Implementation and Evaluation of a Depression Care Model for Homebound Elderly

LWW/JNCQ NCQ200198 November 7, 2012 17:25

J Nurs Care Qual Vol. 28, No. 1, pp. 33–42 Copyright c 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins Implementation and Evaluation of a Depression Care Model for Homebound Elderly

Rose Madden-Baer, DNP, RN, MHSA, BC-PHCNS, CPHQ, CHCE, COS-C; Eleanor McConnell, PhD, RN, GCNS, BC; Robert J. Rosati, PhD; Peri Rosenfeld, PhD; Ilaina Edison, MBA, RN

Depression affects 14% to 46% of homebound elderly and is costly and disabling. Home health agencies face significant challenges delivering effective depression care. In response, an evidence- based depression care model was developed in a home health agency. Twelve-month program evaluation data demonstrated a 2.99 mean reduction in depression scores (P < .0001) on the Geriatric Depression Scale and confirmed that a clinically effective, operationally feasible, and financially sustainable depression care model can be implemented in home health care. Key Words: cognitive behavior therapy, depression, geriatrics, home care

HE PREVALENCE OF DEPRESSION tients with depression studied were not re- T ranges from 14% to 46%1-3 in the home- ceiving treatment and 40% of the patients bound elderly. In addition, homebound el- on treatment were receiving inadequate ther- derly patients are twice as likely to have apy. Other studies indicate that depression is depression compared with those in primary a prevalent comorbidity with heart disease, care.1 Estimates of the direct and indirect cancer, and diabetes, and depression is asso- medical costs of patients with depression are ciated with hastened mortality in these med- approximately 83.1 billion in the year 2000.4 ical conditions.5 According to Carson and However, there remains a lack of screening Vanderhorst,6 there is a 36% to 38% mortal- for patients with depression and an under- ity rate associated with depression and dia- treatment of their diagnosis and symptoms. betes alone over a 2-year period. Depression Bruce and associates1 found that 78% of pa- is associated with increased medical and func- tional disabilities in the homebound elderly and an increased risk for falls, even when con- trolling for antidepressant use.7,8 Importantly, Author Affiliation: Visiting Nurse Service of New customary home care is not sufficient to treat York, New York (Drs Madden-Baer, Rosati, and 9 Rosenfeld and Ms Edison); and Gerontological depression, as Raue and colleagues docu- Nursing Specialty, Duke University School of mented the persistence of depression even Nursing, Durham, North Carolina (Dr McConnell). after 1-month receipt of standard home care The authors declare no conflict of interest. services. Correspondence: Rose Madden-Baer, DNP, RN, MHSA, Two major changes in home health care BC-PHCNS, CPHQ, CHCE, COS-C, Behavioral Health, As- have highlighted the need to implement and sessment Services and Special Projects, Visiting Nurse sustain evidence-based care for patients with Service of New York (VNSNY), New York, NY 10954 ([email protected]). depression. The first occurred in 2008 when the Medicare payment reimbursement sys- Accepted for publication June 20, 2012 tem added psychiatric diagnoses, such as Published online before print: July 20, 2012 depression, to the case mix.10 This adjust- DOI: 10.1097/NCQ.0b013e318265702a ment reflects the federal movement to address 33

34 JOURNAL OF NURSING CARE QUALITY/JANUARY–MARCH 2013

resource management in caring for patients homebound elderly. Researchers also exam- with psychiatric illnesses such as depression ined the collaborative models of “geropsy- and encourage certified home health agencies chiatric teams” of psychiatrists, advanced to provide services. The second significant practice nurses, psychiatric nurses, and so- change took place in January 2010, when the cial workers to deliver psychiatric evalu- Patient Health Questionnaire–2 and the de- ations, psychotherapy, and psychopharma- pression screening, care planning, and inter- cology interventions.24 These overall results vention process questions were added to the showed compelling evidence for the efficacy OASIS C data set (a mandated home health of these practice models and depression care assessment tool).11 model (DCM) teams in reducing depression Certified home health agencies encounter in the homebound elderly. This evidence pro- significant patient, provider, and system chal- vided the impetus for creation of this DCM lenges in trying to provide clinically appropri- and team in a home health agency, using a ate and cost-effective psychiatric care. Barri- combination approach to depression care. ers include the following: (1) lack of patient and provider awareness of depression; (2) pa- DEPRESSION CARE MODEL DESIGN tient reluctance to seek treatment; (3) navi- AND IMPLEMENTATION gation of complex regulatory and reimbursement requirements;12 (4) misconceptions that Organizational setting these programs are under-reimbursed, expen- The Visiting Nurse Service of New York sive in terms of resource utilization, and not fi- (VNSNY) is a large, complex community nancially sustainable; (4) impressions that the health organization in New York City. The VN- processes needed for recruitment and training SNY is the largest not-for-profit home health of qualified, competent staff to deliver psychi- agency in the United States, with an average atric home care nursing are impractical13,14; daily census of 31 000 patients. The organiza- (5) complexities associated with various risk tion provides a vast array of services includ- screening tools and assessments1,11,14,15;and ing traditional and nontraditional home health (6) and logistical challenges related to ex- care, hospice, and various community-based tensive planning, oversight, and follow-up services programs. The organization is cen- required for a coordinated care and com- trally hierarchical but regionally operated and munication model with primary care physi- deployed from 7 regions including the 5 bor- cians, psychiatric specialists, and home health oughs of New York City and 2 metropolitan nurses.8,12 area suburbs. The organization has a strong reputation as a “safety net” provider of care. EVIDENCE BASE FOR THE DEPRESSION CARE MODEL Description of DCM innovation The DCM is a model of care delivered Clinical practice guidelines exist for de- by a team of psychiatric home care nurses, pression, some of which specifically target advanced practice psychiatric nurses, and the elderly. These guidelines state recom- psychiatrists, known in the VNSNY as the mendations for screening for depression and behavioral health (BH) program. This psychi- treatment using cognitive behavioral therapy atric team implements a 3-component model: (CBT) alone, medications alone, or a combi- (1) a team of specialty-trained psychiatric nation of CBT and medications as the most home care nurses screen for depression, us- effective approach.4,16-19 ing a validated instrument: the Geriatric De- Several systematic reviews5,20-23 also sup- pression Scale (GDS); (2) psychiatric home port combination therapy as effective for de- care nurses provide patient visits using CBT pression care, and in some cases, treatment as counseling techniques in addition to pro- efficacy has been reported specifically for viding ongoing monitoring, psychoeducation,

Depression Care Model for Homebound Elderly 35

and medication management; and (3) psychi- matic unhealthy cognitive thoughts that lead atrists and psychiatric nurse practitioners are to depression. Verbal or written patient re- available and provide “in-home” patient psy- sponse to treatment assignments is also in- chiatric evaluations and diagnostic consulta- cluded. tions with recommendations for medications. Access to psychiatric specialists is provided The psychiatrist or psychiatric nurse practi- to the patients and their primary care physi- tioner consultations are then e-faxed to their cians. At the request of the primary care physi- primary care physicians via an electronic med- cian, psychiatric consultation visits are made ical record (EMR) application. by a psychiatrist or psychiatric nurse practi- The clinical care protocol includes a pre- tioner to collect diagnostic information and scribed set of depression care interventions make treatment suggestions including recom- based on the severity level of a patient’s mended psychotropic medications. This DCM assessed depression score at the first eval- component is designed to assist the primary uation visit. The protocol specifies clinical care physician, who by regulation is responsi- visit guidelines that have been internally de- ble for the home care patient’s entire plan of veloped on the basis of a combination of care, in treating the patient’s psychological as in-house expertise and a review of home well as medical needs. This DCM is designed care and depression care best practices. The to routinely bridge the gaps between multiple visit guidelines also include “home care” pro- unrelated providers in primary, psychiatric, gram/visit requirements specified in federal and home care while also giving primary care and state home health agency rules and regu- physicians the “total patient picture” of their lations, such as recording of patient response homebound elderly patient. to treatment and documentation of specific psychiatric assessments, care plans, and orders. EVALUATION The clinical visit guidelines begin with an initial assessment (pretreatment) visit by a Design psychiatric home care nurse to determine el- A retrospective program evaluation was igibility and medical necessity, using a valid conducted from September 2010 to Septem- and reliable depression instrument: the GDS, ber 2011 to address the following question: which is considered effective for detecting de- Can an evidence-based DCM for homebound pression in the elderly (84%-92% sensitivity elderly (aged 65 years and older) be imple- rates and 89%-95% specificity rates).25 Geri- mented in a certified home health agency atric Depression Scale scores are also ob- that is clinically effective, operationally fea- tained at 55 to 60 days (prior to recertifica- sible, and financially sustainable? The study tion) to evaluate continued medical neces- used the RE-AIM model through the assess- sity or at the discharge visit (posttreatment ment of 5 dimensions, which include: reach, visit), whichever comes first. Documentation effectiveness, adoption, implementation, and of continued skilled need and continued med- maintenance.26 The RE-AIM model is a com- ical necessity is required by Medicare for prehensive evaluation framework used to de- continuing skilled home care into a second termine the overall impact of a population- episode. based or public health program. Psychiatric home care CBT counseling visits The model’s reach or ability to increase pa- are typically delivered once a week; however, tient access was examined by assessing refer- the guidelines allow for more frequent visits if ral volume, targeted at 50 per month, which indicated by the patient’s clinical condition or corresponded to a 15% prevalence rate of pa- depression level. Psychiatric home care visits tients under care by VNSNY with suspected include delivering the CBT techniques, which depression during a 2009 analysis. Evaluation are aimed at reducing and eliminating auto- of the model’s adoption was measured by

36 JOURNAL OF NURSING CARE QUALITY/JANUARY–MARCH 2013

referral and admission volume and team trans- Sample fer rates. Referral and admission volume rep- Patients included in the program evalua- resented the program’s acceptance by re- tion were homebound patients aged 65 years ferred patients, their primary care providers, and older who (1) were admitted into the and the BH team’s nurses. Transfer rates re- BH program and treated between September flected the program’s diffusion into “busi- 2010 and September 2011, (2) had a complete ness as usual” and an increased awareness home care OASIS C assessment, (3) received by agency staff that depression may con- the depression care interventions, (4) had a tinue beyond their patient’s medical needs primary or comorbid diagnosis of depression and are best served by psychiatric services and had signs/symptoms of depression as evi- rather than standard home care. These reach denced by a score of 6 or greater on the GDS, and adoption measures were reported quar- or whose condition had not been formally di- terly both by program performance indi- agnosed but had a score of 6 or greater on cator reports and by the program’s intake the GDS, and (5) had a pre- and posttreat- coordinator. ment GDS score. Patients were excluded if Three implementation or fidelity of pro- they were (1) rehospitalized and thus did not cess measures and an effectiveness measure complete the program or (2) did not com- were used to assess clinical performance. The plete the treatment for other reasons such as implementation fidelity measures included patient refusal or moved out of service area. the following: (1) triggered CBT care plans as A power analysis indicated that a sample size the treatment protocol in the clinical system of 57 had 80% power (α = .05, 2-tailed) to for these patients; (2) the presence of a pre- detect an effect size of 38% to 40% and a sam- and posttreatment GDS scores documented ple of 92 had 99% power. The final sample in the EMR; and (3) the number of psychiatric included 191 patients who met all study crite- nursing visits made per episode of care com- ria and thus was adequately powered to assess pared with the recommended episode num- program effectiveness. ber of visits in the clinical protocol. Effective- ness was measured by mean change in GDS scores from pre- to posttreatment. Data collection and analysis In addition, financial sustainability was eval- Data collection and analysis occurred be- uated by maintenance metrics for overall tween September and December 2011 after program margin and the program’s ability institutional review board approval was ob- to provide favorable results with regard to tained for the study. Data were collected and organizational revenue and budget targets. reported by using agency research and fi- These targets include the following: (1) av- nance/operational analysts and 2 work study erage episode margin, defined by average rev- BSN interns. The interns assisted with medical enue versus cost for each episode; (2) overall record abstraction, audits, and data recording program revenue compared with budget tar- to prevent bias and preserve the integrity of gets; (3) the percentage of contribution mar- results. De-identified data reports were pro- gin compared with budget targets; and (4) vided to this evaluator by the analysts and the volume of payer denials. These financial interns. metrics were preset in January 2011 and are The program’s reach was evaluated using reported in financial statements as revenues descriptive statistics on the demographic and and program contribution margin percent- clinical attributes of clients served. Referral ages. The reports are available monthly via the volume was then assessed to evaluate reach financial reporting systems and are available or access to the program. To evaluate the ef- online for review and analysis. Specific indi- fectiveness of the DCM, changes in pre- and cators of each metric are summarized in the posttreatment depression scores on the GDS Table. were obtained and then analyzed for statistical

Depression Care Model for Homebound Elderly 37

Table. Metrics Used in Evaluation

Evaluation Domain Measure How Obtained Notes

Reach Demographics Secondary data from Clinical attributes (eg, agency EMR diagnoses) Intake coordinator’s Referral volume referral database Effectiveness Change in GDS score Data extracted from agency EMR and analyzed for statistical significance Adoption Team transfer rate and Scorecard measures referral and retrieved from the admission yield EMR, operations metrics information systems, and intake coordinator referral database Implementation Activation of CBT care Scorecard data extracted Medical record fidelity plans from the EMR and extraction and Completion of pre- and financial and analysis of fidelity posttreatment GDS operational measures Number of psychiatric information systems Scorecard measures nursing visits per Medical record analyzed for the episode abstraction number of nursing visits per episode Maintenance Program/episode Financial information Monthly performance margin and data on systems indicator reports and variance to revenue financial reports and budget targets Number of payer denials

Abbreviations: CBT, cognitive behavioral therapy; EMR, electronic medical record; GDS, Geriatric Depression Scale.

significance using a paired t test. Agency-wide reasons for the unsuccessful conversion of re- adoption of the model was measured through ferrals to admission such as patient refusal or evaluation of both admission yield volume and patient not identified as depressed on the GDS transfer volume expressed as team transfer instrument at the evaluation visit. The fidelity rates. The team transfer report represents pa- of the program’s implementation included tients under care in one of the other agency (1) activated CBT care plans, (2) completion home care programs (identified and evaluated of the pre- and posttreatment GDS, and (3) as patients with depression) and then subse- the number of nursing visits per episode. Au- quently formally transferred to BH as the pri- dits and data extraction of fidelity measures mary program. The referral/admission volume determined the proportion of records where yield report identified the number of referrals a pre- and posttreatment GDS had been com- made to the program that translated into ac- pleted and a CBT care plan had been opened tual admissions. This report also highlighted as was required by the clinical guidelines.

38 JOURNAL OF NURSING CARE QUALITY/JANUARY–MARCH 2013

The first 2 of the fidelity measures were re- fective disorders/other psychoses (37%), di- ported as proportions on a 3 × 2contin- abetes (27%), gait abnormalities (17%), and gency table, and the psychiatric nursing visits other orthopedic disorders (16%). per episode were reported as a mean. Mainte- nance metrics were evaluated through finan- Effectiveness cial sustainability metrics including episode Change in pre- and posttreatment GDS margin, number of payer denials, and success scores for the study sample was used to with meeting revenue and budget targets. Fi- evaluate effectiveness of the DCM. A paired nancial and operational metrics are routinely t-test analysis found a mean reduction of 3 (SD reported on a program-specific performance = 3.30) points in the GDS score from base- indicator and monthly financial reports. line to posttreatment. This decrease is both clinically and statistically significant (t190 = − 12.57; P < .0001; 95% confidence interval, RESULTS − 3.47 to − 2.53). Distribution of the change in scores (posttreatment GDS score minus Reach baseline score) is presented in the Figure. This During the 12-month evaluation period, decrease represents an overall reduction in 597 referrals were made to the BH program depression severity or elimination of depres- which yielded a total of 546 admissions for sion symptoms for most patients posttreat- the DCM. This referral volume was slightly ment. lower than the targeted referral volume for the 12-month period. Of those admitted to Adoption BH, 220 of the 546 (40%) completed an en- The referrals to admission yield and team tire treatment episode without interruptions transfer rates were used to evaluate DCM such as hospitalizations or facility transfers. In adoption. Study results showed that 91% of addition, 29 of the 220 patients had no docu- the patients referred to the DCM were ac- mented post-treatment GDS score. Therefore, cepted and admitted. The 9% of patients who a descriptive analysis was performed on the were not admitted to the program were ana- remaining sample of 191 patients. lyzed for reasons and included the following Descriptive analysis of the patients served categories: patients refused services, patients by the DCM in the BH program included age, already receiving “in-home” mental health ser- gender, lives with status, common diagnoses, vices, patients not admitted because the pri- and number of comorbidities. The BH pro- mary care physician determined services were gram sample (N = 191) was then analyzed not necessary, and other such as patients mov- in comparison with a corresponding (age 65 ing out of the service area. Team transfer years and older) VNSNY sample (N = 56 281). rates were evaluated as an additional measure The mean age for the BH group was 75 years of adoption. These rates reflect how embed- compared with 79 years for the VNSNY group. ded or diffused the model is within routine The proportion of females in the BH program processes, workflows, and agency operations. was 80% compared with 64% in the VNSNY Overall, the 2011 transfer rate was 29.2% com- group. Likewise, the proportion of persons pared with the agency target of 30%. Actual living alone in the BH sample was greater, transfer rates ranged from 23% to 40% and with 38% living alone compared with 27% in were highest in regions where the DCM was the VNSNY group. Both groups displayed the pilot tested. presence of up to 6 comorbidities, with most patients having 3 or more comorbidities for Implementation both samples. The most common diagnoses Fidelity of implementation measures was according to frequency in BH sample were examined for all patients in the original sam- hypertension (75%), heart disease (30%), af- ple (N = 220) who completed the DCM

Depression Care Model for Homebound Elderly 39

Figure. Distribution of change in Geriatric Depression Scale (GDS) score from baseline to posttreatment.

intervention without interruption. Patients were used to evaluate the ability to maintain who experienced interruptions such as hos- the program. During the study period, pro- pitalizations or refused to complete the inter- gram results were as follows: (1) the average vention were excluded. Of the 220 patients, margin per episode was positive, (2) the net 100% had a pretreatment GDS score and 87% contribution margin for the program was 17%, had a posttreatment GDS score noted in the (3) program revenue exceeded budget targets clinical record. The second implementation by 8%, and (4) there were no reported payer fidelity measure was CBT care plan activa- denials. tion for the sample. The compliance rate for activation noted in the EMR clinical record DISCUSSION was 90%. In the 10% of cases for which the CBT care plan was not activated, 21 records The program evaluation indicated that a had free text documentation of CBT interven- clinically effective, operationally feasible, and tions in the clinical notes. Therefore, docu- financially sustainable evidence-based DCM mented CBT interventions were present in for homebound elderly patients could be 99% of the sample cases. The final implemen- implemented in a home care agency. The tation measure was psychiatric nursing visit DCM program increased patient access to compliance with the clinical visit guidelines. services when there was a previous unmet The number of average psychiatric nursing need, and depression severity was reduced visits per episode was 8.3 during the study or eliminated by the DCM program’s spe- period compared with the DCM target of 8 cific interventions. In comparison, Raue and to 10 visits per episode based on guideline colleagues9 found depression persistence af- parameters. ter 1 month of usual standard home care. Program evaluation also showed that model Maintenance adoption and implementation fidelity grad- The financial sustainability of the DCM in- ually improved over time, with monitor- cluded program margin per episode, ability ing and incremental adjustments to systems, to meet program revenue and budget targets, communications, processes, workflows, and and the number of payer denials, and these EMR integration. These incremental change

40 JOURNAL OF NURSING CARE QUALITY/JANUARY–MARCH 2013

methods are necessary for successful diffusion requirements of the DCM are essential to of innovation and consistent with methods maintain its sustainability. espoused by Greenhlaugh and colleagues27 The evaluation also generated findings that needed to sustain successful evidence-based have implications for future research. Con- practices. Program margin depended largely sistent with prior research, depression was on fidelity to the DCM to ensure quality, effi- found among patients with diagnoses such as ciencies, and cost-effectiveness. heart disease and diabetes and those with mul- Financial and operational results substan- tiple comorbidities. There also was a preva- tiate the sustainability of this program as lence of depression among patients who lived a Medicare-covered service with an overall alone (38%). This is potentially an area that product margin for the program. Further- warrants further study. Also, there was a high more, this program evaluation provides data prevalence of depression among women 75 for program expansion to other psychiatric years or older. Further research on prevalence illnesses and a framework for evaluation of rates is needed among age cohorts for both the impact on these psychiatric conditions. males and females. Thus, creation of this BH program for patients In terms of fidelity measures, overall CBT with depression has proven to be a new op- care plan activation and completion of post- portunity for the organization. The program treatment assessment tools rates improved af- evaluation has shown that DCM interventions ter the pilot and after full integration of those are clinically effective in measurably reducing components into the organization’s EMR. This patients’ depression scores while also being is consistent with other research on the use- reimbursable and having a program margin. fulness of embedding tools and protocols into Measures to support ongoing monitoring of the EMR to foster compliance with evidence- fidelity to the model have been integrated to based practice. In addition, EMR integration assess future team compliance. of the tools and protocols reduces depen- Other methods planned to sustain the dence on free text for audits through facili- model included diffusion of this model into tation of medical record abstraction. Embed- all regions and teams within the organization. ding required assessment tools at time points Sustainability of the program also depends on rather than on demand may further influence continued seamless transfers of patients un- positive clinician behavior. Program transfer der care into the program for care coordina- rates also increased as the program became tion and treatment of their BH needs when routinized into each region’s usual practices. their medical status has stabilized. In addition, Other patient outcomes were examined in extension of the program to direct external subsequent analyses and showed statistically referral sources is another move on the road significant reductions in confusion severity toward sustainability. and disruptive behaviors in the sample. Fur- Sustainability also depends on the pro- ther research should be conducted to deter- gram’s ability to be routinely reimbursed for mine whether these findings can be repli- services, ability to meet revenue targets, and cated. avoidance of denials. Avoidance of denials in Limitations to the study include the fact this model is achieved through strict adher- that the program evaluation was done in only ence to the protocols for risk assessment, 1 large home health agency in an urban set- medical necessity, evidence of patient re- ting. This may affect the generalizability of the sponse to treatment, protocol and documen- findings to all certified home health agencies. tation requirements, and psychiatric home Furthermore, the sample was a nonrandom care rules and regulations. Continual over- sample of all patients who met the inclusion sight and monitoring of fidelity to protocols, criteria, and there was no control group to visit guidelines, and the EMR documentation make comparative analyses. Finally, the lack

Depression Care Model for Homebound Elderly 41

of a posttreatment score for 29 patients in the able DCM, address potential barriers and chal- original sample of 220 patients could poten- lenges, and create a robust evaluation to mea- tially bias the results on the clinical interven- sure the program’s weaknesses and successes. tion’s effectiveness. In doing so, certiﬁed home health agencies can develop similar programs and increase the home health industry’s response to the CONCLUSIONS current yet often overlooked and unmet need for depression care for the homebound el- The ﬁndings derived from this program’s derly. Finally, the ability to demonstrate the implementation experience and program successes of innovative home health models evaluation can inform other home health such as this DCM underscores the importance agencies on how to implement a successful of having home health providers and the nurs- DCM program. The program evaluation serves ing profession’s full engagement in shaping as a guide to build and implement a sustain- future health policy.

REFERENCES

1. Bruce ML, McAvay GJ, Raue PJ, et al. Major depression system: ﬁnal report. https://www.cms.gov/Research- in elderly home health care patients. AmJPsychiatry. Statistics-Data-and-Systems/Statistics-Trends-and- 2002;159(8):1367-1374. Reports/Reports/downloads/Coleman_Final_April_ 2. Raue PJ, Weinberger MI, Sirey JA, Meyers BS, Bruce 2008.pdf. Published 2008. Accessed June 6, 2012. ML. Preferences for depression treatment among el- 11. Centers for Medicare & Medicaid Services. Out- derly home health care patients. Psychiatr Serv. come and Assessment Information Set (OASIS-C). 2011;62(5):532-537. http://www.cms.gov/HomeHealthQualityInits/ 3. Valente S. Detecting and preventing depression Downloads/HHQIOASISVerson12-2.pdf. Published among elderly patients in home health. Home Health 2011. Accessed June 6, 2012. Care Manag Pract. 2005;17(2):101-108. 12. Zeltzer BB, Kohn R. Mental health services for 4. US Preventive Services Task Force. Screening for homebound elders from home health nursing agen- depression in adults: U.S. Preventive Services Task cies and home care agencies. Psychiatric Serv. Force recommendation statement. Ann Intern Med. 2006;57(4):567-569. 2009;151(11):784-792. 13. Friedman B, Delavan RL, Sheeran TH, Bruce ML. The 5. Kang-Yi CD, Gellis ZD. A systematic review of effect of major and minor depression on Medicare community-based health interventions on depres- home healthcare services use. J Am Geriatr Soc. sion for older adults with heart disease. Aging Ment 2009;57(4):669-675. Health. 2010;14(1):1-19. 14. Cunningham PA. The role of the psychiatric nurse in 6. Carson VB, Vanderhorst KJ. OASIS-C, depression home care. Home Healthc Nurse. 2007;25(10):645- screening, and M1730: additional screening is nec- 652; quiz 653-644. essary: the value of using standardized assessments. 15. Ell K, Unutzer J, Aranda M, Sanchez K, Lee PJ. Routine Home Healthc Nurse. 2010;28(3):183-190. PHQ-9 depression screening in home health care: de- 7. Byers AL, Sheeran T, Mlodzianowski AE, Meyers BS, pression, prevalence, clinical and treatment charac- Nassisi P, Bruce ML. Depression and risk for adverse teristics and screening implementation. Home Health falls in older home health care patients. Res Gerontol Care Serv Q. 2005;24(4):1-19. Nurs. 2008;1(4):245-251. 16. Fochtmann I, Gelenberg AJ. Guideline Watch: Prac- 8. Sheeran T, Byers AL, Bruce ML. Depression and in- tice Guideline for the Treatment of Patients With creased short-term hospitalization risk among geri- Major Depressive Disorder. Arlington, VA: American atric patients receiving home health care services. Psychiatric Association; 2005. Psychiatr Serv. 2010;61(1):78-80. 17. Gelenberg AJ, Freeman MP, Markowitz JC. Practice 9. Raue PJ, Meyers BS, McAvay GJ, Brown EL, Keo- Guideline for the Treatment of Patients With Major hane D, Bruce ML. One-month stability of depression Depressive Disorder. Arlington, VA: American Psy- among elderly home-care patients. Am J Geriatr Psy- chiatric Association; 2010. chiatry. 2003;11(5):543-550. 18. Kurlowicz LH, Harvath TA. Depression: evidence- 10. Centers for Medicare & Medicaid Services. Reﬁne- based geriatric nursing protocols for best practice. ment of Medicare’s home health prospective payment In: Capezuti DZE, Mezey M, Fulmer T, eds. Evidence-

42 JOURNAL OF NURSING CARE QUALITY/JANUARY–MARCH 2013

Based Geriatric Nursing Protocols for Best Practice. 23. Wilson KC, Mottram PG, Vassilas CA. Psychother- 3rd ed. New York, NY: Springer; 2008:57-82. apeutic treatments for older depressed people. 19. Piven MLS. Detection of depression in the cogni- Cochrane Database Syst Rev. 2008;(1):CD004853. tively intact older adult. In: Titler MG, ed. Series on 24. Knight MM, Houseman EA. A collaborative model for Evidence-Based Practice for Older Adults.IowaCity, the treatment of depression in homebound elders. IA: The University of Iowa Gerontological Nursing Issues Ment Health Nurs. 2008;29(9):974-991. Interventions Research Center, Research Translation 25. Holroyd S, Clayton AH. Measuring depression in the and Dissemination; 2005. elderly which scale is best. Medscape Mental Health. 20. Ayalon L, Fialova D, Arean PA, Onder G. Challenges 2000;5(5):1-8. associated with the recognition and treatment of de- 26. Glasgow RE, Klesges LM, Dzewaltowski DA, Es- pression in older recipients of home care services. Int tabrooks PA, Vogt TM. Evaluating the impact of health Psychogeriatr. 2010;22(4):514-522. promotion programs: using the RE-AIM framework to 21. Ell K. Depression care for the elderly: reducing bar- form summary measures for decision making involv- riers to evidence-based practice. Home Health Care ing complex issues. Health Educ Res. 2006;21(5):688- Serv Q. 2006;25(1/2):115-148. 694. 22. Steinman LE, Frederick JT, Prohaska T, et al. Rec- 27. Greenhlaugh T, Robert G, Bate R, MacFarlane F, ommendations for treating depression in community- Kyriakidou O. Diffusion of Innovations in Health based older adults. Am J Prev Med. 2007;33(3):175- Services Organisations. Oxford, England: Blackwell; 181. 2007.