Journal of Public Health Medicine
Vol. 22, No. 3, pp. 393±399
Printed in Great Britain
Disabled people and professionals differ in
their perceptions of rehabilitation needs
Paula Kersten, Steve George, Lindsay McLellan, Jenifer A. E. Smith and
Mark A. Mullee
Abstract
Background This study set out to explore whether there are
systematic differences in unmet needs for rehabilitation
services and resources as identi®ed by disabled people and
nominated key professionals.
Methods A cross-sectional interview survey of 87 pairs of
community-dwelling disabled people (aged 16±65) and their
nominated key professionals was conducted in southern
England. The Southampton Needs Assessment Questionnaire (SNAQ) was used to examine unmet needs and the
Of®ce of Population Censuses and Surveys (OPCS) Disability
and Severity Scale to examine level of disability.
Results Eighty-seven pairs of disabled people and their
nominated key professionals participated. People were
severely disabled (median OPCS score 8; interquartile
range 6±9). Disabled participants reported signi®cantly
more unmet needs than did professionals. There was zero
concordance on identi®ed unmet needs between disabled
participants and professionals in 56 per cent of cases.
Concordance between disabled participants and professionals was signi®cantly better if the disabled participant
had reported more unmet needs. There was a trend for
general practitioners to be poorer at reporting disabled
participants' needs than other professionals. There was fair
agreement between disabled participants and professionals
in ®ve areas of unmet need only (in descending order):
adaptations, physiotherapy, equipment, assistance with
`non-care' activities and the use of a day centre. In all the
other areas of unmet need the agreement was poor.
Conclusions Needs for rehabilitation services and resources
(for disabled people) are perceived differently by disabled
people and professionals. Until we know who is right about
rehabilitation needs, it is important to determine both users'
and providers' views when planning and commissioning
services.
Keywords: rehabilitation needs, disabled people, professionals, Southampton Needs Assessment Questionnaire
Background
The Community Care Act 1990 introduced needs led assessments, rather than service led assessments, to decide on the
services required to enable elderly and disabled people to live as
independently as possible at home.1 The de®nition of need,
however, was linked to the availability of resources, which has
meant that needs depend on priorities.2 The relative priority of
need was to be assessed according to eligibility criteria
measuring levels of `risk' and `dependence'. It has been
estimated that the prevalence of disability in Great Britain is 14
per cent for people aged 16±65.3 Considering these large
numbers and the greater emphasis on community care since the
1990s, it might be expected that many community-based
rehabilitation services would be oriented towards meeting
disabled people's objectives and needs. There is some evidence
from the literature that this is not the case, however. In a 1982
Kingston-upon-Thames study general practitioners (GPs) knew
about fewer than 50 per cent of the dif®culties with activities of
daily living that were reported by 37 per cent of disabled
people.4 Unmet needs have also been found for domestic help
(19±54 per cent), holidays or convalescence (14 per cent), extra
aids (18 per cent) and advice on bene®ts (30 per cent).5±7 What
is not certain from work performed so far, however, is whether
such differences are random in nature, or more systematic. This
study set out to explore whether there are systematic differences
in unmet needs for rehabilitation services and resources as
identi®ed by disabled people and nominated key professionals.
Methods
A cross-sectional interview study of disabled people (aged 16±
65) was carried out in Basingstoke and Southampton National
Health Service (NHS) Health Districts in southern England.
1
School of Health Professions and Rehabilitation Sciences, University of
Southampton, High®eld, Southampton SO17 1BJ.
2
Health Care Research Unit, University of Southampton, Mailpoint 805,
Southampton General Hospital, Southampton SO16 6YD.
3
Southampton University Hospital Trust, Highcroft, Romsey Road, Winchester
SO22 5DH.
Paula Kersten,1 Postdoctoral Research Fellow, Health Research Unit
Steve George,1 Senior Lecturer in Public Health Medicine and Director
Lindsay McLellan,1 Europe Professor of Rehabilitation, Rehabilitation
Research Unit
Jenifer A. E. Smith, Consultant in Public Health Medicine2 (now Director,
South and West Cancer Intelligence Unit3)
Mark A. Mullee,2 Senior Research Fellow in Medical Statistics
Address correspondence to Paula Kersten.
E-mail: [email protected]
q Faculty of Public Health Medicine 2000
394
JOURNAL OF PUBLIC HEALTH MEDICINE
People with a primary physical disability, who were living at
home, were included provided that their GPs had given their
consent to invite their patients to participate. People with a
primary disability of blindness or deafness were excluded.
Table 1 Topics of discussion in the Southampton Needs
Assessment Questionnaire
Sampling frame
Disability
Medical condition and anticipated progression
Effect of disability on day-to-day life
Household situation
Marital status
Children
Who the participant lives with
Social and intellectual ful®lment
Employment
Social activities
Going out
Hobbies and recreation
Care situation
Care received or not
What activities carers or family members help with
Services
NHS
Social care services
Private
Disability organizations
Membership
The kind of services these organizations provide to the participant
Home
Suitability
Adaptations
Hardware
Equipment
Wheelchairs
Mobility
Personal mobility (e.g. falls, use of walking aids)
Transport
Community accessibility
Finances
Bene®ts received
Whether all bene®ts to which the participant is entitled are received
Whether more money would help the participant to overcome some
of the effects of his or her disability
Information
Condition
Treatments or care plans
Health care services
Social care services
Rehabilitation services
Sexual relationships or dif®culties
We used two existing disability registers, in Basingstoke and
Southampton, to randomly select disabled people.8 These
registers comprised disabled people, aged 16±65, who had been
identi®ed by community rehabilitation services as being in need
of regular surveillance by formal assessment of their care needs.
These registers must be distinguished, therefore, from those
comprising self-identi®ed individuals.
During the interview, participants were asked to identify the
main health or social care professional who they expected to be
most knowledgeable about their needs. If they were unable to
identify a professional in answer to the above question, they
were asked whether their GP would know them well enough to
identify their needs. The GP therefore acted as a default option.
Those who did not think that their GP would know them well
enough were excluded. A `case' is de®ned in this paper as a
disabled person plus a nominated professional.
Needs assessment
In the absence of an existing valid needs assessment research
tool, a new semi-structured needs assessment questionnaire was
developed during a pilot study (the Southampton Needs
Assessment Questionnaire, SNAQ). It was shown to have
good content and construct validity during the pilot. Its
development and validation will be published separately. The
SNAQ covers topics identi®ed as important to disabled
participants in the pilot study (Table 1). The instrument was
administered to disabled people by face-to-face interview at
home, following consent from the disabled person. If disabled
people had very poor communication skills, or objected to
being interviewed on their own, they were interviewed together
with their carer. During this process of questioning it was
possible to derive those needs that were met and thus enable
participants to think about their unmet needs. At the end of the
interview participants were asked to list all their unmet needs in
order of priority. Level of disability was measured using the
Of®ce of Population Censuses and Surveys (OPCS) Disability
Severity Scale.3
Professionals were interviewed over the telephone. The
questionnaire used for professionals was identical to that used
for disabled people, with the exception that questions were
altered to re¯ect the professional's perspective.
Statistical analysis
The numbers of unmet needs reported by participants did not
follow a normal distribution and the analysis was therefore
conducted using non-parametric statistics. Comparisons
between disabled participants and their professionals on the
Topics discussed in the Southampton Needs Assessment
Questionnaire
numbers of needs reported were made with Wilcoxon Signed
Rank tests. Concordance between participants about the nature
of unmet needs was established as follows. First, the proportion
of unmet needs identi®ed by the disabled person that was also
identi®ed by his or her professional was calculated. Thus, a
concordance score (CS) ranging from zero to one could be
obtained (where zero indicates none of the unmet needs
reported by the disabled person; one indicates all of the unmet
needs reported by the disabled person). In a large proportion of
cases CS equalled zero, however, and we therefore grouped
395
PERCE PTIONS OF RE HABILITATION NEE DS
cases into two categories: cases in which CS was `moderate to
good' (CS > 0), and cases in which CS was `poor' (CS ˆ 0).
We hypothesized that concordance scores between disabled
participants and their professionals would be higher if the
disabled person was older, male, had been disabled longer or
had a progressive deteriorating condition. In addition, we
hypothesized that concordance scores would be lower if the
disabled person had poorer communication skills (OPCS),
poorer intellectual functioning (OPCS), or had identi®ed a GP
as their key professional. Mann±Whitney U-tests and x2 tests
were used to test these hypotheses. We also explored whether
there were systematic differences in the magnitude of selfreported need between the two groups of `poor' and `moderate
to good' CS.
Second, agreement between disabled participants and their
professionals on the nature of the reported unmet needs was
established with kappa9 statistics (and 95 per cent con®dence
intervals of kappa) after dichotomizing the data into whether or
not a person reported an unmet need of a certain type. For this
purpose needs were grouped.
Ethical approval for the study was granted by the local ethics
committees. Baseline data were collected in 1995.
Results
Participation and case characteristics
A random sample of 154 disabled people, who ®tted the
inclusion criteria and for whom GP consent had been granted,
was selected from the disability registers. A postal invitation to
take part in the study was accepted by 96 disabled people. Forty
people declined, and 18 did not respond to the invitation. Out of
the participating 96 disabled people, it was realized at interview
that three had such severe communication dif®culties that they
were unable to communicate their unmet needs effectively and
they were therefore excluded, leaving a potential 93 cases in the
study. Participants and non-participants were similar in age
(median [interquartile range (IQR)] 48 years (39±55) versus 48
(36±56), Mann±Whitney U (MWU) 2035, p ˆ 0:72), gender
(male:female 36:57 versus 27:31, x2 ˆ 0:90, p ˆ 0:34) and
diagnosis (Table 2). The largest single diagnostic group
comprised people with multiple sclerosis.
It now remained to match disabled people to nominated
professionals. Two disabled people could not identify a health
or social care professional who was knowledgeable enough
about their needs to take part, including the GP default, and
were excluded. One nominated professional ceased employment before being available for interview, and three further
nominated professionals declined participation. There were
thus 87 cases where both the disabled person and the
professional consented to take part. The level of disability
was severe in this group of people as measured with the OPCS
Disability and Severity Scale (median 8.0; IQR 6.0±9.0). The
disciplines of the professionals varied greatly (Table 3), the
Table 2 Main diagnosis of participants and non-participants
(numbers, with percentages given in parentheses)
Condition
Multiple sclerosis
Rheumatoid arthritis
Stroke
Traumatic brain damage
Cerebral palsy
Paralysis following neurosurgery
Traumatic spinal cord injury
Motor neurone disease
Poliomyelitis
Brain damage ± from early childhood
Above knee amputation
Back pain
Charcot Marie Tooth syndrome
Friedrich's ataxia
Spina bi®da
Subarachnoid haemorrhage
Brain damage ± after cardiac arrest
Epilepsy
Diabetes mellitus
Cortical atrophy with hydrocephalus
Facio scapulo humeral dystrophy
Spastic paraplegia after measles
Transverse myelitis-tetraplegia
Undiagnosed neurological disorder
Arthrogryphosis
Congenital hydrocephalus
Knee arthrodesis following trauma
Ankylospondylitis
Joint hypermobility syndrome
Epidermiolysis bullosa dystrophica
Hysterical paraplegia
Cerebellar degeneration
Muscular dystrophy
Osteoarthritis
Brain tumour
Cerebellar ataxia
Hydrocephalus
Myasthenia gravis
Fractured pelvis
Congenital bilateral hip dislocation
Sarcoidosis
Meningitis
Spinal carcinoma
Thalidomide
Missing
Non-participants
(n ˆ 70)
13
6
12
6
3
(19)
(9)
(9)
(4)
3 (4)
1 (1)
1 (1)
6 (9)
1 (1)
3
2
2
1
1
1
1
1
1
1
1
1
1
1
(4)
(3)
(3)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
Participants in
main study
(n ˆ 93)
28 (30)
7 (8)
6 (6)
5 (5)
5 (5)
4 (4)
3 (3)
3 (3)
3 (3)
3 (3)
2 (2)
2 (2)
2 (2)
2 (2)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
largest single group comprising GPs. Out of the 24 GPs, 11 had
been chosen only after the disabled person could not
spontaneously think of a professional who would be knowledgeable about his or her needs.
Needs
The most frequently used services were those of the GP (85),
dentist (63), optician and hospital consultant (55), and
occupational therapist (39). Disabled people identi®ed signi®cantly more unmet needs than did their key professionals
(Table 4). A large proportion of participants had multiple
sclerosis, and potentially limited the generalizability of the
396
JOURNAL OF PUBLIC HEALTH MEDICINE
Table 3 Discipline of participating professional (n ˆ 87)
(numbers, with percentages given in parentheses)
Discipline of professional
Frequency
General practitioner
Nurse
Day centre employee
Physiotherapist
Social worker
Occupational therapist
Care attendant scheme employee
Health care services manager physical disabilities
Rehabilitation medicine consultant
Speech therapist
Homecarer
Administrator hyperbaric oxygen
Aromatherapist
Head Cheshire Home
Project worker James Society
(alcoholics and homeless people)
Project worker Stonham Housing Association
(people with disabilities)
24 (28)
10 (11)
10 (11)
8 (9)
8 (9)
7 (8)
6 (7)
6 (7)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
1 (1)
Table 4 Number of unmet needs of disabled people as
reported by disabled people and their professionals
Number of needs
Disabled people
Total
Median
Interquartile range
Range
Wilcoxon signed rank test
351
195
3
2
2±7
1±3
0±12
0±6
Z ˆ 4:77, p < 0:001
hypothesized case characteristics (Table 6). There was a trend
that GPs were poorer than other professionals at identifying
disabled participants' needs [difference in proportion ±20 per
cent, 95 per cent con®dence interval (CI) ±42 per cent to +2 per
cent]. Concordance scores of zero between disabled participants and the 11 `default GPs' occurred in 64 per cent of cases
and between disabled participants and the other 13 GPs in 77
per cent of cases. The concordance scores between these two
groups were not signi®cantly different, however (x2 ˆ 0:07,
p ˆ 0:79).
An exploration of the magnitude of self-reported need
showed that CS were signi®cantly better if the disabled
participant had reported more unmet needs (Table 6).
Table 7 shows that more disabled participants than
professionals identi®ed unmet needs for adaptations, information, wheelchairs, hospital consultant, physiotherapy, equipment and chiropody. By comparison, more professionals than
disabled participants reported unmet needs for improved social
life and improved intellectual ful®lment.
Kappa statistics showed that there was fair agreement in ®ve
areas of unmet need only (in descending order): adaptations,
physiotherapy, equipment, assistance with `non-care' activities
and the use of a day centre. In all the other areas of unmet need
the agreement was poor (Table 7).
Professionals
study. To check whether this was the case, we compared the
number and types of unmet needs reported by them with those
reported by other disabled participants: there were no
signi®cant differences between them [median (IQR) 3 (2±5)
versus 3 (2±7), MWU ˆ 844:0, p ˆ 0:58].
Concordance scores between disabled people and their
professionals were low, with a score of one (complete
agreement) being obtained in four cases only (Table 5).
The two groups of disabled participants and their professionals (those with poor and those with moderate to good
concordance) did not differ signi®cantly on any of the
Table 5 Unmet needs: concordance scores between disabled participants and their professionals (numbers, with
percentages given in parentheses)
Concordance scores (categorized)
Frequency
Concordance score ˆ 1
0 < concordance score < 1
Concordance score ˆ 0
4 (4)
34 (39)
49 (56)
Discussion
An important ®nding in this study was the discrepancy between
the needs and priorities identi®ed by disabled people in face-toface interviews and the identi®ed key professionals in telephone
interviews. Could this difference have been due to the different
method employed? We consider this to be unlikely, as the
wording of the questions was virtually identical. In addition,
this method made it easier for professionals to consult their
notes while responding to questions.
The different perceptions of needs for rehabilitation services
and resources may be explained in several ways. First and
foremost, the subjective experience of disability is different for
each disabled person, carer and professional.10 However, this,
although a cause of random variation, should not cause
systematic differences between groups.
Second, however, there is no clear consensus on the
de®nition of `need'. Need has been de®ned as `the ability to
bene®t from health care'11,12 but, conversely, it can be argued
that care is the antithesis of rehabilitation.13 `Need', therefore,
may be dependent upon the prior experiences and, if viewed in
the context of the health professional, the training of the
observer. The higher levels of concern experienced by
professionals about personal ful®lment and meaningful occupation of time, for instance, might re¯ect an unduly pessimistic
perception of the sources of stress and of satisfaction in the lives
of disabled people. This tends to support Oliver's view that
professionals identify and pursue goals for disabled people that
are inappropriate and possibly oppressive.14
397
PERCE PTIONS OF RE HABILITATION NEE DS
Table 6 Case characteristics and level of concordance about disabled participants' unmet needs
between disabled participants and their professionals
Case characteristics
Continuous data (median, range)
Age
Communication (OPCS)
Intellectual functioning (OPCS)
Duration of disability (years)
Number of unmet needs reported
by disabled participant
Nominal data
Male
Female
Stable or improving condition
Deteriorating condition
Professional is a GP
Professional is not a GP
Statistical
comparison
Zero
concordance
(CS ˆ 0, n ˆ 49)
Moderate to good
concordance
(CS > 0, n ˆ 38)
...................................
MWU
p value
48.0 (21±64)
0 (0±8.5)
0 (0±13.0)
17.0 (1±50)
48.0 (28±64)
0 (0±8.5)
0.5 (0±13.0)
14.5 (1±50)
804.0
910.0
877.0
895.0
0.37
0.83
0.61
0.76
465.5
<0.001
X2
0.001
p value
0.97
0.21
0.65
2.84
0.09
2 (0±12)
%
57.6 (n
55.6 (n
53.1 (n
58.2 (n
70.8 (n
50.8 (n
ˆ 19)
ˆ 30)
ˆ 17)
ˆ 32)
ˆ 17)
ˆ 32)
5 (0±12)
%
42.4 (n
44.4 (n
46.9 (n
41.8 (n
29.2 (n
49.2 (n
ˆ 14)
ˆ 24)
ˆ 15)
ˆ 23)
ˆ 7)
ˆ 31)
CS, concordance score; MWU, Mann±Whitney U.
A third explanation arises from the heterogeneity of the
disabled population.15 Disabilities differ in individuals depending upon their circumstances and their objectives. The same
impairments call for different solutions, therefore, depending
upon their context. For this reason, rehabilitation needs are
impossible to derive from the prevalence of disabling diseases
or even of speci®c impairments. The political and organizational changes brought about by the Community Care Act,1
however, encourage assessors to regard needs in terms of
individual impairment rather than as socially created discrimination. We can expect a mismatch between needs as viewed by
disabled people and practitioners, therefore, merely as a result
of an underlying political agenda.2
A fourth explanation relates to the fact that rehabilitation
services, required to meet disabled people's needs, often cross
the boundaries of traditionally segregated services, such as the
NHS Trusts, local authorities and GP fund-holders. Rehabilitation needs, at an organizational level, may be viewed as a
function of what each organization has traditionally supplied.
Failure of communication between these services, the dif®culties of co-ordinating them and the fact that funding is separate
compounds these problems, however.16
Does this study re¯ect reality?
We did not seek to identify the appropriateness of unmet needs
identi®ed in this study, nor to negotiate an agreement between
differing views. In clinical practice unmet needs are assessed
together with the disabled person (and the carer) and in some
cases with the involvement of a multi-disciplinary team. The
latter does not often happen in real practice, however, and the
method employed here re¯ects more accurately current
practice. This may result in better concordance than that
found here. Concordance between disabled people and professionals was, however, very poor. The fact that concordance
scores were much better when the disabled participant had
reported more unmet needs suggests that professionals were
more in tune with people who were, at least in their own
opinion, more needy. The only other trend predicting poor
concordance was the choice of a GP as the participating
professional. Some GPs were identi®ed as a default option
because participants had not spontaneously been able to
identify another professional who they felt knew their needs.
This may have placed those GPs at a disadvantage. There were,
however, no differences in concordance scores between default
GPs and others. Whether or not our methods led to some
professionals being disadvantaged in this study, GPs will take a
leading role in the purchase of rehabilitation services in the
future. In the past the allocation of health funding has been
caricatured as `the same as last year, plus or minus ®ve percent
for pressure groups'.17 Primary care groups will take over the
role currently occupied by health authorities in undertaking
needs assessment and commissioning care for many services.18
So what should primary care groups do to ensure that they best
represent the views of disabled people? Most obviously,
perhaps, is to ensure that disabled people's views are taken
into account when shaping rehabilitation services. Secondly,
recent work has shown that different professions also have
differing priorities for commissioning of services.19 Given the
current structure of many primary care groups, it is not clear
how the views of professions other than medicine will be
articulated, and it is for primary care groups to resolve how the
398
JOURNAL OF PUBLIC HEALTH MEDICINE
Table 7 Agreement between disabled participants and their nominated key professionals on the presence
of unmet needs
Disabled person
reported the need
Disabled person did
not report the need
Statistical comparison
..........................
............................
...............................
Unmet need of
disabled person
Prof
did
Prof
did not
Prof
did
Adaptations
Assistance with activities
(not care)
Care
Changes in the community
Chiropody
Complementary medicine or
therapy
Counselling
Day centre
Domestic help
Equipment
Finances
Holidays
Information
Intellectual ful®lment
Occupational therapy
Other
Other medical services
Other professionals allied to
medicine
Personal issues
Physiotherapy
Respite care + short breaks
Social life
Social worker
Specialist
Speech and language therapy
Vehicles
Wheelchairs
17
25
4
1
0
0
Prof
did not
kappa*
95% CI
of kappa
5
40
0.30
0.10±0.50
11
10
16
6
6
7
6
0
66
69
65
81
0.21
0
0
0
0±0.55
0±0.43
0±0.29
0±0.77
0
0
1
2
7
1
0
2
1
0
0
1
2
3
4
10
18
11
2
18
7
4
4
9
2
9
2
5
5
4
6
2
18
0
0
3
83
75
80
70
57
71
79
65
61
83
83
74
0
0
0.22
0.12
0.24
0.04
0
0.10
0
0
0
0.08
0±0.96
0±0.50
0±0.82
0±0.53
0±0.50
0±0.48
0±0.65
0±0.44
0±0.29
0±0.96
0±0.96
0±0.56
1
1
6
0
2
0
0
1
3
2
10
4
16
2
4
3
10
6
6
20
3
10
3
5
16
1
0
1
9
4
73
72
62
80
65
83
77
79
69
61
0.07
0.05
0.28
0
0.07
0
0
0.19
0.19
0.04
0±0.54
0±0.51
0±0.57
0±0.70
0±0.43
0±0.96
0±0.58
0±0.77
0±0.56
0±0.37
*Strength of agreement indicated by value of kappa: poor (<0.20), fair (0.21±0.40), moderate (0.41±0.60), good (0.61±0.80), very
good (0.81±1.00).10
views of both professionals and patients are taken into account
when commissioning services.
3 Martin J, Meltzer H, Elliot D. OPCS surveys of disability in Great
Britain. Report 1. The prevalence of disability among adults. London:
HMSO, 1988.
Acknowledgements
4 Patrick DL, Peach H, Gregg I. Disablement and care: a comparison of
patient views and general practitioner knowledge. J R Coll Gen Pract
1982; 32: 429±434.
We would like to extend our thanks to the NHS Research and
Development Programme South and West for funding this
project, and to the subjects and their professionals for
participating in the study. The views expressed, however, are
those of the authors alone.
References
1 Department of Health. Community care in the next decade and beyond.
Caring for the 1990s. London: HMSO, 1989.
2 Audit Commission. The community revolution: the personal social
services and community care. London: HMSO, 1992.
5 Badley EM, Tennant A. A survey of disablement in a British population
using an action-orientated measure, physical independence handicap:
problems with activities of daily living and level of support. Int Disab
Stud 1991; 13: 91±98.
6 Badley EM, Tennant A. Impact of disablement due to rheumatic
disorders in a British population: estimates of severity and prevalence
from the Calderdale Rheumatic Disablement Survey. Ann Rheumat Dis
1993; 52: 6±13.
7 Badley EM, Tennant A. Disablement associated with rheumatic
disorders in a British population: problems with activities of daily living
and level of support. Br J Rheumatol 1993; 32: 601±608.
8 Lovelock R. Friends in deed: three care attendant schemes for the
PERCE PTIONS OF RE HABILITATION NEE DS
399
younger physically disabled in Hampshire. Portsmouth: Social Services
Research and Intelligence Unit, 1981.
14 Oliver M. Understanding disability. From theory to practice. London:
Macmillan, 1996.
9 Altman DG. Some common problems in medical research. In: Practical
statistics for medical research. London: Chapman and Hall, 1991: 403±
409.
15 Warren MD. The prevalence of disability. J R Coll Phys Lond 1989; 23:
171±175.
10 Peters DJ. Disablement observed, addressed, and experienced: integrating subjective experience into disablement models. Disab Rehab
1996; 18: 593±603.
11 Stevens A, Gabbay J. Needs assessment needs assessment... Hlth Trends
1991; 23: 20±23.
12 Stevens A, Raftery J. Introduction. In: Stevens A, Raftery J, eds. Health
care needs assessment. The epidemiology based needs assessment
reviews. Oxford: Radcliffe Medical Press, 1994: 11±30.
13 Nocon A, Baldwin S. Trends in rehabilitation policy. A review of the
literature. London: King's Fund, 1998.
16 Hughes D. Caring for people ef®ciently. Hlth Policy 1993; 25: 81±94.
17 Cochrane, AL. Effectiveness and ef®ciency. Random re¯ections on
health services. London: Nuf®eld Provincial Hospitals Trust, 1971.
18 Department of Health. The new NHS. London: HMSO, 1997.
19 Barclay S, Todd C, McCabe J, Hunt T. Primary care group
commissioning of services: the differing priorities of general practitioners and district nurses for palliative care services. Br J Gen Pract
1999; 49: 181±186.
Accepted on 16 December 1999