Embed Size (px)
Citation preview
This article was downloaded by: [Umeå University Library]On: 24 November 2014, At: 07:02Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK
Aging & Mental HealthPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/camh20
The CORE-OM in an older adult population:Psychometric status, acceptability, and feasibilityM. Barkham Professor a , A. Culverwell b , K. Spindler c & E. Twigg aa Psychological Therapies Research Centre , University of Leeds , Leedsb East Kent NHS Partnership Trust , UKc University of Kent , UKPublished online: 19 Oct 2010.
To cite this article: M. Barkham Professor , A. Culverwell , K. Spindler & E. Twigg (2005) The CORE-OM in an olderadult population: Psychometric status, acceptability, and feasibility, Aging & Mental Health, 9:3, 235-245, DOI:10.1080/13607860500090052
To link to this article: http://dx.doi.org/10.1080/13607860500090052
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe Content. Any opinions and views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions
ORIGINAL ARTICLE
The CORE-OM in an older adult population: Psychometric status,acceptability, and feasibility
M. BARKHAM1, A. CULVERWELL2, K. SPINDLER3, & E. TWIGG1
1Psychological Therapies Research Centre, University of Leeds, Leeds, 2East Kent NHS Partnership Trust,
and 3University of Kent, UK
(Received 7 February 2004; accepted 17 June 2004)
AbstractThere is a need to extend and test the feasibility and acceptability of mental health outcome measures in the older population(i.e., aged 65–100). We present data on the CORE-OM (Clinical Outcomes in Routine Evaluation-Outcome Measure) ona sample of 118 people aged 65–97 presenting for mental health treatment and 214 people aged 65–94 drawn from a non-clinical population. Results show the CORE-OM to be a reliable measure in both samples when the overall mean item isused but the reliability is not as high for the specific domains as psychometrically stable structures. The CORE-OM showedlarge overall differences between the non-clinical and clinical samples indicating that it is equally as sensitive to thesediffering populations across this older age band as with working-age adults. However, the norms for the clinical sample wereconsistently lower than the equivalent clinical norms for a working-age sample. These findings suggest that the collection andcompilation of age-specific norms is crucial in ensuring that appropriately referenced norms are used rather than assumingthat norms are generalizable across the whole adult life-span.
Introduction
The UK Office of National Statistics (ONS) report
on the mental health of older people reports a
prevalence rate for the presence of any neurotic dis-
order of 16%, 10% and 9% for the age bands 60–64,
65–69, and 70–74 respectively (Evans, Singleton,
Meltzer, Stewart & Prince, 2003). These prevalence
rates reflect a decrease from a level of approxi-
mately 20% across the age band 40–54 (Singleton,
Bumpstead, O’Brien, Lee & Meltzer, 2001)
and indicate the need for an understanding of
this continuum from working-age to older adults.
One such area relates to the use of outcome measures
that provide a description of patients and an evalua-
tion of the impact of psychological interventions.
However, the majority of work on measure develop-
ment has been targeted at adults within the age range
of 18–65 with relatively little work focusing on adults
over the age of 65. The traditional cut-off age of 65
does, for many, mark a significant change in people’s
lives and there might be an argument for the devel-
opment of specific measures for older adults. On the
other hand, a developmental life-span approach
would support a continuum spanning into much
later years in which the focus was on understanding
how measures perform with differing age groups.
In order to achieve this aim, practitioners need
appropriate tools. Of particular interest to us was the
need to investigate whether tools and technologies
designed for working age adults could be extended to
older adults appropriately. Our focus in the present
paper was on the CORE Outcome Measure (CORE-
OM), a client-completed outcome measure that has
recently been developed for the adult population and
is currently widely used in the evaluation of the
psychological therapies in the UK (Barkham et al.,
2001; Evans et al., 2000). The CORE-OM was
designed in order to help avoid a ‘state of disarray,
if not chaos’ in measuring outcomes in such settings
(Froyd, Lambert & Froyd, 1996). A programme
of research set out the rationale (Barkham et al.,
1998),development(Evansetal.,2000),psychometric
properties (Evans et al., 2002) and applications
(Barkham et al., 2001) of the CORE-OM, which
provided a single ‘core’ outcome measure for clinical
service evaluation. The population it was designed
for comprised clients seeking and/or receiving
psychological therapies. The full CORE-OM com-
prises 34 items and is generic in its theory base.
Relevant psychometric properties for the CORE-
OM, which have been reported previously (Barkham
et al., 2001; Evans et al., 2002), comprise alphas of
0.94 in both clinical non-clinical samples; test–retest
Correspondence: Professor Michael Barkham, Psychological Therapies Research Centre, 17 Blenheim Terrace, Universityof Leeds, Leeds LS2 9JT, UK. Tel: 0113 343 5699. Fax: 0113 343 1956. E-mail: [email protected]
Aging & Mental Health, May 2005; 9(3): 235–245
ISSN 1360-7863 print/ISSN 1364-6915 online � 2005 Taylor & Francis Group LtdDOI: 10.1080/13607860500090052
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
(rho) stability of 0.90 in a student sample; a mean
score for the clinical population of 1.86 (SD¼ 0.75)
and for a non-clinical population of 0.76
(SD¼ 0.59).
The CORE-OM comprises the domains of well-
being, problems/symptoms, functioning and risk.
It has shown good reliability, validity against longer
and less general measures, and has been shown
to be sensitive to change (Evans et al., 2000). It is
‘copyleft’—that is, it can be reproduced free of
charge but cannot be altered or financial gain made
out of it which, in the current climate of increasing
scale charges, makes it especially valuable to research
and practice within NHS settings. The CORE-OM
is now in widespread use in psychological therapy
services in the UK and has been found acceptable
to both therapists and clients (Barkham et al., 2001;
Evans et al., 2000).
However, the norms for the non-clinical popula-
tion were based on a sample with a mean age of 20.5
years and a range of 14–45 years of age. The norms
for the clinical population were based on a sample
with a mean age of 36 years and range of 16–78 years
of age with only 7 people (0.4%) aged over 65 years.
This reveals an under-representation of the older
adult population, which may undermine the validity
of the CORE-OM for an older client group.
Attempts have been made to address the specific
needs of this age group in relation to other measures
and draw on one of two approaches. The first focuses
on findings about the elderly population. For exam-
ple, administering the Brief Symptom Inventory
(BSI; Derogatis & Melisaratos, 1983) to older
adults yielded a finding that the elderly reported
higher levels of distress on most of the symptom
dimensions when compared with mean scores of a
normative sample of younger adults (Hale, Cochran
& Hedgepeth, 1984). A second approach has focused
on the potential adaptations required to extend
measures designed for adults into the elderly popu-
lation. Studies of the Health of the Nation Outcome
Scales for elderly people (HoNOS 65þ; e.g., Burns
et al., 1999) suggested the need for revisions to the
glossary to better reflect the presentation of illness
and phenomenology in older people (e.g., greater
emphasis on somatic presentation).
The aim of this study, therefore, is three-fold.
First, we sought to determine whether the psycho-
metric properties of the CORE-OM are sufficiently
robust so as to suggest that it is a valid and reliable
instrument to use with an older adult population.
Second, we sought to establish whether a differing
set of norms are appropriate for this age group by
testing the obtained scores against published data.
Third, using practitioner feedback, we sought to
present responses and reactions from a clinical
service in administering the CORE-OM. We
included this latter part because even though the
results from the quantitative stages might show good
psychometric properties, the burden to and reactions
of participants could suggest that administration
of such a measure carries associated problems that
outweigh the benefits.
Method
Participants
The total sample comprised 332 people with n¼ 214
in the non-clinical sample n¼ 118 in the clinical
sample. In the non-clinical sample, 72.4% were
women, 25.2% men, and the gender of 2.3% was not
recorded. The mean age was 78.5 (SD¼ 7.25) with
quartiles of 73, 78, and 85 years of age and an age
range of 65–94 years of age. In the clinical sample,
62.7% were women, 33.9% men, and the gender
of 3.4% was not recorded. The mean age was
75.1 years (SD¼ 6.70) with quartiles of 70, 75, and
80 years of age and an age range of 65–97 years
of age. There were no significant differences
in the proportion of females in the clinical
and non-clinical samples: non-clinical¼ 72.4%
female; clinical¼ 62.7% female; z¼ 1.749; p¼ 0.08
(two-tailed).
All clients in the non-clinical sample for whom
ethnicity was indicated (n¼ 201) were white
European. The clinical sample comprised 90 white
European, one non-white European and 27 clients
for whom no data on ethnicity was provided. A total
of 60.7% (n¼ 130) of the non-clinical sample were
widowed compared with only 33.1% (n¼ 39) of the
clinical sample: this difference was highly significant
(z¼ 3.568, p<0.001). Numbers who were married/
cohabiting were similar for both samples: clinical
31.4% (n¼ 37); non-clinical 26.6% (n¼ 57). A total
of 4.7% (n¼ 10) of the non-clinical sample were
single compared with 6.8% (n¼ 8) of the clinical
sample. A significantly higher proportion (z¼ 2.21,
p<0.05) of the clinical sample were separated or
divorced (10.2%, n¼ 12) compared to the non-
clinical sample (5.1%, n¼ 11). Data on marital
status was not provided for six (2.8%) clients in the
non-clinical sample and 22 (18.6%) clients in the
clinical sample. Professional backgrounds of partici-
pants were provided and showed no significant
differences between the clinical and non-clinical
groups (all p values >0.05) except in the intermediate
class where 35.5% of non-clinical respondents and
20.3% of clinical respondents were categorized
(z¼ 2.885; p<0.05).
Measure
The measure used in the present study was
the CORE-OM, which has been described above.
The CORE-OM scoring followed specific pro-
cedures. Where there were missing items, these were
‘pro-rated’ to minimize loss of data. The criteria
for addressing missing items related to (a) re-scaling
the total or non-risk scores if three or less items have
236 M. Barkham et al.
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
been missed; and (b) re-scaling dimension scores
if only one item is missing. Removing those cases
where either of the inclusion criteria was not met
reduced the numbers of clients for whom valid
data had been received to 187 non-clinical and 101
clinical. Additionally, where scores were investigated
by gender or age and these variables had missing
values, n varied accordingly.
Procedure
Non-clinical sample. The non-clinical sample was
recruited through a number of community organiza-
tions and health facilities: 50% from four different
Age Concern centres; 23% sample of convenience;
15% health centre attendees for chiropody/falls
prevention; and 12% active volunteers recruited
via a Volunteer Co-ordinator. These were initially
contacted by letter asking people for their help in
the study. Two weeks later they were telephoned to
see if their members might be willing to participate.
If they agreed, a time was arranged for the project
worker (KS) to attend. Individuals attending the
centres were approached and the study aims and
requirements were explained. Confidentiality and
anonymity were assured. Participants were invited to
complete the forms at the time in private, with the
assistance of the project worker if required, or to take
them home and complete and then return them
in a pre-paid stamped-addressed envelope (SAE).
The questionnaires took approximately 10 minutes
to complete. Participants were thanked for their
co-operation and invited to keep the participant
information sheet for reference. Further partici-
pants were obtained as a sample of convenience via
friends, relatives, and work colleagues. For this
group, questionnaires were mailed to them with
the information sheet.
Clinical sample. The clinical sample was drawn
from those receiving support from mental health
services. A range of professionals were contacted to
see if they would be willing for people under their
care to be approached as potential participants.
If agreeable, potential clients were identified. Over
a four-month period, all clients identified by four
staff groups as meeting the criteria and attending
the Outpatient Clinic or Day Hospital on targeted
days were approached. The four staff groups
accounted for recruiting the following percentage of
the clinical sample: community mental health team
(CMHT), 20%; psychiatrist, 39%; psychologist, 6%;
and Day Hospital Manager/key worker, 35%. People
were excluded if they presented with cognitive
impairment or language difficulties such that they
might not be able to complete the form reliably, or if
they were acutely distressed at that point. In
addition, participants had to be at an early stage of
their contact with mental health services. Individuals
were approached via their mental health worker. The
study was explained to them and they were invited to
participate. Participants were given the same options
for completing as offered to the non-clinical sample.
Ethical approval was obtained from the East Kent
Local Ethics Committee.
Analyses
Principal component analysis (PCA) was carried out
on the older adult non-clinical and clinical data sets
to determine the factor structure of the CORE-OM
in these two samples. This was deemed more
appropriate than confirmatory factor analysis
because there was no certainty that the older adult
population responded to CORE-OM questions in a
similar fashion to the normative sample gathered
previously. Subsequent to this analysis, we used
Cohen’s (1960) kappa statistic to ascertain the
similarity of this analysis to the analysis of the
original CORE-OM normative data. Kappa is a
chance-corrected coefficient of agreement between
two ‘judges’ (i.e., in this instance factor analyses).
We used effect sizes (ES) to provide an indication
of the difference between samples. An ES gives an
indication of the magnitude of the treatment effect,
or in this case the difference between non-clinical
and clinical samples. To determine cut-off points
between the two samples, we applied the criteria
reported by Jacobson and Truax (1991) for the total
score that would aid assignment of people to either a
clinical or non-clinical population. The formula used
was as follows:
meanclinsdnorm þmeannormsdclinsdnorm þ sdclin
ð1Þ
Investigation of the CORE-OM mean scores showed
that they were non-normally distributed. Hence non-
parametric tests were used to make comparisons
within the data set and with the data from the original
analysis of the normative data.
Results
Missing items
Data was obtained for a total of 214 non-clinical
cases and 118 clinical cases. Table I shows the mean
number of missing items by domain and overall for
clinical and non-clinical older adults and normative
samples, along with confidence intervals for the
difference between older adults and normative
samples. The pattern for the older adults sample
was compared with the normative data sample where
appropriate to look for differences in missing items
between the two samples. There were significant
differences in the number of missing items in the
normative and older adults clinical samples on all
domains ( p<0.05) except problems and overall.
These differences were highly significant between
The CORE-OM in an older adult population 237
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
the non-clinical normative and older adults samples
( p<0.0001 for all domains and overall).
The mean number of missing items was consis-
tently higher in the clinical than in the non-clinical
older adults sample but this did not prove to be
significant when tested (Mann-Whitney p¼ 0.337 for
the full 34 items and >0.05 for individual domains).
The mean number of missing items across all parti-
cipants was 1.42. A total of 66% of the non-clinical
older adult sample (91% in the normative sample)
returned complete forms compared with 73% of the
clinical sample (80% in the normative sample).
In terms of individual items omitted, the most
frequent in the clinical older adults sample was I19
‘I have felt warmth or affection for someone’ which
was omitted by 10.2% of participants. However,
in the non-clinical population, the most common
item to be left out was I31 ‘I have felt optimistic
about the future’. Item 19 was also the most
frequently omitted in the normative clinical sample
(3.7% of respondents) and the non-clinical sample,
although in this sample I30 ‘I have thought I am to
blame for my problems or difficulties’ was equally
often missed.
Internal consistency
Values for Cronbach’s alpha were calculated for
non-clinical and clinical samples and for specific age
bands within each sample (see Table II) and for men
and women.
Data in Table II shows the overall alphas for both
non-clinical and clinical samples to be 0.83 and 0.90
respectively. In the non-clinical sample, the alphas
for the domains are variable with only ‘problems’
>0.70. For the clinical sample, all alphas were >0.70
with the exception of ‘well-being’. Overall, the alphas
for CORE-OM are �0.80 across all age bands (65–
69 years, 70–79 years, and �80 years) although the
values decrease monotonically. The alphas for the
domains show clear age effects for the clinical sample
whereby the reliability for well-being >0.70 for
people aged <69 but <0.50 for the other two age
bands, and the reliability for risk <0.50 for people
aged >80. When reliability analysis was performed
splitting the sample by gender, very similar
overall alphas were obtained for men and women
but with the alphas for risk >0.70 for men in both
samples but 0.08 (non-clinical) and 0.51 (clinical)
for women.
Structure
A principal component analysis (PCA) was carried
out on the older adult non-clinical and clinical data
sets to determine the factor structure of the CORE-
OM in these two samples.
For both samples, initial analysis yielded 11 items
whose eigenvalues exceeded 1. However, a scree plot
of the eigenvalues suggested an ‘elbow’ at three
factors for both samples which were in turn rotated
from a principal components analysis. An oblique
rotation was used as there was an expectation that
there would be a correlation between items, and
results suggested that the three factors could be
similarly labelled in both data sets as ‘negatively
worded experiences’, ‘positively keyed experiences’,
and ‘areas of risk’. The factor loadings >0.40 are
presented in Tables III and IV for the non-clinical
and clinical samples respectively. The italicised items
represent those factors that loaded on different
factors when compared with the original factor
analyses carried out in the development of the
CORE-OM. This phenomenon is more pronounced
in the non-clinical sample but it is the clinical
sample, which is of particular interest. In the clinical
sample, two items which loaded as ‘positively keyed’
in the original analysis (I26 & I30), are now in
Table I. Data on missing items from older adult and normative data sets for clinical and non-clinical samples.
Clinical
Older adults (n¼118) Normative (n¼ 890) Difference
Domain Mean SD Mean SD 95% CI p value
Well-being items (4) 0.16 0.60 0.06 0.33 �0.01–0.21 0.046
Symptoms items (12) 0.50 1.66 0.17 0.69 0.03–0.64 0.079
Functioning items (12) 0.64 1.74 0.22 0.91 0.10–0.74 0.001
Risk items (6) 0.33 1.00 0.11 0.54 0.03–0.41 0.005
Total missing items (34) 1.64 4.80 0.56 2.19 0.19–1.96 0.031
Non-clinical
Older adults (n¼214) Normative (n¼ 1106) Difference
Mean SD Mean SD 95% CI p value
Well-being items (4) 0.15 0.43 0.06 0.43 0.03–0.15 <0.0001
Symptoms items (12) 0.44 1.42 0.20 1.34 0.03–0.44 <0.0001
Functioning items (12) 0.43 1.39 0.22 1.32 0.01–0.42 <0.0001
Risk items (6) 0.29 0.91 0.10 0.69 0.06–0.32 <0.0001
Total missing items (34) 1.31 3.84 0.58 3.73 0.17–1.29 <0.0001
238 M. Barkham et al.
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
the ‘areas of risk’ (I26) and ‘negatively worded
experiences’ (I30), which are more appropriate
categories. Item 21 loaded appropriately as a
positively keyed item rather than as ‘negatively
worded experiences’ in the original analysis, while
I29 loaded as an ‘areas of risk’ item rather than a
‘negatively worded experiences’.
Table V presents the overall kappa values as well as
the values for proportions of agreement between the
factor analyses on the three components. The overall
Table III. PCA of non-clinical sample.
Component
1 2 3
% Variance accounted for 19.93 7.50 6.57
Cumulative % variance accounted for 19.93 27.42 33.99
1 2 3
W HI 17. I have felt overwhelmed by my problems 0.78
P HI 15. I have felt panic or terror 0.78
P HI 23. I have felt despairing or hopeless 0.72
F HI 33. I have felt humiliated or shamed by other people 0.67
P HI 28. Unwanted images or memories have been distressing me 0.65
P LO 27. I have felt unhappy 0.63
P LO 20. My problems have been impossible to put to one side 0.61
F LO 25. I have felt criticized by other people 0.51
W HI 14. I have felt like crying 0.50
P HI 11. Tension and anxiety have prevented me doing important things 0.50
F HI 1. I have felt terribly alone and isolated 0.50
P LO 2. I have felt tense, anxious or nervous 0.43
R HI 22. I have threatened or intimidated another person 0.40
P HI 13. I have been disturbed by unwanted thoughts and feelings
F HI 10. Talking to people has felt too much for me
P LO 30. I have thought I am to blame for my problems and difficulties
F HI þ 32. I have achieved the things I wanted to 0.71
F LO þ 21. I have been able to do most things I needed to 0.60
W LO þ 31. I have felt optimistic about the future 0.59
F LO þ 12. I have been happy with the things that I have done 0.58
W LO þ 4. I have felt OK about myself 0.55
F LO þ 19. I have felt warmth or affection for someone 0.53
F HI þ 7. I have felt able to cope when things go wrong 0.45
F LO þ 3. I have felt I have someone to turn to for support when needed
F HI 26. I have thought I have no friends
R LO 24. I have thought it would be better if I were dead �0.75
R HI 16. I made plans to end my life �0.71
R HI 34. I have hurt myself physically or taken dangerous risks with my health �0.61
R LO 9. I have thought of hurting myself 0.40 �0.53
P HI 5. I have felt totally lacking in energy and enthusiasm �0.42
P LO 18. I have had difficulty getting to sleep or staying asleep �0.41
P LO 8. I have been troubled by aches, pains or other physical problems �0.41
F LO 29. I have been irritable when with other people
R HI 6. I have been physically violent to others
Component Correlations
Component 2 0.24
Component 3 �0.21 0.09
Table II. Coefficient alpha for non-clinical and clinical older adults’ data.
Non-clinical (n¼ 139) Clinical (n¼ 84)
Domain
Alpha
(95% CI for alpha)
Alpha by age group
Alpha
(95% CI for alpha)
Alpha by age group
<69
(n¼24)
70–79
(n¼71)
80þ
(n¼44)
<69
(n¼ 24)
70–79
(n¼ 71)
80þ
(n¼44)
Subjective well-being 0.55 (0.41–0.66) 0.69 0.52 0.49 0.47 (0.26–0.63) 0.77 0.40 0.26
Problems/Symptoms 0.76 (0.70–0.81) 0.77 0.79 0.68 0.84 (0.78–0.89) 0.90 0.86 0.77
Functioning 0.64 (0.54–0.72) 0.74 0.57 0.65 0.74 (0.65–0.82) 0.86 0.69 0.74
Risk 0.57 (0.45–0.67) 0.05 0.76 0.38 0.72 (0.62–0.80) 0.78 0.75 0.48
Non-risk items 0.83 (0.79–0.87) 0.86 0.82 0.79 0.89 (0.85–0.92) 0.94 0.88 0.87
All items 0.83 (0.79–0.87) 0.86 0.83 0.80 0.90 (0.87–0.93) 0.94 0.89 0.87
The CORE-OM in an older adult population 239
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
agreement levels for the non-clinical and clinical
samples compared with the normative samples
were 0.57 (95% CI 0.34–0.80) and 0.81 (95% CI
0.64–0.98) respectively.
Considering the proportions of agreement for each
component compared with the original factor anal-
yses, only the third component in the non-clinical
sample is below an acceptable kappa level (usually
considered to be 0.5).
Comparison of clinical versus non-clinical olderadult scores
Figure 1 presents a dual histogram of CORE-OM
mean scores, which shows clearly the differences in
distributions of the two populations. Both are
unimodal but the modal value is substantially lower
for the non-clinical population compared with the
clinical population. A Mann-Whitney U-test shows
that the two samples are significantly different
(U¼ 1554, n¼ 288, p<0.0001).
Table VI presents the means and standard devia-
tions for male and female participants in the older
adults sample together with differences between the
groups represented by p values, confidence intervals,
and effect size (ES). In the non-clinical sample,
Table IV. PCA of clinical sample.
Component
1 2 3
% variance accounted for 25.17 8.97 7.19
Cumulative % variance accounted for 25.17 34.14 41.33
1 2 3
P HI 13. I have been disturbed by unwanted thoughts and feelings 0.76
P HI 23. I have felt despairing or hopeless 0.74
P HI 5. I have felt totally lacking in energy and enthusiasm 0.74
W HI 17. I have felt overwhelmed by my problems 0.73
P LO 20. My problems have been impossible to put to one side 0.73
P LO 2. I have felt tense, anxious or nervous 0.72
P HI 11. Tension and anxiety have prevented me doing important things 0.71
P LO 27. I have felt unhappy 0.69
P HI 15. I have felt panic or terror 0.59
F HI 1. I have felt terribly alone and isolated 0.56
W HI 14. I have felt like crying 0.50
P LO 8. I have been troubled by aches, pains or other physical problems
P HI 28. Unwanted images or memories have been distressing me
P LO 30. I have thought I am to blame for my problems and difficulties
F HI 10. Talking to people has felt too much for me
P LO 18. I have had difficulty getting to sleep or staying asleep
R HI 16. I made plans to end my life 0.73
R HI 34. I have hurt myself physically or taken dangerous risks with my health 0.67
R LO 9. I have thought of hurting myself 0.61
R HI 6. I have been physically violent to others 0.59
F HI 33. I have felt humiliated or shamed by other people 0.56
F LO 25. I have felt criticized by other people 0.50
F HI 26. I have thought I have no friends 0.50
R LO 24. I have thought it would be better if I were dead 0.50
R HI 22. I have threatened or intimidated another person 0.45
F LO 29. I have been irritable when with other people 0.43
F LO þ 12. I have been happy with the things that I have done 0.74
F LO þ 3. I have felt I have someone to turn to for support when needed 0.50 0.62
W LO þ 4. I have felt OK about myself 0.62
F LO þ 21. I have been able to do most things I needed to 0.60
F HI þ 32. I have achieved the things I wanted to 0.57
W LO þ 31. I have felt optimistic about the future 0.56
F LO þ 19. I have felt warmth or affection for someone 0.50
F HI þ 7. I have felt able to cope when things go wrong 0.40
Component Correlations
Component 2 0.30
Component 3 0.25 0.07
Table V. Kappa and proportion of agreement levels between
PCAs based on older and normative data sets for non-clinical and
clinical samples.
Components
in PCA
Non-clinical sample Clinical sample
Proportion
of agreement 95% CI
Proportion
of agreement 95% CI
1 0.67 0.43–0.85 0.83 0.58–0.96
2 0.67 0.35–0.89 0.80 0.44–0.97
3 0.30 0.08–0.65 0.70 0.35–0.92
Kappa 0.57 0.34–0.80 0.81 0.64–0.98
240 M. Barkham et al.
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
the mean score for women was consistently higher
than for men with the exception of ‘risk’. The
differences between men and women for the overall
CORE-OM scores (as well as for the Problems and
Functioning domains specifically) were statistically
significant within the non-clinical sample although
the ES difference was small (i.e., 0.25). In the clinical
sample, there were no significant differences between
men and women and the overall ES difference
was <0.10.
Age was considered as a continuous variable and
correlated with the CORE-OM mean overall score.
There was a significant (rho¼ 0.382, p<0.0001)
relationship in the non-clinical population between
age and overall mean score but not in the clinical
population (rho¼ 0.072, p¼ 0.446). This is pre-
sented graphically by the notched box-plots in
Figure 2 in which CORE-OM mean scores increase
monotonically across the three age bands for non-
clinical in contrast to the clinical group in which
there is only a slight increase across the three age
bands.
Comparison of older adults versus normative scores
Table VII presents the means and standard devia-
tions for the non-clinical and clinical older age
samples compared with the published norms
together with differences between the groups again
represented by p values, confidence intervals, and
effect size. Older adults scored significantly lower
overall (as well as for the Functioning and Risk
domains) with an ES approaching 0.30. There were
significant differences between the clinical samples
on all domains and overall with the latter ES
difference being 0.46.
Population-specific item functioning
In comparing older and normative clinical samples,
we sought to test the differential functioning of
individual items on the CORE-OM as a possible
consequence of the population from which the
sample had been drawn. Whilst we knew that older
clinical adults overall scored lower than the norma-
tive clinical adults, we predicted that for specific
items, the older adults might actually score higher (or
the difference should be lowest) when compared with
scores of adults below age 65. We predicted this
effect would occur on items relating to: (a) physical
aspects (i.e., I have been troubled by aches, pains
or other physical problems’, and ‘I have felt totally
lacking in energy and enthusiasm’); and (b) the
future (i.e., I’ve felt optimistic about the future).
Table VIII shows the item means and the difference
between them for the two clinical samples. The
mean difference for all items was 0.31 (SD¼ 0.32).
Table VI. Means and standard deviations for men and women in non-clinical and clinical older age samples.
Non-clinical sample
Female (n¼137) Male (n¼ 47) Differences
Dimension Mean SD Mean SD p value 95% CI for difference Effect size d
Well-being 0.92 0.70 0.79 0.78 0.152 �0.12–0.39 0.18
Problems 0.83 0.50 0.67 0.58 0.010 �0.03–0.35 0.31
Functioning 0.61 0.42 0.57 0.62 0.047 �0.15–0.24 0.08
Risk 0.05 0.14 0.09 0.30 0.674 �0.13–0.05 �0.21
All non-risk items 0.75 0.39 0.64 0.51 0.022 �0.05–0.27 0.26
All items 0.63 0.33 0.54 0.44 0.026 �0.06–0.22 0.25
Clinical sample
Female (n¼61) Male (n¼ 37) Differences
Dimension Mean SD Mean SD p value 95% CI for difference Effect size d
Well-being 2.26 0.79 1.98 0.84 0.090 �0.07–0.62 0.35
Problems 2.14 0.83 1.94 0.74 0.188 �0.12–0.52 0.25
Functioning 1.37 0.67 1.45 0.68 0.608 �0.37–0.19 �0.12
Risk 0.28 0.36 0.36 0.63 0.634 �0.31–0.14 �0.17
All non-risk items 1.83 0.66 1.74 0.66 0.459 �0.18–0.36 0.14
All items 1.55 0.57 1.50 0.60 0.524 �0.19–0.30 0.09
0
1
2
3
Mea
n C
OR
E-O
M s
core
0102030Count
0 10 20 30Count
ClinicalNon-clinical
POPULATION
Figure 1. Dual histogram showing distributions of CORE-OM
scores for clinical and non-clinical older adult populations.
The CORE-OM in an older adult population 241
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
There were four items on which older adults had
a higher mean score: the greatest difference being
on I8 ‘I have been troubled by aches, pains and
other physical problems’, with a difference >2 SDs
( p<0.05). The mean differences for the other three
items were all >1 SD unit.
Clinical cut-off points
Using data reported in Table VII, the overall cut-off
point for older men and women inclusive was 0.952.
Because the cut-off points for the normative sample
were calculated separately for men and women,
we did the same for this sample using data from
Table VI. The gender specific cut-off points
were 0.946 (rounded to 0.95) for men and 0.967
(rounded to 0.97) for women. Applying these
gender-specific cut-offs to the data resulted in
16.3% (30/184) of the non-clinical sample scoring
above the clinical cut-off while 14.3% (14/98) of the
clinical sample had a mean score below the clinical
cut-off.
Clinical reflections on use in routine practice
In considering the acceptability of the CORE-OM
for use with older adults, we drew upon the
experiences of the clinical workers in administering
the measure. Comments from the project worker in
collecting the normative data suggested that people
often requested help in completing the form. They
were not used to forms and even less used to rating
Table VII. Means and standard deviations for non-clinical and clinical older age samples compared with published norms.
Non-clinical sample
Normative (n¼ 1077) Older adults (n¼187) Differences
Dimension Mean SD Mean SD p value 95% CI for difference Effect size d
Well-being 0.90 0.82 0.88 0.72 0.743 �0.14 to �0.09 0.02
Problems 0.90 0.72 0.79 0.52 0.503 �0.20 to �0.02 0.16
Functioning 0.85 0.64 0.60 0.48 <0.001 �0.33 to �0.17 0.40
Risk 0.20 0.46 0.06 0.19 <0.001 �0.18 to �0.10 0.32
All non-risk items 0.88 0.65 0.72 0.42 0.073 �0.23 to �0.08 0.26
All items 0.76 0.58 0.60 0.36 0.030 �0.22 to �0.09 0.29
Clinical sample
Normative (n¼ 835) Older adults (n¼101) Differences
Dimension Mean SD Mean SD p value 95% CI for difference Effect size d
Well-being 2.35 0.96 2.13 0.83 0.010 �0.40 to �0.04 0.23
Problems 2.30 0.87 2.06 0.79 0.004 �0.41 to �0.08 0.28
Functioning 1.86 0.84 1.39 0.67 <0.001 �0.62 to �0.33 0.57
Risk 0.63 0.75 0.31 0.47 <0.001 �0.43 to �0.22 0.44
All non-risk items 2.12 0.81 1.78 0.66 <0.001 �0.48 to �0.20 0.43
All items 1.86 0.75 1.52 0.58 <0.001 �0.46 to �0.21 0.46
<=69 70–79 80+
Age group
0.0
0.5
1.0
1.5
2.0
CO
RE
Out
com
e M
easu
re m
ean
scor
e
Non-clinical population
<=69 70–79 80+
Age group
0
1
2
3
CO
RE
Out
com
e M
easu
re m
ean
scor
e
Clinical population
Figure 2. Notched box-plots for the non-clinical and clinical samples for CORE-OM scores across the three age bands.
242 M. Barkham et al.
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
scales. Some had a tendency to get into a fixed
pattern of responding. Surprisingly, the font size was
not problematic. People rarely kept their ticks within
the boxes marked on the CORE-OM, which is a
problem if you want to optically scan the ques-
tionnaire.
It became clear that a reminder about the time
scale (i.e., an emphasis on ‘the last week’ and not an
overall rating of their lives) was needed, particularly
for those items, which invite a life review perspective
(i.e., item 12, ‘I’m happy with the things I’ve done’
and 32 ‘I have achieved the things I’ve wanted to’).
As a result of the above, in clinical practice we
regularly include a ‘practice rating’ or two, in order
to show clients how to use the scale. We also make a
specific reminder to think in terms of how they have
felt over the seven days prior to completing the form.
We have found that it is possible to use the CORE-
OM with people who have a very mild degree of
cognitive impairment, but have not considered it
appropriate (in terms of acceptability, feasibility, and
internal consistency/validity) for those with more
significant cognitive impairment.
In clinical practice our impressions are that people
do not find any of the items either particularly
surprising or distressing. We use it not only as a
pre-therapy measure but also as a tool for enabling
further discussion, and in this way it is possible to
query any seemingly contradictory responses and to
explore any omitted items. It is typically introduced
to clients during the first or second meeting with an
invitation to complete it at home and bring to the
next session. The rationale offered included the
following:
. It is a tool that allows them to describe how they’ve
been feeling without the therapist having to ask
them a barrage of questions;
. We are interested in their own views, not their
spouse’s, and there are no right or wrong answers;
. It is sometimes easier to work out how you’ve been
feeling quietly, in your own time and in the privacy
of your own home;
. Some people find it easier to describe their feelings
on paper rather than say it directly to someone;
. It is a way of measuring changes – for them (to see
what has/hasn’t got easier over time) and for us to
see if what we have been doing is helpful to people.
Hence it provides part of the socialization to therapy
process, so often commented on as an area of work
that needs particular attention and time when work-
ing with older clients (Knight, 1996) as well as acting
Table VIII. Comparison of mean item CORE-OM scores for normative and older adults’ clinical samples.
Item
Normative
mean
Older adults
mean
Difference in
mean scores
8 I have been troubled by aches, pains or other physical problems 2.18 2.71 �0.53***
5 I have felt totally lacking in energy and enthusiasm 2.56 2.81 �0.25*
11 Tension and anxiety have prevented me doing important things 2.07 2.16 �0.09
31 I have felt optimistic about the future 2.52 2.60 �0.08
1 I have felt terribly alone and isolated 1.88 1.86 0.02
4 I have felt OK about myself 2.44 2.31 0.13
16 I made plans to end my life 0.47 0.31 0.16
19 I have felt warmth or affection for someone 1.52 1.36 0.16
2 I have felt tense, anxious or nervous 2.67 2.50 0.17
15 I have felt panic or terror 1.48 1.30 0.18
17 I have felt overwhelmed by my problems 2.19 2.01 0.18
23 I have felt despairing or hopeless 1.98 1.80 0.18
6 I have been physically violent to others 0.28 0.09 0.19
21 I have been able to do most things I needed to 1.67 1.45 0.22
27 I have felt unhappy 2.68 2.45 0.23*
32 I have achieved the things I wanted to 2.55 2.27 0.28*
7 I have felt able to cope when things go wrong 2.22 1.91 0.31**
34 I have hurt myself physically or taken dangerous risks with my health 0.52 0.21 0.31*
18 I have had difficulty getting to sleep or staying asleep 2.37 2.05 0.32*
20 My problems have been impossible to put to one side 2.67 2.34 0.33*
26 I have thought I have no friends 1.53 1.17 0.36*
22 I have threatened or intimidated another person 0.50 0.13 0.37***
24 I have thought it would be better if I were dead 1.23 0.86 0.37**
9 I have thought of hurting myself 0.78 0.32 0.46***
10 Talking to people has felt too much for me 1.77 1.31 0.46***
3 I have felt I have someone to turn to for support when needed 1.89 1.42 0.47**
13 I have been disturbed by unwanted thoughts and feelings 2.35 1.88 0.47***
28 Unwanted images or memories have been distressing me 2.18 1.69 0.49**
14 I have felt like crying 2.25 1.75 0.50***
33 I have felt humiliated or shamed by other people 1.26 0.70 0.56***
12 I have been happy with the things that I have done 2.16 1.42 0.74***
29 I have been irritable when with other people 2.02 1.14 0.88***
25 I have felt criticized by other people 1.89 0.92 0.97***
30 I have thought I am to blame for my problems and difficulties 2.44 1.32 1.12***
The CORE-OM in an older adult population 243
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
as an outcome measure. Using the measure in this
way has resulted in only one client refusing to
complete the form in over three years of use in one
service.
Discussion
The aim of the present study was to investigate how
the CORE-OM performed in an older adult popula-
tion given that this group represented less than 1%
of the normative sample employed in the original
analyses. Our questions focused on the psychometric
status of the CORE-OM in an older adult population
and the extent of differences between this older adult
population and the normative adult population in
regard to internal consistency, structure, and norms.
Overall, the results show the CORE-OM to be a
reliable and structurally sound outcome measure to
use with older adults. The finding that the values of
Cronbach’s alpha for the CORE-OM in this sample
ranged from 0.80–0.94 across age bands for non-
clinical and clinical samples compared with the alpha
of 0.94 reported in the normative sample suggests
that there is no appreciable lessening of internal
consistency when the CORE-OM is administered in
an older adult population. There was a consistent
finding of the domain of ‘well-being’ yielding the
lowest alpha and this appeared to decline further in
the age groups over 70. It would not be unreasonable
to suppose that people’s sense of ‘well-being’ altered
significantly once past the age of 70, at least in terms
of the items used to tap this domain in people aged
20–70.
The results of the PCA showed a similar structure
to that obtained in the original analyses (see Evans
et al., 2002). The kappa statistic yielded a satisfactory
agreement level of items loading onto the same factor
in the older adult as compared with the normative
data. The finding that the agreement was stronger
for the clinical rather than the non-clinical older
adult sample was logical since the CORE-OM was
designed as a clinical measure. The finding that
the weakest agreement was on the third factor within
the non-clinical sample was also logical in that at
this level it would be expected that there would be
increasingly less commonality between participants
to pull items into a single coherent factor.
Evidence of the CORE-OM being sensitive to
aging effects was obtained from the elevated scores
on selected items by older adults in the clinical
sample. We correctly predicted higher scores on
items relating to aspects of physical (aches and pains,
lacking in energy) and future (optimism) domains
of life, thereby showing that selected items were
performing selectively in this particular population.
However, an additional item also corresponded with
clinical intuition. The older adults in the clinical
sample identified tension and anxiety as preventing
them from doing important things (rather than
tension and anxiety per se). This may correspond
with a sense of increasing frailty in negotiating
aspects of everyday life, which were previously
unproblematic (e.g., transport, general mobility)
and giving rise to specific concerns (e.g., fear of
falling). These four items were unique in countering
the trend of mean item scores to be lower for the
older adults in the clinical sample. The differential
response in the clinical population of older as
opposed to working-age adults to selected items
argues strongly for comparisons only to be made
against age-related norms and thereby controlling
for their effect.
The key areas of differences between the older
adult and normative data related to missing items
and norms. The finding that there were significantly
greater numbers of items not completed in the older
adult samples might have implications for its accept-
ability. However, recall that the mean number of
items omitted for non-clinical and clinical samples
was less than two items, which is still below the
threshold of three items set in the original article for
excluding the form. On balance, this data, whilst
yielding higher omitted items, does not appear to
indicate the need to modify the length of the CORE-
OM for administration in an older adult sample.
The other key difference in the adult samples
relates to the norms. Although the trend remained
for women to score higher then men, the overall
mean scores were significantly lower for the older
adults in both clinical and non-clinical samples.
The finding that women score higher than men
regardless of age has been reported with other
outcome measures; for example, with the Brief
Symptom Inventory (BSI; de Leo, Frisoni, Rozzini
& Trabucchi, 1993). Our results extend the con-
tinuum of findings regarding the scores of older
adults compared with other adults. For example, de
Leo et al. (1993) found no difference in BSI scores
between adults and older adults while Hale et al.
(1984) found older adults to score higher on the BSI.
Chester (2001) reported findings with the BSI,
which were closer to those of Hale et al. (1984),
dissimilar to de Leo et al. (1993), and distinct from
the originally published norms of Derogatis and
Melisaratos (1983).
However, the results from the present study
suggest that the well-being of older adults declines
progressively, as represented by higher total CORE-
OM scores, in both clinical and non-clinical samples
across age bands >69 and specifically in the non-
clinical sample. This trend may be associated with
findings from the literature which suggest that in
older people there are lower levels of severe mental
health problems but a higher prevalence of psycho-
logical distress (Watts et al., 2002). Overall, findings
from the present study are consistent with the
literature showing decreasing psychological health
with advancing age.
The finding of lower clinical scores for older adults
as compared with the normative data necessitates
244 M. Barkham et al.
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
a lower cut-off point between clinical and non-
clinical samples. From this there are important
practice implications. In effect, these revised norms
are lower by 0.24 and 0.32 respectively, a difference
sufficient to result in false positives (i.e., grouping
some older adults as being in the clinical population
when they are not). Accordingly, we recommend that
the parameters and cut-off points reported here
should be applied to people aged 65 and over when
using the CORE-OM as a screening tool.
Overall, the present findings suggest that the
CORE-OM might be an acceptable and reliable
tool to use with older adults. Its copyleft status (i.e.,
free to use but not to modify or make profit from)
makes it a strong candidate for routine use in
everyday practice settings. In turn, this would help
to build a more relevant and rigorous practice-based
evidence for designing more appropriate and respon-
sive health care for older adults. This process would
complement policy initiatives driven by the more
traditional evidence-based practice paradigm and
might then yield a more robust basis for treatment
and service-delivery decisions (Barkham & Mellor-
Clark, 2003).
Acknowledgements
We would like to thank all participants in this
research for their co-operation. Authors affiliated to
PTRC received support from the NHS Priorities
& Needs R&D Levy via Leeds Community and
Mental Health Teaching Trust.
References
Barkham, M., Evans, C., Margison, F., McGrath, G.,
Mellor-Clark, J., Milne, D. et al. (1998). The rationale for
developing and implementing core batteries in service settings
and psychotherapy outcome research. Journal of Mental Health,
7, 35–47.
Barkham, M., Margison, F., Leach, C., Lucock, M.,
Mellor-Clark, J., Evans, C., et al. (2001). Service profiling
and outcomes benchmarking using the CORE-OM: Towards
practice-based evidence in the psychological therapies. Journal
of Consulting and Clinical Psychology, 69, 184–196.
Barkham, M., & Mellor-Clark, J. (2003). Bridging evidence-based
practice and practice-based evidence: Developing a rigorous
and relevant knowledge for the psychological therapies. Clinical
Psychology & Psychotherapy, 10, 319–327.
Burns, A., Beevor, A., Lelliott, P., Wing, J., Blakey, A., Orrell, M.,
et al. (1999). Health of the nation outcome scales for elderly
people (HoNOS 65þ). British Journal of Psychiatry, 174,
424–427.
Chester, G. A. (2001). Normative data for the brief symptom
inventory for mature and independent living adults. Dissertation
abstracts international: Section B: the sciences & engineering,
62(4-B), 2108. US: Univ Microfilms International.
Cohen, J. (1960). A coefficient of agreement for nominal scales.
Educational and Psychological Measurement, 20, 37–46.
de Leo, D., Frisoni, G. B., Rozzini, R., & Trabucchi, M.
(1993). Italian community norms for the brief symptom
inventory in the elderly. British Journal of Clinical Psychology,
32, 209–213.
Derogartis, L. R., & Melisaratos, N. (1983). The brief symptom
inventory: An introductory report. Psychological Medicine, 13,
595–605.
Evans, C., Mellor-Clark, J., Margison, F., Barkham, M.,
McGrath, G., Connell, J. et al. (2000). Clinical outcomes in
routine evaluation: The CORE-OM. (2000). Journal of Mental
Health, 9, 247–255.
Evans, C., Connell, J., Barkham, M., Margison, F.,
Mellor-Clark, J., McGrath, G. et al. (2002). Towards a
standardised brief outcome measure: Psychometric properties
and utility of the CORE-OM. British Journal of Psychiatry, 180,
51–60.
Evans, O., Singleton, N., Meltzer, H., Stewart, R., & Prince, M.
(2003). The mental health of older people. Office of National
Statistics. London: HMSO.
Froyd, J. E., Lambert, M. J., & Froyd, J. D. (1996). A review of
practices of psychotherapy outcome measurement. Journal of
Mental Health, 5, 11–15.
Hale, W. D., Cochran, C. D., & Hedgepeth, B. E. (1984). Norms
for the elderly on the brief symptom inventory. Journal of
Consulting and Clinical Psychology, 52, 321–322.
Jacobson, N. S., & Truax, P. (1991). Clinical significance: A
statistical approach to defining meaningful change in psy-
chotherapy research. Journal of Consulting and Clinical
Psychology, 59, 12–19.
Knight, B. G. (1996) Psychotherapy with the older adult (Second
Edition). Thousand Oaks, CA: Sage.
Singleton, N., Bumpstead, R., O’Brien, M., Lee, A., &
Meltzer, H. (2001). Psychiatric morbidity among adults living in
private households, 2000. London: The Stationary Office.
Watts, S. C., Bhutani, G. E., Stout, I. H., Ducker, G. M.,
Cleator, P. J., McGarry, J., et al. (2002). Mental health in
older adult recipients of primary care services: Is depression
the key issue? Identification, treatment and the general
practitioner. International Journal of Geriatric Psychiatry, 17,
427–437.
The CORE-OM in an older adult population 245
Dow
nloa
ded
by [
Um
eå U
nive
rsity
Lib
rary
] at
07:
02 2
4 N
ovem
ber
2014
