Family Practice Vol. 16, No. 2, 190-195
© Oxford University Press 1999
Questions you didn't ask? COOP/WONCA Charts in clinical work and research
Department of General Practice and Community Medicine, University of Oslo, P.O. Box 1130, Blindern, 0318 Oslo,
a The Ullensaker Survey, Jessheim and
b GlaxoWellcome, Norway.
Bentsen BG, Natvig B and Winnem M. Questions you didn't ask? COOP/WONCA Charts in clinical work and research. Family Practice 1999; 16: 190–195.
Received 12 January 1998;
Revised 19 May 1998;
 |
Abstract |
|---|
 Top Abstract IntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
Objective. COOP/WONCA Functional Assessment Charts are widely in use in research and clinical work. They originally evolved from Dartmouth COOP Functional Assessment Charts. Our objective is to describe our experiences with COOP/WONCA Charts and to provide data on reliability and reference data from a normal population.
Method. A test–retest study comprising 40 randomly selected patients attending GPs' surgeries was conducted. Eight GPs took part in this study. The response rate for the questionnaire was 100%. In a second study, 245 asthma patents who consulted 53 different GPs were consecutively included; 171 of these repeated completion of the charts after 6 months. Mean scores from a random sample of 2864 persons in the total population of the municipality of Ullensaker in Norway are given as reference values. There was a postal return rate of 64%.
Results and conclusions. The reliability of the charts is good, and nine out of ten of the patients gave clinically meaningful answers after 2–3 days. The scores of the 171 asthma patients were nearly identical after 6 months. The reference values are presented. A standardized method for reliability testing is discussed, and the use of COOP/WONCA Functional Assessment Charts is commented on.
Keywords. Clinical work, functional status assessment, outcome study, reference values, standard for reliability testing..
|
Introduction |
|---|
|
TopAbstractIntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
One of the important changes in health care in recent years has been the move towards the assessment of health status and outcome. Diagnosis of a health problem is no longer the only aim of health care. A patient's quality of life and the perceived outcome of medical interventions have also come into focus. One example is teaching of the ‘patient-centred method':1 the patient's agenda as well as the doctor's agenda is better presented in a real dialogue than in traditional history taking. A second is the new problem-based curricula at many medical schools. We know that a diagnosis in itself does not usually tell us about the consequences of a disease for the patient. Nor do biological or physiological tests. Respiratory physiological tests do not necessarily correlate well with the quality of life of asthma patients,2 nor does the diagnosis of chronic diseases say anything about a person's quality of life. Several health problems may often affect the functional capacity of a person at the same time.3,4
Health Related Quality of Life Questionnaires have been developed in increasing numbers during recent years.5–9 These questionnaires are intended both for use in daily clinical practice and for research. They include general, generic inventories, condition- or disease-specific questionnaires, and special function questionnaires.5–9 In busy general practice there has been a need for an instrument to obtain additional information about a patient's functional capacity and the outcome of interventions. The general questionnaires intend to give a basic assessment. In our opinion the best of them are COOP/WONCA Charts.
|
The COOP/WONCA Charts |
|---|
|
TopAbstractIntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
The requirements for an optimal general, generic questionnaire suitable for use in primary health care have been listed by several authors (Table 1
).10–14
|
COOP/WONCA Functional Assessment Charts meet these needs.9–14,16 The original Dartmouth COOP Function Charts were developed by Gene Nelson at Dartmouth Medical School in co-operation with about 200 GPs. In 1988 the WONCA (World Organization of Colleges, Academies and Academic Associations of General Practitioners/Family Physicians) Classification Committee revised the charts.
The assessment is self-reported and simple, comprising six charts (Fig. 1). Five of the charts present different domains: physical fitness, feelings, daily activities, social activities and overall health. The variables seem to be acceptably independent, assessed by correlation coefficients between five of the COOP/ WONCA Charts.17 The sixth chart asks about change in health status. Each chart has a five-point scale: 1 = very good, 2 = good, 3 = moderate, 4 = bad and 5 = very bad.10 In some of the translations ‘Overall health' has been modified giving ‘fair' the score of ‘3' in order to balance the scale and to make it suitable for parametric testing.16 Each step of the scale is illustrated with a drawing. An individual's score gives a profile of health status.
|
The questionnaire is clinically relevant and gives important information about somatic, psychological and social functional capacity. One study18 showed that in two of three encounters in general practice, it gave information of importance or some importance to the doctor. The original charts have been tested against other ‘Quality of life questionnaires' in use, and validity was good.11,12 It is easy to use. The self-reported questionnaire takes no more time than measuring blood pressure: 3–5 minutes. The questions are easy to understand.11,12,16
Scores can be directly compared with the individual patient's earlier scores or with reference values.12,16,17 Statistical analysis is not needed in daily practice. However, in research and outcome studies this may be necessary. Assistance from interviewing or complicated technology are not required, and the system is low-cost.
The questionnaire is usually not limited by the age, gender, ethnic, religious or cultural characteristics of the patient. It gives an international standard, is in use in many countries and has been translated into many languages.
Compared with other instruments that are recommended for use in general practice, we find it to be the best.5–9
The reliability of COOP/WONCA Charts
Not enough testing has been carried out on the reliability and validity of functional status or quality of life inventories. This is the general conclusion of McDowell and Newell,5 Bowling6,7 and McSweeny and Creer.8
In Measuring Functional Status with the COOP—WONCA Charts,12 published in 1995 by van Weel and co-workers, it is stated that there has been no known reliability study of the revised charts, which differ from the original Dartmouth COOP Function Charts. Results from studies of the original charts are given.12,13 One British study has been published, but the conclusions are reserved.15 The charts are in wide use in Norway for research purposes, but are now also easily available for all health workers in Norway. We felt that there was a need for reliability testing of the Norwegian edition of the charts.16 A test–retest study will be presented. An outcome study of asthma patients completing COOP/ WONCA Charts after an interval of 6 months will also be presented. Results from a population survey in the municipality of Ullensaker, which gives reference values for possible comparisons, will also be given.
|
Methods and results |
|---|
|
TopAbstractIntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
A test–retest study of COOP/WONCA Charts
Forty randomly selected patients attending eight different GPs' surgeries were included in the study by the physician.16 The patients completed the COOP/ WONCA Charts in the waiting room before the consultation. When they left the surgery, they were handed a second copy of the questionnaire, and asked to repeat the procedure after 2–3 days. The patients were first told about the request to repeat the questionnaire when they left the office. Because of a misunderstanding, four completed the second copy of the form after 6–7 days. The return rate of the questionnaire was 100%.
Twenty-three gave identical answers, and 17 scored better or worse on Chart E (‘Change in health'; see Fig. 1
). This is reasonable because they attended a doctor's surgery partly for acute health problems. The reasons for encounter were respiratory disorders (33%), musculoskeletal (25%), mental (20%), cardiovascular (8%) and neurological (8%).
Of the 17 whose answer on chart E indicated a ‘Change in health', 12 showed a corresponding change in one of the other variables. For example, a patient who answered ‘Better' also had a ‘better score' for at least one other variable. This correspondence between scores on Charts E and A–D or F can be called reliability consistency. Even among the remaining five patients, plausible explanations could be found. For example, one patient had indulged in a drinking spree between test and retest, and the answers were therefore meaningless. However, the results were clinically valid for 87.5% of the participants, or 90% if the ‘drinker' was excluded (see Table 2).
|
A long-term study of COOP/WONCA Charts
The period of time which the charts ask the patient to consider is "the last two weeks". What are the results when we repeat completion of the charts after a longer interval? In another Norwegian study, 245 patients consulted 53 different GPs for asthma.16 They were consecutively included in the study. They completed the COOP/WONCA Charts together with a condition-specific questionnaire. This was repeated after 6 months. One hundred and seventy-one (70%) of the cohort returned and completed the charts a second time. This is not a test–retest study, but has the character of an outcome study. There was practically no change in medication. The mean COOP/WONCA scores of the cohort for the different variables were nearly identical on both occasions. This was despite the fact that nearly half the responders (48%) indicated a change in answer to another question: "During the last six months has your asthma worsened, been unchanged, or improved?". Table 3
depicts a surprising stability in the mean scores of a cohort of asthma patients. The mean scores from this study showed much worse values for all variables than that of a random ‘normal' population.16
|
Reference values for comparison
The scores for different variables can be compared with an individual's or a group's earlier scores. Each variable has to be assessed separately. However, the scores can also be compared with the mean scores of a representative population like the one in the Ullensaker Survey.16 Ullensaker represents a typical average Norwegian municipality similar to the neighbouring municipality of Nes in Akershus. Socially, economically and professionally Nes is close to the national average.4 The Ullensaker Survey initially comprised a random sample of 398 persons.17 In a new, expanded, postal, randomized survey in the municipality, 2864 persons were included.16 Initially the sample represented all persons in certain age groups within the municipality: 24–26, 34–36, 44–46 years of age, etc. The return rate for the COOP/WONCA questionnaire was 64%. Table 4
shows the mean scores for the different variables by age groups and gender. It is essential when comparing mean scores to make corrections for age and gender.
|
|
Discussion |
|---|
|
TopAbstractIntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
A standard method for reliability testing?
The best way of testing for reliability is to compare answers given by the same person to a questionnaire on two different occasions.9,19–21,23 This is also called reproducibility. Assessment of stability requires that a questionnaire is completed on three or more occasions,21 but this was not done in the present study. However, between two timepoints we must expect changes in the health status of individuals. The ideal ‘stable' sample does not exist. To determine the extent of change, we need to test convergent validity using other inventories simultaneously. The designs and contents of reliability tests are described in insufficient detail in the literature. It has been difficult to find a method to test ‘reliability consistency' when this includes a real change in the patient's health status, as in this test–retest study.
Several universal or specific questions about the COOP/WONCA Charts can be raised. Table 5 lists some of the problems and biases of reliability tests.
|
Revision of the charts
The COOP/WONCA Charts are now under revision and two questions can be asked.
- What is understood by the term "the last two weeks"? One special day, for example, the last one, the worst one or an overall assessment? Probably individuals answering will have the same concept both at test and retest.
- Could the text be better formulated? There have been suggestions for simplification and a more common wording. Translations into other languages, differences in culture and the concept of ‘health', raise other problems. Direct translations from English to other languages do not always address the real concepts. For example the American "extremely good health" as used in the original charts is not in use in Norway.
These questions could be subjects for studies and assessment.
The number of subjects
The number of individuals included in a test–retest study should be at least 40. This seems to be sufficient for a reliability study. Donner and Eliaziw23 pointed out that this is enough " ... to provide an 80 per cent power for testing contours of coefficient of interclass correlation, derived from a one-way analysis of variance model. This operates well at the level of 5% of significance for selected values". However, surgery patients must be expected to show change in health status. Increasing the sample to, for example, 60 patients would have increased the power of the tests used.
The optimal condition for testing is when patients are stable and unchanged. However, this is impossible in daily clinical practice. One important use of the COOP/WONCA Charts is in daily clinical practice. This represents the reality. The presented study shows that this is possible.
The interval between test and retest
The interval between the test and the retest can be relatively short. This can make the results more reliable by removing many external, changing factors. An interval of 2–3 days between the tests was chosen for this study. This interval cannot be more than 2 weeks because of the wording in the charts. It cannot be as short as 2 hours because it is too easy to recall the questions and answers. Living conditions during such a short interval are as stable as possible. There is also a question about motivating patients to participate in follow-up for a research project. Observed participation in follow-up was, as mentioned earlier, 100% in the presented study. However, it was hoped that the ‘sealed envelope' with the second questionnaire, given to the patient when leaving the surgery, and not opened before the stated interval, would ensure a real break in time and memory.
Structure and statistical analysis
The COOP/WONCA Charts represent two types of scales: the Likert, ordinal scales for the different variables, and the variable for change. The results of the test–retest study seem to show not only good reliability but also good criterion validity and an acceptable discriminatory power to record change. The term ‘sensitivity' is already used in other contexts. This is especially true of the ‘change' variable. Scales of 4–7 steps have been recommended.20 However, scales of five steps, as in the COOP/WONCA Charts, seem to be preferred.10,19–22 Statistical analyses of studies such as those presented here may be necessary. However, a sensible, sensitive test for inclusion of natural, documented change such as that observed in the present test–retest study seems to be lacking.
There is a worry with one of the COOP/WONCA Charts—the skewed scale of ‘Overall health'. It should be symmetrical. It must be stressed that skewed scales are not fit for parametric testing. The use of score 3 as the ‘neither/nor score' will make it possible to use mean scores. Therefore, some of the translations of the English edition of the charts are corrected and changes are made to the ‘Overall health' scale. This is now recommended by the WONCA International Classification Committee.
The results of the presented test–retest study confirm the qualities of the COOP/WONCA Charts. The charts demonstrate good criterion validity, and also good reliability and discriminatory power on the individual level.
The presented asthma study depicts unchanged mean scores after an interval of 6 months. Would not a worsening of the cohort's scores have been expected, or is the time span too short? Could this be the result of a Hawthorne effect?
The use of the COOP/WONCA Functional Assessment Chart
In daily clinical practice the chart scores seem to give important information additional to that collected in an ordinary consultation.18
When assessing the outcome of medical intervention the charts can give a base for measurement of effect. The individual health worker can assess her/his own work. The use of the charts can also give a base for discussion in quality assurance colleague groups.
One important use of the charts is in assessment of effect of pharmaceutical drugs or other forms of interventions. Does the testing of a drug or treatment demonstrate improvement of functional capacity/quality of life? If the timespan before retest exceeds 2 weeks, an additional question must be asked: "Has your overall status changed since you last filled in the charts?"
The charts can also be used in epidemiological research and in descriptions of the natural history of disease. This can include repeated completion of the charts over time.
It is obvious that such a simple, general inventory may from time to time need additional, more detailed questions. Condition-specific questionnaires may fill this gap. However generic, global questionnaires should always be used when assessing functional capacity or quality of life, even when condition/disease-specific inventories are also used.
The six charts presented are the base for this questionnaire. The WONCA Classification Committee is working together with experts in this field and the Dartmouth COOP project on an expansion with six optional, additional charts: pain, fatigue, sleep, self-care, coping and mental functioning. This expansion is a main task for WONCA during the coming years. It seems that the combination of simple questions, illustrations and numerical scores is the strength of the COOP/ WONCA Charts.
|
Conclusion |
|---|
|
TopAbstractIntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
The COOP/WONCA Charts have been tested for reliability. A Norwegian test–retest study showed that reliability is good and that reliability consistency is impressive.16 A long-term study of asthma patients showed that as a cohort mean scores were nearly the same after an interval of 6 months.16 We find the use of COOP/WONCA Charts feasible both in clinical practice and in research, including outcome studies. Reference values of mean scores from a representative Norwegian population are given. The charts add new dimensions to the International Classification of Primary Care.24
This article outlines the principle steps needed to conduct a test–retest study. This is suggested as a standardized method for reliability testing of COOP/WONCA Charts. The test–retest should be performed for all editions of COOP/WONCA Charts regardless of language or culture. A sample of a ‘normal', representative population can give reference values.12,16
|
Acknowledgments |
|---|
Thanks to all those involved in these studies. Practical and economic support was given by Glaxo Norway for realization of the study of asthma patients. Support was provided by The Norwegian Research Council for Sciences and Humanities for realization of the Ullensaker Survey.
|
Notes |
|---|
Parts of this article have been published previously: Bentsen BG, Natvig B, Winnem M. Assessment of own functional status. COOP–WONCA Charts in clinical work and research. [In Norwegian]. Tidsskr Nor Laegeforen 1997; 117: 1790–1793.
|
References |
|---|
|
TopAbstractIntroductionThe COOP/WONCA ChartsMethods and resultsDiscussionConclusionReferences |
|---|
1 Levenstein JH, McKracken E, McWhinney IR, Stewart MA, Brown JB. The patient-centred clinical method. 1. A model for the doctor-patient interaction in family medicine. Fam Pract 1986; 3: 24–30.
2 Rutten-van Mölken MPH, Custers F, Van Doorslaer EKA et al. Comparison of four instruments in evaluating the effects of salmeterol on asthma quality of life. Eur Respir J 1995; 8: 888–896.[Abstract]
3 Bentsen BG. Assessment of disability in a population. [In Norwegian.] Tidsskr Nor Lægeforen 1961; 81: 295–302.
4 Bentsen BG. Illness and general practice. A survey of medical care in an inland population in south-east Norway 1952–1955. In Universitetsforlaget, 1970. 2nd edn. Oslo: Scandinavian University Books, 1986.
5 McDowell I, Newell C. Measuring Health. A Guide to Rating Scales and Questionnaires. Oxford: Oxford University Press, 1987.
6 Bowling A. Measuring Health. A Review of Quality of Life Measurement Scales. Milton Keynes: Open University Press, 1991.
7 Bowling A. Measuring Disease. A Review of Disease-specific Quality of Life Measurement Scales. Buckingham: Open University Press, 1995.
8 McSweeny AJ, Creer TL. Health-related quality-of-life assessment in medical care. Disease-a-Month, 1995; 41: 3–71.
9 Hutchinson A, Bentzen N, König-Zahn C. Cross Cultural Health Outcome Assessment: A User's Guide. EC/BIOMED project 1993–1995. European Research Group on Health Outcomes, 1997.
10 WONCA Classification Committee. Functional status measurement in primary care. In Lipkin M (ed.). Frontiers of Primary Care. New York: Springer-Verlag, 1990.
11 Nelson EC, Landgraf JM. Hayes JW et al. The COOP Function Charts: a system to measure patient function in physicians offices. In: Lipkin M (ed.). WONCA Classification Committee: Functional Status Assessment Measurement in Primary Care. New York: Springer-Verlag, 1990: 79–131.
12 van Weel C, Touw-Otten FWMM, van Duijn NP, Meyboom-de Jong. Measuring Functional Status with the COOP/WONCA Charts. den Haag: WONCA, ERGHO, NCH. CIP-Gegevens Koninklijke Bibliotheek, 1995.
13 Nelson EC, Wasson JH, Johnson DJ, Hays RD. Dartmouth COOP Functional Health Assessment Charts: brief measures for clinical practice. In Spilker B (ed.). Quality of Life and Pharmacoeconomics in Clinical Trials. 2nd edn. Philadelphia: Lippincott-Raven, 1996: 161–168.
14 Mæland JG, Lærum E. Measuring quality of life in general practice. Scan J Prim Health Care 1992; 10: 1–2.
15 Kinnersley P, Peters T, Stott N. Measuring functional health status in primary care using the COOP–WONCA Charts: acceptability, range of scores, construct validity, reliability and sensitivity to change. Br J Gen Pract 1994; 44: 545–549.[Web of Science][Medline]
16 Bentsen BG, Natvig B, Winnem M. Assessment of own functional capacity. COOP–WONCACharts in clinical work and research. [In Norwegian.] Tidsskr Nor Lægeforen 1997; 117: 1790–1793.
17
Bruusgaard D, Nessøy I, Rutle O, Furuseth K, Natvig B. Measuring functional status in a population survey. The Dartmouth COOP Functional Health Assessment Charts/WONCA used in an epidemiological study. Fam Pract 1993; 10: 212–218.
18 Landgraf JM, Nelson EC, Dartmouth COOP Primary Care Network. Summary of the WONCA/COOP international health assessment field trial. Aust Fam Physician 1992; 21: 255–269.[Medline]
19 Altman DG. Practical Statistics in Medical Research. London: Chapman and Hall, 1991.
20 Streiner DL, Norman GR. Health Measurement Scales. A Practical Guide to their Development and Use. Oxford: Oxford University Press, 1991.
21 Dunn G. Design and Analysis of Reliability Studies: the Statistical Evaluation of Measurement Errors. London: Edward Arnold, 1992.
22 Cox EP III. The optimal number of response alternatives for a scale: a review. J Marketing Res 1980; 17: 407–422.
23 Donner A, Eliaziw M. Sample size requirements for reliability studies. Stat Med 1987; 6: 441–448.[Web of Science][Medline]
24 Lamberts H, Wood M. International Classification of Primary Care. Oxford: Oxford University Books, 1987.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. I. Brox and O. Froystein Health-related quality of life and sickness absence in community nursing home employees: randomized controlled trial of physical exercise Occup. Med., October 1, 2005; 55(7): 558 - 563. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||