Development and Validation of a Self-Report Symptom Inventory to Assess the Severity of Oral-Pharyngeal Dysphagia

UNIVERSITY OF NEW SOUTH WALES Thesis/Project Report Sheet

Surname or Family name: WALLACE First name: KAREN Other name/s: LEE Abbreviation for degree as given in the University calendar: MSc. School: St George Clinical School Faculty: Medicine Development and validation of a self-report symptom inventory to assess the severity of oral-pharyngeal dysphagia.

Abstract 350 words maximum:

Background/Aims: Objective assessment of symptonnatic change in the severity of oral-pharyngeal dysphagia is hampered by the lack of a valid and reliable means of measuring symptom severity. The aim of this thesis was to develop and evaluate the validity and reliability of a self-report inventory to measure symptomatic severity of oral-pharyngeal dysphagia.

Methods: Test-retest reliability; face-, content- and construct-validity of a prototype, visual analogue scale, inventory was assessed in 45 patients who had stable, neuromyogenic dysphagia. Discriminant validity was estimated in 11 patients with Zenker's diverticulum who underwent surgery.

Results: Normalised scores varied over time by -0.5% ± 17.6% (95% 01: -9.2% - 8.2%). Factor analysis identified a single factor (dysphagia), to which 18 of 19 questions contributed significantly, which accounted for 56% of total variance (p < 0.0001). Following deletion of 2 questions with poor face validity and patient compliance, this proportion rose to 59%; mean test-retest change was -2% (95% 01: -11% - 7%); and total score correlated highly with an independent global assessment severity score (r = 0.7; p < 0.0001). A mean 70% reduction in score (p < 0.0001) was observed following surgery in patients with Zenkers.

Conclusion: Applied to patients with neuromyogenic dysphagia the 17 question inventory demonstrates: strong test-retest reliability over two weeks as well as face, content, and construct validity. Discriminant validity (responsiveness) has been demonstrated in a population with a correctable, structural cricopharyngeal disorder. Responsiveness of the instrument to treatment in neuromyogenic dysphagia remains to be quantified.

Declaration relating to disposition of project report/thesis

I am fully aware of the policy of the University relating to the retention and use of higher degree project reports and theses, namely that the University retains the copies submitted for examination and is free to allow them to be consulted or borrowed. Subject to the provisions of the Copyright Act 1968, the University may issue a project report or thesis in whole or in part, in photostat or microfilm or other copying medium.

1 also authorise the publication by University Microfilms of a 350 word abstract in Dissertation Abstracts international (applicable to doctorates only).

Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing to the Registrar. Requests for a longer period of restriction may be considered in exceptional circumstances if accompanied by a letter of support from the Supervisor or Head of School. Such requests must be submitted with the thesis/project report.

FOR OFFICE USE ONLY Date of completion1 o of trequirement requirementss fo TOr Award amm "ir^^ cii-O sputv Principal Development and Validation of a Self-report Symptom

Inventory to Assess the Severity of Oral-pharyngeal

Dysphagia.

Ms Karen Wallace

Master of Science

2000 UNSW ? 2 MAY 2001 LIBRARY Dedication

To all the people who have walked with me along

my path and to those who are yet to come.

~ Life -

Author Unknown

Experiences have continued to highlight and enrich my life. Encounters with different people have allowed me to refine and shape my conception of the world and myself, and have helped me to remain open to ongoing changes occurring within and without me. By reflecting and re-examirung my own attitudes and beliefs, I can see more clearly who I am and where I stand Table of Contents

DEDICATION I

TABLE OF CONTENTS II

CERTIFICATE OF ORIGINALITY V

ACKNOWLEDGMENTS VI

PUBLICATION ARISING FROM THIS THESIS VII

ABSTRACT VIII

CHAPTER 1 BACKGROUND 1

1.1 RATIONALE OR NEED FOR MEASUREMENT TOOLS IN DYSPHAGIA RESEARCH 1

1.1.1 The magnitude and cost of the oropharyngeal dysphagia I

1.1.2 Why is it important to measure severity of dysphagia? 2

1.2 PROJECT AIM 4

1.3 AVAILABLE MEASUREMENT TOOLS TO QUANTIFY DYSPHAGIA SEVERITY 5

1.3.1 Measures of efficacy and function 5

L3.I.1 Videofluoroscopy 5

1.3.1.2 Videomanometry 6

1.3.1.3 Scintigraphy 7

1.3.2 Measures of symptom severity 7

1.3.2.1 Clinician's assessnoent - history and examination 7

1.3.2.2 Symptom questionnaires and inventories 9

1.4 EXPECTATIONS OF A NEW SWALLOW INVENTORY 16

1.4.1 Summary 18

1.5 MECHANISM OF ORAL-PHARYNGEAL DYSPHAGIA AND RELEVANCE TO SWALLOW

INVENTORY DESIGN 20

1.5.1 Introduction 20

1.5.2 Mechanism ofNomial & Ahiormal Swallow 20 1.5.3 Disease Aetiology 23

1.6 TREATMENTS, OUTCOMES AND RESPONSIVENESS TO THERAPY 24

1.6.1 Importance of natural history 24

1.7 SUMMARY 26

CHAPTER 2 METHODS 28

2.1 INTRODUCTION - FUNDAMENTALS OF INVENTORY OR QUESTIONNAIRE DESIGN 28

2.2 THE INVENTORY 30

2.3 PATIENTS AND CONTROLS 32

2.4 DATA COLLECTION 33

2.5 ANALYTICAL METHODS 34

2.5.1 Test-retest Reliability 35

2.5.2 Validity Measures 36

2.5.2.1 FACE VALIDITY 36

2.5.2.2 CONTENT VALIDITY 37

2.5.2.3 CONSTRUCT VALIDITY 39

2.5.2.4 DISCRIMINANT VALIDITY 39

CHAPTER 3 RESULTS

3.1 INVENTORY DEMOGRAPHICS 41

3.2 RELIABILITY AND VALIDITY TESTING 41

3.2.1 Predictive validity 41

3.2.2 Test-retest Reliability 42

3.2.3 Face validity 42

3.2.4 Content Validity 43

3.2.5. Construct Validity 43

3.2.6 Discriminant Validity 44

3.3 PERFORMANCE OF THE MODIFIED INVENTORY 44

CHAPTER 4 DISCUSSION, CONCLUSIONS AND IMPLICATIONS 46

4.1 DISCUSSION AND CONCLUSIONS 46

4.2 IMPLICATIONS FOR FUTURE RESEARCH 52 III REFERENCES: 53

LIST OF TABLES:

TABLE 1: CURRENT LITERATURE-CRITICAL EVALUATION 65 TABLE 2: MECHANISMS OF OROPHARYNGEAL DYSPHAGL^ 72 TABLE 3: AETIOLOGY OF ORAL-PHARYNGEAL DYSPHAGIA 73 TABLE 4: COMPONENT QUESTIONS OF THE PROTOTYPE SWALLOW INVENTORY 74 TABLE 5: AETIOLOGY OFNEUROMYOGENIC DYSPHAGL\ IN GROUP 1 75 TABLE 6: SUMMARY OF THE FACTOR ANALYSIS MATRIX WITH COMMUNALITY SUMMARY IN GROUP 1 PATIENTS (N = 45) 76

LIST OF FIGURES:

FIGURE 1 .77 FIGURE 2 .78 FIGURES .79 FIGURE 4 .80

FIGURE 5 .81

FIGURE 6 .82 Certificate of Originality

I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgment is made in the thesis. Any contribution made to the research by others, with whom I have worked at

UNSW or elsewhere, is explicitly acknowledged in the thesis.

I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.

Karen Wallace

Signed: Acknowledgments

My thanks and gratitude for gently persisting throughout the life of this thesis must go to my extraordinarily patient co-supervisor Associate Professor Ian Cook. His support, energy and mentorship during the writing of this thesis gave me the fortitude to persist till the end. To Rohan Williams, who dealt with small but significant problems that one faces by being external to the primary location of the work in the later stages, I would like to pass on heartfelt thanks and wish him well with his PhD. Hopefully, I will be in a position to return the favour in the near future. To my other co-supervisor. Associate Professor David de Carle, I would like to thank for his public support with the Faculty of Medicine at the University of New South Wales and his faith that the commitment to complete the task at hand would be fulfilled.

In thanking my family, I must confess that they too have taken a leap of faith in my fulfilling a commitment made so long ago. To my parents. Bill and Der\ise, who have instilled in me the values to grasp life with both hands, step out and take responsibility for any endeavour I choose to pursue in life. To my siblings, Linda and Andrew, who have been there in my life standing beside me. In more recent times to my loving, understanding, and tolerant husband Craig, who has lived through the ups and downs of this thesis.

To all the patients, many of whom were experiencing a traumatic time in their own lives, who extended the generosity to help a fellow man; I thank you as without this conunitment the research would not have been possible.

As this landmark step in life passes, it is now time to move onto the next chapter... Publication arising from this thesis Wallace, K.L., S. Middleton, 1. J. Cook (2000). "Development and validation of a self- report symptom inventory to assess severity of oral-pharyngeal dysphagia." Gastroenterology 118:678-687 Abstract

Background/Aims: Objective assessment of symptomatic change in the severity of oral-pharyngeal dysphagia is hampered by the lack of a valid and reliable means of measuring symptom severity. The aim of this thesis was to develop and evaluate the validity and reliability of a self-report inventory to measure symptomatic severity of oral-pharyngeal dysphagia.

Results: Normalised scores varied over time by -0.5% ± 17.6% (95% CI: -9.2% -

8.2%). Factor analysis identified a single factor (dysphagia), to which 18 of 19 questions contributed significantly, which accounted for 56% of total variance

(p < 0.0001). Following deletion of 2 questions with poor face validity and patient compliance, this proportion rose to 59%; mean test-retest change was -

2% (95% CI: -11% - 7%); and total score correlated highly with an independent global assessment severity score (r = 0.7; p < 0.0001). A mean 70% reduction in score (p < 0.0001) was observed following surgery in patients with Zenkers.

(responsiveness) has been demonstrated in a population with a correctable, structural cricopharyngeal disorder. Responsiveness of the instrument to treatment in neuromyogenic dysphagia remains to be quantified. Chapter 1 Background

1.1 Rationale or need for measurement tools in dysphagia research

1.1.1 The magnitude and cost of the oropharyngeal dysphagia

Oral-pharyngeal dysphagia is a common multi-dimensional symptom complex carrying significant morbidity, mortality and cost (Winstein 1981; Groher 1986;

Siebens, Trupe et al. 1986; Gordon, Hewer et aL 1987; Horner, Massey et al.

1988; Barer 1989; Bloem, Lagaay et al. 1990; Lindgren and Janzon 1991; Johnson,

McKenzie et al. 1993; Croghan, Burke et al. 1994). The epidemiological evidence on the prevalence of dysphagia in the wider community is limited.

The studies, which have examined oropharyngeal dysphagia, report an incidence of between 16-22% in people over the age of 50(Kjellen and Tibbling

1981; Bloem, Lagaay et al. 1990; Lindgren and Janzon 1991).

Oral-pharyngeal dysphagia generally has a neurogenic or myogenic aetiology.

The commonest cause of oral-pharyngeal dysphagia is stroke. One third of stroke patients overall and 67% of those with brainstem stroke have dysphagia

(Horner, Massey et al. 1988). Of the patients with stroke related dysphagia, the mortality rate within 6 months is 45 - 68%, this being largely due to dysphagia- related nutritional and pulmonary complications (Barer 1989; Schmidt, Holas et al. 1994). It is estimated that 12% of patients in acute care hospitals and up to

60% of nursing home occupants have feeding difficulties (Siebens, Trupe et al.

1986) of whom a substantial proportion have oral-pharyngeal dysphagia. Other commonly affected populations include head injuries, Parkinson's disease and 1 Alzheimer's in which the prevalence of oral-pharyngeal dysphagia is 20 - 40%

(Horner, Massey et al. 1988; Barer 1989).

The consequences of oral-pharyngeal dysphagia can be severe: malnutrition, aspiration, choking, pneumonia and death with an overall 12 month mortality of 45 - 62% (Croghan, Burke et al. 1994; Cowen, Simpson et al. 1997). Among the subset of patients with post-stroke dysphagia, aspiration pneumonia occurs in 43 - 50% during the first year. Radiological demonstration of aspiration carries a 45% 12-month mortality (Johnson, McKenzie et al. 1993; Croghan,

Burke et al. 1994). In light of our ageing population, oral-pharyngeal dysphagia is a large and growing problem. The cost implications are significant and the resulting complications significantly increase the chance of institutionalisation and hospital readmission following stroke (Smithard, O'Neill et al. 1996).

1.1.2 Why is it important to measure severity of dysphagia?

Oral-pharyngeal symptoms are complex and varied, with many diseases producing similar symptoms. In addition to bolus hold-up and difficulties in clearing a swallowed bolus from the mouth and pharynx, affected patients' report a range of associated symptoms, all of which may contribute to the overall severity of dysfunction. For example poor oral containment, post-nasal regurgitation and aspiration are not necessarily manifestations of bolus arrest but may be equally important contributors to feeding disability and quality of life. The measurement of dysphagia severity is important in making management decisions, prognosticating, and in the objective evaluation of treatment efficacy.

Due to the lack of a truly objective, reliable and valid means of measuring overall symptom severity, the symptomatic outcome assessment to aid in therapeutic intervention for oral-pharyngeal dysphagia has largely been qualitative (Cook and Kahrilas 1999). There is currently no available questionnaire or inventory with proven reliability, stability and validity designed to quantify specifically the symptomatic severity of oral-pharyngeal dysphagia (Fleming and Weaver 1987; Sonies, Parent et al. 1988; Lindgren and

Janzon 1991; Nathadwarawala, Nicklin et al. 1992; DePippo, Holas et al. 1994;

Nathadwarawala, McGroary et al. 1994; Wintzen, Badrising et al. 1994; Deary,

Wilson et al. 1995; Gustafsson and Theorell 1995; Litvan, Narahari et al. 1997;

Young and Durant-Jones 1997; Hinds and Wilds 1998; Murry, Madasu et al.

1998). 1.2 Project Aim

The aim of this project was to design an inventory to measure the severity of oral-pharyngeal dysphagia and to determine its reliability, and validity and to examine its sensitivity to change in response to therapy. 1.3 Available measurement tools to quantify dysphagia severity

1.3.1 Measures of efficacy and function

Current quantitative and qualitative assessment techniques are vital in the assessment of swallowing dysfunction, however they are limited by the fact that they examine specific aspects of oral-pharyngeal dysfunction (Logemann

1983; Logemann 1991; Cook and Kahrilas 1999). The methods available to the clinician include: videofluoroscopy (modified barium swallow); pharyngeal manometry combined with videofluoroscopy, and nasolaryngoscopy. In addition, less frequently adopted methods have been used primarily in the research setting and include: pharyngeal and cricopharyngeal EMG recordings, pharyngeal scintigraphy to assess oral and pharyngeal clearance.

1.3.1.1 Videofluoroscopy

Of the currently available techniques, videofluoroscopy, is the most useful and widely adopted measurement tool in the assessment of the mechanical consequences of oral-pharyngeal dysphagia (Logemann 1991; Logemann, Roa

Pauloski et al. 1992). Videofluoroscopy has become the standard measurement tool of choice since the 1970's. Prior to this cineradiography and plain films were utilised. Frequently it is referred to as the "modified barium swallow"

(Logemann 1983). The examination is structured so that functional impairment in the swallow mechanism can be analysed. The patients are asked to swallow various standard volumes and consistencies of a contrast material. The image is framed in a lateral position such that it includes oropharyruc, palate, proximal oesophagus & upper airway. The images are recorded onto videotape for future review, in various speeds, thus enabling detection of a swallowing

dysfunction. This technique can quantify specific measures of oral-pharyngeal

dysfunction from which the clinician or therapist derives an impression of overall dysphagia severity (Logemann 1983). However, this tool cannot quantify longitudinal changes in overall swallow efficiency, functional disability, or symptomatic response to therapy and therefore has some limitations in providing an overall measure of the patients' level of dysfunction

(Ott, Hodge et al. 1996).

1.3,1.2 Videomanometry

This enhances the features of the individual measurement tools of both videofluoroscopy and manometry. By utilising a timing device which synchronously imprints a time code on both recording media, the pressures generated by specific pharyngeal and cricopharyngeal motor functions can be correlated and related to the bolus trajectory. This interaction being the movement of contrast with the manometric pressures recorded during the swallow. This technique can enable a variety of subcategories of swallow dysfunction to be detected. For example failed upper oesophageal sphincter

(UOS) relaxation can be differentiated from inadequate opening by this combined modality (Dantas, Cook et al. 1990; All, Wallace et al. 1996; Williams,

Wallace et al. 1997). While this technique is proving to be a useful research tool its place in the clinical assessment of the dysphagic patients remains to be clearly established (Cook and Kahrilas 1999). 1.3.1.3 Scintigraphy

This is a less utilised research tool which may become more popular, as it has

the advantage of being able to quantify accurately the amount of residual which

remains in either the oral or pharyngeal cavities after the swallow (Cook,

Weltman et al. 1994). Measurements of timing within the phases of the swallow have also been attempted, although it is recognised that these measurements

are somewhat different to the tinning during videofluoroscopy (Hamlet, Muz et al. 1989; Cook, Weltman et al. 1994). The technique shows promise but has not been adopted widely in the clinical workup of the dysphagic patient.

1.3.2 Measures of symptom severity

1.3.2.1 Clinician's assessment - history and examination

The majority of patients, except those with silent aspiration, will seek medical advice regarding swallowing difficulties. Dysphagia is rarely psychogenic and symptoms should always be investigated fully to identify an underlying orgai\ic cause. In one study of 23 patients labelled as psychogenic dysphagic subsequently assessed in a tertiary referral centre, 15 (65%) were found to have a physical explanation for their symptoms (Ravich, Wilson et al. 1989).

A careful history taken by the physician or therapist should lead to a diagnosis in up to 85% of cases (Duranceau, Lafontaine et al. 1987). The perception of patients in defining and describing the symptoms is highly accurate. The report of phenomena such as retention of food in the mouth, inability to control the bolus in the mouth, delayed or inability to initiate a swallow or pharyngeal response and aspiration are almost universally recognised by patients and reported as such. It is reported that patient's identification and localisation of symptoms was up to 99% accurate when compared to radiographic images

(Kahrilas 1989). In assessing the patient's history previous neurogenic, myogenic, movement disorders and a drug history must be considered.

A physical examination is necessary to detect clues to possible imderlying neuromyogenic disorders and to assess weight and nutritional status, which are important parameters of dysphagia severity. Neurological function, muscle wasting, palpation of the head and neck, inspection of dentition, and signs of prior surgery or radiotherapy are important (Cook and Kahrilas 1999).

The lack of a reliable and validated "gold standard" to measure the severity of oral-pharyngeal dysphagia, has lead to the adoption in most cases of the next best option, this being the clinician's impression based on experience and supported by investigations such as videoradiography. The validity of such an approach is difficult to establish. Currently the diagnosis and management of oral-pharyngeal dysphagia is an inexact science but the clinician's assessment is generally accepted as current best practice by reasonable evidence and the weight of opinion (Cook and Kahrilas 1999). 1.3.2.2 Symptom questionnaires and inventories

The assignment of numerical indices or scores to a subjective set of characteristics, in which the observer or individual determines the size of such an index, cannot be considered complete until the tool has been validated. The validation of any tool of measurement involves the use of a series of reliability and validity assessments. These assessments establish the amount of objectivity that can be associated with the tool (McDowell and Newell 1996). To be of value a measurement tool in the form of a questionnaire or inventory must adhere to minimal principles of questionnaire/inventory design (DeVellis

1991). In assessing the current literature of questionnaires or inventories measuring aspects of dysphagia several features were used as parameters in a critical evaluation. These features included: the main focus of the paper (ie:

What is being measured?), comparison to control group, test-retest reliability, validity testing of any sort, and comparison to an independent objective or

"gold standard"(Table 1).

A MEDLINE search from Jan 1966 to June 1999 revealed 25 papers utilising some form of swallow questionnaire in humans. Of the 25 papers, 17 related specifically to oropharyngeal dysphagia (Fleming and Weaver 1987; Sonies,

Parent et al. 1988; Layne, Losinski et al. 1989; Lindgren and Janzon 1991;

Tibbling and Gustafsson 1991; Nathadwarawala, Nicklin et al. 1992; DePippo,

Holas et al. 1994; Nathadwarawala, McGroary et al. 1994; Wintzen, Badrising et al. 1994; Deary, Wilson et al. 1995; Gustafsson and Theorell 1995; Bartolome,

Prosiegel et al. 1997; Litvan, Narahari et al. 1997; Young and Durant-Jones 1997; 9 Hinds and Wilds 1998; Murry, Madasu et al. 1998; O'Neil, Purdy et al. 1999). Of the 17 studies, when critically evaluated using the above criteria, none of the existing questiormaires had all of the desirable features. Only 9 of the papers attempted formal validation (Sonies, Parent et al. 1988; DePippo, Holas et al. 1994; Wintzen, Badrising et al. 1994; Deary, Wilson et al. 1995; Litvan, Narahari et al. 1997; Hinds and Wilds 1998; Murry, Madasu et al. 1998; O'Neil, Purdy et al. 1999). Four of these 17 were epidemiological studies, which simply utilised a questionnaire as a screening device to detect dysphagia in the general population (Sonies, Parent et al. 1988; Lindgren and Janzon 1991; Tibbling and Gustafsson 1991; Gustafsson and Theorell 1995). Although there was no evidence presented in these papers to indicate that the questionnaires utilised had been previously validated, therefore even on the basis that they were only used for detection of dysphagia, there objectivity remains unknown.

A comparison with a ''gold standard" was the focus of validation attempts in some of these previous studies. "Gold standards" selected by the investigators included: interventions by speech and language pathologists, barium swallows, ultrasound, and comparison with other inventories (eg: hospital anxiety and depression scales, head and neck QOL) (Sonies, Parent et al. 1988; DePippo, Holas et al. 1994; Wintzen, Badrising et al. 1994; Deary, Wilson et al. 1995; Litvan, Narahari et al. 1997; Hinds and Wilds 1998; Murry, Madasu et al. 1998). Other aspects of validity that were tested include: comparison to controls (Wintzen, Badrising et al. 1994; Litvan, Narahari et al. 1997; Murry, Madasu et al. 1998); test-retest reliability (Gustafsson and Theorell 1995; O'Neil, Purdy et al. 1999) and validity assessment (Deary, Wilson et al. 1995; Litvan,

Narahari et al. 1997; Hinds and Wilds 1998).

Several of the papers reviewed applied questionnaires to detect the presence of dysphagia within a specific disease population, but did not quantify severity

(Nathadwarawala, Nicklin et al. 1992; DePippo, Holas et al. 1994;

Nathadwarawala, McGroary et al. 1994; Wintzen, Badrising et al. 1994; Litvan,

Narahari et al. 1997; Young and Durant-Jones 1997; Hinds and Wilds 1998).

Three studies simply utilised a questionnaire to catalogue patient characteristics without attempting to quantify dysphagia severity (Fleming and Weaver 1987;

Layne, Losinski et al. 1989; DePippo, Holas et al. 1994). Consequently of the 17 papers critically reviewed, only 3 utilised or developed a questionnaire that was specifically structured to assess dysphagia severity or change in severity over time (Deary, Wilson et al. 1995; Murry, Madasu et al. 1998; O'Neil, Purdy et al.

1999).

Murry et al aimed to assess both the quality of life and swallowing in patients with advanced head and neck cancer following treatment with intra-arterial chemo-radiation (Murry, Madasu et al. 1998). The swallow questionnaire utilised within the study focused on symptoms of dysphagia, overall swallowing status, and weight. The primary aim of the research was not to assess the questionnaire, but use the information collected to assess the patient's response to chemo-radiation. In light of this aim, the authors have not formally attempted to validate their swallow questionnaire, although comparisons were made with the previously validated "Head and Neck Radiotherapy

Questionnaire Quality of Life (HNRQ QOL)" to aid in the assessment of the questionnaires' potential validity. The patient's were divided into three groups dependent upon the site of their disease: hypopharynx, larynx, and oropharynx. Both questionnaires were also given to a group of age-matched controls, although this data is not referred to in the paper. Overall the swallow questionnaire mirrored the results of the HNRQ QOL in all disease groups, with the hypopharyngeal patients showing the most rapid recovery in QOL and swallowing. Neither the HNRQ QOL nor the swallow questionnaire correlated with weight loss. The concluding remarks recommended that the psychosocial recovery should be a consideration in the overall patient care plan.

Deary et al specifically designed a questionnaire to assess the symptom of globus pharyngis (Deary, Wilson et al. 1995). Globus is a sensory motor abnormality, which is characterised by a sensation of non-painful lump or fullness in the throat and, in the majority, is not associated with true dysphagia.

Interestingly, this sensation is usually alleviated by eating (Thompson and

Heaton 1982; Cook 1991). The questionnaire developed was labelled "Glasgow

Edinburgh Throat Scale (GETS)". Validation was undertaken in two ways, the first being "principal components analysis" which serves to sub-categorise individual questions within the scale into groups. The groups represent clusters of questions that essentially measure different facets of the one feature.

This analysis revealed 3 factors, nominally labelled dysphagia, globus, and pain/swelling. The robustness of the questiormaire was further tested in 12 comparing it to a previously validated anxiety scale (Hospital Anxiety and

Depression Scale (HADS)). The aim in conducting this assessment was to establish the relationship between the throat symptoms and concurrent feelings of anxiety and depression. The results of the paper have shown that the scale developed had high validity when assessed by factor analysis. The presence of three factors indicates that the questionnaire has satisfactorily identified three areas of symptom concern as described by the patient. The authors advocate that the GETS enables users to establish a quick, quantitative estimate of the presence of and severity of common throat complaints, especially globus. Being self-report, it also may add important clinical information to clinical history. It is intended that the GETS be further validated to include those patients with oral-pharyngeal dysphagia and phar)mgitis. Although the appropriateness of this particular tool to oropharyngeal dysphagia is doubtful as the inventory was applied in an Ear, Nose and Throat (ENT) clinic to patients presenting with the sensory disturbance of globus. Only 30% reported of the patient's in the study had actual dysphagia and only 3 of the 10 component questions in the

GETS were related to dysphagia per se.

The only published work which aimed to develop a scale designed specifically to measure severity of dysphagia which might be applicable to the majority of oral-pharyngeal dysphagia patients, is that of O'Neil et al (O'Neil, Purdy et al.

1999). The tool identified as the "Dysphagia Outcome and Severity

Scale"(C>OSS) endeavours to include a wide variety of neuromyogenic aetiologies. The severity scale however, is primarily based upon the results of a 13 modified barium swallow (MBS). The overall aim in formulating the scale was to achieve consistency among health professionals who were assessing the evidence provided by the MBS and other means of data collection (eg: chart review, history as described by the patient, and clinical bedside evaluation). In developing the scale the authors expressed a need to incorporate factors beyond the results of the MBS. In examining the literature the authors identified three areas that they considered important for their scale to be useful across a wide variety of patients and within one patient over time. They established three factors that were candidates to be incorporated into the scale: level of independence, level of nutrition, and diet modifications. Limited validity analysis was undertaken after the finalisation of the scale. However in refining the final scale, the development was based on the following steps: identification of the need for a tool; characterisation of elements within the tool; review of previously collected data to establish link between MBS and nutritional status; and pilot trial. The authors of the study did not mention any prior validation that may have occurred in refining the scale within the current paper.

Following finalisation, the scale was assessed for intra and inter-rater reliability.

This revealed a high correlation between (82-100%) and within raters (86-100%).

This was compared in a gross fashion to the reliabilities achieved on the functional independence measure (FIM) scale. The concluding remarks indicate that the DOSS is advocated as a holistic and objective tool that could be utilised to measure the natural history and outcome of dysphagia, with future usage in determining efficacy of treatment. The forum of current published literature in oral-pharyngeal dysphagia has not satisfied the basic criteria of a measurement tool, in that they have failed to provide evidence of reliability and validity. The inherent lack of a specific tool that adheres to the principles of questionnaire/inventory development is one of the main impediments in evaluating treatments for oral-pharyngeal dysphagia.

Of the three papers Peary, Wilson et al. 1995; Murry, Madasu et al. 1998;

O'Neil, Purdy et al. 1999) which attempt to validate the usage of their tools, two do not provide conclusive evidence that the process was completed or successful (Murry, Madasu et al. 1998; O'Neil, Purdy et al. 1999). Deary et al

(1995) however provided a high level of evidence that the tool developed and evaluated was valid and reliable in the population of patients with globus. The usage of such a tool in oral-pharyngeal dysphagia would not to be feasible in the current format, as only 30% of the questions had any relevance to dysphagia.

Essentially any tool developed for measuring a parameter must be able to provide consistent results on repeated measures and reflect its intended theoretical concept (Carmines and Zeller 1979). Reliability and validity testing establish the objectivity of a measurement tool. This enables users of these tools to be confident in the tool's ability to captiire the information for which it was designed. The principles of objective evaluation of tools that collect subjective data have been used in behavioural science research for a number of years.

Extensive research in the behavioural sciences has enabled researchers to define various aspects of both reliability and validity. This refinement has lead to an 15 understanding that it is not satisfactory to establish one facet of validity, but various aspects need to be evaluated to ensure that a tool is truly valid and therefore users can be confident with the results of the tool.

2.4 Expectations of a new swallow inventory

In the development of this inventory we sought primarily to evaluate the symptoms for which most patients seek medical advice. The primary ideas and constructs for the inventory were devised using the clinical judgement and experience of a clinician and a speech language pathologist. The symptoms of oral-pharyngeal are diverse and varied and do not pertain to a particular aetiology, but rather can be defined in terms of three areas: anatomical, type of dysfunction and bolus consistency. It was from this foundation that the initial questions were formulated. The robustness of symptoms can be evaluated in a limited fashion from vast amounts of data previously collated on the assessments and their relationship to symptoms. It is well known that aspiration per se, does not correlate with actual findings on videofluoroscopy.

In the case of aspiration, it is the imder-reporting by patients that makes this an imreliable or less robust symptom.

The most robust of measures are those with unequivocal outcomes; for example death, pneumonia, nutritional deficiency. The assessment of the efficacy of treatment using these end-points can present a bias approach that fails to encompass the situation of patients. Many patients may never reach an unequivocal outcome or in most cases require extensive medical intervention prior to this to preverit an unequivocal outcome. The costs of maintaining and executing treatments where the efficacy is not high, but in most cases backed by reasonable evidence of biological plausibility and weight of clinical opinion are enormous. Attempts must be made to develop tools that although are not based on traditional unequivocal outcomes, do accurately assess the current position of the dysphagic patient.

In consideration of symptoms of dysphagia, it was reasonable to assume that those symptoms that are most closely related to mechanical disturbances would tend to be robust. The decision about the robustness of a symptom needs to be objectively assessed in any new tool by evaluating the reliability and validity of the component within the tool. Whilst the quality of life of the patient is an important factor in the schema of treating the whole body, it was felt that the initial step taken should be that of having an effective tool to which quality of life may be added as an appendix in future work. As evident by the current hterature, this has not been achieved previously. It could be hypothesised that quality of life is closely linked with severity of dysfunction in the case of swallowing, as the results from Murry et al indicate a close relationship with the results of their swallow questionnaire (Murry, Madasu et al. 1998).

In developing the current inventory consideration was given to create a tool which fulfilled the basic criteria of "what clinicians in the field of dysphagic research considered relevant for the assessment of dysphagia severity", therefore ultimately aiming to reduce the morbidity and mortality. The 17 primary intentior\ was to develop a tool that yielded an index or measure of severity. This index should be sensitive to the function of as many of the mechanical components of swallowing as possible and therefore sensitive in recording change with respect to effectiveness of the therapy. Responsiveness to currently available treatments is extremely variable and unpredictable. In many instances the response being dependent upon on several factors including the imderlying cause of the dysphagia, the severity and nature of the mechanical dysfunction, the degree of cognitive dysfunction, and the prognosis of the imderlying disease (Cook and Kahrilas 1999).

1.4.1 Summary

The primary objective of this work was to develop a measurement tool, derived from symptoms, the basis for which we believe lay within recognised underlying mechanisms of dysfunction. This was based on a reasonable, but unproven belief, that disorders within each of these categories of dysfunction might impact on serious and highly relevant clinical end-points such as death, pneumonia and nutritional deficiency. Hence, the clusters of questions derived were based 6n the belief that they were representative of specific underlying pathophysiological mechanisms. An alternative approach might have been to derive questions from a sample of patient opinions about which particular aspects of their dysphagia bother them most or which might influence their quality of life. Again, although unproven, we felt such an index, while representing a highly relevant measure of patient dissatisfaction, might be less likely to correlate with arguably more important clinical end-points such as 18 death, pneumonia and nutritional deficiency. A tool capable of accurately measuring longitudinal change in symptom severity is also highly desirable if, as intended, the inventory is to be applied to outcome research. 1.5 Mechanism of oral-pharyngeal dysphagia and relevance to swallow inventory

design

1.5.1 Introduction

The complex process that constitutes a swallow involves a sequence of events

encompassing various anatomical regions and physiological functions. As oral- pharyngeal aetiology plays a small role in the determining the symptom profile of patients, the swallow inventory was designed in the first instance to be used in patients with neuromyogenic disease. The natural history of stroke-related dysphagia, the commonest neuromyogenic cause, is such that spontaneous improvement can be expected in up to 50% of patients (Winstein 1981; Tibbling and Gustafsson 1991). The swallow inventory is of most value in those patients who do not experience a spontaneous improvement, but have long-term deficit

(30-40%).

1.5.2 Mechanism of Normal & Abnormal Swallow

When functioning normally the swallow is a well-ordered group of events, which involve both voluntary and involuntary sequences. The smooth effortless sequence almost belies the complex nature of the process (Dodds

1989).

The sensory cues that elicit the pharyngeal swallow are not precisely known but, experimentally, entry of fluid or food into the hypopharyruc in the sensory receptive field of the superior laryngeal nerve seems to be the crucial stimulus

(Aviv, Kim et al. 1996; Pouderoux, Logemann et al. 1996; Dua, Ren et al. 1997).

Swallowing can also be initiated solely by volitional effort, providing that there is food or fluid within the oral cavity. Therefore, the required afferent signal for initiation of the swallow response is comprised of a mixture of both peripheral sensory input from oropharyngeal afferents and superimposed control from higher nervous system centres (Martin and Sessle 1993). Neither element of the afferent signal is capable of initiating the swallow without some contribution from the complementary element, as evidenced by the inability to swallow during sleep in which the higher centres are disconnected (Issa 1994), or with deep anaesthesia to the oral cavity (Doty 1968), in which case peripheral afferents are deactivated. However, the relative contribution of the two afferent elements responsible for initiating a particular swallow varies with the circumstance.

Once initiated, the pharyngeal swallow results in a transient reconfiguration of the pharynx. Mechanistically, this involves several closely coordinated actions:

1) elevation and retraction of the soft palate with closure of the nasopharynx, 2)

UOS opening, 3) laryngeal closure at the level of the laryngeal vestibule, 4) tongue loading (ramping), 5) tongue pulsion, and 6) phar5T:\geal clearance.

These events occur in close synchrony. A fundamental aspect of deglutitive pharyngeal reconfiguration is in transforming the oropharynx from a respiratory to a swallow pathway by opening the inlet to the oesophagus and sealing the inlet to the larynx. Lar5mgeal vestibule closure, and hence airway protection during swallow, are achieved by larjmgeal elevation and anterior tilting of the arytenoid cartilages against the base of the epiglottis (Logemann,

Kahrilas et al. 1992). Laryngeal elevation is quantifiable fluoroscopically by 21 vertical movement of the tracheal air column and its persistence above the

critical height necessary to achieve closure of the laryngeal vestibule.

Videofluoroscopic studies done concurrently with manometry have shown that

UOS relaxation occurs during swallow-associated laryngeal elevation and that

relaxation precedes opening of the sphincter by about 0.1 sec (Kahrilas, Dodds

et al. 1988). After relaxation, UOS opening results from anterior traction caused

by contraction of the supra and infrahyoid musculature, evident

fluoroscopically by anterior hyoid movement (Cook, Dodds et al. 1989; Jacob,

Kahrilas et al. 1989).

The functional elements of the pharyngeal swallow responsible for bolus

transport are the propulsive phase of tongue action and the propagated

contraction of the pharyngeal constrictors. Bolus propulsion relies heavily on

deglutitive tongue action. The geometric complexity of the tongue requires

biplane or even three dimensional imaging to visualise its function which

largely consists of deforming the central groove to create a bolus cavity while

maintaining a seal peripherally and then rapidly expelling the contents of that bolus cavity into the opened oesophageal inlet (Shawker, Sonies et al. 1984;

Wein, Bockler et al. 1991; Kahrilas, Lin et al. 1993; Kahrilas, Lin et al. 1996). In

coordination with the expulsion phase of tongue activity, the propagated

pharyngeal contraction progresses from the superior to middle to inferior

pharyngeal constrictors at about 15 cm/sec. This results in the stripping of

almost all residue from the pharynx and hypopharynx except for trace amounts which may be left in the valleculae or pyriform sinuses, thereby minimising the

22 chance that aspiration will occur with resumed respiration (Kahrilas, Logemann

et al. 1992). The overall pattern of the pharyngeal swallow is similar among bolus volumes except the period of reconfiguration into an alimentary conduit

is prolonged while a larger bolus chamber is formed by accentuated tongue

deformation (Kahrilas, Lm et al. 1996). Following bolus clearance, the process

of pharyngeal reconfiguration is reversed allowing the resumption of

respiration; the entire pharyngeal swallow occurs in about 1 second.

Due to the many individual components involved in the swallowing process,

the capacity for several mechanisms to be affected during disease is highly

likely. As several aetiologies produce similar dysfunctions, it is more

appropriate to classify dysfunction by mecharustic components (Table 2).

1.5.3 Disease Aetiology

The aetiology of oral-pharyngeal dysphagia is varied, as swallowing involves

the interaction of major body systems. The natural division in anatomical areas

enables the aetiology of dysphagia to be classified into three main anatomical

areas: neurogenic (including central and peripheral nervous systems),

myogenic and neuromuscular interface. A fourth division, though not

anatomically area based, is structural disorders (intrinsic and

extrinsic)(Duranceau, Lafontaine et al. 1987; Cook 1991).

Oropharyngeal dysphagia can be the presenting or accompanying

manifestation of one of a large number of systemic diseases (Table 3), a 23 multidisciplinary strategy is frequently necessary to achieve optimal management. Neurological impairment may be accompanied by the patient's limited cogrutive and physical competence, which can compromise their ability to understand and cooperate with the process of investigation and therapy.

The general aim of management is to identify and treat an underlying primary disease where possible and then try to compensate or circumvent the specific mechanical disturbances responsible for the dysphagia (Cook and Kahrilas

1999). A single strategy is not appropriate for all cases nor indeed for dysphagia caused by a single disease process because the mechanical disturbances are generally neither homogeneous nor specific for a particular disease. In general logical, generic management principles are widely applicable while specific surgical or rehabilitative swallow therapies should be tailored to the individual case.

1.6 Treatments, outcomes and responsiveness to therapy

1.6.1 Importance of natural history

Patients who experience dysphagia after a head injury, stroke or spinal cord injury may experience recovery in the first 3 months. However, this leads to many others who are afflicted with a dysfunction that may never fully recover.

These patients are disadvantaged, as there is lack of tools to measure the efficacy of the current regime of treatments. The current treatments being offered to patients are: 1) stirgery; 2) swallow behaviour modification; or 3) dietary alteration. Unfortunately many of these treatments, although biologically sound, have never been examined in the light of a randomised controlled trial (RCT) which provides the strongest evidence of efficacy.

Cricopharyngeal myotomy is the most common surgical avenue in treating oropharyngeal dysphagia when the cause is structural defect (eg: diverticulum

(Zenker's), post-cricoid stenosis, webs). This reduces upper oesophageal sphincter (UOS) tone and enables resistance to flow to be decreased. The procedure has yielded good-excellent results in 80-100% of cases treated. The use of myotomy in neurogenic dysphagia is variable (50-100%, average 63%; mortality average 1.8%). In contrast to the structural aetiologies, those neurogenic patients undergoing myotomy is based on uncontrolled and methodologically weak evidence (Cook and Kahrilas 1999).

Behavioural modification and dietary alteration form the basis of treatment for those patients who are not suitable for surgery. The evidence for such treatments is variable, with the majority being inconclusive. Only two RCT exist which examine the efficacy of these two modalities of treatment. Both papers examined therapies with patients effected by stroke. DePippo et al

(1994) randomised patients stroke patient with mild dysphagia to one of three treatment arms; including dietary modification and swallowing therapy, due to ethical constraints there was no untreated control group. Based on the outcome measures of pneumonia, death, caloric-nitrogen deficiency and dehydration, only 15% of the total population developed any of these complications within

12 months, with no significant difference between groups. It is impossible to 25 conclude that this trial lacks efficacy, as there is a possibility of a type II error due to the lack of patient numbers and an untreated control group. In the other

RCT, a comparison was made between two types of dietary modifications in a group of patients who had a stroke. The control group received a traditional puree and non-altered fluids, whilst those in the treatment arm received a diet consisting of soft foods with thickened fluids. The outcome measure was the incidence of aspiration pneumonia, which was reduced by 80% in the treatment group after 6 months. The dropout rate was not accounted, those being patient who died or commenced non-oral feeding, and therefore these may have biased the results (Groher 1987). The remainder of other studies showed reasonable evidence, but lacked efficacy. However this lack of efficacy remains to be proven conclusively, and the weight of consistent low-grade evidence suggest that swallow therapy should be used. The definitive outcome measures such as death and pneumonia are extremely important, however it is crucial to determine whether those patients who have not reached the outcomes have a response to these treatments.

1.7 Summary

In this ageing population, oral-pharyngeal dysphagia is a costly problem that is going to enlarge in the years to come. The patient populations effected by oral- pharyngeal dysphagia are varied and wide with no single aetiology being responsible. The development of effective treatments and more importandy tools to test the efficacy are more significant in these times where evidence based medicine determines the spending and value of the health dollar.

Currently there is no reliable and validated measurement tool to assess the symptomatic severity of oral-pharyngeal dysphagia nor changes in response to therapy. Chapter 2 Methods

2.1 Introduction -fundamentals of inventory or questionnaire design

Visual analogue scales (VAS) have been used to measure responses in

questionnaire in the 1920's. The justification for its development and use comes from research that established peoples responses to subjective phenomena are far more consistent and accurate, when made on an a ratio based scale, rather than to nominal numerical estimates. The illustrative concept of a VAS is an easily explained scale with the majority of persons using the scale able to grasp its meaning in a short span of time. The VAS provides a scale that can cope with the wide variety of answers possible, unlike an categorical scale that is limited by the categories provided (McDowell and Newell 1996).

The VAS has been used in a variety of forms, however the 100mm line in the horizontal format has been shown to produce a more uniform distribution of scores (Scott and Huskisson 1976). The length of the VAS is optimal at 100mm, as lines shorter than these tend to produce a greater error of variance. The perpendicular end-points at the end of the line influence users to limit marking beyond this. This visual barrier provides an important limitation and signifies the limit of the score. A further development of the standard VAS has been to include text descriptor scale above the line, the disadvantage of this is that users may be influenced to position their marks at the site of the text. Therefore creating an erroneous distribution of scores. This can be alleviated by evenly placing all text along the entire length of the line. Depending upon the population that the scale was being designed for this could create confusion and vastly influence mark position (Wewers and Lowe 1990; Cline, Herman et al. 1992; Mottola 1993).

Despite the obvious advantages of using the VAS to establish scores for a variety of subjective phenomena eg: convenient, ease of use and rapid administration. There have been reported difficulties using the VAS. These limitations mainly stem from users imable to conceptually grasp the method.

This can be overcome by the careful explanation, preferably by example in verbal and written capacities and the repeated usage of the VAS, as in longitudinal studies. This problem was further noted in using the VAS in an elderly population. Other concerns arise from the re-production of the questionnaire or inventory. Researchers need to be aware that the line is subject to distortion when copied on low-grade photocopiers. Similarly the angle at which the user views the line (eg: completion of a scale in bed), may lead to perceived variation in the length of the line. A consideration must be given to the cognitive and physical limitations of the users, as the VAS may not provide an accurate assessment if these are significantly affected (Wewers and Lowe

1990; Cline, Herman et al. 1992; Mottola 1993). In using the VAS as part of the development of a new measurement, it is important to recognise these limitations exist and explore their influence through various reliability and validity testing of the tool. In assessing any new tool of measurement, researchers must evaluate both the reliability and validity of the tool. This provides objective evidence that the tool is useful and is measuring the factors for which it was designed. There are several facets in evaluating new tools of measurement. These are based on the following metric principles that evaluate the quality of an instrument: 1) to what extent does the tool measure what it is intended to measure and 2) in achieving this goal, how much error is there (McDowell and Newell 1996). In assessing the objective characteristics of the swallow inventory evaluations of reliability (using test-retest method and four aspects of validity (face, content, construct and discriminant) were made (Bombardier and Tugwell 1982).

2.2 The Inventory

The primary ideas and constructs were derived, primarily from clinical judgement, by an experienced clinician and a speech language pathologist. A larger ''focus group" approach was not utilised in the formulation of the questionnaire in favour of its distribution to 32 internationally recognised authorities in the field of dysphagia management and research in order to estabhsh its face vaUdity (Anastasia 1982; Bombardier and Tugwell 1982).

A range of questions were considered which addressed symptoms referable to combinations of three broad variables: i) anatomical region; ii) type of dysfunction; and iii) swallowed bolus consistency. The anatomical regions included; a) oral cavity (including the lips, tongue and palatoglossal closure mechanism); b) glottis; and c) pharynx. Within each anatomical region, the potential functional disturbances were considered as follows: a) oral cavity: oral bolus containnaent, such as labial spill and pre-swallow spill over tongue base; lingual manipulation of oral bolus; difficulty with swallow initiation; poor oral bolus clearance, b) glottis: aspiration symptoms, c) pharynx: deglutitive velopharyngeal incompetence; incomplete pharyngeal clearance. The influence of variations in swallowed bolus consistency was factored in by the duplication of some questions which were common in respect of region and dysfunction but differed in respect of swallowed bolus consistency (eg thin liquids, liquids, puree consistency and solids). The variable consistency was included because of the recogrused variation in difficulty encountered among differing consistencies and because stratification by consistency might provide the therapist with a measure of longitudinal change in response to particular feeding strategies (Groher 1987; Bisch, Logemann et al. 1994). Two questions were included which aimed to measure the overall severity of dysphagia and to gauge the impact of it on patient's quality of life.

From these concepts we formulated a prototype 19 question, self-report inventory with a maximum possible total score of 1708 with a visual analogue scale appeared immediately beneath all but two questions (Q12, 13) (Table 4).

Each visual analogue scale was a non-graduated, horizontal 100mm line anchored at each end by extreme statements representing normal function to the left and extreme dysfunction to the right (eg: does not occur < >occurs all the time; no difficulty < > extreme difficulty). There were no intervening qualifiers implying a continuum of severity and unlimited range of responses.

31 Two questions (Q12,13) yielded single integer scores from 0-5 and 0-3 based on consumption times for an ''average meal" and "a scoop of ice cream" respectively. An explanation of the aims and mode of completion together with an example, were printed at the head of the inventory. Patients were instructed to mark a single 'X' across the horizontal visual analogue scale at the point which they felt best represented the severity of the particular dysfunction, thus yielding a score of 0 - 100 for each corresponding to distance in mm from the origin of the visual analogue scale. The written instructions were also delivered verbally by one investigator who was available during the completion of the inventory to address conceptual problems if necessary. No attempt was made to guide the patient as to where on the visual analogue scale they should make their mark.

2.3 Patients and Controls

The inventory was administered to two groups of patients with oropharyngeal dysphagia and a group of comparably aged controls without dysphagia.

Patient group 1 comprised 48 patients who had confirmed oral-pharyngeal dysphagia with neuromyogenic aetiology and in whom the severity of dysphagia had been stable for at least 3 months. Patients were excluded from this group if they had: a coexistent structural disorder of the cricopharyngeus or pharynx (eg: tumour, Zenker's diverticulum, circumferential cricopharyngeal stenosis, cervical web); prior head and neck surgery that might have influenced swallow function; inadequate cognitive or language skills to comprehend study requirements. In this group we evaluated the reliability and validity (face, content, construct) of the inventory. Patient group 2 comprised 11 patients with dysphagia due to a Zenker's diverticulum (6M:5F; mean age 66 ±

13 yrs; range 36 - 84yrs) who had been referred for surgery. The purpose of this particular group was to evaluate the sensitivity of the inventory in detecting a change in symptom severity in response to one form of treatment (surgery) recognised to be highly effective (discriminant validity).

In order to assess predictive validity, the questionnaire was administered to 19 controls without dysphagia who were selected at random from the urology and gynaecology wards and who had not experienced dysphagia at any time in the previous 12 months. An individual who was bUnded to the status of the patients scored inventories.

2,4 Data Collection

All patients in Group 1, underwent history and physical examination and a videofluoroscopic examination of swallowing. Additional investigations were performed as required depending upon the suspected underlying aetiology of the neuromyogeruc disorder. Each patient was then assigned a global assessment score that was based upon all the above information and without knowledge of the subsequent inventory scores. The purpose of the global assessment score was to use it as an independent comparator against which the inventory total score was correlated to estimate its construct validity. The 33 global assessment score ranged from 0 - 10, (0 = no dysfunction; 10 = severe dysphagia, total dependence upon non-oral feeding and unable to swallow saliva). The dysphagia inventory was administered to all patients in Group 1 at the time of clinical assessment and again to a subgroup of 18 of these 48 patients one to two weeks later. In patients within Group 2, inventory scores and global assessment scores were obtained at initial consultation and again 6 weeks after surgery for the Zenker's diverticulum. Those completing the inventory on a second occasion did so without access to their responses to the first inventory.

2.5 Analytical Methods

If a patient failed to complete at least 15 of 19 questions, the inventory was excluded from further analysis. Where a patient omitted between one and four questions an estimated score for each omitted question was calculated based on the total score divided by the total possible score for the questions answered.

Estimated scores for individual questions were only used for factor analysis calculations, which requires a complete data set for each patient. For each question within the inventory and for the total score we evaluated: a) test-retest reliability; b) four types of validity (face, content, construct, and discriminant)

(Bombardier and Tugwell 1982). 2.5.1 Test-retest Reliability

The test-retest reliability of a clinical measurement tool is a measure of its ability to yield consistent scores over time given that the clinical status of the patient remains stable (Anastasia 1982; Bombardier and Tugwell 1982). The evaluation of reliability allows us to gauge the amount of chance variation among the scores within the inventory. This variation will always be present within the inventory, but we expect the variation to be similar in magnitude on each testing occasion (Carmines and Zeller 1979).

We evaluated the test-retest reliability of individual question scores and the inventory total score between baseline testing and retesting one to two weeks later in the first 18 consecutive patients within Group 1. This subset was tested as it was determined (and subsequently confirmed) to be of sufficient size to yield adequate statistical power and, if inadequate reliability could not be proven early, the questionnaire would require revision. For each question and the total score, the retest score was subtracted from the test score, retaining the positive or negative direction of the sign, to yield a delta value. The data was then normalised, by dividing the delta scores by the highest score attained by the patient for that particular question, before calculating the 95% confidence interval for the group mean delta value. In assessing test-retest reliability, data were normalised because, a priori, it was equally likely that scores would increase or decrease. In this component of analysis, the magnitude, not the polarity of change, was relevant. In order to render a positive delta of equal absolute value comparable to that of a negative delta of the same absolute 35 value, we expressed the delta as a proportion of the highest value in this

equation.

2.5.2 Validity Measures

The validity of an inventory is defined as its ability to evaluate the attributes for

which it was designed to measure (Anastasia 1982). In this study we

determined measures of four types of validity: face, content, construct and

discriminant (Bombardier and Tugwell 1982).

2.5.2.1 Face Validity

Face validity is defined as the extent to which the inventory has the "visual appearance'' that it is evaluating the functions for which it was designed to measure (Anastasia 1982). The individual components of an inventory must appear to be weighted in a sensible and acceptable fashion as to assess the disease activity for which the tool was designed to measure (Bombardier and

Tugwell 1982). We assessed face validity by distributing the inventory to 32 internationally recognised authorities in the field of dysphagia management and research. Each was asked to rate the clinical relevance of each question to the assessment of oral-pharyngeal dysphagia severity on a scale of low, moderate, or high. Frequency distributions for each question and each rating were determined. We arbitrarily defined a question as having poor face validity if more than 30% of the experts rated the question as having low relevance. We also assessed patient compliance with inventory completion by quantifying the proportion of each question not completed by the patients. Poor compliance with an individual question was deemed present if more than 5% of. patients failed to answer the question.

2.5.2.2 Content Validity Content validity seeks to review whether the relative importance and choice of items within the inventory is appropriate for the intended use of the measurement tool (Carmines and Zeller 1979; Bombardier and Tugwell 1982). Factor analysis, an accepted method of detecting any inter-relationships which exist among component questions within the inventory, was used to evaluate statistically the content validity of the inventory among patients in Group 1 (DeVellis 1991; McDowell and Newell 1996). Using the principal components method (orthotran/varimax rotation), we performed factor analysis on the data yielding an oblique solution (Rummel 1970). Each factor derived from such analysis is assigned an appropriate title which best reflects the content of the factor. All data analysis and statistical tests were performed on a computer (Macintosh Esi, Apple, Cuppertino, Califorrua) and using statistical software (StatView V4.5, Abacus Concepts, California). The computer software adopted the "75% variance rule'' that is, it determined the number of factors required in accounting for 75% of the variance of the data. The output from factor analysis is a matrix of factor loadings. It is generally accepted that a factor loading greater than 0.3 is significant (Johnson and Wichem 1992; Loehlin 1992). This figure may need to by adjusted slightly downward as the sample size and 37 number of variables increase. It may need to be adjusted upward as the number of factors increases. For the purposes of this analysis, we selected a loading ratio of 0.60 as an appropriately conservative cutoff. That is, for an individual question in the inventory to be considered a part of a particular factor, the rotated loading must be greater than 0.60 and, in addition, it must not be highly represented in any other factor. Statistical inferences on the distribution within the factor analysis matrix were made using Bartlett's Chi- square test. A statistically significant chi-square value indicates that the collection of coefficients in the correlation matrix differ from 0 and that the relationships among the coefficients are not random. As an adjunct to examination of the individual factors, factor analysis provides a communality summary. The loading attached to an individual question within the communality summary table provides a measure of the variance of each question which can be accounted for by the combination of all the factors derived by the solution which overall accounts for 75% of the total variance of the questionnaire. As a benchmark, the total variance that each question contributes should be more than 0.60.

We assessed the inventory for internal consistency to determine whether patients simply marked the VAS in the same regions of every scale; a situation which would yield high correlations between adjacent questions. Scores for each question were correlated using a Pearson's product moment correlation with scores for every other question to yield a matrix of correlation coefficients.

Inspection of this matrix demonstrated that the correlations among individual questions were few and were randomly distributed. That is, there was no pattern of correlation between adjacent questions. Hence, we did not need to incorporate further internal checks such as reversal of order or direction of the questions.

2.5.2,3 Construct Validity

Construct validity is a measure of the inventory's ability to produce consistent results that reflect the true clinical state of the patient. In other words, whether the index agrees with expected results based on our underlying hypothesis

(Bombardier and Tugwell 1982). In this case, we h)^othesised that the clinician's overall grading of dysphagia severity correlates with the swallow inventory score. The total score, the sum of imweighted component scores, was correlated with global assessment score using linear regression analysis.

2.5.2.4 Discriminant Validity

Discriminant validity is a measure of the tool's ability to distinguish clinically significant differences in therapeutic responses between patients and within patients over time and should be tested against the clinician's and/or patient's global judgement of change in severity over time (Bombardier and Tugwell

1982). Discriminant validity is important if the inventory is to be used to measure longitudinal change as a measure of efficacy of a treatment. We assessed discriminant validity in Group 2 patients with Zenker's diverticulum by applying the inventory before and 6 weeks after corrective surgery in these 39 patients. Inferences regarding the magnitude of treatment-related changes in total inventory score and in global assessment score, were made independently using Student's paired t-test and the relative magnitude of changes registered by the two techniques was also compared. All data are expressed as mean ± SD unless stated otherwise. Chapter 3 Results

3.1 Inventory demographics

Seventy three percent of the 48 patients completed all 19 questions. Three patients were excluded because they failed to complete at least 15 questions, leaving 45 patients for analysis (27M:18F; mean age 70 ± 13yrs; range:30 -

96yrs). The aetiology of dysphagia in these 45 patients is shown in Table 5. The underlying diagnoses in these patients encompassed a wide range of neurogenic and myogenic disorders; the most prevalent being stroke (31%).

Question 13, which was not answered by 11% of patients, with a further 4% indicating that they "didn't eat ice cream", was the only question deemed to have low patient compliance (Figure 1).

3.2 Reliability and Validity testing

3.2.1 Predictive validity

The mean age of non-dysphagic controls (62 ± 16yrs; range: 31 - 94yrs; 8M:11F), did not differ significantly from that of either patient group with dysphagia.

Mean total score in patient controls showed a very consistent grouping (67+ 63) which was significantly lower than the group mean score in either the neuromyogenic patient population (789 ± 413; p < 0.0001) (Figure 2) or the

Zenker's group (875 ± 385; p < 0.0001). 3.2.2 Test-retest Reliabiliti/

The mean time interval between baseline and repeat administration of the

inventory was 12 ± 2 days. For individual questions the absolute mean delta values for raw scores ranged from 0 to 9.7 (95% CI for latter -22.1 to 2.8) (Figure

3). For the normalised scores, the mean, absolute delta values ranged from

0.9% to 17.2% (95% CI for latter -42.1% to 7.6%). Confidence intervals for all

delta values straddled zero, indicating that no retest score differed significantly

from its baseline score. Mean change in the raw total score was -12 ± 108 (95%

CI: - 65.5 to 42.3) and the mean change in normalised total score was -0.5% ±

17.6% (95% CI: -9.2% to 8.2%).

3.2.3 Face validity

Twenty five of 32 (78%) authorities in the field of dysphagia responded to the face validity survey. Fourteen of 19 (74%) questions were rated as moderately or highly relevant by more than 80% of respondents (Figure 4). Six questions

(Q2,5,9,10,ll/14) were rated as highly relevant by more than 80% of respondents. Questions 2 and 5 relate to difficulty handling different bolus

consistencies; while questions 9,10 and 11 relate to bolus obstruction and

aspiration of solids and liquids. Question 14, addressed the symptom of post-

nasal regurgitation (Table 4). Three questions (Q13,17,18) were rated as having

low relevance by more than 30% of experts polled. These questions related to:

time taken to eat a scoop of ice-cream (Q13); complaints of drooling saliva from

the mouth (Q17); and perceived severity of swallowing problem today (Q18). 3.2.4 Content Validity

The factor analysis matrix with oblique solution, calculated for the 19 individual questions, yielded a four factor solution, meaning that it identified 4 factors which accounted for 75% of the total variance within the data (Table 6).

The first 2 factors accounted for 64.2% of the total variance. All questions, with the exception of Q17, contributed significantly to factor 1, which accounted for

55.5% of the total variance within the inventory. The distribution of the loadings within the matrix was highly significant (p < 0.0001; Bartlett's Chi square) indicating that the relationships which have been realised within the analysis are not random but have a defined pattern. Eighteen of 19 questions contribute significantly to the variance of the phenomenon within factor 1

("dysphagia"). Moreover, none of these 18 questions contributed to any significant degree to the variance of any of the other 3 factors. The loadings for each of the 18 questions are relatively comparable (range 0.61 to 0.87) which indicates that each question has equal influence over factor 1. Question 17

(relating to drooling) was the sole contributor to factor 2. The communality

summary is also shown in Table 3. Only question 14 had a low (ie < 0.6)

communality loading. For example, Q1 with the communality loading of 0.73

means that 73% of the variance in response to Q1 is explained by the 4 major

factors identified by the solution.

3.2.5. Construct Validity

The total inventory score showed a highly significant positive linear correlation

with global assessment score (r = 0.69; p = 0.0001) (Figure 5). 43 3,2.6 Discriminant Validity

Of the 11 patients with Zenker's diverticulum; one failed to complete 2 questions and another patient failed to complete one question. The mean total inventory score pre-operatively was 608 ± 271 which fell significantly by an average of 70% post-operatively to a mean of 182 ± 134 (p < 0.0001) (Figure 6).

Nine of 11 patients had a fall in score of greater than 60%. The corresponding post-operative reduction in global assessment score averaged 89% with all but one patient having a post-operative fall of 75% or greater (Figure 6).

3.3 Performance of the Modified Inventory

On the basis of the above data, two questions were removed from the prototj^e inventory to yield a modified, 17 question inventory with a maximum possible score of 1700, upon which the above reliability and validity analysis was repeated. Of the 3 questions which had been judged by expert opinion to have low relevance to the assessment of dysphagia severity (Q13,17,18). Question

13 also had a low patient compliance and Q17 did not contribute significantly to the variance of the single major factor ("dysphagia") which emerged from the factor analysis matrix. That is, responses to Q17 were not a good measure of dysphagia severity. Accordingly, questions 13 and 17 were deleted from the inventory. Although question 18, rating severity of ''swallowing problem today", was ranked of low relevance by 32% of experts, this question was retained because it made a highly significant contribution to the major factor identified by factor analysis (Table 6). Additionally, in the final modified inventory, individual scores for Q12 were multiplied by 20, thus converting the range of possible scores for this question from 0 - 5 to 0 - 100; consistent with the remaining 16 questions.

Repeat test-retest reliability analysis on the modified total score revealed very little difference from that of the prototype inventory. For the modified inventory the mean normalised total score delta value was -2% ± 17% (95% CI:

-11% to 7%). Repeat validity analysis yielded no substantive changes to the individual weighting of each question in the factor analysis matrix. However, when compared to the prototype inventory, the proportion of the inventory's total variance, which was accoimted for by factor 1 increased from 55.5% to

59%. Repeat linear regression analysis between modified total inventory score and the global assessment score yielded a very slight increase in the correlation coefficient (r = 0.70, p < 0.0001). Chapter 4 Discussion, Conclusions and Implications

4.1 Discussion and Conclusions

One of the current major impediments to the evaluation of therapy for oral- pharjnigeal dysphagia, is the lack of a measurement tool of proven reliability and validity with which to assess longitudinal changes in symptom severity.

The MEDLINE search from January 1966 to June 1999 revealed 25 papers utilising some form of swallow questionnaire in humans. Only 17 related specifically to oropharyngeal dysphagia (Fleming and Weaver 1987; Sonies,

Parent et al. 1988; Layne, Losinski et al. 1989; Lindgren and Janzon 1991;

Tibbling and Gustafsson 1991; Nathadwarawala, Nicklin et al. 1992; DePippo,

Holas et al. 1994; Nathadwarawala, McGroary et al. 1994; Wintzen, Badrising et al. 1994; Deary, Wilson et al. 1995; Gustafsson and Theorell 1995; Bartolome,

Prosiegel et al. 1997; Litvan, Narahari et al. 1997; Young and Durant-Jones 1997;

Hinds and WUds 1998; Murry, Madasu et al. 1998; O'Neil, Purdy et al. 1999).

Four of these 17 were epidemiological studies simply utilised a questionnaire as a screening device to detect dysphagia in the general population (Sonies, Parent et al. 1988; Lindgren and Janzon 1991; Tibbling and Gustafsson 1991;

Gustafsson and Theorell 1995). Seven papers describe application of a questionnaire to detect dysphagia within specific patient populations who may suffer from dysphagia but did not use it to quantify dysphagia severity

(Nathadwarawala, Nicklin et al. 1992; DePippo, Holas et al. 1994;

Nathadwarawala, McGroary et al. 1994; Wintzen, Badrising et al. 1994; Litvan,

Narahari et al. 1997; Young and Durant-Jones 1997; Hinds and Wilds 1998). Three studies simply utilised a questionnaire to catalogue patient characteristics without attempting to quantify dysphagia severity (Fleming and Weaver 1987;

Layne, Losinski et al. 1989; DePippo, Holas et al. 1994). Hence, of these 17 papers only 3 utilised a questionnaire to assess dysphagia severity or change in severity over time (Deary, Wilson et al. 1995; Murry, Madasu et al. 1998; O'Neil,

Purdy et al. 1999). Only one of the three papers measured test-retest reliability

(O'Neil, Purdy et al. 1999) and only one aimed specifically to measure any component of validity of the questionnaire (Deary, Wilson et al. 1995). The appropriateness of this particular tool to oropharyngeal dysphagia is doubtful as the inventory was applied in an ENT clinic to patients presenting with the sensory disturbance of globus in whom only 30% reported actual dysphagia and only 3 of the 10 component questions were related to dysphagia per se.

(Deary, Wilson et al. 1995). The only paper that specifically aimed to evaluate dysphagia severity was O'Neil et al (1999), but validity testing was not comprehensive enough to conclude that this was a robust tool.

In the present study a modified 17 component inventory with a maximum possible score of 1700 was proven to have good test-retest reliability with total score demonstrating a mean variation of only 2% over time in a stable population with neurogenic dysphagia. The inventory also satisfied a number of criteria for validity. Four factors within it accounted for 75% of its variance and the dominant ''dysphagia" factor accounted for 59% of its variance. Good construct validity was confirmed by its strong correlation (r = 0.70; p <0.0001) with clinically determined severity as measured by a global assessment score. 47 We approached the design of this inventory from the standpoint of chnical judgement and experience combined with an understanding of the known mechanical disturbances to the oropharyngeal swallow (Cook and Kahrilas

1999). The inventory achieved a high degree of face validity as judged by authorities in the field, at least in part, vindicated this approach. We attempted to attach symptoms that are readily identified by the dysphagic patient to each of these mechanical disturbances. The intention in this approach was to yield an index of severity that was sensitive to alterations in the function of the mechanical components of swallowing. An instrument so derived should have a high sensitivity for the detection of dysphagia although this was not the primary aim of this instrument. More importantly, we hoped that the ir\strument constructed on this basis would be highly sensitive to change in response to effective therapy (discriminant validity). Indeed, the predictive validity data would suggest the tool is highly sensitive for the detection of dysphagia (Figure 2). The highly significant fall in mean score in patients treated for a fixed disorder (Zenkers) confirms the high discriminant validity of the questionnaire as predicted. This comprehensive approach to the construct of the questioimaire based on mechanical dysfunction should pick up dysfunctional components which might correlate with outcomes of prognostic importance (eg nutritional deficiency, pneumonia and death) (Logemann 1987;

Logemann, Roa Pauloski et al. 1992; Smithard, O'Neill et al. 1996) although this remains unproven. An alternative approach to the inventory construction might have been to poll patient perceptions of specific symptom's that bothered them most. Arguably such an approach might have yielded an index which correlates better with patient quality of life, although this concept also remains unproven.

The analyses of the test-retest reliability of the inventory indicated that the individual questions and the total score were stable over time with a mean change in the total score of the modified inventory of only -2% (95% CI: -11% to 7%). The important implication from this confidence interval is that any intervention-induced improvement in dysphagia severity, and which resulted in a reduction in total score of more than 11% or a deterioration (increase) of at least 7% as measured by this inventory in a comparable population, would achieve statistical significance (Mantha, Thisted et al. 1993). However, changes of this magnitude may not necessarily equate with therapeutic efficacy. As the responsiveness of the instrument, specifically in neuromyogenic dysphagia remains to be determined. Furthermore, statistically significant improvement may not equate with clinically sigruficant improvement, which is best measured indirectly by its impact on either clinical management or its impact on patient well being and quality of life (Sackett, Haynes et al. 1985). None the less, this data would prove useful in estimating approximate sample sizes in any such prospective efficacy study if the patient population were considered comparable to that used in the present study.

The major current problem in validating such a questionnaire is that no validated ''gold standard" exists against which construct validity can be 49 established. It was for this reason that we approached the validation process from multiple angles, only one of which involved construct validation. The most objective validation technique used purely mathematical approach, factor analysis, to establish content validity independent of subjective evaluation

(Carmines and Zeller 1979; DeVellis 1991; McDowell and Newell 1996). All but question 17 (relating to drooling from the mouth) contributed significantly to the variance of factor 1. We labelled factor 1 "dysphagia" because it accounted for 56% and 59% of the variance of the prototype and modified inventories respectively, and because each of the 17 questions in the modified inventory made a significant contribution to this factor. Furthermore, the loadings for each question in factor 1 were in a similar range (0.61 - 0.87). This indicates that each question had roughly equal influence over the variance of factor 1, thus obviating the need to weight any question within the inventory prior to calculating a total score. It could be argued that Q17, being the sole significant contributor to factor 1, which in turn accounted for 8.7% of total variance, is a good reason to retain this question. However, Q17 contributed only to a single phenomenon and, while drooling seems to be a common symptom, it neither satisfied our criterion of face validity nor did it contribute significantly to the primary factor of dysphagia. Indeed dropping Q13 and Q17 increased the proportion of the variance accounted for by factor 1 from 55% in the prototype to 59% in the modified inventory.

The difficulty in establishing construct validity of a tool designed to measure symptom severity is whether a suitable "gold standard'' exists against which it can be compared. Videofluoroscopy can detect and quantify aspects of dysfunction but in itself does not provide an objective numeric index of overall symptom severity (Logemann 1983). In the present study we used an approximation to a ''gold standard" for symptom severity by combining all the clinical and radiological information to derive a global assessment score which not only took into account the perceived severity of videofluoroscopic dysfunction but the overall clinical picture. Given these limitations which beset the validation of any tool designed to quantify symptom severity, the inventory score correlated highly with this surrogate standard, consistent with good construct validity.

Discriminant validity is important if the inventory is to be used to measure longitudinal change in response to treatment in order to establish the efficacy of that treatment. The present study found that the inventory demonstrated a 70% improvement in scores in response to treatment of patients with dysphagia due to a structural disorder (Zenker's). This change approximates the 89% change in global assessment score and is consistent with the published response rates for surgery (80-90%) (Cook and Kahrilas 1999). This data needs to be interpreted with caution because the sensitivity of this inventory to change in response to surgery or to other treatments in neuromyogenic dysphagia remains unknown. Unfortunately, at the present time, there is a lack of good controlled data, which either proves or disproves symptomatic improvement in response to any form of therapy for neuromyogenic dysphagia (Cook and Kahrilas 1999) making it difficult to evaluate discriminant validity in such a population.

51 Instead, we chose the Zenker's group in order to estimate the level of change that might be predicted from a highly efficacious treatment in a clearly defined patient population. None the less this approximation of a maximal expected change in score, together with the knowledge that this inventory can detect a statistically significant improvement in symptom severity as small as 11%, are important parameters in the design of future efficacy studies in oral-pharyngeal dysphagia.

4.2 Implications for future research

The development and validation of the swallow inventory will enable many avenues of research to be initiated that expand the usage of the inventory in different aetiologies and also in the assessment of various treatments of dysphagia. The next major steps that should be taken in the application of this inventory to dysphagia research should include:

1. Establish whether the inventory is sensitive to changes in symptom severity, either due to spontaneous improvement or following specific therapy, specifically in patients with neuromyogenic dysphagia.

2. Determine whether any such changes do correlate (as intended) with other clinically relevant endpoints including death, aspiration pneumonia and nutritional deficiency. In this manner, we could determine whether baseline symptom scores have prognostic significance and whether improvement in scores over time infers a lower risk of occurrence of these endpoints.

3. Determine whether scores and changes in scores correlate with quality of life measures. References:

Ali, G. N., K. L. Wallace, et al. (1996). ''Mechanisms of oral-pharyngeal dysphagia in patients with Parkinson's disease/' Gastroenterology 110: 383-392.

Anastasia, A. (1982). Psychological Testing. New York, Macmillian Publishing

Co.

Aviv, J. E., T. Kim, et al. (1996). 'Tibreoptic endoscopic evaluation of swallowing with sensory testing (FEEST) in healthy controls."

Dysphagia 2: 87-92.

Barer, D. H. (1989). "The natural history and functional consequence of dysphagia after hemisphere stroke." J Neurol Neurosurg Psychiat 52:236-241.

Bartolome, G., M. Prosiegel, et al. (1997). "Long term functional outcome in patients with neurogenic dysphagia." Neurorehabilitation 9(3): 195-204.

Bisch, E, M., J. A. Logemann, et al. (1994). 'Thar3nigeal effects of bolus volume, viscosity, and temperature in patients with dysphagia resulting from neurologic impairment and in normal subjects." Journal of Speech & Hearing

Research 37(5): 1041-59. Bloem, B., A. Lagaay, et al. (1990). ''Prevalence of subjective dysphagia in community residents aged over 87." Br Med J 300: 721-722.

Bombardier, C. and P. Tugwell (1982). "A methological framework to develop and select indices for clinical trials: statistical and judgemental approaches." J

Rheumatol 9(5): 753-757.

Carmines, E. G. and R. A. Zeller (1979). Reliability and validity assessment.

Beverly Hills and London, Sage Publications.

Cline, M. E., J. Herman, et al. (1992). "Standardization of the visual analogue scale." Nursing Research 41(6): 378-380.

Cook, I. J. (1991). "Globus - real or imagined?" Gullet 1:68-73.

Cook, I. J. (1991). Normal and disordered swallowing: new insights. Practical

Issues in Gastrointestinal Motor Disorders. J. Dent. London, Bailliere Tindall,

W.B. Saunders. 5(2): 1-24.

Cook, L J., W. J. Dodds, et al. (1989). "Opening mechanisms of the human upper esophageal sphincter." Am J Physiol 257: G748-G759. Cook, I. J. and P. J. Kahrilas (1999). "American Gastroenterological Association technical review on management of oropharyngeal dysphagia/'

Gastroenterology 116:455-478.

Cook, L J. and P. J. Kahrilas (1999). ''Medical position statement on management of oropharyngeal dysphagia/' Gastroenterology 116:452-454.

Cook, I. J., M. D. Weltman, et al. (1994). "Influence of aging on oral-pharyngeal bolus transit and clearance during swallowing: a scintigraphic study." Am J

Physiol 266: G972-G977.

Cowen, M. E., S. L. Simpson, et al. (1997). ''Survival estimates for patients with abnormal swallowing studies." Journal of General Internal Medicine 12(2): 88-

94.

Croghan, J. E., E. M. Burke, et al. (1994). "Pilot study of 12-month outcomes of nursing home patients with aspiration on videofluoroscopy." Dysphagia 9(3):

141-146.

Dantas, R. O., I. J. Cook, et al. (1990). "Biomechanics of cricopharyngeal bars."

Gastroenterology 99:1269-1274.

Deary, I. J., J. A. Wilson, et al (1995). "Globus pharyngis: development of a symptom assessment scale." Journal of Psychosomatic Research 39(2): 203-213. 55 DePippo, K, L., M. A. Holas, et al. (1994). "The Burke Dysphagia Screening Test: validation of its use in patients with stroke." Arch Phys Med Rehab 75:1284- 1286.

DeVellis, R. F. (1991). Scale development - Theory and applications. Newbury Park, Sage Publications.

Dodds, W. J. (1989). "The physiology of swallowing." Dysphagia 3:171-178.

Doty, R. W. (1968). Neural organization of deglutition. Handbook of Physiology, Alimentary canal. Sect 6, Vol 4. C. F. Code. Washington DC, American Physiology Society. 4:1861-1902.

Dua, K. S., J. Ren, et al. (1997). "Coordination of deglutitive glottal function and pharyngeal bolus transit during normal eating." Gastroenterology 112: 73-83.

Duranceau, A., E. R. Lafontaine, et al. (1987). "Oropharyngeal dysphagia and operations on the upper esophageal spincter." Surgery Annual 19:317-362.

Fleming, S. M. and A. W. Weaver (1987). "Index of dysphagia: A tool for identif3dng deglutition problems." Dysphagia 1:206-208.

Gordon, C., R. L. Hewer, et al. (1987). "Dysphagia in acute stroke." Br Med J 295: 411-414. Groher, M. E. (1986). 'The prevalence of swallowing disorders in two teaching

hospitals/' Dysphagia 1: 3-6.

Groher, M. E. (1987). "Bolus management and aspiration pneumonia in patients

with pseudobulbar dysphagia." Dysphagia 1: 215-216.

Gustafsson, B. and T. Theorell (1995). "Adaptedness and coping in dysphagic

students." Dysphagia 10: 86-92.

Hamlet, S. L., J. Muz, et al. (1989). "Pharyngeal transit time: assessment with videofluoroscopic and scintigraphic techniques." Dysphagia 4:4-7.

Hinds, N. P. and C. M. Wilds (1998). "Assessment of swallowing and referral to

speech and language therapists in acute stroke." Q J Med 91:829-835.

Horner, J., E. W. Massey, et al. (1988). "Aspiration following stroke: Clinical

correlates and outcome." Neurology 38:1359-1362.

Issa, F. G. (1994). "Gustatory stimulation of the oropharynx fails to induce

swallowing in the sleeping dog." Gastroenterology 107:650-656.

Jacob, P., P. Kahrilas, et al. (1989). "Upper esophageal sphincter opening and

modulation during swallowing." Gastroenterology 97:1469-1478. Johnson, E. R., S. W. McKenzie, et al. (1993). "Aspiration pneumonia in stroke/'

Arch Phys Med Rehabil 74(9): 973-6.

Johnson, R. and D. Wichern (1992). Applied Multivariate Statistical Analysis.

New Jersey, Prentice-Hall.

Kahrilas, P., J. Logemarm, et al. (1992). "Pharyngeal clearance during

swallowing: a combined manometric and videofluoroscopic study."

Gastroenterology 103:128-136.

Kahrilas, P. J. (1989). The anatomy and physiology of dysphagia. Dysphagia:

diagnosis and treatment. D. W. Gelfand and J. E. Richter. New York, Igaku-

Shoin: 11-28.

Kahrilas, P. J., W. J. Dodds, et al. (1988). "Upper esophageal sphincter function

during deglutition." Gastroenterology 95: 52-62.

Kahrilas, P. J., S. Lin, et al. (1996). "Oropharyngeal accommodation to swallow

volume." Gastroenterology 111: 297-306.

Kahrilas, P. J., S. Lin, et al. (1993). "Deglutitive tongue action: volume

accommodation and bolus propulsion." Gastroenterology 104(1): 152-162. Kjellen, G. and L. Tibbling (1981). "Manometric and oesophageal function, acid perfusion test and symptomatology in a 55 year old general population/' Clin

Physiol 1: 405-415.

Layne, K. A., D. S. Losinski, et al. (1989). "Using the Fleming index of dysphagia to establish prevalence.'' Dysphagia 4: 39-42.

Lindgren, M. D. and L. Janzon (1991). "Prevalence of swallowing complaints and clinical findings among 50-79 year old men and women in an urban population." Dysphagia 6:187-192.

Litvan, I., S. Narahari, et al. (1997). "Characterizing swallowing abnormalities in progressive supranuclear palsy." Neurology 48(6): 1654-62.

Loehlin, J. C. (1992). Latent variable models: an introduction to factor, path and structural analysis. Hillsdale, New Jersey, Lawrence Erlbaum Associates.

Logemarm, J. (1991). Approaches to management of disordered swallowing.

Practical Issues in Gastrointestinal Motor Disorders. J. Dent. London, Bailliere

Tindall, W.B Saunders. 5(2): 260-280.

Logemann, J., P. Kahrilas, et al. (1992). "Closure mechanisms of lar3mgeal vestibule during swallow." Am J Physiol 262: G338-G344. Logemann, J. A. (1983). Evaluation and treatment of swallowing disorders. San

Diego, College Hill Press.

Logemann, J. A. (1987). "Criteria for studies of treatment for oral-pharyngeal dysphagia/' Dysphagia 1:193-199.

Logemann, J. A., B. Roa Pauloski, et al. (1992). "Impact of the diagnostic procedure on outcome measures of swallowing rehabilitation in head and neck cancer patients.'' Dysphagia 7:179-186.

Mantha, S., R. Thisted, et al. (1993). "A proposal to use confidence intervals for visual analog scale data for pain measurement to determine clinical significance

[published erratum appears in Anesth Analg 1994 May;78(5):1035]/' Anesthesia

& Analgesia 77(5): 1041-7.

Martin, R. E. and B. J. Sessle (1993). "The role of the cerebral cortex in swallowing." Dysphagia 8(3): 195-202.

McDowell, L and C. Newell (1996). Measuring Health - A guide to rating scales and questiormaires. New York, Oxford University Press.

Mottola, C. A. (1993). "Measurement strategies: The visual analogue scale."

Decubitus 6(5): 56-58. Murry, T., R. Madasu, et al. (1998). ''Acute and chronic changes in swallowing and quality of life following intra-arterial chemoradiation for organ preservation in patients with advanced head and neck cancer/' Head & Neck

20:31-37.

Nathadwarawala, K. M., A. McGroary, et al. (1994). "Swallowing in neurological outpatients: Use of a timed test." Dysphagia 9:120-129.

Nathadwarawala, K. M., J. Nicklin, et al. (1992). "A timed test of swallowing capacity for neurological patients." J Neural Neurosurg Psychiatry 55: 822-825.

O'Neil, K., M. Purdy, et al. (1999). "The dysphagia outcome and severity scale."

Dysphagia 14:139-145.

Ott, D. J., R. G. Hodge, et al. (1996). "Modified barium swallow: clinical and radiographic correlation and relation to feeding recommendations." Dysphagia

11:187-190.

Pouderoux, P., J. A. Logemann, et aL (1996). "Pharyngeal swallowing elicited by fluid infusion: role of volition and vallecular containment." Am J. Physiol. 270;

G347-G354

Ravich, W. J., R. S. Wilson, et al. (1989). "Psychogenic Dysphagia and Globus:

Reevaluation of 23 Patients." Dysphagia 4: 35-38. 61 Rummel, R. J. (1970). Applied Factor Analysis. Evanston, Northwestern

University.

Sackett, D., R. Haynes, et al. (1985). Clinical epidenuology. A basic science for clinical medicine. Toronto, Little Brown and Company.

Schmidt, J., M, Holas, et al. (1994). "Videofluoroscopic evidence of aspiration predicts pneumonia and death but not dehydration following stroke."

Dysphagia 9(1): 7-11.

Scott, J. and E. C. Huskisson (1976). "Graphic representation of pain." Pain 2:

175-184.

Shawker, T., B. Sonies, et al. (1984). "Soft tissue anatomy of the tongue and floor of the mouth: an ultrasound demonstration." Brain Language 21:335-350.

Siebens, H., E. Trupe, et al. (1986). "Correlates and consequences of eating dependency in institutionalized elderly." J Am Geriratr Soc 34:192-198.

Smithard, D. G., P. A. O'Neill, et al. (1996). "Complications and outcome after acute stroke: does dysphagia matter?[ published erratum appears in Stroke

1998 Jul;29(7):1480-l].Stroke 1996;27:1200-1204." Stroke 27:1200-1204. Sonies, B. C., L. J. Parent, et al. (1988). "Durational aspects of the oral- pharyngeal phase of swallow in normal adults/' Dysphagia 3:1-10.

Thompson, W. and K. Heaton (1982). "Heartburn and globus on apparently healthy people." CMA Journal 126: 46-48.

Tibbling, L. and B. Gustafsson (1991). "Dysphagia and its consequences in the elderly." Dysphagia 6: 200-202.

Wein, B., R. Bockler, et al. (1991). "Temporal reconstruction of sonographic imaging of disturbed tongue movements." Dysphagia 6:135-139.

Wewers, M. E. and N. K. Lowe (1990). "A critical review of visual analogue scales in the measurement of clinical phenomena." Research in Nursing &

Health 13: 227-236.

Williams, R. B., K. L. Wallace, et al. (1997). "Upper esophageal sphincter (UES) compliance and opeiung characteristics in neurogenic dysphagia with failed relaxation." Gastroenterology 112(4): A852.

Winstein, C. J. (1981). "Neurogenic dysphagia: frequency, progression, and outcome in adults following head injury." Phys Ther 63(12): 1992-1997. Wintzen, A. R., U. A. Badrising, et al. (1994). ''Dysphagia in ambulant patients with Parkinson's disease: common, not dangerous." Can J Neurol Sci 21(1): 53-

56.

Young, E. C. and L. Durant-Jones (1997). ''Gradual onset of dysphagia: A study of patients with oculopharyngeal muscular dystrophy." Dysphagia 12:196-201. Table 1: Current Literature - Critical Evaluation

Citation Subject type Type of Questionnaire Main focus of Compared Test-retest Any validity Questionnaire Comments and number Dysphagia Type/Aim paper to controls Reliability Testing compared to independent objective measure or "gold standard" Bartolome G Neurogenic Oral- Catalogue of Long-term No No No No Advocate 1997 dysphagia pharyngeal patient records outcome of regular (n=63) patients with follow-up & neurogenic incorporate dysphagia video- using swallows etc. questionnaire & notes. Deary IJ 1995 Globus Sensory Self-report Validate a No No Principal Compared Population patients Oral- 10 item questionnaire components with Hospital presenting (n=105) pharyngeal Modified visual that assesses analysis Anxiety and with globus; analogue scale symptoms of revealed 3 depression only 30% ^obus. factors: sc^es had Dysphagia dysphagia. Globus ?relevance to Pain/ dysphagic swelling patients. Depippo KL Stroke Orai- Case-note review Validate Burke No No No Compared to Assessed 1994 patients pharyngeal + assessment of Dysphagia development of screening with no meal screening test medical test on prior oral- consumption inCVA complication overall pharyngeal 7 item. patients. during hospital outcome + anomaly Detection of admission eg: individual (n=139) dysphagia in pneumonia item failure susceptible population. Fleming S NA NA Catalogue of Simple No No No No No 1987 care-givers description of validation observations the attempted questionnaire Gustafsson B School Unselected Quest 1: Management of No Yes No No Identified 1995 children Self-report dysphagia by 114 (n=1295) 20 item -Y/N gaining data on dysphagjc examined every day children swaUow eating situation Examined difficulties, and psychologica QOLetc 1 aspects of Quest 2: dysphagia Self-report and the 20 item coping 1-5 point scale strategies. Aimed to assess which eating desires develop and habits independent (psychological) of clirucal Phone interview: intervention Interview 50 multichoice or fixed answer Hinds NP Acute stroke Oral- Interview Use of various No No PPVandNPV Intervention by Concluded 1998 (n=115) pharyngeal 11 items assessments of water test Speech and that use of (Able to sit Y/N answers swallow & Swallow Language the H20 test up with aid, Detection of questionnaire Question- Therapist(SLT) is a good GCS>13) dysphagia in & validation of naire bedside susceptible timed H20 screening population swallow test test. Foimd low sensitivity between swallow questionnair e and intervention ofSLT. LayneKA Neuro- Unselected Interview with Prevalence of No No No No Epidemiolog 1989 psychiatric patients primary nurse dysphagia in ical survey inpatients carer and review hospital Patients who (majority of patient notes. population. were men) classified as (n=513) ''possible" for dysphagia severity were repeated at 4mths to assess deterioration

(Used Fleming index) Lindgren S General Unselected Self-report Epidemiologica No No No No Used to 1991 population 44 items 1 population establish the (n=473) Y/N answers study with extent of Detection of respect to swallowing dysphagia in swallowing complaints community complaints andGORD including in gastro- commimity oesophageal reflux disease. Litvan 11997 Progressive Oral- Self-report Investigate Yes No Compared Barium Showed that supra- pharyngeal 20 item whether with barium swallow & coughing or nuclear Y/N ar\swers swallow swallow & Ultrasound choking. palsy (n=27) questionnaire ultrasoimd food Detection of or oral spreading dysphagia in examination over mouth, susceptible could predict excessive population abnormalities saliva observed in accurately ultrasound or predict and barium abnormal swallow tests. swallow in PSP patients. Murry T Head & neck Oral- Self-report Compared Yes, No No Weight, results Showed that 1998 cancer pharyngeal 14 item - oi\ly 11 HNRQ(QOL). however ofHNRQ HNRQ-QOL (n=58) (stratified items scored with ability to data not (QOL). has a strong by site: Oro- Yes/No/ swallow & utilised Assessed correlation pharynx, Sometimes change during patients prior with Hypo- DxRT to, end of swallow pharynx, Severity treatment questionnair Larynx) (wkT), & 6mth e (SQ), even post treatment. out to 6mths. Weight did not correlate with HNRQ- QOL or SQ. Dxsite important in HNRQ-QOL &SQ. Nathad- Neuro- Oral- Self-report Validate a No No No No Questionnair warawala myogenic pharyngeal 16item,Y/N& swallowing e used only 1992 dysphagia 0-4 scale speed test in to select (n=81) Detection of neuro- patients for dysphagia in myogenic suitability susceptible dysphagia. for study of population Used timed questionnaire swallowing as part of test validation. Nathan- Neurology Unselected Self-report Validate a No No No No Questionnair warawala outpatients patients 16 items swallowing e used only 1994 (n=90) Y/N answers & speed test. to select 0-4 scale Used patients for Detection of questionnaire suitability dysphagia in as part of for study of susceptible pop validation. timed swallowing test OlsFeiU 1999 135 Oral- Observer rating Develop a scale No Yes- No Most of the Trys to Neurogenic pharyngeal on aU available to rate severity inde- scale based on incorporate dysphagics information and functional pendent MBS - which swallow including level. observer maybe severity on modified barium considered MBS with swaUow (MBS) "gold standard" overall (Scale 1-7) functional status in one score. Sonies B 1988 Non- Unselected Screening Timing of oral- N/A No No Yes - subjects Questionnair dysphagic questionnaire pharyngeal underwent eused adult (n=47) phase of swallow study purely as a swallowing in screening normal adults tool Tibbing L General Unselected Self-report Epidemiologica No No No No Determined 1991 population Selection of 1 population dysphagia (n=796) dysphagic survey. prevalence individuals in general population WintzenAR Parkinsons Oral- Self-report Assess Yes-8 No No Compared Descriptive 1994 disease (PD) pharyngeal 13 item (Y/N) frequency of spouse with video- statistics. (n=16) Detection of subjective and controls swallow Did not dysphagia in objective observations validate susceptible dysphagia in questionnair population. PD. e Young EC Occulo- Oral- Self-report Explore the No No No No Mentioned 1997 pharyngeal pharyngeal or interview aspects of use of muscular 21 item (4 OPMD. swallow dystrophy sections). questionnair (OPMD) Detection of e provided (n=5) dysphagia in as an susceptible appendix - population poorly explained in paper. Table 2: Mechanisms of Oropharyngeal Dysphagia

Oral Phase Dysfunction Mechanism Aetiology Drooling Poor lip closure "^Facial muscle weakness Poor oral clearance Lingual dysfunction ^Central lesion ^Myopathy Delayed swallow initiation ^Afferent or central lesion Premature bolus spill Incompetent glossopalatal ^Myopathy closure ^Palatal surgery Pharyngeal Phase Post-nasal regurgitation Velopharyngeal incomptence "Central *Xth nerve ^Myopathy Laryngeal penetration or Reduced laryngeal elevation '^Suprahyoid muscle aspiration dysfunction Incomplete epiglottal closure ^Suprahyoid muscle c-f dysfunction *Tumour Impaired closure vocal cords ^Medullary or Xth nerve lesion Impaired pharyngeal "Central lesion clearance "Myopathy Impaired pharyngeal Absent or delayed pharyngeal "Central lesion propulsion response Impaired tongue base motion "Central lesion "Myopathy Impaired constrictor m. action "Central lesion "Myopathy Increased outflow Failed UOS relaxation "Medullary lesion resistance Loss of UOS compliance "Cricopharyngeal fibrosis Impaired hyoid traction "Suprahyoid muscle dysfunction Table 3: Aetiology of oral-pharyngeal dysphagia

Body System or Region Disease Central nervous system Stroke Extrapyramidal syndromes (Parkinson's, Huntington's chorea, Wilson's disease) Head trauma Brainstem tumours Alzheimer's disease t Multiple sclerosis Cerebral palsy Amyotrophic lateral sclerosis Drugs (phenothiazines, benzodiazepines) Peripheral nervous system Spinal muscular atrophy Guillain-Barre syndrome Poliomyelitis Post-polio syndrome Diphtheria Drugs (botulinum toxin, procainamide, cytotoxics) Myogeruc Myasthenia gravis Botulism Dermatomyositis Polymyositis Mixed connective tissue disease Sarcoidosis Thryotoxic myopathy Paraneoplastic syndromes Myotonic dystrophy Occulopharyngeal muscular dystrophy Drugs (amiodarone, alcohol, cholesterol-lowering drugs) Structural disorders Posterior (Zenker's) pharyngeal diverticulum Lateral pharyngeal diverticulum Cricopharyngeal bar Cricopharyngeal stenosis Cervical web Oropharyngeal tumours Head and neck surgery Radiotherapy Cervical osteophytes Table 4: Com-ponent Questions of the Prototype Swallow Inventory

1. How much difficulty do you have swallowing at present ? 2. How much difficulty do you have swallowing THIN liquids ? (eg Tea, soft drink, beer, coffee) 3. How much difficulty do you have swallowing THICK liquids ? \ (eg milkshakes, soiips, custard) 4. How much difficulty do you have swallowing SOFT foods ? (eg momays, scrambled egg, mashed potato) 5. How much difficulty do you have swallowing HARD foods ? (eg steak raw fruit, raw vegetables) 6. How much difficulty do you have swallowing DRY foods ? (eg bread, biscuits, nuts) 7. Do you have any difficulty swallowing your saliva ? 8. Do you ever have difficulty starting a swallow ? 9. Do you ever have a feeling of food getting stuck in the throat when you swallow? 10. Do you ever cough or choke when swallowing solid foods ? (eg bread, meat or fruit) 11. Do you ever cough or choke when swallowing liquids ? (eg coffee, tea, beer) *12.How long does it take you to eat an average meal ? How long would it take you to eat a scoop of ICE-CREAM ? 14. When you swallow does food or liquid ever go up behind your nose or come out of your nose ? 15. Do you ever need to swallow more than once for food to go down ? 16. Do you ever cough up or spit out food or liquids DURING a meal ? 17. Do you ever drool saliva from your mouth ? 18. How do you rate the severity of your swallowing problem today ? 19. How much does your swallowing problem interfere with your enjoyment or quality of life? All responses, except to Q12 & Q13, made by marking a visual analogue scale (not shown) imder each question.

* Possible responses: <15mins, 15-30mins, 30-45mins, 45-50mins, >60mins and Unable to swallow at all. Possible responses: <3mins, 3-5mins, 5-lOmins, >10mins and I don't eat ice- cream. Table 5: Aetiology ofNeuromyogenic DysjjJtagia in Group 1

Diagnosis Category Neuromyogenic patient group (N = 45) CVA brainstem 12 CVA other 2 Parkinsons disease 12 Movement disorder - other (eg: 4 dystonia) Myopathy 3 Myasthenia gravis 1 MND 6 Brainstem lesions - other (eg: 2 tumour, surgery) Idiopathic 3 Table 6: Summary of the Factor Analysis Matrix with Communality Summary in

Group 1 Patients (n = 45)

Note: Items which appear in bold and itaUcs exceed 0.60, indicating a significant contribution to the variance of the factor represented by the respective columns.

Factor 1 Factor 2 Factor 3 Factor 4 Communality Summary Question "Dysphagia" "Drooling" Loadings No 1 0.802 -0.135 -0.257 -0.027 0.73 2 0.799 0.089 0.429 0.103 0.84 3 0.831 -0.337 -0.084 0.184 0.84 4 0.761 -0.275 0.148 -0.151 0.70 5 0.726 0.295 -0.002 -0.391 0.77 6 0.753 0.345 0.203 -0.134 0.74 > 7 0.774 0.092 -0.267 0.392 0.83 8 0.672 0.315 -0.172 0.238 0.64 9 0.822 -0.039 -0.192 0.181 0.75 10 0.836 0.141 0.288 -0.27 0.87 11 0.716 0.137 0.509 0.203 0.83 12 0.732 -0.472 0.032 0.064 0.76 13 0.614 -0.411 0.091 0.272 0.63 14 0.621 -0.257 0.175 0.117 0.50 15 0.761 0.024 -0.404 -0.233 0.80 16 0.793 -0.12 0.007 -0.325 0.75 17 0.3 0.763 -0.018 0.314 0.77 18 0.866 0.117 -0.148 -0.206 0.83 19 0.788 0.114 -0.264 -0.015 0.70 Variance % 55.5 8.7 5.8 5.2 64.2% 12-

10^

Proportion of patients (%) 5 -

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Question number

Figure 1:

Patient compliance with question completion: proportion of patients who fail to complete each question. 1800 1 P < 0.0001

1600

1400

1200

Inventory 1000 Total Score 800

600

400

200

Controls Neuromyogenic (n = 19) Dysphagia (n = 45)

Figure 2; Group mean total score in patients with neuromyogenic dysphagia is highly sigruficant greater than that of controls without (p < 0.0001). 25

5 Delta Score 0 + (mm) -5

-10- -151

-20-

-25 T 1 I 1 I r- T 1 1 1 1 1 1 1 I 1 1 1 1— 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Inventory question number Figure 3: Group mean delta scores and 95% confidence intervals (CI) for each of the 19 questions derived by subtracting retest score from baseline score. No significant change over time w^as noted. The mean delta values for raw scores ranged from 0 to 9.7, and mean change in total score was -11.6 (95% CI, -65% to 42.3%). With the exception of Q12 and Q13, the maximimi possible score for each question is 100. Average change in normalised total score was - 0.5% (95% CI, -9.2% to 8.2%). Relevance to 19 Oral-pharyngeal 18 mmmmmm Dysphagia 17 • Low 16 ^ Moderate 15 • High 14 v/jmm 13 12 11 Question 10 W////////K number 9 8 7 6 5 4 3 2 1 f mm^mmm.\ 1 1 1 1 1 1 1 1 1 0 10 20 30 40 50 60 70 80 90 100 Ranking (% experts polled)

Figure 4:

Experts' opinion as to the relevance of each individual question to the assessment of severity of oral-pharyngeal dysphagia to estimate the face validity of the inventory. Sixteen of 19 (84%) questions rated were considered to have moderate or high relevance. Inventory Total Score

Clinician Global Assessment score

Figure 5:

Correlation between global assessment score and the total score of the prototype inventory. The highly significant correlation coefficient (r = 0.69; p < 0.0001) is consistent with construct validity. 1800- B 1600-

1400-

1200-

1000- Inventory Global Assessment Total score Score

Pre Surgery Post Surgery Pre Surgery Post Surgery Zenker's Diverttculum Patients Zenker's Diverticulum Patients (n=:11) (n = 11}

Figure 6: Changes in (A) total inventory score and (B) global assessment score after surgery in 11 patients with Zenker's diverticulum. Total inventory score decreased by a mean of 70% after surgery (p < 0.0001). A concurrent decrease in mean global assessment of 89% was noted (p < 0.0001). Addendum - Clarification and comments in response to issues raised by reviewer

Part B.2 Methods

1. No. The approach stemmed from consideration of the underlying mechanisms of dysphagia and how they might translate into symptoms. It was felt this was reasonable, as tliere is no evidence that neither the age of the patient nor the aetiology of the dysfunction would translate into different symptoms given common mechanisms.

2. The choice of 32 internationally recognised dysphagia authorities was purely arbitrary.

3. The clinician, A/Prof. Ian Cook, is a gastroenterologist with a proven track record in the clinical evaluation and research of oropharyngeal dysphagia. This was indicated in the title of the publication in the journal Gastroenterology. The speech language pathologist (Ms Paula Bowman) was in fulltime employment for at least 5 years in a teaching hospital and in this role had extensive responsibilities for the assessment and management of patients with neuromyogenic dysphagia.

4. There were no specific reasons for choosing precisely 19 questions. The questions were "generated'' to address the symptoms believed likely to emanate from each of the specified mechanisms of oral and pharyngeal dysfunction. Therefore, to some extent, the selected number of questions was "arbitrary''.

5. The clinical stability of the 48 patients for 3 months was not tested formally. This was based on reasonable clinical evidence that after an acute event, any substantive change, particularly over a 14 day period, after an initial 3 months of natural recovery, would be most unlikely.

6. The issue of sample size calculations is extremely difficult with a study of this type. This is because, a priori, when designing an instrument from scratch, neither the sensitivity nor the variance of the scores is known. In the determination of sample size these two variables are important. As we adopted what we thought was a fairly logical approach, ie, if the instrument did not reach statistical significance with the 18 patients in the test-retest reliability, it was not likely to be clinical significant either.

7. We agree with the reviewer that it would have been better if more than one independent clinician made an assessment of these patients. However, this would have proved difficult logistically and another clinician with a comparable degree of clinical experience with this specific patient population was not readily available to us. Furthermore we felt that blinding the clinician to the questionnaire scores would be sufficient to minimise bias. 8. The patients chosen (N=18) for test/retest reliability measures were not chosen randomly. Rather, we selected the first 18 patients who could be tested on a second occasion.

Part B.3 Results, Findings and Data Analysis

1. Inventories with less than 15 questions answered were excluded from all analysis. However, patients who completed between 15-18 questions (ie: omitted between 1 and 4 questions) were utilised for the factor analysis only.

2. Terminology and jargon surrounding various types of validity are confusing. After reviewing the literature in this area we defined construct validity as . .measurement tools ability to produce consistent results that reflect the true state clinical state of the patient " (Bombardier & Tugwell 1982). Construct validity is more loosely defined type of criterion validity. Construct validity is usually used when the investigator does not have a set of criterion that are fully valid. Criterion validity by definition requires a much more stringent approach (Cronbach & Meehl 1955)

We agree with the reviewer that the inventory is less likely to be as relevant or even discriminatory at the severe end of the spectrum than say at the mild or moderate severity level. However, in assessing longitudinal change many patients either start or finish at the severe end of the spectrum making a scale of this type able to accurately quantify severity throughout the scale important. An important supplementary question for future research would be to better evaluate the construct validity and sensitivity to change at the mild to moderate end of the spectrum. In lieu of a better "gold standard'" we do not believe that the addition of more than one clinician to provide the global assessment score would solve this problem. It may reduce the variants of the GAS score by doing so would not necessarily transform a "bronze" into a "gold" standard. Nonetheless, it is highly likely that the inclusion of multiple assessors would be a valuable alteration to the methods in future studies evaluating further aspects of this questionnaire. However, in doing so, it would need to be recognised that there is no available data "calibrating" a clinician as an assessor nor do we know the test/retest reliability of a clinician in this role.

Reference: Cronbach, L.J. and Meehl, P.E. (1955). "Construct Validity in Psychological Tests." Psychological Bulletin, 52: 281-302. Correct-ed. ^ Hp- I r

J?-

vir i- , ^

jr •t' yir

p mr

'ytih

(. A*

- vtV k' i Jk <

<> ¥ 1 i ff St ,4

3 V jf A-

Vif

1 c -r . % V. T

> If i^I > V f I ^ Vilf 4

ij Wf E , J \ V ^^wmar / ^ -Si-