hutchinson A, Marshall M. Research methods used in developing and applying quality indicators in primary care. Bmj 2003;326(7393):816-819.

79. Mainz J. Developing evidence-based clinical indicators: a state of the art methods primer. Int J Qual Health Care 2003;15 Suppl 1:i5-11.

80. Fitch K, Bernstein SJ, Burnand B, LaCalle JR, Lazaro P, McDonnell J, et al. The RAND/UCLA Appropriateness Method User's Manual. Santa Monica, CA: RAND Corporation; 2001.

81. Stelfox HT, Straus SE. Measuring quality of care: considering conceptual approaches to quality indicator development and evaluation. J Clin Epidemiol 2013;66(12):1328-37.

82. Aggarwal M, Hutchison B. Toward a primary care strategy for Canada. Canadian Foundation for Healthcare Improvement 2012.

83. Donabedian A. Evaluating the quality of medical care. Milbank Memorial Fund Quarterly 1966;44:166-206.

84. Donabedian A. Methods for deriving criteria for assessing the quality of medical care. Med Care Rev 1980;37(7):653-98.

85. Teljeur C, Smith SM, Paul G, Kelly A, O'Dowd T. Multimorbidity in a cohort of patients with type 2 diabetes. Eur J Gen Pract 2013;19(1):17-22.

86. WHO. The global burden of disease. 2004 update. Geneva: World Health Organization; 2008. http://www.who.int/healthinfo/globalburden_disease/GBD_report_2004update_full.pdf Accessed Aug 25, 2014.

87. Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27(5):1047-53.

88. Canadian Diabetes Association. An economic tsunami: The cost of diabetes in Canada. http://www.diabetes.ca/CDA/media/documents/publications-and- newsletters/advocacy-reports/economic-tsunami-cost-of-diabetes-in-canada-english.pdf. 2010. Accessed November 23, 2015.

89. Buchanan D, Tourigny-Rivard MF, Cappeliez P, Frank C, Janikowski P, Spanjevic L, et al. National Guidelines for Seniors’ Mental Health: The Assessment and Treatment of Depression. Canadian Journal of Geriatrics 2006;5, (2 Suppl.):S52-8.

90. Blazer DG, Hybels CF, Pieper CF. The association of depression and mortality in elderly persons: a case for multiple, independent pathways. J Gerontol A Biol Sci Med Sci 2001;56(8):M505-9.

227

91. Unutzer J, Patrick DL, Diehr P, Simon G, Grembowski D, Katon W. Quality adjusted life years in older adults with depressive symptoms and chronic medical disorders. Int Psychogeriatr 2000;12(1):15-33.

92. Wittchen HU, Holsboer F, Jacobi F. Met and unmet needs in the management of depressive disorder in the community and primary care: the size and breadth of the problem. J Clin Psychiatry 2001;62 Suppl 26:23-8.

93. Kilbourne AM, Schulberg HC, Post EP, Rollman BL, Belnap BH, Pincus HA. Translating evidence-based depression management services to community-based primary care practices. Milbank Q 2004;82(4):631-59.

94. Birtwhistle R, Morkem R, Peat G, Williamson T, Green ME, Khan S, et al. Prevalence and management of osteoarthritis in primary care: an epidemiologic cohort study from the Canadian Primary Care Sentinel Surveillance Network. CMAJ Open 2015;3(3):E270-5.

95. Peat G, McCarney R, Croft P. Knee pain and osteoarthritis in older adults: a review of community burden and current use of primary health care. Ann Rheum Dis 2001;60(2):91- 7.

96. Chow CK, Teo KK, Rangarajan S, Islam S, Gupta R, Avezum A, et al. Prevalence, awareness, treatment, and control of hypertension in rural and urban communities in high-, middle-, and low-income countries. Jama 2013;310(9):959-68.

97. Lloyd-Jones D, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K, et al. Heart disease and stroke statistics--2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2009;119(3):480-6.

98. Caminal J, Starfield B, Sanchez E, Casanova C, Morales M. The role of primary care in preventing ambulatory care sensitive conditions. Eur J Public Health 2004;14(3):246-51.

99. World Health Organization. Innovative care for chronic conditions: building blocks for action. Global report. World Health Organization, 2002.

100. Seddon ME, Marshall MN, Campbell SM, Roland MO. Systematic review of studies of quality of clinical care in general practice in the UK, Australia and New Zealand. Qual Health Care 2001;10(3):152-8.

101. Harteloh PP. The meaning of quality in health care: a conceptual analysis. Health Care Anal 2003;11(3):259-67.

102. Campbell SM, Hann M, Hacker J, Durie A, Thapar A, Roland MO. Quality assessment for three common conditions in primary care: validity and reliability of review

228 criteria developed by expert panels for angina, asthma and type 2 diabetes. Qual Saf Health Care 2002;11(2):125-30.

103. Higgins JP, Deeks JJ. Selecting studies and collecting data. Higgins J, Green S, editors. England: John Wiley & Sons Ltd. 2008.

104. de Koning J, Smulders A, Klazinga NS. The Appraisal of Indicators through Research and Evaluation (AIRE) instrument. Academic Medical Center, Amsterdam, The Netherlands; 2006. doi: 10.1111/ acem.12591 PMID: 25676528.

105. Smeulers M, Verweij L, Maaskant JM, de Boer M, Krediet CT, Nieveen van Dijkum EJ, et al. Quality indicators for safe medication preparation and administration: a systematic review. PLoS One 2015;10(4):e0122695.

106. Strudwick K, Nelson M, Martin-Khan M, Bourke M, Bell A, Russell T. Quality indicators for musculoskeletal injury management in the emergency department: a systematic review. Acad Emerg Med 2015;22(2):127-41.

107. Pasman HR, Brandt HE, Deliens L, Francke AL. Quality indicators for palliative care: a systematic review. J Pain Symptom Manage 2009;38(1):145-56.

108. IOM (Institute of Medicine). 2012. Toward quality measures for population health and the leading health indicators. Washington, DC: The National Academies Press. Washington, DC: The National Academies Press.

109. Patel MM, Brown JD, Croake S, Lewis R, Liu J, Patton L, et al. The Current State of Behavioral Health Quality Measures: Where Are the Gaps? Psychiatr Serv 2015;66(8):865- 71.

110. Donabedian A. The quality of care. How can it be assessed? Arch Pathol Lab Med 1988;121(11):1145-50.

111. Orwin RG. In H. Cooper and L.V. Hedges (eds.) The Handbook of Research Synthesis. Russell Sage Foundation. 1994:150-151.

112. Shekelle P, Vijan S. Quality indicators for the care of diabetes mellitus in vulnerable elders. J Am Geriatr Soc 2007;55 Suppl 2:S312-7.

113. Nicolucci A, Greenfield S, Mattke S. Selecting indicators for the quality of diabetes care at the health systems level in OECD countries. Int J Qual Health Care 2006;18 Suppl 1:26-30.

114. Martirosyan L, Braspenning J, Denig P, de Grauw WJ, Bouma M, Storms F, et al. Prescribing quality indicators of type 2 diabetes mellitus ambulatory care. Qual Saf Health Care 2008;17(5):318-23.

229

115. Perez-Cuevas R, Doubova SV, Suarez-Ortega M, Law M, Pande AH, Escobedo J, et al. Evaluating quality of care for patients with type 2 diabetes using electronic health record information in Mexico. BMC Med Inform Decis Mak 2012;12:50.

116. Majumdar SR, Johnson JA, Bowker S, Booth JL, Dolovich L, Ghali W. A Canadian consensus for the standardized evaluation of quality improvement interventions in Type 2 Diabetes. Canadian Journal of Diabetes 2005;29(3):220-29.

117. (NQMC) NQMC. Diabetes long-term complications admission: percentage of admissions for a principal diagnosis of diabetes with long-term complications per 100,000 population, ages 18 years and older. In. Rockville MD: Agency for Healthcare Research and Quality (AHRQ).

118. (NQMC) NQMC. Diabetes short term complications admission: percentage of admissions for a principal diagnosis of diabetes with short-term complications per 100,000 population, ages 18 years and older. In. Rockville MD: Agency for Healthcare Research and Quality (AHRQ).

119. Abrahamyan L, Boom N, Donovan LR, Tu JV. An international environmental scan of quality indicators for cardiovascular care. Can J Cardiol 2012;28(1):110-8.

120. Min LC, Mehrotra R, Fung CH. Quality indicators for the care of hypertension in vulnerable elders. J Am Geriatr Soc 2007;55 Suppl 2:S359-65.

121. Watson K, Fung CH, Budoff M. Quality indicators for the care of ischemic heart disease in vulnerable elders. J Am Geriatr Soc 2007;55 Suppl 2:S366-72.

122. Kerr EA, Asch SM, Hamilton EG, McGlynn EA. Quality of Care for Cardiopulmonary Conditions: A Review of the Literature and Quality Indicators. RAND monograph report: MR-1282-AHRQ. 2000.

123. Burge FI, Bower K, Putnam W, Cox JL. Quality indicators for cardiovascular primary care. Can J Cardiol 2007;23(5):383-8.

124. Drozda J, Jr., Messer JV, Spertus J, Abramowitz B, Alexander K, Beam CT, et al. ACCF/AHA/AMA-PCPI 2011 performance measures for adults with coronary artery disease and hypertension: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Performance Measures and the American Medical Association-Physician Consortium for Performance Improvement. J Am Coll Cardiol 2011;58(3):316-36.

125. Campbell SM, Ludt S, Van Lieshout J, Boffin N, Wensing M, Petek D, et al. Quality indicators for the prevention and management of cardiovascular disease in primary care in nine European countries. Eur J Cardiovasc Prev Rehabil 2008;15(5):509-15.

230

126. Doubova SV, Lamadrid-Figueroa H, Perez-Cuevas R. Use of electronic health records to evaluate the quality of care for hypertensive patients in Mexican family medicine clinics. J Hypertens 2013;31(8):1714-23.

127. Iovieno N, Tedeschini E, Bentley KH, Evins AE, Papakostas GI. Antidepressants for major depressive disorder and dysthymic disorder in patients with comorbid alcohol use disorders: a meta-analysis of placebo-controlled randomized trials. J Clin Psychiatry 2011;72(8):1144-51.

128. Levkovitz Y, Tedeschini E, Papakostas GI. Efficacy of antidepressants for dysthymia: a meta-analysis of placebo-controlled randomized trials. J Clin Psychiatry 2011;72(4):509- 14.

129. Pedrelli P, Iovieno N, Vitali M, Tedeschini E, Bentley KH, Papakostas GI. Treatment of major depressive disorder and dysthymic disorder with antidepressants in patients with comorbid opiate use disorders enrolled in methadone maintenance therapy: a meta-analysis. J Clin Psychopharmacol 2011;31(5):582-6.

130. Patten SB, Kennedy SH, Lam RW, O'Donovan C, Filteau MJ, Parikh SV, et al. Canadian Network for Mood and Anxiety Treatments (CANMAT) clinical guidelines for the management of major depressive disorder in adults. I. Classification, burden and principles of management. J Affect Disord 2009;117 Suppl 1:S5-14.

131. Nakajima GA, Wenger NS. Quality indicators for the care of depression in vulnerable elders. J Am Geriatr Soc 2007;55 Suppl 2:S302-11.

132. Hermann RC, Palmer H, Leff S, Shwartz M, Provost S, Chan J, et al. Achieving consensus across diverse stakeholders on quality measures for mental healthcare. Med Care 2004;42(12):1246-53.

133. Veterans Health Administration/Department of Defense. VHA/DOD Performance Measures for the Management of Major Depressive Disorder in Adults, Version 2.0. Washington, DC, 2000.

134. Hermann RC, Mattke S, Somekh D, Silfverhielm H, Goldner E, Glover G, et al. Quality indicators for international benchmarking of mental health care. Int J Qual Health Care 2006;18 Suppl 1:31-8.

135. Snow V, Lascher S, Mottur-Pilson C. Pharmacologic treatment of acute major depression and dysthymia. American College of Physicians-American Society of Internal Medicine. Ann Intern Med 2000;132(9):738-42.

136. MacLean CH, Pencharz JN, Saag KG. Quality indicators for the care of osteoarthritis in vulnerable elders. J Am Geriatr Soc 2007;55 Suppl 2:S383-91.

231

137. Pencharz JN, MacLean CH. Measuring quality in arthritis care: the Arthritis Foundation's Quality Indicator set for osteoarthritis. Arthritis Rheum 2004;51(4):538-48.

138. American Academy of Orthopaedic Surgeons/Physician Consortium for Performance Improvement®. Osteoarthritis Physician Performance Measurement Set. American Medical Association, June, 2006 http://www.ama-assn.org/apps/listserv/x- check/qmeasure.cgi?submit=PCPI Accessed November, 2014.

139. Boulkedid R, Abdoul H, Loustau M, Sibony O, Alberti C. Using and reporting the Delphi method for selecting healthcare quality indicators: a systematic review. PLoS One 2011;6(6):e20476.

140. Man MS, Chaplin K, Mann C, Bower P, Brookes S, Fitzpatrick B, et al. Improving the management of multimorbidity in general practice: protocol of a cluster randomised controlled trial (The 3D Study). BMJ Open 2016;6(4):e011261.

141. Measuring and improving quality of care : A report from the american heart Association/American college of cardiology first scientific forum on assessment of healthcare quality in cardiovascular disease and stroke. Stroke 2000;31(4):1002-12.

142. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross- sectional study. Lancet 2012;380(9836):37-43.

143. Bayliss EA, Edwards AE, Steiner JF, Main DS. Processes of care desired by elderly patients with multimorbidities. Fam Pract 2008;25(4):287-93.

144. Bayliss EA, Bonds DE, Boyd CM, Davis MM, Finke B, Fox MH. Understanding the context of health for persons with multiple chronic conditions: moving from what is the matter to what matters 10.1370/afm.1643. Ann Fam Med 2014;12.

145. Boyd C, Fortin M. Future of Multimorbidity Research: How Should Understanding of Multimorbidity Inform Health System Design? Public Health Rev 2010;32.

146. Werner RM, Greenfield S, Fung C, Turner BJ. Measuring quality of care in patients with multiple clinical conditions: summary of a conference conducted by the Society of General Internal Medicine. J Gen Intern Med 2007;22(8):1206-11.

147. Nutting PA, Rost K, Smith J, Werner JJ, Elliot C. Competing demands from physical problems: effect on initiating and completing depression care over 6 months. Arch Fam Med 2000;9(10):1059-64.

148. Ryan M, Scott DA, Reeves C, Bate A, van Teijlingen ER, Russell EM, et al. Eliciting public preferences for healthcare: a systematic review of techniques. Health Technol Assess 2001;5(5):1-186.

232

149. Normand SL, McNeil BJ, Peterson LE, Palmer RH. Eliciting expert opinion using the Delphi technique: identifying performance indicators for cardiovascular disease. Int J Qual Health Care 1998;10(3):247-60.

150. Tran CTT, Lee DS, Flintoft V, Higginson L, Grant FC, Tu K. CCORT/CCS quality indicators for acute myocardial infarction care. Can J Cardiol 2003;19(1):38-45.

151. Hutchings A, Raine R, Sanderson C, Black N. An experimental study of determinants of the extent of disagreement within clinical guideline development groups. Qual Saf Health Care 2005;14(4):240-5.

152. Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. Journal of Advanced Nursing 2000;32(4):1008-1015.

153. Young JM, Masya LM, Solomon MJ, Shepherd HL. Identifying indicators of colorectal cancer care coordination: a Delphi study. Colorectal Dis 2013;16(1):17-25.

154. Barnsley J, Berta W, Cockerill R, MacPhail J, Vayda E. Identifying performance indicators for family practice: assessing levels of consensus. Can Fam Physician 2005;51:700-1.

155. Hsu C, Sandford BA. The Delphi Technique: Making Sense of Consensus. Practical Assessment Research & Evaluation 2007;12(10).

156. Canadian Diabetes Association Clinical Practice Guidelines Expert Committee. Canadian Diabetes Association 2013 Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada. Can J Diabetes 2013;37(suppl 1):S1-S212.

157. Petrosyan Y, Bai YQ, Koné Pefoyo AJ, Gruneir A, Thavorn K, Maxwell C, et al. The relationship between diabetes care quality anddiabetes-related hospitalizations, and the modifying role of comorbidity. Canadian Journal of Diabetes 2016(In press).

158. Hughes LD, McMurdo ME, Guthrie B. Guidelines for people not for diseases: the challenges of applying UK clinical guidelines to people with multimorbidity. Age Ageing 2012;42(1):62-9.

159. Watts G. Why the exclusion of older people from clinical research must stop. Bmj 2012;344:e3445.

160. Hochberg MC, Altman RD, April KT, Benkhalti M, Guyatt G, McGowan J, et al. American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee. Arthritis Care Res (Hoboken) 2012;64(4):465-74.

233

161. American Geriatrics Society Panel of the Pharmacological Management of Persistent Pain in Older Persons. Pharmacological management of persistent pain in older persons. J Am Geriatr Soc 2009;57(8):1331-46.

162. Kim J, Lee J, Shin CM, Lee DH, Park BJ. Risk of gastrointestinal bleeding and cardiovascular events due to NSAIDs in the diabetic elderly population. BMJ Open Diabetes Res Care 2015;3(1):e000133.

163. American Geriatrics Society updated Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc 2012;60(4):616-31.

164. Lee SJ, Boscardin WJ, Stijacic Cenzer I, Huang ES, Rice-Trumble K, Eng C. The risks and benefits of implementing glycemic control guidelines in frail older adults with diabetes mellitus. J Am Geriatr Soc 2010;59(4):666-72.

165. Mullen PM. Delphi: myths and reality. J Health Organ Manag 2003;17(1):37-52. 166. Okoli C, Pawlowski S. The Delphi method as a research tool: An example, design considerations and applications. Information and Management 2004;42(1):15-29.

167. Landeta J. Current validity of the Delphi method in social sciences. Technological Forecasting and Social Change 2006;73(5):467-482.

168. Goodman CM. The Delphi technique: a critique. J Adv Nurs 1987;12(6):729-34. 169. Campbell SM, Hann M, Roland MO, Quayle JA, Shekelle PG. The effect of panel membership and feedback on ratings in a two-round Delphi survey: results of a randomized controlled trial. Med Care 1999;37(9):964-8.

170. Corrigan JM, Donaldson MS, Kohn LT, Maguire SK, Pike KC. Crossing the Quality Chasm. A New Health System for the 21st Century. Washington, DC: Institute of Medicine, National Academy of Sciences, National Academy Press, 2001.

171. Fortin M, Bravo G, Hudon C, Vanasse A, Lapointe L. Prevalence of multimorbidity among adults seen in family practice. Ann Fam Med 2005;3(3):223-228.

172. Valderas JM, Starfield B, Sibbald B, Salisbury C, Roland M. Defining comorbidity: implications for understanding health and health services. Ann Fam Med 2009;7(4):357-63.

173. Salisbury C. Multimorbidity: redesigning health care for people who use it. Lancet 2012;380(9836):7-9.

174. Freund T, Kunz CU, Ose D, Peters-Klimm F. Patterns of multimorbidity in primary care patients at high risk of future hospitalization 10.1089/pop.2011.0026. Popul Health Manag 2012;15.

234

175. Zhao Y, Wright J, Guthridge S, Lawton P. The relationship between number of primary health care visits and hospitalisations: evidence from linked clinic and hospital data for remote Indigenous Australians. BMC Health Services Research 2013;13:466.

176. Barkin RL, Beckerman M, Blum SL, Clark FM, Koh EK, Wu DS. Should nonsteroidal anti-inflammatory drugs (NSAIDs) be prescribed to the older adult? Drugs Aging 2010;27(10):775-89.

177. Arboe B, Ulrik CS. Beta-blockers: friend or foe in asthma? Int J Gen Med 2013;6:549-55.

178. Wami WM, Buntinx F, Bartholomeeusen S, Goderis G, Mathieu C, Aerts M. Influence of chronic comorbidity and medication on the efficacy of treatment in patients with diabetes in general practice. Br J Gen Pract 2013;63(609):e267-73.

179. Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care 2002;25(3):512-6.

180. Hux JE, Tang M. Patterns of prevalence and incidence of diabetes. In: Hux, J. E., Booth, G.L., Slaughter, P.M., et al. Diabetes in Ontario: an ICES Practice Atlas.Toronto, ON. Institute for Clinical Evaluative Sciences 2003:1.1-1.18.

181. Fiest KM, Jette N, Quan H, St Germaine-Smith C, Metcalfe A, Patten SB, et al. Systematic review and assessment of validated case definitions for depression in administrative data. BMC Psychiatry 2014;14:289.

182. Lix L, Yogendran M, Burchill C, Metge C, McKeen N, Moore D, et al. Defining and Validating Chronic Diseases: An Administrative Data Approach. Winnipeg, Manitoba Centre for Health Policy. 2006.

183. Marshall M, Klazinga N, Leatherman S, Hardy C, Bergmann E, Pisco L, et al. OECD Health Care Quality Indicator Project. The expert panel on primary care prevention and health promotion. Int J Qual Health Care 2006;18 Suppl 1:21-5.

184. OECD: Health at a Glance 2009. In OECD Indicators. Paris, France: OECD; 2009. 185. Kralj B. Measuring Rurality - RIO2008_BASIC:Methodology and Results. Available at: https://www.oma.org/Resources/Documents/2008RIO-FullTechnicalPaper.pdf. Accessed September 17, 2013.

186. Gruneir A, Forrester J, Camacho X, Gill SS, Bronskill SE. Gender differences in home care clients and admission to long-term care in Ontario, Canada: a population-based retrospective cohort study 10.1186/1471-2318-13-48. BMC Geriatr 2013;13.

187. Wooder SD. Primary care compensation models. Ontario Medical Association 2011.

235

189. Krein SL, Heisler M, Piette JD, Makki F, Kerr EA. The effect of chronic pain on diabetes patients' self-management. Diabetes Care 2005;28(1):65-70.

190. Laiteerapong N, Huang ES, Chin MH. Prioritization of care in adults with diabetes and comorbidity. Ann N Y Acad Sci 2011;1243:69-87.

191. American Diabetes Association. Standards of medical care in diabetes--2011. Diabetes Care;34 Suppl 1:S11-61.

192. Calderon-Larranaga A, Abad-Diez JM, Gimeno-Feliu LA, Marta-Moreno J, Gonzalez-Rubio F, Clerencia-Sierra M, et al. Global health care use by patients with type-2 diabetes: Does the type of comorbidity matter? Eur J Intern Med 2015;26(3):203-10.

193. Kalyani RR, Saudek CD, Brancati FL, Selvin E. Association of diabetes, comorbidities, and A1C with functional disability in older adults: results from the National Health and Nutrition Examination Survey (NHANES), 1999-2006. Diabetes Care 2010;33(5):1055-60.

194. Nichols GA, Joshua-Gotlib S, Parasuraman S. Glycemic control and risk of cardiovascular disease hospitalization and all-cause mortality. J Am Coll Cardiol 2013;62(2):121-7.

195. Malabu UH, Vangaveti VN, Kennedy RL. Disease burden evaluation of fall-related events in the elderly due to hypoglycemia and other diabetic complications: a clinical review. Clin Epidemiol 2014;6:287-94.

196. Vu HT, Keeffe JE, McCarty CA, Taylor HR. Impact of unilateral and bilateral vision loss on quality of life. Br J Ophthalmol 2005;89(3):360-3.

197. Menec VH, Sirski M, Attawar D, Katz A. Does continuity of care with a family physician reduce hospitalizations among older adults? 10.1258/135581906778476562. J Health Serv Res Policy 2006;11.

198. Saultz JW, Lochner J. Interpersonal continuity of care and care outcomes: a critical review. Ann Fam Med 2005;3.

199. Worall G, Knight J. Continuity of care is good for elderly people with diabetes. Retrospective cohort study of mortality and hospitalization. Canadian Family Physician 2011;57:e16-20.

200. Flaherty JH, Perry HM, 3rd, Lynchard GS, Morley JE. Polypharmacy and hospitalization among older home care patients. J Gerontol A Biol Sci Med Sci 2000;55(10):M554-9.

236

201. Sganga F, Landi F, Ruggiero C, Corsonello A, Vetrano DL, Lattanzio F, et al. Polypharmacy and health outcomes among older adults discharged from hospital: results from the CRIME study. Geriatr Gerontol Int 2014;15(2):141-6.

202. Hohl CM, Dankoff J, Colacone A, Afilalo M. Polypharmacy, adverse drug-related events, and potential adverse drug interactions in elderly patients presenting to an emergency department. Ann Emerg Med 2001;38(6):666-71.

203. Hilmer SN, Gnjidic D. The effects of polypharmacy in older adults. Clin Pharmacol Ther 2009;85(1):86-8.

204. Tatti P, Pahor M, Byington RP, Di Mauro P, Guarisco R, Strollo G, et al. Outcome results of the Fosinopril Versus Amlodipine Cardiovascular Events Randomized Trial (FACET) in patients with hypertension and NIDDM. Diabetes Care 1998;21(4):597-603.

205. Niskanen L, Hedner T, Hansson L, Lanke J, Niklason A. Reduced cardiovascular morbidity and mortality in hypertensive diabetic patients on first-line therapy with an ACE inhibitor compared with a diuretic/beta-blocker-based treatment regimen: a subanalysis of the Captopril Prevention Project. Diabetes Care 2001;24(12):2091-6.

206. Patel A, MacMahon S, Chalmers J, Neal B, Woodward M, Billot L, et al. Effects of a fixed combination of perindopril and indapamide on macrovascular and microvascular outcomes in patients with type 2 diabetes mellitus (the ADVANCE trial): a randomised controlled trial. Lancet 2007;370(9590):829-40.

207. Berl T, Hunsicker LG, Lewis JB, Pfeffer MA, Porush JG, Rouleau JL, et al. Cardiovascular outcomes in the Irbesartan Diabetic Nephropathy Trial of patients with type 2 diabetes and overt nephropathy. Ann Intern Med 2003;138(7):542-9.

208. Cheng J, Zhang W, Zhang X, Han F, Li X, He X, et al. Effect of angiotensin- converting enzyme inhibitors and angiotensin II receptor blockers on all-cause mortality, cardiovascular deaths, and cardiovascular events in patients with diabetes mellitus: a meta- analysis. JAMA Intern Med 2014;174(5):773-85.

209. Harel Z, Silver SA. ACE inhibitors are associated with a reduction in all-cause mortality versus angiotensin II receptor blockers in patients with diabetes mellitus 10.1136/ebmed-2014-110048 Evidence Based Medicine 2014

210. Bell AD, Roussin A, Cartier R, Chan WS, Douketis JD, Gupta A, et al. The use of antiplatelet therapy in the outpatient setting: Canadian Cardiovascular Society guidelines. Can J Cardiol 2011;27 Suppl A:S1-59.

211. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. Bmj 2002;324(7329):71-86.

237

212. Antithrombotic Trialists' Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002;324:71–86.

213. Armitage J, Bowman L. Cardiovascular outcomes among participants with diabetes in the recent large statin trials. Curr Opin Lipidol 2004;15(4):439-46.

214. Collins R, Armitage J, Parish S, Sleigh P, Peto R. MRC/BHF Heart Protection Study of cholesterol-lowering with simvastatin in 5963 people with diabetes: a randomised placebo-controlled trial. Lancet 2003;361(9374):2005-16.

215. Shepherd J, Blauw GJ, Murphy MB, Bollen EL, Buckley BM, Cobbe SM, et al. Pravastatin in elderly individuals at risk of vascular disease (PROSPER): a randomised controlled trial. Lancet 2002;360(9346):1623-30.

216. Zhang W, Nuki G, Moskowitz RW, Abramson S, Altman RD, Arden NK, et al. OARSI recommendations for the management of hip and knee osteoarthritis: part III: Changes in evidence following systematic cumulative update of research published through January 2009. Osteoarthritis Cartilage 2010;18(4):476-99.

217. Hernandez-Diaz S, Rodriguez LA. Association between nonsteroidal anti- inflammatory drugs and upper gastrointestinal tract bleeding/perforation: an overview of epidemiologic studies published in the 1990s. Arch Intern Med 2000;160(14):2093-9. 218. Richy F, Bruyere O, Ethgen O, Rabenda V, Bouvenot G, Audran M, et al. Time dependent risk of gastrointestinal complications induced by non-steroidal anti-inflammatory drug use: a consensus statement using a meta-analytic approach. Ann Rheum Dis 2004;63(7):759-66.

219. American Geriatrics Society 2015 Updated Beers Criteria for Potentially Inappropriate Medication Use in Older Adults. J Am Geriatr Soc;63(11):2227-46.

220. Fortin M, Stewart M, Poitras ME, Almirall J, Maddocks H. A systematic review of prevalence studies on multimorbidity: toward a more uniform methodology. Ann Fam Med 2012;10(2):142-51.

221. Austin PC, van Walraven C, Wodchis WP, Newman A, Anderson GM. Using the Johns Hopkins Aggregated Diagnosis Groups (ADGs) to predict mortality in a general adult population cohort in Ontario, Canada. Medical care 2011;49(10):932-939.

222. Menendez ME, Neuhaus V, van Dijk CN, Ring D. The Elixhauser comorbidity method outperforms the Charlson index in predicting inpatient death after orthopaedic surgery. Clin Orthop Relat Res 2014;472(9):2878-86.

223. Fried TR, O'Leary J, Van Ness P, Fraenkel L. Inconsistency over time in the preferences of older persons with advanced illness for life-sustaining treatment. J Am Geriatr Soc 2007;55(7):1007-14.

238

224. Fried TR, Tinetti M, Agostini J, Iannone L, Towle V. Health outcome prioritization to elicit preferences of older persons with multiple health conditions. Patient Educ Couns 2011;83(2):278-82.

225. Christensen B, Howren M, Hillis S. Patient and physician beliefs about control over health: association of symmetrical beliefs with medication regimen adherence. J Gen Intern Med 2010;25(5):397-402.

226. Cheraghi-Sohi S, Bower P, Kennedy A, Morden A, Rogers A, Richardson J, et al. Patient priorities in osteoarthritis and comorbid conditions: a secondary analysis of qualitative data. Arthritis Care Res (Hoboken) 2012;65(6):920-7.

227. Mulley AG, Trimble C, Elwyn G. Stop the silent misdiagnosis: patients' preferences matter. Bmj 2012;345:e6572. 228. Salisbury C. Multimorbidity: time for action rather than words. Br J Gen Pract 2013;63(607):64-5.

229. Meyers JL, Parasuraman S, Bell KF, Graham JP, Candrilli SD. The high-cost, type 2 diabetes mellitus patient: an analysis of managed care administrative data. Arch Public Health 2013;72(1):6.

230. Canadian Coalition for Seniors’ Mental Health. National guidelines for seniors’ mental health: The assessment and treatment of depression. Toronto, ON: Canadian Coalition for Seniors’ Mental Health; 2006.

239

APPENDICES

Appendix 1. Search strategies

Appendix 1.1 Search strategy for diabetes quality indicators

Ovid MEDLINE(R), Ovid MEDLINE(R) In-Process

# Searches

quality of health care/ or guidelines as topic/ or exp quality assurance, health care/ or benchmarking/ or report card/ or exp quality indicators, health care/ or exp "outcome and process assessment (health care)"/ or ((quality adj2 (healthcare or health care)) or ((("health 1 care" or healthcare) adj3 benchmarking) or (best adj practice adj analysis) or benchmark$) or ((quality adj3 indicator? adj3 healthcare) or (quality adj2 indicator) or (indicator$ adj measure$)) or ((measure? or assessment?) adj4 (outcome or process))).ti,ab. exp Diabetes Mellitus, Type 1/ or diabetes.mp. or exp Diabetes Complications/ or exp 2 Diabetes Mellitus, Type 2/ or exp Diabetes Insipidus/ or exp Diabetes Mellitus/

3 1 and 2 limit 3 to (english language and humans and yr="2000 - current" and ("all aged (65 and 4 over)")

240

Appendix 1.2 Search strategy for hypertension / ischemic heart disease quality indicators

Ovid MEDLINE(R), Ovid MEDLINE(R) In-Process

# Searches

quality of health care/ or guidelines as topic/ or exp quality assurance, health care/ or benchmarking/ or report card/ or exp quality indicators, health care/ or exp "outcome and process assessment (health care)"/ or ((quality adj2 (healthcare or health care)) or ((("health 1 care" or healthcare) adj3 benchmarking) or (best adj practice adj analysis) or benchmark$) or ((quality adj3 indicator? adj3 healthcare) or (quality adj2 indicator) or (indicator$ adj measure$)) or ((measure? or assessment?) adj4 (outcome or process))).ti,ab. exp Coronary Artery Disease/di, dt, pc, rh, th [Diagnosis, Drug Therapy, Prevention & 2 Control, Rehabilitation, Therapy]

exp Hypertension, Renal/ or exp Hypertension/ or exp Hypertension, Portal/ or exp 3 Hypertension, Renovascular/ or hypertension.mp. or exp Hypertension, Malignant/ 4 ((systolic or diastolic or arterial) adj3 pressur*).ti,ab. 5 exp Antihypertensive Agents/ or exp Hypertension/ or exp Blood Pressure/ 6 2 or 3 or 4 or 5 7 1 and 6 limit 7 to (english language and humans and yr="2000 - current" and ("all aged (65 and 8 over)")

241

Appendix 1.3 Search strategy for depression quality indicators

Ovid MEDLINE(R), Ovid MEDLINE(R) In-Process

# Searches

quality of health care/ or guidelines as topic/ or exp quality assurance, health care/ or benchmarking/ or report card/ or exp quality indicators, health care/ or exp "outcome and process assessment (health care)"/ or ((quality adj2 (healthcare or health care)) or ((("health 1 care" or healthcare) adj3 benchmarking) or (best adj practice adj analysis) or benchmark$) or ((quality adj3 indicator? adj3 healthcare) or (quality adj2 indicator) or (indicator$ adj measure$)) or ((measure? or assessment?) adj4 (outcome or process))).ti,ab. exp Depression/di, dt, mo, pc, th [Diagnosis, Drug Therapy, Mortality, Prevention & 2 Control, Therapy]

exp Depressive Disorder/di, dt, mo, pc, th [Diagnosis, Drug Therapy, Mortality, Prevention 3 & Control, Therapy]

4 2 or 3 5 1 and 4 limit 5 to (english language and humans and yr="2000 - current" and ("all aged (65 and 6 over)")

242

Appendix 1.4 Serch strategy for osteoarthritis quality indicators

Ovid MEDLINE(R), Ovid MEDLINE(R) In-Process

# Searches

quality of health care/ or guidelines as topic/ or exp quality assurance, health care/ or benchmarking/ or report card/ or exp quality indicators, health care/ or exp "outcome and process assessment (health care)"/ or ((quality adj2 (healthcare or health care)) or ((("health 1 care" or healthcare) adj3 benchmarking) or (best adj practice adj analysis) or benchmark$) or ((quality adj3 indicator? adj3 healthcare) or (quality adj2 indicator) or (indicator$ adj measure$)) or ((measure? or assessment?) adj4 (outcome or process))).ti,ab. osteoarthritis.mp. or exp Osteoarthritis, Hip/ or exp Osteoarthritis/ or exp Osteoarthritis, 2 Spine/ or exp Osteoarthritis, Knee/

3 arthritis, degenerative.mp. or exp Osteoarthritis/

4 2 or 3 5 1 and 4 limit 5 to (english language and humans and yr="2000 - current" and ("all aged (65 and 6 over)")

243

Appendix 2. The AIRE instrument: domains and items

1. Purpose, relevance and organizational context

1. The purpose of the indicator is described clearly.

2. The criteria for selecting the topic of the indicator are described in detail.

3. The organizational context of the indicator is described in detail.

4. The quality domain the indicator addresses is described in detail.

5. The health care process covered by the indicator is described and defined in detail.

2. Stakeholder involvement

1. The group developing the indicator includes individuals from all relevant

professional groups.

2. Considering the purpose of the indicator, all relevant stakeholders have been

involved at some stage of the development process.

3. The indicator has been formally endorsed.

3. Scientific evidence

1. Systematic methods were used to search for scientific evidence.

2. The indicator is based on recommendations from an evidence-based guideline or

studies published in peer-reviewed scientific journals

3. The supporting evidence has been critically appraised.

4. Additional evidence, formulation, usage

244

1. The numerator and denominator are described in detail.

2. The target patient population of the indicator is defined clearly.

3. A strategy for risk adjustment has been considered and described.

4. The indicator measures what it is intended to measure (validity).

5. The indicator measures accurately and consistently (reliability).

6. The indicator has sufficient discriminative power.

7. The indicator has been piloted in practice.

8. The efforts needed for data collection have been considered.

9. Specific instructions for presenting and interpreting results.

245

Appendix 3. Full list of extracted indicators

Appendix 3a. Full list of extracted indicators for ambulatory care for diabetes

Category Indicator

Process of care 1 or 2 HbA1c testing per year Self-monitoring of blood glucose Annual LDL cholesterol testing Annual screening for microalbuminuria Annual eye exam Routine foot examination Regular measurement of blood pressure Patients with systolic blood pressure >140 and prescribed any antihypertensive drug Patients prescribed a second antihypertensive drug from a different class if systolic blood pressure remained >140 with first class of antihypertensive drug Patients without hypertension with albuminuria prescribed ACE inhibitor or ARB Patients with hypertension and history of ischaemic heart disease or myocardial infarction prescribed b-blocker Patients with HbA1c >7% and prescribed any oral antihyperglycaemic agent or insulin Patients with high cardiovascular risk who are prescribed a statin Percentage diabetes patients with high cardiovascular risk who are prescribed ASA Smoking cessation counseling Weight/body mass index (BMI) control Nutritional counseling provided by the nutrition service Advise to practice aerobic physical exercise of moderate intensity, at least 150 minutes per week, unless contraindicated Registration of adherence to dietary recommendations Registration of adherence to aerobic physical exercise

246

Overweight/obese (BMI≥ 25 kg/m2) patients who received metformin, unless contraindicated Short-term outcomes HbA1c <7% or fasting glucose ≤130 mg/dl in the last 3 measurements Total cholesterol levels<200 mg/dl in the last measurement Blood pressure <130/80 mmHg in the last 3 measurements Overweight/obese (BMI≥ 25 kg/m2) patients who lost ≥5% body weight in the last year Long-term outcomes Lower extremity amputation rates Renal disease in people with diabetes Cardiovascular mortality in patients with diabetes

247

Appendix 3b. Full list of extracted indicators for ambulatory care for hypertension

Category Indicator

Process of care Use of Beta-blockers if comorbid ischemic heart disease Use of ACE inhibitors if history of HF or left ventricular hypertrophy/c Pharmacotherapy if not responsive to lifestyle modification/BP above target level with target organ damage or independent CV risk factors First-line pharmacotherapy for diabetes patients include an ACE inhibitors Parenteral hypertensive therapy in a monitored hospital setting in malignant hypertension Non-pharmacological intervention for newly diagnosed (diet, exercise, weight loss, reduced alcohol) Documentation of alcohol intake within 3 months in newly diagnosed Documentation of an intervention/ suspected reason if persistent systolic BP above goal Documentation of change in therapy/repeat counseling if BP above goal Documentation of plan of care if BP above target levels Appropriate BP measurement frequency and levels Cardiovascular disease risk assessment in newly diagnosed patients with hypertension Renal function assessment in newly diagnosed patients with hypertension Short-term outcome BP target level achievement

248

Appendix 3c. Full list of extracted indicators for ambulatory care for chronic ischemic heart disease

Category Indicator

Process of care Use of beta-blockers Use of ACE inhibitors/ARBs Use of statins Use of antiplatelet therapy Risk assessment if BMI ≥25 Lipid profile assessment Blood glucose level assessment ECG in newly diagnosed patients with ischemic heart disease Referral to exercise testing and/or coronary angiography Revascularization offered in significant LM or three-vessel disease Smoking cessation counseling

249

Appendix 3d. Full list of extracted indicators for ambulatory care for major depression

Category Indicator

Structure indicator Written guidelines are in place to ensure that, where services are not provided locally, GPs can refer patients outside their locality. Specialist services are based on locally agreed written service plans and agreements which include the range, quality, and volume of mental health services, including depression There is a demonstrable commitment to promote continuous professional and practice development in primary care; practices are offered protected time for GPs and nurses to attend appropriate training courses There is a range of collaborative initiatives in place with other key agencies demonstrating effective partnerships There is an agreed definition of depressive disorders which is explicit and standard within the practice There is a written complaints procedure which is prominently displayed regarding the provided care There is a clear referral and feedback procedure for the practice counsellor Patients are able to make a routine appointment to see a general practitioner within 2 days A member of the primary health care team is available as a point of contact for all patients to talk to in an emergency; clear written practice protocols are in place for obtaining specialist help in an emergency/crisis situation There is equity of access to talking treatments regardless of ethnic origin, age, place of residence, socioeconomic status, and sex There is good access to integrated and community- based mental health services out of hours. There are locally agreed written standards and protocols for the delivery of out of hours care for mental health problems. There is evidence of monitoring to ensure that out of hours standards are met There are agreed written protocols and guidelines, based on best available evidence, for prescribing and monitoring psychotropic medication Confidential discussions take place in private. There is an appropriate (i.e.

250

private, quiet, relatively non-clinical) room for counselling/visiting mental health staff The confidentiality of medical records is protected and ensured at all times; where practicable, patient consent is sought before giving information to carers Process of care Patient education at least once during the measurement period regarding depression, depression treatment, prescribed medication and coping strategies Patients with depression are treated as individuals with individual needs and not as a “diagnosis of depression”. Treatment plans are individually tailored for each patient Staff treat all patients with depression registered with the practice with respect, courtesy and consideration irrespective of age, sex, religious/cultural beliefs, or diagnosis. Staff are aware that patients with depression may be concerned about feelings of stigmatisation and are treated in a way to minimise these feelings. Management time is available to support and lead change in service development; patients are not made to feel that they are wasting health professional’s’ time Staff are aware of the potential impact of a depressive disorders on patient behaviour Patient’s views about their condition are explicitly sought to help treatment adherence Patients are as fully involved as practicable in the formulation and delivery of their care and in any decisions about referral; where practicable, patients are informed of the reasons for referral to specialists or other professionals Details of currently prescribed maintenance drugs are prominently recorded in the medical record. Medical records, including computerized records, are up to date and summarized Patients who were on repeat maintenance drugs and offered regular reviews of their medication including monitoring for possible side effects and interactions with other drugs Physical symptoms in patients with depression are taken seriously and not automatically considered as psychosomatic; assessment takes into account language barriers, the needs of people with disabilities, ethnic, cultural and religious preferences. Patients were screened for depression using an age appropriate standardized tool and had follow-up plan documented, during the initial

251 primary care evaluation and annually Patients who have in the medical record at least three of the nine DSM-IV target symptoms for major depression were documented within 2 weeks of diagnosis Patients were assessed for suicidal ideation at initial evaluation Patients were evaluated for substance dependence or abuse for men hypothyroidism for women, within 1 month or in the prior 3 months Patients were offered antidepressant treatment, psychotherapy, or electroconvulsive therapy within 2 weeks after diagnosis No drug is prescribed unless the health professional understands the potential efficacy and side effects; prescribing for depression is based on up to date evidence and, where available, local management protocols Patients who were not responding to first line drug treatment at the therapeutic dosage and were asked about adherence Patients who were experiencing difficulties undertaking withdrawal from medication and were offered referral to a mental health worker Patients were prescribed antidepressants using tertiary amine tricyclics, MAOIs (unless atypical depression is present), benzodiazepines, or stimulants (except methylphenidate) as first- or second-line therapy Patients were prescribed anticholinergic antidepressants as first- or second-line therapy Choice of medication is based on individual patient factors including the desirability of sedation, previous response to a drug treatment including adverse reactions, co-morbid psychiatric or medical conditions, concurrent drug treatment, and relative risk of medication in overdose If a patient with a diagnosis of major depression or dysthymia is taking an selective serotonin reuptake inhibitor (SSRI), then an MAOI should not be used for at least 2 weeks after termination of the SSRI, and vice versa Patients who responded to antidepressant medication and remained on an antidepressant treatment for at least 3 months (12 weeks) Patients who had no meaningful symptom response after 6 weeks of psychotherapy treatment (without medication) and for whom the medication treatment has been initiated, a patients was referred to a psychiatrist by the 8th week of depression treatment Patients who had no meaningful symptom response after 6 weeks of drug treatment and the drug dose was optimized or changed, or a patient was referred to a psychiatrist by the 8th week of depression treatment Patients who responded only partially after 12 weeks of psychotherapy treatment (without medication) and who had been prescribed a medication by the 16th week of treatment, or ECT had been considered; or a patients

252

has been referred to a psychiatrist Patients who responded only partially after 12 weeks of drug treatment and who had been switched to a different medication class, or a second medication had been added to the first, or psychotherapy was added Patients who responded to antidepressant medication, remained on the drug at the same dose for at least 6 months Patients who experienced three or more episodes of depression and received maintenance antidepressant medication with the same type and dose of medication for at least 24 months, with at least four office or telephone visits for depression during that period Patients who had a systematic symptom assessment at 12 weeks following diagnosis, or if in remission by week 12, then a systematic symptom assessment is performed at the time of remission. Patients who had been referred to a psychiatrist or received treatment with a combination of an antidepressant and an antipsychotic If a patient is newly treated for depression, then degree of response to at least two of the nine DSM-IV target symptoms for major depression and, if he or she is taking antidepressant medications, medication side effects should be documented at the first follow-up visit to the same physician or to a mental health provider within 4 weeks of treatment initiation Patients with severe depression are offered regular appointments to monitor and follow up treatment, symptoms, side effects and adherence Outcome indicators Patients with an initial PHQ-9 score greater than nine who achieve remission at six months as demonstrated by a six month (+/- 30 days) PHQ-9 score of less than five. Patients with an initial PHQ-9 score greater than nine who achieve remission at twelve months as demonstrated by a twelve month (+/- 30 days) PHQ-9 score of less than five. Patients with an initial PHQ-9 score greater than nine who achieve a response at six months as demonstrated by a six month (+/- 30 days) PHQ-9 score that is reduced by 50% or greater from the initial PHQ- 9 score. Patients with an initial PHQ-9 score greater than nine who achieve a response at twelve months as demonstrated by a twelve month (+/- 30 days) PHQ-9 score that is reduced by 50% or greater from the initial PHQ-9 score.

253

Appendix 3e. Full list of extracted indicators for ambulatory care for osteoarthritis

Category Indicator

Structure All professionals managing patients with osteoarthritis at a primary care setting should have continuous access to education on important preventive and therapeutic strategies in the management of osteoarthritis If a patient is diagnosed with osteoarthritis has been referred to an orthopedic surgeon, then the waiting time from first referral should not exceed three months Process of care Patients for whom a physical examination (visual inspection, palpation, range of motion) of the involved joint was performed during the initial visit Patients with a documentation of assessed pain annually and when new to a primary care Patients with a documentation of assessed functional status annually and when new to a primary care Patients who have been given individually tailored education about the natural history, treatment and self-management of the disease at least once Patients who have been advised annually to lose weight Patients who have been advised at least annually to lose weight and the benefit of weight loss on the symptoms of osteoarthritis have been explained Patients who received referral to a weight loss program Patients who has no contraindication to exercise and are physically and mentally able to exercise was offered an individualized exercise program or was referred to a relevant health professional Patients with a documentation of the assessment for use of anti- inflammatory or analgesic over-the-counter medications Patients who were prescribed acetaminophen first, unless there is a documented contraindication to use Patients who have been advised of the risk of liver toxicity measurement period If oral pharmacologic therapy for osteoarthritis is changed from acetaminophen to a different oral agent, then there should be evidence that the patient has had a trial of maximum dose acetaminophen (suitable for age and comorbid conditions) Patients who are taking a NSAID with a documented assessment for gastrointestinal and renal risk factors Patients who are taking a NSAID in combination with misoprostol or

254

proton pump inhibitor Patients for whom the need for ambulatory assistive devices has been assessed Patients who have failed to respond to pharmacological/ non- pharmacological treatment and have been referred to an orthopedic surgeon Outcome Patients who have reported improvement of functional ability by 20% indicators within three months after initiation/change of pharmacological/ non- pharmacological treatment Patients who have reported reduction of pain level by 20% within three months after initiation/change of pharmacological/ non-pharmacological treatment

255

Appendix 4. Email text – Delphi Round I

Dear (name of potential participant),

I am contacting you to request your participation in a Delphi study that aims to develop a set of quality indicators to assess overall quality of care of older diabetes patients with comorbid conditions. You have been identified as an expert in (general practice/geriatric practice/clinical pharmacy/research). I would like to invite you to participate in my PhD study which is supervised by Dr. Walter Wodchis at the University of Toronto. The Thesis

Committee members are Dr. Jan Barnsley, Dr. Barbara Liu and Dr. Kerry Kuluski.

This study is using a Delphi technique to define the most appropriate set of quality indicators for assessing quality of overall care of older diabetes patients with comorbid concordant and discordant chronic conditions. As you are likely aware, while diabetes occurs mostly in conjunction with other comorbid conditions, most diabetes management programs and guidelines are entirely focused on a single condition and do not address the challenges that primary care physicians and patients face when managing multiple concomitant conditions.

Various quality measures have been developed for assessing care for single diseases.

However, adherence to disease-specific measures for patients with multiple chronic conditions may lead to the unintended consequence of delivering inappropriate care.

Therefore, it is crucial to identify measures that would address the heterogeneity and scope of care for a particular individual with particular types of co-existing conditions to improve the quality of care of people with multimorbidity.

256

Your participation will be anonymous, and will consist of responding to 2-3 electronic questionnaires, and each round will take approximately 20 minutes to complete. After completion of all Delphi rounds you will be given a cheque for $200 to compensate you for any disruption to your practice.

I have attached an information letter to this email. If you have any questions or concerns please contact me at [email protected].

Thank you for your consideration,

Walter P Wodchis, PhD Yelena Petrosyan, PhD(c) Associate Professor Institute of Health Policy, Management and Evaluation Institute of Health Policy, Management and University of Toronto Evaluation Email address: [email protected] University of Toronto Email address: [email protected]

257

Appendix 5. Participant information sheet – Delphi Round I

Methodological assessment and selection of indicators in the context of various

combinations of concurrent chronic conditions

We would like to invite you to take part in a Delphi consensus study. Before you decide whether or not you would like to take part, it is important for you to consider why the research is being done and what it will involve. Please read this information sheet carefully.

Background

As you are likely aware, various quality measures have been developed for assessing care for single diseases. However, adherence to disease-specific measures for patients with multiple chronic conditions may lead to the unintended consequence of delivering inappropriate care that is not aligned with the patient’s goals and preferences. Therefore, it is crucial to identify measures that assess the quality of overall care among older diabetes patients with comorbidities in order to improve their care.

This project is focused on older adults because they are more likely than younger individuals to have comorbid chronic conditions that can be complex and difficult to manage. Diabetes has been chosen as a condition of interest due to the high burden of co-existing chronic conditions in this group of patients. Four chronic conditions among diabetes patients, including hypertension, ischemic heart disease, depression and osteoarthritis, were selected

258 for the purpose of this project, since they have been identified as the most common and costly conditions among people diagnosed with diabetes. Next, these chronic conditions were grouped in five disease combinations that represented most prevalent clusters of concurrent conditions across multimorbidity groupings based on the prior research results. Finally, the defined five disease combinations were categorized into three groups by comorbidity type, including diabetes-concordant (have similarities with diabetes with respect to the disease management plan), diabetes-discordant (are not directly related to diabetes in their pathogenesis and management plan) and both types of comorbid conditions, in particular:

 Concordant conditions:

o Diabetes with comorbid hypertension;

o Diabetes with comorbid hypertension and ischemic heart disease;

 Discordant conditions:

o Diabetes with comorbid depression;

o Diabetes with comorbid depression and osteoarthritis;

 Both types of conditions:

o Diabetes with comorbid osteoarthritis and hypertension.

First, a systematic review was conducted to identify a set of published quality indicators for assessing care for five selected disease categories, including indicators for management of single diseases as well as co-existing chronic conditions. The results showed that most of the existing indicators have been developed for assessing care for single diseases. There is limited number of indicators regarding management of two co-existing conditions, and, in particular, they relate to management of concordant conditions. The indicators that are

259 available in administrative databases have been extracted for the purpose of this project. The list of candidate indicators for each particular disease combination will be critically appraised by the expert panel by means of Delphi technique. The final set of appropriate indicators, as judged by the Delphi Panel, will be used in empirical analyses for assessing overall quality of care of older people with selected combinations of the concurrent conditions.

Purpose of the study

This study is using a Delphi technique to select the most appropriate set of indicators for assessing overall quality of care of older diabetes patients with comorbid concordant and discordant chronic conditions.

Methods

The Delphi technique will be used for the purpose of this project. The Delphi technique is a structured process that uses questionnaires to gather information and reach a consensus. It is anticipated that this process will take 2-3 rounds. On the first round, you will be asked to rank the indicators based on the criteria, including: meaningfulness, potential for improvements in clinical practices and overall value of inclusion for process indicators, and importance, modifiability and overall value of inclusion for outcome indicators, on a 5-point

Likert scale, with respect to each selected disease combination. You will also be asked to

260 suggest additional items that you think are important in terms of management of selected disease combinations.

Participation will be via an internet based survey. If you prefer, a regular mail option will be available and you will be mailed a questionnaire with a stamped envelope for reply. After the results of the first round have been analyzed, you will be sent rankings of the items. Items that have received poor rankings will not be included. Although suggested additional items will be included, it is expected that there will be fewer items to rate following the first round.

Each round will take approximately 20 minutes to complete, and there will be 2-3 rounds.

This Delphi process will take place from October, 2015 to January, 2016. In order to allow timely conclusion of the study we would respectfully request a response time of 2 weeks for completion of each round.

Confidentiality

No personal information will be collected and survey responses will be collated anonymously. All responses received in the study will be strictly confidential, and your identity will not be divulged. Direct quotes to free‐text answers will be used as part of the study report or later Delphi iterations, but these will be not be traceable back to you. Any information that could be used to identify you or the organization you are affiliated with will be removed.

261

Compensation

After completion of all three rounds, you will be given a cheque for $200 to compensate you for any disruption to your practice. If you withdraw from the study before the end of the third

Delphi round, you will not receive compensation.

Your participation in the study

Participation in this study is voluntary and there are no risks to you from being involved.

You may withdraw from the study at any time. There are no consequences if you withdraw from the study. Because the study involves iterations that quickly return aggregated results to participants as part of the study design, data cannot be withdrawn (as they will have already been used to compile aggregated results and returned to study participants).

The input that you provide will facilitate the development of quality indicators applicable to administrative database for assessing overall quality of care for older diabetes patients with comorbidities. The results of this study will be submitted to an academic peer-reviewed journal for publication and shared with the participants of the panel. The emergent indicators will be used in an empirical study of care provided to patients with diabetes. The study results will help inform policy and decision makers about potential improvements in the quality of care for older diabetes patients with comorbidities.

262

Your participation

Thank you for reading this information sheet and for considering taking part in this research. If you are happy to proceed please indicate you have given consent and complete the following survey.

If you choose to participate, you have 3 options for submitting your questionnaire: 1. Complete on the Web at http://fluidsurveys.com/s/MultimorbidityQualityIndicators/?code=fsszx6fczr; or 2. Complete the word document provided and send via email to [email protected], or 3. Contact [email protected] for a paper copy of the questionnaire and a stamped envelope.

We are asking you to fill out the rating form and return them no later than October 21, 2015.

We are pleased to enclose herewith the materials you will need for the first round ratings:

 Survey questionnaire;  Criteria and instructions for rating indicators;  List of quality indicators to be rated with detailed specification, by each selected disease category.

If you have any questions about your rights as a research participant, you may contact the

Office of Research Ethics ([email protected] or 416-946-3273).

If you have any questions regarding this research, please contact me at the email address

provided below.

Thank you in advance for your time and cooperation.

263

Appendix 6. Survey questionnaire, Delphi Round I

Survey questionnaire

Please circle or highlight ONE number below for each candidate indicator rating how much you agree or disagree with each statement, including meaningfulness, potential for improvements in clinical practices and overall value of inclusion for process indicators, and importance, modifiability and overall value of inclusion for outcome indicators. A score of one will indicate the lowest rating and a score of five will indicate the highest possible rating.

You are also invited to comment on each item using the dedicated "comment box", and/or to add items, including both process and outcome indicators, for each of the disease combinations selected for this project.

Only indicators applicable to administrative databases have been considered for this project.

Please consider all diseases presented in various combinations as if they are at moderate severity.

The final set of appropriate indicators, as judged by the Delphi panel, will be used in empirical analyses assessing the overall quality of care for older people with the selected combinations of concurrent conditions.

264

1. Quality indicators to assess the quality of care of older adults with controlled diabetes with comorbid hypertension (“1” is the lowest rating and “5” is the highest rating)

Controlled diabetes with comorbid hypertension

Potential for improvements Overall value of ID Process Indicator Meaningfulness Comment box in clinical inclusion practices

HbA1c testing 1 every 6 month 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

LDL-cholesterol

testing once per 2 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 year

Eye examination 3 every two years 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

265

Microalbumin

testing once per 4 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 year

Statin therapy 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

ACEI or ARB 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 therapy 6

Antiplatelet

therapy 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

This is where you can list any other indicators for this disease combination

266

Controlled diabetes with comorbid hypertension

Outcome Overall value ID Importance Modifiability Comment box Indicator of inclusion Hospital

admission rate for

8 diabetes long- 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

term

complications Hospital admission rate for 9 diabetes short- 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 term complications

Lower-extremity

10 amputation rate 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Cardiovascular

11 mortality rate 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

This is where you can list any other

indicators for this

disease combination

267

268

2. Quality indicators to assess the quality of care of older adults with controlled diabetes comorbid with hypertension and chronic ischemic heart disease (“1” is the lowest rating and “5” is the highest rating)

Controlled diabetes comorbid with hypertension and chronic ischemic heart disease

Potential for Process improvements Overall value of ID Meaningfulness Comment box Indicator in clinical inclusion practices

HbA1c testing 1 every 6 month 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

LDL- cholesterol

testing once per 2 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 year

Eye examination 3 every two years 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Microalbumin

testing once per 4 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 year

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

269

5 Statin therapy

ACEI or ARB therapy 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 6

Beta-blockers therapy 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 7

This is where you can list any other

indicators for this disease combination

Controlled diabetes comorbid with hypertension and chronic ischemic heart disease

Outcome Overall value ID Importance Modifiability Comment box Indicator of inclusion Hospital

admission rate for

8 diabetes long- 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

term

complications Hospital 9 admission rate for 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 diabetes short-

270

term complications

Lower-extremity 10 amputation rate 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Cardiovascular 11 mortality rate 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

This is where you can list any other indicators for this disease combination

271

3. Quality indicators to assess the quality of care of older adults with controlled diabetes with comorbid osteoarthritis (moderate pain and moderate functional disability) (“1” is the lowest rating and “5” is the highest rating)

Controlled diabetes with comorbid osteoarthritis

Potential for Overall value of ID Process Indicator Meaningfulness improvements in Comment box inclusion clinical practices

HbA1c testing 1 every 6 month 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

LDL- cholesterol 2 testing once per 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 year

Eye examination 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 every two years

Microalbumin

4 testing once per 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

year

Acetaminophen as 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 first-line therapy

272 Non-selective 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 NSAID therapy 6

Non-selective

NSAID in

combination with 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 misoprostol or 7 proton pump

inhibitors

Cox-selective 8 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 NSAID therapy

This is where you can list any other indicators for this disease combination

Controlled diabetes with comorbid osteoarthritis

Outcome Overall value of ID Importance Modifiability Comment box Indicator inclusion Hospital admission rate for diabetes 9 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 long-term complications

Hospital admission

rate for diabetes 10 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 short-term

complications

11 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

273

Lower-extremity amputation rate

Cardiovascular 12 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 mortality rate

This is where you can list any other indicators for this disease combination

274

Controlled diabetes with comorbid osteoarthritis and moderate depression

Potential for Process Overall value of ID Meaningfulness improvements in Comment box Indicator inclusion clinical practices

HbA1c testing 1 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 every 6 month

LDL- cholesterol

2 testing once per 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

year

Eye examination 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 every 1-2 years

Microalbumin 4 testing once per 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 year

Acetaminophen 5 as first-line 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 therapy

Non-selective 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 NSAID therapy 6

Non-selective NSAID in

combination with 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 misoprostol or 7 proton pump inhibitors

275

Cox-selective

8 NSAID therapy 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Interval between SSRIs and monoamine 9 oxidase inhibitors 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 therapy SSRIs and MAOs therapy Use of tri/tetracyclic antidepressant or 12 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 benzodiazepine or MAO inhibitors

At least 3

months 13 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 antidepressant

treatment

At least 6

months 14 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 antidepressant

treatment

This is where you can list any other indicators for this disease combination

276

Controlled diabetes with comorbid osteoarthritis and moderate depression

Overall value of ID Outcome Indicator Importance Modifiability Comment box inclusion Hospital admission rate 12 for diabetes long-term 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 complications Hospital admission rate 13 for diabetes short-term 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 complications

Lower-extremity 4 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 amputation rate

Cardiovascular 15 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 mortality

This is where you can list any other indicators

for this disease combination

277

4. Quality indicators to assess the quality of care of older adults with controlled diabetes with comorbid osteoarthritis (moderate pain and moderate functional disability) and moderate depression (“1” is the lowest rating and “5” is the highest rating)

278

5. Quality indicators to assess the quality of care of older adults with controlled diabetes comorbid with osteoarthritis (with moderate pain and moderate functional disability) and hypertension (“1” is the lowest rating and “5” is the highest rating)

Controlled diabetes with comorbid osteoarthritis and hypertension

Potential for Overall value of ID Process Indicator Meaningfulness improvements in inclusion Comment box clinical practices

HbA1c testing every 6 1 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 month

LDL-cholesterol 2 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 testing once per year

Eye examination every 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1-2 years

Microalbumin testing 4 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 once per year

Statin therapy 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

ACEI or ARB therapy 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

279

Beta-blocker therapy 7

8 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Antiplatelet therapy

9 Acetaminophen as 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 first-line therapy

Non-selective NSAID 10 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 therapy

Non-selective NSAID in combination with 11 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 misoprostol or proton pump inhibitors

Cox-selective NSAID 12 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 therapy

This is where you can list any other

indicators for this disease combination

Controlled diabetes with comorbid osteoarthritis and hypertension

Overall value of ID Outcome Indicator Importance Modifiability Comment box inclusion 13 Hospital admission 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

280

rate for diabetes long- term complications

Hospital admission

14 rate for diabetes short- 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

term complications

Lower-extremity 15 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 amputation rate

Cardiovascular 16 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 mortality rate

This is where you can list any other

indicators for this

disease combination

281

Appendix 7. Criteria for rating quality indicators

Criteria for rating process indicators

Meaningfulness This is a meaningful measure of the quality of care we deliver to the patient aged 65 and older diagnosed with this disease combination Not at all Somewhat Moderately Very Extremely meaningful meaningful meaningful meaningful meaningful (rating=1) (rating=2) (rating=3) (rating=4) (rating=5)

Potential for improvements in clinical practices It is possible to improve the care that impacts this indicator in patients aged 65 and older diagnosed with this disease combination Not at all Somewhat Moderately Very possible Extremely possible possible possible (rating=4) possible (rating=1) (rating=2) (rating=3) (rating=5)

Overall value of inclusion Overall Assessment – Considering your ratings on all dimensions, rate this process measure overall for inclusion in the context of this disease combination Do not include Little reason Could include Should Must include (rating=1) to include (rating=3) include (rating=5) (rating=2) (rating=4)

282

Criteria for rating outcome indicators

Importance This outcome is an important indicator of the quality of care of the patient aged 65 and older diagnosed with this disease combination Not at all Somewhat Moderately Very Extremely important important important important important (rating=1) (rating=2) (rating=3) (rating=4) (rating=5)

Modifiability This outcome is potentially modifiable by improvements in patient’s care Not at all Somewhat Moderately Very Extremely modifiable modifiable modifiable modifiable modifiable (rating=1) (rating=2) (rating=3) (rating=4) (rating=5)

Overall value of inclusion Overall Assessment – Considering your ratings on all dimensions, rate this outcome measure overall for inclusion in the context of this disease combination Do not Little reason to Could include Should Must include include include (rating=3) include (rating=5) (rating=1) (rating=2) (rating=4)

283

Appendix 8. Detailed information related to candidate indicators

For example,

Definition: The percentage of patients aged 65 and over with a diagnosis of diabetes who had at least 2 glycated hemoglobin (HbA1c) (at least 3 months apart) tests performed during the measurement year.

Denominator: The number of patients aged 65 and over diagnosed with diabetes prior to the measurement year

Numerator: Is a subset of the denominator: the number of patients in the denominator who had at least 2 glycated hemoglobin (HbA1c) (at least 3 months apart) tests performed during the measurement year.

Inclusion criteria: Diabetes type I and type II

Exclusion criteria: Patients with gestational or steroid-induced diabetes

Quality domain: on-going monitoring

Source of data: ICD-9/ICD-10 and feecodes

Rationale/supporting evidence:

1. American Diabetes Association. Standards of medical care in diabetes – 2007. Diabetes Care. 2007;30(suppl 1):S4-S41. 2. Sacks DB, Bruns DE, Goldstein DE, et al. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem. 2002;48:436-472. 3. Rohlfing CL,Wiedmeyer HM, Little RR, et al. Defining the relationship between plasma glucose and HbA(1c): analysis of glucose profiles and HbA(1c) in the Diabetes Control and

284

Complications Trial. Diabetes Care. 2002;25:275-278. 4. Consensus statement on the worldwide standardisation of the HbA1c measurement. American Diabetes Association, European Association for the Study of Diabetes, International Federation of Clinical Chemistry and Laboratory Medicine, and the International Diabetes Federation. Diabetologia. 2007; 50:2042-2043. 5.Davidson J. Strategies for improving glycemic control: effective use of glucose monitoring. Am J Med. 2005;118(suppl9A):27S-32S. 6. Halter JB. Geriatric patients. In: Lebovitz HE. Therapy for Diabetes Mellitus and Related Disorders, 3rd Ed. Alexandria, VA: American Diabetes Association, 1997, pp 34–240. 7. Vijan S, Hofer TP, Hayward RA. Estimated benefits of glycemic control in microvascular complications in type 2 diabetes. Ann Intern Med 1997;127:788–795. 8. Selvin E, Marinopoulos S, Berkenblit G et al. Meta-analysis: Glycosylated hemoglobin and cardiovascular disease in diabetes mellitus. Ann Intern Med 2004;141:421–431. 9. NQMC. Diabetes mellitus: percent of patients with 2 HbA1c's in the last year (at least 3 months apart). In. Rockville MD: Agency for Healthcare Research and Quality (AHRQ).

285

Appendix 9. Respondent comments

1. Older adults with controlled diabetes with comorbid hypertension

Indicator 1st Round Comments HbA1c testing every 6- “Maybe lower frequency” months “The frequency of HbA1c test should be guided by how well a

patient controls his/her diabetes. For a patient with poorly controlled diabetes or who is making changes in a therapeutic regimen, testing frequency of every 3 month may be appropriate” “I think this might lead to more reasonable prescription of hypoglycemic” “Too often if the diabetes is well managed, no change in medication or health status” “Important Indicator to monitor care” “The problem is that without an assessment of frailty, it is but a number. What is most important is setting the A1C target first, and using it as a guide for therapy. The more frail, the less tight the glucose ought to be. Similarly, the more frail, the less important A1C may be, especially if random glucose levels are reasonable most of the time. For example, what is the value and meaning of an A1C of 7% in someone with hypertension and moderately severe dementia?” “Value decreases as patients age” “This is in line with the CDA guidelines” LDL-cholesterol testing “Evidence of statin benefit in the elderly is not closely tied to once per year lipid/cholesterol levels, period”

286

“Lower frequency” “Not a major factor in older people” “Very little usefulness in the geriatric population” “Again, why test cholesterol in someone with frailty? Frailty assessment is more important. A non-frail 87 year old with an MI would benefit from a statin. a frail 73 year old with dementia would probably not” “Once they are on a statin, they are on a statin. Little need for testing, other than to measure compliance. Substantially can improve cardiovascular outcomes by assessing and treating over testing” “Guidelines on value of treat to target for LDL and frequency of lipid testing are changing, less valid to me as indicator” Eye examination every “I doubt this can be retrieving form HA Data. Often done 1-2 years outside the usual care by usual professionals

Critical for quality of life. First cataract replacement can reduce falls” “Visual impairment a risk factor for falls and impedes self-care. Early detection of the retinopathy is important as it can be treated” “Hard to get this in the chart b/c it is done as outpatient so I find our data really unreliable. not really the family doctor's "fault" or in our control to order/ensure completed” Microalbumin testing “Potentially good, but might lead to overaggressive treatment once per year and falls Low value in elderly patients if stable” “Same comment regarding need to assess frailty. Chronological age is not a useful determinant of treatment. Biological age (frailty) is. Depends on how well controlled”

287

“I remain unconvinced that statins reduce mortality in diabetics, Statin therapy unless patient has had a prior MI or stroke” “I am not convinced of its value in the elderly” “Similar for testing, not much benefit, lots of harm” “Value could be questioned” “The therapy should be very individual and based on patient test results and clinical manifestations. It would be hard to measure the clinical relevance and appropriateness. There may be some contraindications too” “The results would be very hard to interpret” “The statin therapy would be meaningful and has potential for improvement in clinical practices for patients with diabetes and dyslipidemia” ACE I or ARB therapy “ACE therapy (esp ramipril/ perindopril) is mortality reducing, arb therapy is not, unless patient has systolic HF” “Important and helpful, if dizziness and falls are asked for to tailor therapy. Potential for over treatment however” “ACE inhibitors may have additional benefits over other medications. Sumukadas showed benefits on muscle function CMAJ 2007. Ohrui (Neurology 2004) showed potential benefits on cognition. HYVET had ACE for BP” “These are important aspects but the results would be very hard to interpret as it still patient specific and there are no benchmarks to compare with” “But only for those who are eligible” Antiplatelet therapy “As with statins, mortality reducing only in diabetics with prior MI/stroke” “Assuming that the patient does not have previous cardiac events and only diabetes + HTN as a risk factor. This indicator would be highly” “Difficult at a population level without knowing a patient characteristics (ex. gender, concomitant

288

conditions)” “This would be indicated only if concurrent vascular issues” “Indicator frequency is very patient and situation dependent. Targets for indicator inclusion and assessment should NOT be set as this is not a true measure of the individual physician/patient relationship” “Not everyone needs ASA” “Antiplatelet therapy should be prescribed for diabetic patients with ACS” “Depends on age” “Again guidelines for primary prevention are changing” Hospitalization for “Pneumonia, MI, falls, delirium in the elderly are increasingly diabetes long-term going to be managed in the home, which as a trend over time complications affects the value of change in admission rates- all cause admission rate more useful than either long term or short term” “I was unsure why the stems changed here to importance and modifiability; I answered based on the stems given in this question. the groundwork might have been laid at a much earlier age” “Complications often multifactorial. Not a common thing, but an important one” “Hard to change since here are many factors contributing to this” “I think avoiding hospitalization is a key outcome. However I'm a bit confused about this one - I"m not sure how much I can control hospitalization for diabetes complications if this is in "controlled diabetics". if controlled (ie: a1c at target) is it that i'm not managing the other issues that lead to hospitalization (?renal ?vascular). If you take out controlled then the indicator makes more sense to me” Hospitalization for “Much more of an IDDM problem

289 diabetes short-term particularly for falls and hypoglycemia” complications “Usually only for type one diabetics, and rare in elderly”

“Not the most common reason for ED visits” “Here I'm assuming you mean hypoglycemia/diabetic ulcers - maybe more room for Family doctor to modify these” Lower-extremity “Once again, issues might have started long before, but this is amputation rate one of the most important things to prevent Again, multifactorial” “DM1 or DM2? “ Cardiovascular mortality “Frailty is what dominates prognosis in some. Recent trials do rate not demonstrate major impact of control on mortality”

“This outcome measure may not be specific or attribute to diabetes” Oral hypoglycemic use

Baseline ECG

MRI head/heart Suggestion from the 1st Delphi round – “Given that most strokes/MIs in diabetics are silent, should we use MRI head (rule out stroke) , MRI heart (rule out ischemic disease) to determine which diabetics 'need' statins and ASA, etc.” All-cause mortality

Ocular complications due to diabetes Urinary/skin and soft tissue infections

290

Indicator 2nd Round Comments

LDL-cholesterol testing “Difficult as importance wanes with advancing age. Much more once per important for a 65 than an 85 year old” year “Once per year is vague. I would extent timeframe. Also, if they are on statin and adherence is fairly certain checking this often may be wasteful” “Would suggest 16-18 months vs yearly” “Lower frequency would be more appropriate” “Should be on statin, not have LDL measured” “Frailty must first be assessed to provide context for meaningfulness” “Not clear if the indicator is: at least one measure of LDL cholesterol. Multiple testing may be seen as a negative indicator if there is no specific reason to justify it” “Many are statin intolerant” Statin therapy “Though it is recommended in the guidelines the evidence for statin use in older people is not completely clear. The physicians may decide on case-by case basis. If measured the results will be hard to interpret” “Frailty...”

Antiplatelet therapy “No longer really indicated” “Despite guideline recs, I am unconvinced that evidence is so strong for ASA in DM for primary prevention” “Not everyone needs to be on ASA” “Frailty...”

Hospitalization for “Data would have to be aggregated up so much as to make diabetes long-term meaningless” complications “Usually these will be in the company of comorbidity and frailty. No need to split hairs”

Hospitalization for “Need clear definitions here - would be interesting to think diabetes short-term about ones that are "ambulatory care sensitive" complications

Cardiovascular mortality “It will be hard to attribute the death to diabetes” rate “Depends on frailty status...”

291

Oral hypoglycemic use “Tricky with those who are diet controlled or metformin intolerant or on other antiglycemics /insulin” “The treatment options are based on different factors that physicians may consider. The results may be very hard to interpret and it may not show the supporting clinical details. In terms of data can be captured only the prescription but not the actual use” "Baseline" usually means at time of diabetes diagnosis (or at least that is the recommendation in DM guidelines), so will be Baseline ECG difficulty to implement — low score. “According to Diabetes guidelines the baseline resting ECG should be performed in individuals age>40, duration of diabetes, organ failure and cardiac risk factors. It would be meaningful only in this target population. Is this feasible to measure; is it these data available in admin database?” MRI head/heart “I have to admit as a primary care physician I don't know anything about the evidence for utility of this test” “Routine in all older adults with diabetes?” “Should not be included” “It won't be normal regardless of cognitive status” All-cause mortality “Therapy dependants on risk, not just do status and hypertension” “Very little of this attributable to physician care” “It would be hard to attribute mortality rate to diabetes and/or its complications” “Interesting but too many variables involved to understand role with DM + HTN” “Depends on frailty status” Ocular complications “If we can tease out the specific DM + HTN complications due to diabetes (vascular, renal, ocular, cvd) that would be interesting, not sure how to do this. If we can't then just including ocular seems not that useful to me” “Important functionally” Urinary/skin and soft “I think the serious ones would be captured under "short-term" tissue infections complications?” “UTIs are ubiquitous and multifactorial. Skin infects – perhaps”

292

2. Older adults with controlled diabetes comorbid with hypertension and chronic ischemic heart disease

Indicator 1st Round Comments HbA1c testing every 6- “Prevent over treatment” months “Familiarity with measure and ease of measurement make it

attractive. The problem is that it is not meaningless in older persons without a measure of frailty” “As per last question- I think for stable DM this may be too frequent monitoring of A1c” LDL- cholesterol testing “Will treat elderly with this disease combination with low dose once per year statin regardless of cholesterol” “Unclear value in the elderly” “Familiarity with measure and ease of measurement make it attractive. The problem is that it is not meaningless in older persons without a measure of frailty” Eye examination every “Most feared complication for patients” 1-2 years “Again validity of this measure is limited by difficulties in

getting this data into the EMR/patient chart as done by outside clinics” “Familiarity with measure and ease of measurement make it Microalbumin testing attractive. The problem is that it is meaningless in older persons once per year without a measure of frailty” “With IHD, they may be more likely to have kidney damage” Statin therapy “Depends on how long from acute event. Likely remote event less relevant for continuing statin therapy if even minor symptoms occur” “The problem is that it is meaningless in older persons without a measure of frailty” ACEI/ARBs “ACE, not ARB”

293

“Beta-blocker benefit for IHD is now tiny, and only for 30 days Beta-blockers therapy post stent/plasty/MI; beta-blockers no longer indicated for HTN beyond age 75 (CHEP). I think it is important to get a lot of elderly diabetics OFF beta blockers, unless they have systolic heart failure” “Beta-blockers in stable cardiac disease may or may not be standard and depending on patient level characteristics this may be a difficult combination in a patient with diabetes. This would be difficult to measure at a population level beta blockers can be a two edged word. Likely depends on how recently acute vascular events occurred” “Recent data raises questions on need for long-term use” “Depends on whether they had MI or not. ASHD without MI = not as much evidence” Hospital admission rate “Likely will die of MI first, and once again issues might have for diabetes long-term started years ago” complications “Frailty and geriatric syndromes are more likely in this case to be reasons for hospitalization” Hospital admission rate “Not a common occurrence” for diabetes short-term “This would be very useful to monitor at practice level. Proper complications outpatient treatment may reduce the incidence and admissions for short term complications” Lower-extremity “Multifactorial and not so common” amputation rate

Bariatric surgery Suggestion from the 1st round: “There is a growing pile of good evidence that bariatric surgery can cure type 2 diabetes, and that it hugely reduces all-cause mortality, even when BMIs are not 'that high'. I think we will all be recommending bariatric surgery to 65 year old diabetics

294

very soon, especially if they also have hypertension, ischemic heart disease, or a combination”

Indicator 2nd Round Comments LDL- cholesterol testing “Q 18 months better than q yearly once per year statins, not LDL” “Depends on frailty status”

“Q 18 months better than q yearly Microalbumin testing frailty...” once per year

“Lots of alternatives to beta blockers” Beta-blockers therapy “Many not tolerant of bb, with "stable" IHD they may be off BB and on something else b/c of side effects of BB” “Less important over time unless heart failure or cardiomyopathy is present”

Hospital admission rate “They have them already, in all likelihood” for diabetes long-term complications

295

Hospital admission rate “Depends on how long the physician was following the patient” for diabetes short-term complications Lower-extremity “Depends on how long the physician was following the patient” amputation rate

Antiplatelet “Especially if secondary prevention. I would stratify” therapy “Frailty…”

Hospital admissions for “Any CV admission might be preferable” heart failure “Not sure about this - do we need to put in admission for Coronary Artery Disease as well? Again would be interesting to see ICES ambulatory care sensitive indicators and if CAD or CHF are considered in this category” “Prevalent outcome” “Undertreatment of CAD may lead to HF” All-cause mortality rate “Frailty…” Bariatric surgery “I would do this if obesity presents - presence of IHD not really the key factor here” “Are seniors ever offered this?”

296

3. Older adults with controlled diabetes with comorbid osteoarthritis

Indicator 1st Round Comments HbA1c testing every 6 “The hypoglycemia is also important to measure in this age month group” LDL- cholesterol testing “Again, why test cholesterol in someone with frailty? Frailty once per year assessment is more important. A non-frail 87 year old with an MI would benefit from a statin. A frail 73 year old with dementia would probably not” Eye examination “Falls prevention”

Microalbumin testing once “Frailty…” per year Acetaminophen as first-line “Can you capture this?” therapy “Safest approach”

“Tylenol not very effective in OA”

|Pain is grossly undertreated and is this risk factor for sedentary life, and thus worse outcomes” “Treatment for osteoarthritis should be dependent from diabetes. It may be difficult to measure the use of acetaminophen using administrative data since it is an OTC drug” Hospital admission rate for “They are more likely to be admitted with falls and fractures” diabetes long-term “Maybe to limit to older patients with diabetes and look at the complications results by co-morbidity?”

Use of non-selective “Important but renal issues very important as well, obviously NSAIDs in combination best avoided” with misoprostol or proton “Misoprostol not used anymore” pump inhibitors “Is this common anymore?” “This is not specific for Diabetes patients” Use of non-selective “Depends on severity of arthritis rather than concomitant NSAIDs diabetes”

297

“Not in elderly (for more than one week)” “In terms of "what to try? Maybe better to avoid?" “Treatment for osteoarthritis should be dependent from diabetes. It may be difficult to measure the use of acetaminophen using administrative data since it is an OTC drug” “I'm assuming this would be to avoid” “NSAID therapy? Negative indicator?” “Not clear…to avoid?” Use of cox-selective “Probably to be avoided in case of unsuspected vascular disease” NSAIDs “Treatment for osteoarthritis should be dependent from diabetes.

The use of these medications may be driven by other factors besides diabetes” “Again is this to use instead of the non-selective or to avoid?” “Not clear. i'm not that happy with drug related process outcomes as family physician - i think we need to focus on all of the nonpharmacologic ways to manage arthritis” “Not clear…to avoid?”

Indicator 2nd Round Comments HbA1c testing every 6 month “Useful for determining excessive testing” “I think A1C should be in all the categories, not just here. I think A1C yearly is ok with "controlled" diabetics” “depends on frailty” LDL- cholesterol testing “As before, so dependent on how old” once per year “Frailty…” Microalbumin testing once “Frailty…” per year

298

Acetaminophen as first-line “Effectiveness of acetaminophen has been questioned” therapy “Difficult to measure at scale”

“Not sure how you will measure this - maybe better to have

"no NSAIDS" as the indicator” “The treatment options are based on multiple factors”

Cardiovascular mortality rate “Would have considered this for HTN & stable CHD more than OA, since the former impact this as well as DM”

“Again I'm not sure why this is here (? because of NSAIDS) I might be inclined to put this specific outcome within the DM + IHD or DM + HTN group but not here” Use of non-selective NSAIDs “So dependent on circumstances in combination with Could be hard to measure. Non selective NSAID dugs are misoprostol or proton pump available out of counter” inhibitors Use of non-selective NSAIDs “Negative” indicator

“So dependent on prior med trials that would be difficult to interpret” “Only relevant if longer Rx maybe” “Could be hard to measure. Non selective NSAID dugs are available out of counter” Use of cox-selective NSAIDs “Negative” indicator” “I'd be worried about CV risk with these and would use naprosyn & ppi ahead of colecoxib|” “Need to justify use regardless” Use of topical NSAIDs “This would probably be not feasible as well. How would this treatment be tracked?” “Depends on frailty” Statin therapy “Not sure what the intent is here - they should be on Use of opioids optiods? They shouldn't be?”

299

“The prescription of these medications is based on indications and is not routinely recommended” ACE inhibitor therapy “Would have suggested this was a better indicator for DM+HTN or DM+stable CHD , less important in OA concomitantly as no role in this condition” “Not relevant for DM + OA” Referral for home care “Need to specify. Do you mean OT?” “Interesting idea to have this in - I like the idea of trying to capture the community based services in the care of these patients - should it be just the OA + DM or any of the comorbidities? Should it just be referral to home care or to any allied health as many of these (social worker, foot care nurses, dieticians, DM teams) play an important role in care of these comorbid patients” “Depends on frailty!!!! and function!!!!” Joint replacement therapy “Should stratify by age sex SES as disparate referral patterns need addressing” “Interesting” “Unless really ill or really frail, should this not be offered regardless? This is an odd one” ED/hospitalizations for fall “Consider fractures”

300

4. Older adults with controlled diabetes with comorbid osteoarthritis and moderate depression

Indicator 1st Round Comments HbA1c testing every 6 “The regular testing is important and recommended for all month patients with diabetes including those with depression.

Recommend measuring in the larger diabetes population and reporting by different stratification, including comorbidity to see if there are any patterns in certain groups and target the interventions towards improving the practices for specific sub groups” LDL- cholesterol testing “Frailty” once per year “The therapies should be individualized and would be hard Acetaminophen as first-line to interpret the results” therapy At least 3 months “If respond, usual recommendation is treat for a year, antidepressant treatment Important for a good trial before giving up” (acute phase) “Dose may be relevant as well”

What is the guideline for the duration of the treatment? “The need to use these drugs depends on the severity of depression” “Not sure - evidence on use of anti-depressants for mild- moderate suggests may not need rx. many have side effects/non-compliance even when started” At least 6 months “Can comment these are very clinical. Overall the antidepressant treatment treatments indicators are hard to interpret as there are many (continuation phase) other factors that may influence the treatment choice and

duration. The need to use these drugs depends on the severity of depression” Lower-extremity amputation rate Cardiovascular mortality “Especially in depressed patients who also have ischemic

301 rate heart disease”

Use of non-selective “This is not specific treatment for diabetes patients. It is NSAIDs in combination relevant for all patients” with misoprostol or proton pump inhibitors Use of non-selective “Negative indicator!” NSAIDs “Avoid?” Use of cox-selective “Negative indicator!” NSAIDs “Maybe avoid?” Use of tri/tetracyclic “Really? I give SSRIs or SNRIs to just about all of my antidepressants, elderly patients with this triad; I never give tricyclics benzodiazepines, Z-drugs, (mirtazapine is my safer alternative), or benzodiazepines” or MAOIs “I am assuming this is a negative indicator (i.e. looking for decreases in this process). I have difficulties when you combine medication classes with "and", because TCA may be used for diabetic neuropathy (which isn't discussed here but could be used) but unlikely for depression (which is much higher doses)” “Benzos a bad choice, MAO inhibitors likely require a specialist These drugs are used less have high rate of side effect. SSRI SNRI more useful” “As in "avoid" if possible” “The need to use these drugs depends on the severity of depression not clear on this - is this drugs to avoid?” “So dependent on circumstances, particularly risks of tricyclics over emphasized” “Split up. TCA less patient issues and wider range of appropriate uses other than depression. e.g. chronic pain” “Have difficulty with this as TCA may be used

302

appropriately in low doses for diabetic neuropathy and may therefore not a negative indicator” “I'm assuming this would be to avoid” “Negative indicator?” “Not clear…to avoid?”

Indicator 2nd Round Comments “Difficult to measure” Acetaminophen as first-line therapy At least 3 months “Depends on diagnostic accuracy of depression” antidepressant treatment “Evidence poor in elderly for benefit and side effect more (acute phase) likely”

“The length of antidepressant treatment may depend on whether it is the first prescription” “Little evidence for anti-depressants for mild-moderate (which is majority)” “Can be measure the filled prescriptions but not the actual use” “At proper dose” At least 6 months “The length of antidepressant treatment may depend on antidepressant treatment whether it is the first prescription. (continuation phase) again at proper dose - 6 months or three months at low dose

is poor care regardless”

Lower-extremity “Could be a good thing depending on frailty” amputation rate

303

Cardiovascular mortality “Frailty” rate Use of tri/tetracyclic “As well, MAOI can be used in patients with severe antidepressants, depression. The use of these benzodiazepines & z-drugs can benzodiazepines, Z-drugs, increase risk of fall and should not be used Long term” or MAOIs “Interesting - I would align with something like the Beer's criteria” “Unless justified”

“Not sure there is solid evidence for any specific class over Use of SSRI or SNRI another. Recent reports of fewer falls with tricyclics!” “Would refer back to depression guidelines. As long as they are on recommended first line treatment” “Depends - really, what is important is evidence that the Use of topical NSAIDs prescriber used the WHO stepwise approach” “Depends on caregiver, living arrangement. Referral to Referral for home care community support services could be adequate”

Hospital admissions for “Uncommon” depression

All-cause ED visits “I would look at those ED visits with low acuity CTAS scores not all - I wouldn't put this just for this particular combination of comorbidities” Joint replacement rate “Odd... didn't think this was a problem”

304

5. Older adults with controlled diabetes comorbid with osteoarthritis and hypertension

Indicator 1st Round Comments HbA1c testing every 6 month “Frailty”

LDL-cholesterol testing once “Frailty” per year Microalbumin testing once “Frailty” per year “Trial of stopping statin to see if pain/disability gets Statin therapy better?”

“Frailty” ACEI/ARB therapy “ACE, not ARB”

“Continue only if prior mi/stroke Antiplatelet therapy Frailty” Acetaminophen as first-line therapy Hospital admission rate for “As mentioned before these outcomes are important to diabetes long-term measure but you can consider including as stratification to complications larger population cohort in the study”

Cardiovascular mortality “You can also consider he hospitalization for CAD as well” Use of non-selective NSAIDs “Negative indicator!” “Avoid?” Use of cox-selective NSAIDs “Negative indicator!” “Avoid?”

305

Indicator 2nd Round Comments HbA1c testing every 6 month “Frailty”

LDL-cholesterol testing once “Frailty” per year Microalbumin testing once per “Frailty” year Cardiovascular mortality Use of non-selective NSAIDs “Even more important here with hypertension to avoid NSAIDs” Use of cox-selective NSAIDs

“Should be consistent and include this in the HTN + End-stage renal disease rate DM group if we include it here”

“Depends on frailty” “ACE inhibitors, metformin, bariatric surgery reduce mortality in NIDDM. aspirin and statins only reduce mortality in NIDDM patients who have had stroke, MI, or peripheral vascular disease. ESRD and amputation rates are important and expensive NIDDM secondary outcomes that can be measured just as accurately as mortality. big pharma has been very effective at 'proving' every new NIDDM treatment reduces some easier to fudge outcome measure- this is thinly veiled marketing”

306

Appendix 10. Email text – Delphi Round II

Dear (name of potential participant),

Thank you for participating in the first round of the Delphi study that aims to develop indicators for assessing the quality of overall care of older diabetes patients with comorbid conditions. After round 1, consensus was reached for 13 indicators that are not represented in

Round 2. Consensus was defined as having at least 73% of panelists rated a given indicator as 4 or 5.The remaining indicators are included in the Round II survey together with new or revised indicators that have been suggested by the panel. In this second round, we included ratings for all indicators suggested by the panel, except for those that cannot be measured using health administrative data (e.g., smoking status, BP or weight measurement). The results of this project will be used to measure quality of care of patients with selected disease combinations using Ontario health administrative data.

Attached, please find the document with the frequency of ratings from all 15 respondents to the first round of the survey, together with a reminder of your own responses and median of criterion “Overall value of inclusion” for each candidate indicator. After you have read the results from the first round, please follow the link to complete the second round of the web- based questionnaire. Please consider whether, in the light of your colleagues’ responses, you would like to alter your answers. We are pleased to enclose herewith criteria and instructions for rating indicators you will need for the second round ratings. The questionnaire will take approximately 20 minutes to complete.

307

You have 3 options for submitting the Round 2 questionnaire:

1. Complete on the Web at http://fluidsurveys.com/surveys/hsprnpiloto/quality-

indicators-round-2/?code=pmqtr67krz or

2. Complete the word document provided and send via email to

[email protected] or

3. Contact [email protected] for a paper copy of the questionnaire and

a stamped envelope.

Please submit your completed questionnaire by Friday, January 8th.

Remember that your responses are strictly confidential to the research team and that your participation in this project is entirely voluntary. However, we would like to stress the importance of having everyone who responded to the first round to complete the remaining couple of rounds. After completion of all Delphi rounds you will be given a cheque for $200 to compensate you for any disruption to your practice.

If you have any questions or concerns please contact me at [email protected].

Thank you for your consideration, Walter P Wodchis, PhD Yelena Petrosyan, PhD(c) Associate Professor Institute of Health Policy, Management and Evaluation Institute of Health Policy, Management and University of Toronto Evaluation Email address: [email protected] University of Toronto 308 Email address: [email protected]

Appendix 11. Feedback material for Round II

Please see below the results of rating indicators for assessing care of older adults with different comorbidities. The table illustrates the frequency of panel responses in Round 1

(upper line) for each indicator with respect to each disease combination, together with a circle indicating your own response in bottom line, as well as median of the criterion

“Overall value of inclusion” for each candidate indicator.

For example,

Frequencies Potential for Median of & own improvements Overall value of overall value Indicator Meaningfulness response in clinical inclusion of inclusion practices (min, max)

Freq. of HbA1c 1 4 5 5 2 8 4 2 10 3 responses testing 4 (3, 5) every 6

month 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Own

response

In the preceding example, two panelists rated the overall value of inclusion of the indicator a

3, ten panellists rated it a 4, and three panellists rated it a 5. This particular panellist rated it a

5. The median of overall inclusion of this indicator with min and max values equals to 4 (3,

5).

309

Appendix 12. Quality indicators that did not reach consensus, or were not included in this study

Diabetes with Diabetes Diabetes Diabetes Diabetes comorbid with with with with hypertensio comorbid comorbid Indicator comorbid comorbid n and osteoarthriti osteoarthriti hypertensio osteoarthriti ischemic s and major s and n s heart depression hypertension disease LDL-cholesterol No No No No No testing once per consensus consensus consensus consensus consensus year Statin therapy No consensus

Antiplatelet No therapy consensus

Baseline No electrocardiograph consensus y MRI of head/heart Consensus to reject

All-cause mortality No No No No consensus consensus consensus consensus

Urinary/skin/soft No tissue infections consensus Beta-blocker No therapy consensus

Bariatric surgery Consensus rate to reject

Use of No No No acetaminophen as consensus consensus consensus

310 first-line therapy Non-selective No No No NSAIDs in consensus consensus consensus combination with misoprostol or proton pump inhibitors Use of topical No NSAIDs consensus

No No Statin therapy consensus consensus

No Use of opioids consensus

No Use of ACEs consensus inhibitors

Referral for home No care consensus

Joint replacement No No therapy consensus consensus

ED visits/hospital No No admissions for fall consensus consensus

At least 3 months No antidepressant consensus treatment At least 6 months No antidepressant consensus treatment Use of SSRI or Consensus SNRI to reject

Use of tricyclic No antidepressants consensus

311

Use of topical No No NSAIDs consensus consensus

No Use of opioids consensus

No No Referral for home consensus consensus care

Lower-extremity No No amputation rate consensus consensus

No Hospital admission consensus for depression

No All-cause ED visits consensus

Beta-blocker Consensus therapy to reject

Antiplatelet No therapy consensus

End-stage renal No disease consensus

Indicators that are not amenable to measurement using Ontario health administrative data

312

Microalbumin testing X X X X X

Serum creatinine X testing Waist circumference X

BP measured in past X 6 or 12 months Body mass index X X

Self-management X X X X teaching Medical adherence X X X X assessed Exercise assessed X

Percentage of patients with most recent HbA1c level X >9.0% ( poor control) Percentage of patients with most recent blood X pressure <140/90 mmHg Quality of life X X

313

Appendix 13. Study Time Frame Definitions

Project Time Frame Definitions

Max Follow-up Date Accrual Window

Look-back Window Observation Window

Index Event Date(in which to look for outcomes)

Accrual Start/End Dates April 1st, 2010 - March 31st, 2014 Max Follow-up Date March 31st, 2014 Look-back window April 1, 2001 – April 1st, 2010

314

Appendix 14. Comorbid chronic conditions

Condition ICD 9 / OHIP ICD 10

Rheumatoid arthritis 714 M05-M06 Osteoporosis 733 M81 M82 Other mood disorders 300, 309 F38—F42, F431, F432, F438, F44, F450, F451, F452, F48, F530, F680, F930, F99

Psychiatric conditions 291 292 295 297 298 299 F04 F050 F058 F059 F060 F061 F062 F063 other than mood 301 302 303 304 305 306 F064 F07 F08 F10 F11 F12 F13 F14 F15 disorders and dementia 307 313 314 315 319 F16 F17 F18 F19 F20 F21 F22 F23 F24 F25 F26 F27 F28 F29 F340 F35 F36 F37 F430 F439 F453 F454 F458 F46 F47 F49 F50 F51 F52 F531 F538 F539 F54 F55 F56 F57 F58 F59 F60 F61 F62 F63 F64 F65 F66 F67 F681 F688 F69 F70 F71 F72 F73 F74 F75 F76 F77 F78 F79 F80 F81 F82 F83 F84 F85 F86 F87 F88 F89 F90 F91 F92 F931 F932 F933 F938 F939 F94 F95 F96 F97 F98 Dementia 290, 331 (OHIP) / (DAD: F00, F01, F02, F03, G30 046.1, 290, 294, 331.0, 331.1, 331.5, 331.82) ODB subclnam =: ‘CHOLINESTERASE INHIBITOR’ Renal failure 403,404,584,585,586,v451 N17, N18, N19, T82.4, Z49.2, Z99.2 Asthma 493 J45 Cancer 140-239 (broad algorithm C00-C26, C30-C44, C45-C97 from ICD table) Cardiac Arrythmia 427.3 (DAD) / 427 I48.0, I48.1 (OHIP)

315

CHF 428 I500, I501, I509 COPD 491, 492, 496 J41-J44 Stroke 430, 431, 432, 434, 436 I60-I64

316

Appendix 15. Sensitivity analyses

Appendix 14a. Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension only

Model 1 Model 2 Model 3 All-cause Hospitalizations for Hospitalizations for Characteristic hospitalisations diabetes short-term diabetes long-term

*AOR (95% CI) complications complications *AOR (95% CI) *AOR (95% CI) HbA1c testing annually No tests Ref. Ref. Ref. 1 or 2 HbA1c tests 0.84 (0.80-0.86) 0.92 (0.56-1.26) 1.04 (0.94-1.15) 3 or more HbA1c tests 0.77 (0.72-0.83) 0.83 (0.48-1.28) 0.99 (0.89-1.10) Annual eye examination No Ref. Ref. Ref. Yes 0.62 (0.59-0.68) 0.64 (0.44-0.93) 0.61 (0.57-0.65) Use of oral hypoglycemic drugs No Ref. Ref. Ref. Yes 0.67 (0.63-0.70) 0.75 (0.48-1.15) 1.10 (1.01-1.19) Use of ACE-inhibitors No Ref. Ref. Ref. Yes 1.14 (1.08-1.19) 1.30 (1.10-1.51) 1.40 (1.30-1.51) Use of ARBs No Ref. Ref. Ref. Yes 0.90 (0.84-0.95) 0.95 (0.78-1.16) 1.10 (1.08-1.20) COC index COC≤0.70 Ref. Ref. Ref. COC>0.70 0.42 (0.40-0.44) 0.41 (0.39-0.43) 0.44 (0.41-0.46) Number of prescribed 1.21 (1.20-1.22) 1.11 (1.09-1.12) 1.21 (1.20-1.22) drugs *Adjusted for age, sex, income quintile, RIO index, duration of diabetes, duration of hypertension, number of primary care visits, primary care models

317

Appendix 13b. Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid hypertension and chronic ischemic heart disease only

Characteristic Model 1 Model 3 All-cause hospitalisations Hospitalizations for *AOR (95% CI) diabetes long-term complications *AOR (95% CI) HbA1c testing annually No tests Ref. Ref. 1 or 2 HbA1c tests 0.85 (0.77-0.94) 1.03 (0.90-1.18) 3 or more HbA1c tests 0.76 (0.68- 0.85) 1.04 (0.90-1.21) Annual eye examination No Ref. Ref. Yes 0.70 (0.64-0.75) 0.74 (0.72-0.79) Use of oral hypoglycemic drugs No Ref. Ref. Yes 0.81 (0.74-0.88) 1.60 (1.42-1.72) Use of ACE-inhibitors No Ref. Ref. Yes 1.15 (1.05-1.25) 1.25(1.12-1.40) Use of ARBs No Ref. Ref. Yes 1.00 (0.90-0.11) 0.99 (0.87-1.23) Use of antiplatelet drugs No Ref. Ref. Yes 1.36(1.32-1.40) 1.80 (1.68-1.92) Use of statins No Ref. Ref. Yes 0.92 (0.88-1.04) 1.14 (0.97-1.34) COC>0.60 Ref. Ref. COC≤0.60 0.52 (0.50-0.54) 0.51 (0.49-0.53) Number of prescribed 1.18 (1.16-1.20) 1.18 (1.16-1.21) drugs *Adjusted for age, sex, income quintile, RIO index, duration of diabetes, duration of hypertension, duration of ischemic heart disease, number of primary care visits, primary care models

318

Appendix 14c. Associations between process measures and the likelihood of hospitalizations among older adults with diabetes with comorbid osteoarthritis only

Model 2 Model 1 Hospitalizations for All-cause hospitalisations Measure diabetes long-term *AOR (95% CI) hospitalisations

*AOR (95% CI) HbA1c testing annually No tests Ref. Ref. 1 or 2 HbA1c tests 0.81 (0.71-0.81) 1.02 (1.01-1.03) 3 or more HbA1c tests 0.78 (0.69-0.82) 1.03 (1.01-1.06) Annual eye examination No Ref. Ref. Yes 0.70 (0.68-0.75) 0.72 (0.70-0.74) Use of oral hypoglycemic drugs No Ref. Ref. Yes 0.59 (0.52-0.66) 1.23 (1.20-1.28) Use of NSAIDs No Ref. Ref. Yes 0.71 (0.65-0.78) 0.59 (0.51-0.64) COC index COC≤0.63 Ref. Ref. COC>0.63 0.43 (0.41-0.44) 0.42 (0.40-0.44) Number of prescribed drugs 1.20 (1.18-1.22) 1.22 (1.19-1.24) *Adjusted for age, sex, income quintile, RIO index, duration of diabetes, duration of osteoarthritis, number of primary care visits, primary care models

319

Appendix 16. DIN/PIN codes of drugs

ACE inhibitors (Angiotensin 02019892 02291894 02300052 02233006 02300095 Converting Enzyme 02352257 02300001 02299968 02300133 00670901 Inhibitors) 02020025 02291878 02300036 02300680 02300079 02352230 02299984 02299933 02300117 00851795 02019906 02291908 02300060 02233007 02300109 02352265 02300028 02299976 02300141 00670928 02019884 02291886 02300044 02233005 02300087 02352249 02299992 02299941 02300125 00708879 02221829 02251515 02387387 02295482 02331101 02420457 02301148 02295369 02308363 02299372 02310503 02287692 02291398 02221853 02251582 02387417 02295512 02331144 02420481 02421321 02301172 02247919 02287943 02255332 02310546 02291436 02247947 02221837 02251531 02387395 02331128 02420465 02421305 02301156 02247917 02287927 02255316 02310511 02287706 02291401 02247945 02221845 02251574 02387409 02295504 02331136 02420473 02421313 02301164 02247918 02287935 02255324 02310538 02291428 02247946 02281112 02325381 02420503 02421348 02343932 02425548 02290340 00885843 02273918 00885851 02290332 00885835 00893625 00546305 02163594 01942999 00893595 00695661 02163551 01942964 00893609 00546283 02163578 01942972 00893617 00546291 02163586 01942980 02266008 02331004 01907107 02262401 02294524 02247802 02266016 02331012 01907115 02262428 02294532 02247803 02123274 02123282 02246624 01947672 02248500

320

02290995 01947680 02248501 02291002 01947699 02248502 02291010 01947664 02248499 02290987 02231459 02231460 02239267 09853960 02217503 02394480 02271451 02361558 02422514 02274841 02285126 02292211 00839396 02294249 09857286 02289202 02285088 02049376 09854010 02217511 02394499 02271478 02361566 02422522 02274868 02285134 02292238 00839418 02294257 09857287 02289229 02285096 02049384 02217481 09853685 02394472 02271443 02361531 02422506 02274833 02285118 02292203 00839388 02294230 09857272 02289199 02285061 02049333 ARBs (angiotensin receptor 02353512 02403358 02445980 02354845 02182882 blockers) 02398850 02422484 02405768 02368293 02309777 02404486 02313359 02424983 02357976 02426617 02379058 02403323 02445964 02354829 02182815 02398834 02422468 02405733 02368277 02309750 02404451 02313332 02424967 02380838 02426595 02353504 02403331 02445972 02354837 02182874 02398842 02422476 02405741 02368285 02309769 02404478 02313340 02424975 02357968 02426609 02426617 02399105 02365367 02365359 02365340 02239090 02239091 02239092 02311658 02379260 02379279 02379287 02379295 02379295 02376547 02376539 02376520 02386526 02386534 02386518 02386496 02379147 02379120 02379155 02379139 02391228 02391171 02391198 02391201 02380706 02380714 02380692 02380684 02326973 02326957 02326965 02392267 02417340 02366320 02366339 02366312 02371510 02371529 02371537 02371545 02414201 02414228 02414244 02414236 02337487

321

02337509 02337495 02337517 02270528 02244781 02244782 02289504 02383551 02383543 02383535 02383527 02312999 02313006 02313014 02344564 02363100 02363062 02363119 02356759 02356740 02356767 02356775 02356651 02356678 02356686 02356643 02386968 02386976 02386984 02406098 02406101 02406128 02237925 02237923 02237924 02328070 02328089 02328100 02418215 02418193 02418207 02423006 02422980 02422999 02347318 02347296 02347326 02317079 02317060 02317087 02406810 02406829 02406837 02316412 02316404 02316390 02328496 02328461 02328488 02315971 02316005 02315998 02315998 02427095 02427109 Oral hypoglycemic drugs 02403277 02403269 02403250 02416794 02333864 02333856 02333872 02305062 02268493 02257734 02257726 02167786 02438283 02438275 02421828 02421836 02162849 02099233 02380196 02380218 02380722 02380730 02378620 02378639 02378868 02378841 02388766 02388774 02148765 02229656 02045710 02242589 02223562 02269058 02269031 02242931 02242974 02246821 02246820 02379767 02389169 02389185 02389177 02355671 02355663 02355698 02424258 02424266 02424274 02321483 02321475 02321491 02239925 02239926 02239924 02354926 02354934 02354942 02357453 02357461 02357488 02303442 02303450 02303469 02242573 02242572 02242574 02302942 02302950 02302977 02384906 02384914 02384922 02302861 02302888 02302896 02397307 02365529 02365537 02326477 02326485 02326493 02298279 02298287 02298295 02274914 02274922 02274930 02303124 02303132

322

02303140 02339587 02391600 02339595 02375877 02375869 02375850 02297906 02297914 02297922 02245274 02245273 02245272 02295377 02295393 02295385 02273772 02273756 02273764 02273128 02273136 02273101 02269589 02269597 02269619 01913654 01913662 02224550 02224569 00808741 02236734 02248008 02248009 01913689 01913670 02190893 02190885 Antiplatelet drugs 02412942 02252767 02398591 02416387 02303027 02415550 02422255 02408910 02351536 02348004 02238682 02330555 02379813 02359316 02293161 02388065 00010340 00010332 00216666 00229296 02405083 02240351 02367858 02240352 Statins 02295318 02295296 02295261 02295288 02407280 02407264 02407256 02407272 02310899 02310929 02310902 02310910 02288362 02288346 02288354 02288370 02391074 02391082 02391058 02391066 02230714 02230713 02230711 02243097 02373246 02392976 02373203 02392933 02373211 02392941 02373238 02392968 02399377 02399393 02399385 02399407 02313723 02313715 02313758 02313707 02350327 02350335 02350297 02350319 02417960 02417936 02417944 02417952 02324970 02324946 02324954 02324962 02220172 02220180 02248572 02248573 00795860 00795852 02243127 02243129 02246542 02246543 02246013 02246014 02247056 02243506 02243506 02243507 02248184 02248182 02248183 02330954 02330962 02330970 02317478 02317486 02317451 02257106 02257092 02257114 02247655 02247656 02247657 00893749 02222051 00893757 02284448 02284456 02284421 02246930

323

02246931 02246932 02247858 02247010 02247009 02247008 02338009 02337975 02337983 02337991 02339781 02339773 02339765 02339803 02247163 02247162 02265540 02247164 02391287 02391279 02391260 02391252 02413108 02413078 02413051 02399180 02399172 02399164 02399199 02397781 02397803 02397811 02397838 02381265 02381281 02397803 02397838 02397811 02381265 02381281 02381273 02381303 02378523 02378531 02378566 02378558 02382644 02382652 02382660 02382679 02338726 02338734 02338750 02338742 02354624 02354616 02354608 02354632 02247015 02247012 02247013 02247014 02247011 02405180 02405172 02405164 02405156 02405148 02248104 02248105 02248106 02248107 02248103 02375591 02375605 02375613 02375621 02375648 02375052 02375044 02375036 02375060 02375079 02372940 02372932 02372975 02372967 02372959 02246582 02246583 02246585 02246584 02246737 02269260 02269252 02269279 02269295 02269287 02329174 02329182 02329158 02329166 02329131 02265907 02265893 02265885 02250179 02250187 02250160 02250152 02250144 00884340 00884359 02240332 00884324 MAO inhibitors (mono 00476552 01919598 02230641 02123312 02231036 amine oxide inhibitors) 02068087 00476552 01919598 Benzodiazepines 00865400 00865397 02243611 02243612 02400111 02400138 02400146 02400154 02137534 02229814 02137542 02229813 02417642 02417650 02417634 01913484 01913492 00723770 00548359 00548367 00813958 00522724 00522988 00522996 00012637 00012645 00012629 02177897 02177889 02270641

324

02230950 02230951 02239025 02239024 02048701 02048736 02207818 02103737 00382825 00382841 00382841 02238162 09853340 09853430 00405337 00405329 00362158 00013293 00013277 00013285 02410761 02410753 02410745 02041472 02041456 02041464 00655740 00655759 00655767 02041413 02041448 02041421 00637742 00711101 00637750 00728187 00728195 00728209 00402680 00402745 00402737 02043688 02043661 02043653 Gaba receptor agonist 02436159 02434946 02391678 02370433 02245077 02218313 02271931 02271958 02216167 01926799 02356805 02406969 02406977 02386771 02386798 02391716 02391724 02296616 02238596 02251469 02251450 02243426 02240606 02267926 02267918 02246534 02242481 02008203 02257572 02386909 02386917 Tetracyclic antidepressants 02436159 02434946 02391678 02370433 02245077 02218313 02271931 02271958 02216167 01926799 02356805 02406969 02406977 02386771 02386798 02391716 02391724 02296616 02238596 02251469 02251450 02243426 02240606 02267926 02267918 02246534 02242481 02008203 02257572 02386909 02386917 Non-steroidal anti- 02243433 02239753 02261774 02239355 00881635 inflammatory drugs 00839175 00839183 00808547 00808539 02302624 (NSAIDs) 02302616 02261960 00514012 00514004 02091194 02162814 02158582 02048698 02231504 02231505 02261901 02261944 00782459 00590827 02441020 01940414 02231508 02231506 02261928 02261936 00632724 00632732 02356783 02247265 02420988 02039486 02039494 00587699 00576131 02232317

325

02232318 02142031 02142023 02244680 02244681 02017628 02017636 00647942 00600792 01912038 01912046 02100517 02100509 00016039 00016047 00337420 00337439 00016233 00594466 01934139 02231799 02231800 00790427 01926403 00790435 00842664 01926365 01926381 00761680 02172577 01926373 01926411 02015951 01968300 02245821 02369362 02229080 02162660 02238639 02244563 02240867 02240868 02083531 02083558 02246700 02246701 02246699 02243432 02241024 02162792 02162423 02162415 02294710 02294702 02243314 02243313 02243312 02162431 02162458 02017237 00592277 00600806 00522651 00522678 02162482 02162474 02162490 00299413 00627097 00589861 00565350 02162725 02162717 00784354 01940309 00642894 00642886 00525596 00525618 00695718 00695696 00632716 02154463 00456888 00432369 00745588 00745596

326