Original Papers

Measuring health status: a new tool for clinicians

and epidemiologists
S. M. HUNT, MA, Ph.D Impact Profile;9 the Cornell Medical Index;'0 the McMaster
Health Index Questionnaire; 1" the General Index of Well-
J. McEWEN, FFCM, MFOM being'2 and many others. 13-17
There are a number of criticisms which can be made of existing
S. R McKENNA, BA, Ph.D measures of self-reported health status, although not all the
following comments apply to each instrument. First, they are
SUMMARY The development and validation of a short and often long and complicated with ambiguous statements;
simple measure of perceived health problems is described. secondly, scoring and weighting for seriousness often reflect the
Extensive testing with selected groups, including the elderly, values of the physician not those of the lay person; thirdly, the
the chronically ill, pregnant women, fracture victims, and focus of the measures may be on too narrow an area, for example
a random sample of the community has established the face, disability; and fourthly, where the answers are summed to a single"
content and criterion validity, and the reliability of the score or index this can be derived in many different ways and
instrument. The Nottingham Health Profile is intended as a involve the addition of scores from areas not logically connected,
for example physical mobility and appetite.
standardized tool for the survey of health problems in a A tool for the survey of populations, as Culyer has pointed
population, but is equally valid and useful as a means of out,'8 should not be too sophisticated because of the difficulty
evaluating the outcome of medical and/or social interven- of interpreting responses and standardizing scores. It must be
tions and as an adjunct to the clinical interview. sensitive enough for the assessment of the health needs of the
population and specific enough for the evaluation of health care
provision for special groups. It must also be understood by a
Introduction large majority of potential respondents, be short and simple to
THE increasing cost of the provision of health services, answer, cheap to administer and score and, above all, be valid
often in the face of meagre evidence about the efficacy of and reliable.
many interventions, together with doubts about the way in which Although 'quality of life' is now widely determined, it is often
resources are allocated has led to attempts to find efficient and difficult to know what is being measured since there are no
reliable means of assessing health needs and outcomes. agreed criteria for what constitutes quality of life and such in-
Consequently, in recent years many writers have called for the struments lack validity. It seems more appropriate for those in-
development of 'sociomedical' or 'subjective' indicators. In volved in health care to consider a 'health profile' which records
general, it has been hoped that such indicators would be capable the perceived health (or departures from health) of individuals
of measuring the health status of whole populations at a par- or groups. The relationship between 'objective' and 'subjective'
ticular time; of providing reliable repeated measures; and of is often regarded as a methodological problem, yet it may be
assessing the efficacy of health care practice accurately.1' 2White more useful to consider the two aspects as being essential to our
suggested that the model for policy planning based on tradi- knowledge of human beings and their reactions. Most so-called
tional indicators should be replaced by a cybernetic model using objective criteria involve clinical judgements about normal func-
information about health status as a basis for problem defini- tioning being essential to a high quality of life but evidence is
tion, resource allocation and service organization.3 accumulating which shows that people judge their experiences
Even 50 years ago MacKenzie suggested that subjective percep- in relation to their expectations. Certain limitations and
tions could be indicators of the onset of disease4 and such disabilities seem normal after an adjustment period.
perceptions have been found to be excellent predictors of Comparisons of value in health-related activities must allow
mortality5 and to be key factors in adjusting to major illness.6 the perceptions of the patient an equal, if not greater, place than
Moreover, it is perceived, and not necessarily actual, problems clinical evaluations. The subjective assessment of the patient may
which lead to demand for health care, although several in- allow more successful interpretations of the impact that disease
vestigators have found perceived health status to be an accurate and treatment have on his or her quality of life, whereas objec-
reflection of so-called objective measures.78 Subjective indices tive indicators may merely be projections of professional mores.
of health also widen theoretical frameworks of aetiology to in-
clude perceived occupational stress, domestic strife, sexual con- The Nottingham Health Profile
flicts and so on.
There have been several attempts to develop standard measures Development of the profile
of self-assessed health, particularly in the USA, and to come In 1975 work started in the Department of Community Health
to terms with the problems of definition, measurement, at Nottingham University on the development of a measure for
weighting, reliability, validity, sensitivity and applicability which the quality of life Statements were collected from over 700 people
are endemic to such endeavours. Examples include the Sickness describing the typical effects of ill-health - social, psychological,
behavioural and physical - for example, 'I sleep badly'. 'I've
S.M. Hunt, Senior Research Fellow, Research Unit in Health and lost interest in sex', 'I find it hard to walk about'
Behavioural Change, University of Edinburgh; J. McEwen, Senior Lec- An initial pool of 2200 statements enabled key concepts to
turer, Department of Community Medicine, King's College School of be identified and after checking for redundancy, colloquialisms
Medicine and Dentistry, University of London; S.P. McKenna, Senior
Research Officer, MRC/ESRC Social and Applied Psychology Unit, and ambiguity, the number of statements was reduced to 138.
University of Sheffield. Combinations of these statements were used in a number of
© Journal of the Royal College of General Practitioners, 1985, 35, small and large scale studies between 1976 and 1978, using diverse
185-188. patient populations, and the number of statements was further

Journal of the Royal College of General Practitioners, April 1985 185

S.M. Hunt, J. McEwen and S.P. McKenna Original Papers

Part I oftheprofile. Part 1 of the profile comprises 38 statements

Listed below are some problems people may have in their which met the stringent criteria detailed above and which best
daily life. reflected problems with health. These problems fall into six areas:
Look down the list and put a tick in the box
for any problem you have at the moment.
F
under YES sleep, physical mobility, energy, pain, emotional reactions and
social isolation. The first page of the questionnaire is shown
Tick the box under NO for any problem you do not have. in Figure 1 and this illustrates how the profile statements from
the six areas are randomly distributed. Within each area
Please answer every question. If you are not sure whether to statements have been weighted for severity using the Thurstone
say yes or no, tick whichever answer you think is more true method of paired comparisons, with a sample of 215 members
at the moment. of the general public. Thus the weights reflect the perceived
YES NO severity of the items from the point of view of the patient.23
I'm tired all the time El E] Hunt and McEwen24 give details of the methodology of this
I have pain at night El El stage.
Things are getting me down El El Part 2 of the profile. Part 2 of the profile consists of seven
statements relating to those areas of daily life most often affected
YES NO by health: paid employment, jobs around the house, social life,
I have unbearable pain El LI personal relationships, sex life, hobbies and interests, and
I take tablets to help me sleep El El
holidays.
I've forgotten what it's like to enjoy myself El El On both parts of the profile respondents are required to answer
yes if the statement applies to them and no if it does not. In
YES NO part 1 positive answers are given the appropriate weighting and
I'm feeling on edge El El the higher the score on any section the greater the number and
I find it painful to change position El El
severity of perceived problems in that area. The maximum score
on any section is 100.
I feel lonely El Lii The content and weighting of the statements on the profile
are thus designed to determine and quantify, directly, the distress
YES NO experienced by the respondents in a way which is consistent with
I can only walk about indoors El E everyday life, and lay, rather than professional, values.
I find it hard to bend [I] El The profile has been tested for face, content and criterion
Everything is an effort El El validity25 with diverse groups of people. It has been found to
differentiate successfully between elderly people who do not con-
sult general practitioners, those who are physiologically 'fit' and
Figure 1. The first page of the Nottingham Health Profile. those with chronic illness.26 A comparison of individuals who
consulted their general practitioner more often than average with
reduced to 82. By relating the scores on the questionnaire to those who had had no contact with a doctor in the previous six
medical information and independent assessments of the well- months showed that scores on every section of the profile dif-
bcing of the patients as well as to other standardized measures ferentiated between these two groups with a high level of
such as disability scales, statements were found to be reliable statistical significance. Moreover, days of absence from work
and could be used to distinguish between different degrees of through ill-health were also significantly related to profile scores.
disability; they were also sensitive to changes with time. In ad- Age and sex differences were in agreement with other studies
dition, the statements could be used to distinguish between concerning their effects on perceived health and consultation
physical and mental disorders. 19-22 rates.27 Projects have been undertaken to measure the perceived
In 1978 a further grant was obtained from the Social Science health of men who could be presumed to be in good physical
Research Council to develop the existing instrument into a health - firemen and mine-rescue workers. Section scores on
population survey tool. This meant refining the criteria by which the profile showed that, as expected, both groups had low
statements were chosen for inclusion in the questionnaire. In scores.28,29
particular it was necessary to have statements which were ex- The profile was also used to monitor a group of women
pressed at a reading level understandable by a great majority throughout pregnancy and scores reflected those physical and
of the population, which were short and easy to answer, and emotional changes which are well authenticated in the literature.
the meaning of which was commonly understood. Statements In addition, an evaluation of the effect of minor surgery was
which met these standards were tested on patient and non-patient carried out30 and an assessment of the effects of a fractured
groups and those statements which proved satisfactory were limb on patients and their families.3' Table 1 shows some com-
retained. parative scores for Part 1 of the profile.
Table 1. Mean scores on Part 1 of the Nottingham Health Profile for selected groups.
Pregnant women Patients Patients
at with with Patients
Mine- minor peripheral with
rescue 'Fit' 18 37 non-acute Fracture vascular Chronically osteo-
workers elderly weeks weeks conditions victims disease ill elderly arthrosis
Energy 1.0 4.1 31.4 39.6 24.2 25.8 30.3 38.0 63.2
Pain 1.4 1.1 2.1 11.2 15.9 26.5 22.6 29.2 70.8
Emotional reactions 1.3 3.3 15.7 15.7 14.7 13.7 13.9 15.1 21.3
Sleep 4.2 0.7 11.3 28.3 18.7 28.0 24.7 32.1 48.7
Social isolation 0.4 1.3 6.4 6.2 5.1 8.0 9.2 12.8 12.5
Physical mobility 0.5 1.9 7.3 26.0 7.3 27.6 22.0 29.2 54.8

186 Journal of the Royal College of General Practitioners, April 1985

S.M. Hunt, J. McEwen and S.P. McKenna Original Papers

To test the reliability of the instrument it was necessary to to social and economic stresses among younger people in lower
find groups of people who could be expected to give fairly con- socioeconomic groups, with some adaptation and resignation
sistent responses gathered on two different occasions, where the occurring after middle age.35
time lag between test and re-test was long enough to avoid con- This study indicated that the profile would be of value as an
tamination from one occasion to the other. Accordingly, two epidemiological tool as well as being an aid to clinical evaluation.
groups of patients were chosen, one group with osteoarthrosis32
and a second group with peripheral vascular disease.33 In both Limitations of the profile
cases little change would be expected in the objective condition The Nottingham Health Profile has some limitations of which
of such patients over the projected four weeks between prospective users should be aware. The items on Part 1 repre-
administrations of the questionnaire. A successful postal survey sent rather severe problems. This was found to be necessary in
was carried out using the test-re-test method and gave high cor- order to avoid picking up large numbers of false positives.
relation coefficients between the two sets of scores for both However, it does mean that some milder forms of distress may
groups, as shown in Tables 2 and 3. These studies indicate that not show up on the profile. Members of 'normal' populations,
the profile is a valid and reliable indicator of subjective health or those with minor ailments may affirm very few statements.
status in physical, social and emotional areas. This makes it difficult to compare their scores or to evaluate
change. In pre- and post-intervention studies improvement in
Table 2. Reliability coefficients (Spearman's r) for each area of Part condition for those who score zero on the first occasion cannot
1 of the Nottingham Health Profile.
be demonstrated on the profile. Scoring on Part 1 involves six
Patients with outcomes plus a further seven scores if Part 2 is used. Analysis
Patients with peripheral can, therefore, become cumbersome if large numbers of other
osteoarthrosis vascular variables need to be taken into account. The profile investigates
Area concerned (n = 58) disease (n = 93) negative aspects of health only, since all the items refer to prob-
Energy 0.77 0.77 lems. Therefore, it cannot be used to assess positive feelings of
Pain 0.79 0.88 well-being, as zero scores do not necessarily indicate a total
Emotional reactions 0.80 0.75 absence of distress.
Sleep 0.85 0.85
Social isolation 0.78 0.77 Advantages and uses of the profile
Physical mobility 0.85 0.79
The profile has some important advantages. It is sensitive to
All correlation coefficients are significant (P <0.01). change with time and different patterns of scores can be a useful
indication of particular problems being experienced by patients.
Table 3. Reliability coefficients (Cramer's o) for each statement in
For example, in a study of cancer patients, although pain and
Part 2 of the Nottingham Health Profile. physical mobility scores were relatively low as a consequence
of symptom-relieving medication, the profile recorded high levels
Patients with of emotional distress and sleep disturbance which indicated the
Patients with peripheral need for a more psychological type of intervention. On the other
Area concerned in osteoarthrosis vascular hand, scores on pain and physical mobility may be high, while
each statement (n = 58) disease (n = 93) sleep and emotional distress scores are low, suggesting that the
Paid employment 0.86 0.55 patient is able to adapt to his illness without excessive anxiety.
Jobs around the home 0.85 0.64 A recent study of heart transplants has shown that the pro-
Social life 0.59 0.61 file could be a useful adjunct in evaluative studies of the cost
Family relationships 0.64 0.89 of medical interventions.36 The profile is currently being used
Sex life 0.84 0.85 in studies on stroke, myocardial infarction, cancer and multiple
Hobbies/interests 0.44 0.86 sclerosis as an outcome measure.
Holidays 0.71 0.72
From an epidemiological standpoint the profile can be used
All correlation coefficients are significant (P <0.01). to record patterns of perceived health in a community. Its most
recent use in this respect was in a survey of health in London,
A population survey where it was found to be a highly reliable and satisfactory in-
Having established the validity and reliability of the profile it dicator of variations in health experiences.37 A project is cur-
was necessary to gauge its usefulness and acceptability as a survey rently under way to examine patterns of perceived health among
tool. Previous studies using it as a postal questionnaire had the unemployed.38
yielded response rates ranging from 72 per cent to 93 per cent, From a clinical perspective, as the principal problems of our
but these were from highly motivated patient groups. Follow- time tend to be those of a chronic and intractable kind, the treat-
ing the Black Report,34 a study was set up to use the profile to ment of which may be associated with a variety of side-effects,
examine social class differentials in perceived health by taking some assessment of the quality of life of the patients and their
a random sample from the records of a group practice in Not- levels of distress and discomfort would seem to be a vitally
tingham. Questionnaires, together with a covering letter and pre- necessary addition to the usual outcome measures both in clinical
paid reply envelope, were posted to 3200 patients. trials and in assessing needs for counselling and support.
A response rate of 68 per cent was obtained. Results showed The use of the profile in epidemiological studies is to provide
that social class differentials in perceived health reflect overall information, not readily available in routinely collected statistics,
patterns of morbidity as calculated from routinely collected vital concerning the experience of people at the community level.
statistics. Younger people (aged 20-44 years) of both sexes in Morbidity surveys are time-consuming and expensive, but the
social classes 4 and 5 achieved significantly higher scores on profile can provide a cheap, quick and easy means of assessing
emotional reactions, sleep, social isolation and energy than did those experiences and effects on daily life which are known to
respondents of the same age in classes 1, 2 and 3. These social be associated with the demand for services.
class differences were not significant in older age groups. These The profile questionnaire takes only a few minutes to com-
findings were interpreted as suggesting a greater vulnerability plete and is acceptable and understood by a majority of

Journal of the Royal College of General Practitioners, April 1985 187

S.M. Hunt, J. McEwen and S.P. McKenna Original Papers
respondents. The statements are easy to score and compute and 18. Culyer AJ. Need, value and health status measurement. In:
are particularly suited for analysis using the Statistical Package Economic aspects of health services. Culyer AJ, Wright KS,
(eds). York: Martin Robertson, 1978.
for the Social Sciences and other statistical packages. Graphical 19. Martini CHM, McDowell I. Health status: patient and physi-
presentation of profile scores aids assessment of specific areas cian judgements. Health Serv Res 1976; 11: 508-515.
of dysfunction. 20. McDowell I, Martini CHM. Problems and new directions in
Since the profile does not ask directly about symptoms, it is the evaluation of primary care. Int J Epidemiol 1976; 5:
more likely to identify people who are in distress or at risk, but 247-250.
who do not see their problems being specifically related to health. 21. Martini CHM, McDowell I. The evaluation of strategies for the
The profile can be used to measure general perceived health improvement of life style: an example from two locomotor
disorders. Paper read at a conference of the International
status or specific conditions of ill-health. Above all, it provides Epidemiological Association, Puerto Rico, September 1977.
a measure of the perceptions of patients and thus can be regarded 22. McDowell I, Martini CHM, Waugh W. A method for self-
as a direct reflection of need and possible demand, and an assessment of disability before and after hip replacement
accurate guide to the efficacy of health care in affecting how operations. Br Med J 1978; 2: 875-879.
people feel. It has proved its usefulness in a wide variety of 23. McKenna SP, Hunt SM, McEwen J. Weighting the seriousness
of perceived health problems using Thurstone's method of
medical and non-medical settings and with a wide range of paired comparisons. Int J Epidemiol 1981; 10: 93-97.
clinical conditions. In addition, it may be of value in the ex- 24. Hunt SM, McEwen J. The development of a subjective health
ploration of theoretical aspects of the relationship between indicator. Social Health and Illness 1980; 2: 231-246.
pathology and phenomenology. 25. Backett EM, McEwen J, Hunt SM. Health and quality of life.
Applied appropriately, the Nottingham Health Profile pro- Report of the Social Science Research Council. London: SSRC,
vides a much needed additional tool for clinical and 1981.
epidemiological research. 26. Hunt SM, McKenna SP, McEwen J, et aL A quantitative
approach to perceived health status: a validation study. J
Epidemiol Community Health 1980; 34: 281-286.
References 27. Hunt SM, McKenna SP, McEwen J, et al. The Nottingham
1. Abel-Smith B. Value for money in health services. London: Health Profile: subjective health status and medical consulta-
Heinemann, 1976. tions. Soc Sci Med 1981; 15A: 221-229.
2. World Health Organization. Statistical indicators for the plan- 28. McKenna SP, Hunt SM, McEwen J, et al. Looking at health
ning and evaluation of public health programmes, Tech Rep from the consumer's point of view. Occupational Health 1980;
Series 472. Geneva: WHO, 1971. 32: 350-355.
3. White K. Contemporary epidemiology. Int J Epidemiol 1976; 29. McKenna SP, Hunt SM, McEwen J. Mine-rescue workers: their
3: 295-301. perceived health and absence from work. Occupational Health
4. MacKenzie J. The basis of vital activity. London: Faber and 1981; 33: 70-74.
Gwyn, 1926. 30. Hunt SM, McKenna SP, McEwen J, et al. Subjective health
5. Singer E, Garfinkel R, Cohen SM, Srole L. Mortality and assessments and the perceived outcome of minor surgery. J
mental health: evidence from the midtown Manhattan restudy. Psychosom Res 1984; 28: 105-114.
Soc Sci Med 1976; 10: 517-521. 31. McKenna SP, McEwen J, Hunt SM, et al. Changes in the
6. National Heart and Lung Institute. Proceedings of the Heart
perceived health of patients recovering from fractures. Public
Health 1984; 98: 97-102.
and Lung Institute Working Conference on Health Behaviour. 32. Hunt SM, McKenna SP, Williams J. Reliability of a population
Weiss SM (ed). Pub. No. (NIH) 76-868. New York: DHEW, survey tool for measuring perceived health problems: a study
1976. of patients with osteoarthrosis. J Epidemiol Community
7. Palmore F, Luikart C. Health and social factors related to life Health 1981; 35: 297-300.
satisfaction. J Health Soc Behav 1972; 13: 68-73. 33. Hunt SM, McEwen J, McKenna SP, et al. Subjective health of
8. Maddox G, Douglas EB. Aging and individual differences: a patients with peripheral vascular disease. Practitioner 1982;
longitudinal analysis of social, psychological and physiological 226: 133-136.
indicators. J Gerontol 1975; 29: 555-563. 34. Department of Health and Social Services. Inequalities in
9. Bergner M, Bobbitt RA, Kressel S, et al. The Sickness Impact health. Report of a research working group. Black D (Chair-
Profile: a conceptual formulation and methodology for the man). London: DHSS, 1980.
development of a health status measure. Int J Health Serv 35. Hunt SM, McEwen J, McKenna SP. Perceived health and social
1978; 6: 393-415. inequalities. Effective Health Care 1985; in press.
10. Broadhan K, Erdmann AH, Lorge I, Wolfe HG. The Cornell 36. Buxton M. The economics of heart transplant programmes:
Medical Index and Health Questionnaire as a diagnostic tool. measuring the benefits. In: Measuring the social benefits of
JAMA 1951; 145: 152-157. medicine. Teeling Smith G (ed). London: Office of Health
11. Chambers LW, Segovia J. Indexes of health: lay and pro- Economics, 1983.
fessional perspectives of physical, social and emotional 37. Curtis S. Intra-urban variations in health and health care: the
function. Paper presented at the sociology of health care comparative need for health care survey of Tower Hamlets and
session of the learned society meetings in London Ontario, Redbridge Health Research Group. London: Queen Mary Col-
May 1978. lege, 1984.
12. Wan TTH, Liveratos B. Interpreting a general index of subjec- 38. McKenna SP, Fryer D. Perceived health during lay-off and
tive wellbeing. Milbank Mem Fund Q 1978; 56: 531-556. early unemployment. Occupational Health 1984; 36: 201-206.
13. Grogono AW, Woodgate DI. Index for measuring health.
Lancet 1971; 2: 1024-1026.
14. Rosser R, Watts V. The development of a classification of
symptoms and sickness and its use to measure the output of a Acknowledgements
hospital. In: Impairment, disability and handicap. Lees D, This research was supported by grant number HR 6157/1 awarded by
Shaw S (eds). London: Heinemann for SSRC, 1974. the Social Science Research Council. The work was carried out in the
15. Kaplan RM, Bush JW, Berry CC. Health status: types of Department of Community Health, University of Nottingham.
validity and the index of well-being. Health Serv Res 1976; 11:
478-507.
16. Ware JE. Scales for measuring general health perceptions.
Health Serv Res 1976; 11: 396-415. Address for correspondence
17. Rosser RM, Kind P. A scale of valuations of states of illness: a Dr J. McEwen, King's College Hospital School, Department of Com-
social consensus. Int J Epidemiol 1978; 7: 347-357. munity Medicine, Denmark Hill, London SE5 8RX.

188 Journal of the Royal College of General Practitioners, April 1985