The interest in studying quality of life in persons with schizophrenia started as concern was increasing about the role played by the chronically mentally ill in the community. This was evident after the de-institutionalisation process which took place in the 1960s and 1970s in western countries (Reference BachrachBachrach, 1970; Reference LambLamb, 1979). Although there is no consensus between researchers on the definition of the concept of ‘quality of life’, it seems that certain patients' characteristics, such as family and social relations, safety, employment and finances, have been identified as the main determinants of quality of life in mentally ill people (Reference Lehman, Ward and LinnLehman et al, 1982; Reference LehmanLehman, 1983a ; Reference Sullivan, Wells and LeakeSullivan et al, 1991). An additional, and fundamental, issue concerning the evaluation of quality of life is the fact that the primary dimension of this construct differs markedly depending on whether objective or subjective criteria are used. Thus, for example Corten (Reference Corten, Mercier and PelcCorten et al, 1994) emphasised subjective satisfaction as a major determinant of quality of life in psychiatric patients. We could therefore conclude that the quality of life of a patient depends not only on his/her personal characteristics and objective life conditions, but also on his/her subjective perception of life circumstances in a variety of domains, often referred to as ‘well-being’ or ‘ life satisfaction’.
Taking these issues into consideration, the Lancashire Quality of Life Profile (LQoLP) (Reference Oliver, Huxley and BridgesOliver et al, 1996) was developed from Lehman's Quality of Life Interview (Reference Lehman, Ward and LinnLehman et al, 1982; Lehman, 1983 Reference Lehman a Reference Lehman b ), combining objective and subjective measures in several life ‘domains’. Oliver has assessed the initial psychometric properties of the LQoLP in chronic psychiatric patients. Construct, content and criterion validity were found acceptable and the evaluation of the internal consistency of the sub-scales was considered good (Oliver et al, 1996a, Reference Oliver, Huxley and Priebe 1997 ). However, test-retest reliability assessment was not performed. A subsequent reliability study conducted by Reference Hansson, Svensson and BjörkmanHansson et al in 1998 reported satisfactory results regarding the instruments' reliability and internal consistency. Nevertheless, the significance of their findings might be questioned, given the limited number of cases (29) included in the study. Thus a thorough and comprehensive verification of the cross-cultural applicability and psychometric properties of the instruments was needed. This verification process has been carried out as a component of the European Psychiatric Services: Inputs Linked to Outcome Domains and Needs (EPSILON) Study, a multi-centre study conducted in five European countries, whose aims are described below.
EPSILON STUDY
Aims
The aims of the EPSILON Study are :
-
(a) To produce standardised versions of five assessment instruments in key areas of mental health service research in five European languages (Danish, Dutch, English, Italian and Spanish), following a rigorous conversion process from the original version into the other four languages by: (i) accurate and independent translation and backtranslation from the original into the other four languages; (ii) checks of cross-cultural applicability using focus groups; and (iii) assessment of instrument reliability. Full details of these procedures are given by Becker et al (Reference Becker, Knapp and Knudsen1999) and elsewhere in this supplement (Reference Becker, Knapp and KnudsenBecker et al, 2000; Reference Knudsen, Vázquez-Barquero and WelcherKnudsen et al, 2000; Reference McCrone, Leese and ThornicroftMcCrone et al, 2000; Reference Ruggeri, Lasalvia and Dall'AgnolaRuggeri et al, 2000; Reference Schene, Koeter and van WijngaardenSchene et al, 2000).
-
(b) To obtain and compare data from five regions in different European countries, each with their particular system of health care, about social and clinical variables, characteristics of mental health care and its costs. The results of this are now being prepared for publication.
-
(c) To test both instrument-specific and cross-instrument hypotheses; full details of this stage of the study will be published in due course.
The specific aim of the present paper is to describe the development of a European Version of the LQoLP, the LQoLP-EU, at the same time verifying its reliability and internal consistency in representative samples of people with schizophrenia in the five European centres participating in the EPSILON Study. The paper should be read in close conjunction with other related papers in this series, which give more detailed accounts of key related aspects of the study (Becker et al, Reference Becker, Knapp and Knudsen1999, 2000; Reference Knudsen, Vázquez-Barquero and WelcherKnudsen et al, 2000; Reference Schene, Koeter and van WijngaardenSchene et al, 2000).
Outcome scales included in the reliability study
The reliability study included the conversion of five scales from their original language into the other four study languages. The scales are: Camberwell Assessment of Need - European Version (CAN-EU), Client Socio-Demographic Service Receipt Inventory - European Version (CSSRI-EU), Involvement Evaluation Questionnaire - European Version (IEQ-EU), Lancashire Quality of Life Profile - European Version (LQoLP-EU), and the Verona Service Satisfaction Scale - European Version (VSSS-EU).
Lancashire Quality of Life Profile - European Version (LQoLP-EU)
The LQoLP was originally developed by Oliver et al (Reference Oliver, Huxley and Bridges1996) from Lehman's Quality of Life Interview (Lehman, 1983 Reference Lehman a , Reference Lehman b ; Reference Lehman, Ward and LinnLehman et al, 1982). It is a structured self-report interview (to be administered by trained interviewers) comprising 105 items. It includes the following nine domains: work and education (7 items); leisure and participation (8 items); religion (4 items); finances (7 items); living situation (12 items); legal status and safety (5 items); family relations (7 items); social relations (6 items); and health (10 items). The subjective components of these ‘domains’ are evaluated on a seven-point Life Satisfaction Scale (LSS). In addition, the interview allows the assessment of the following additional areas: (a) positive and negative affect (10 items) with the Bradburn (Reference Bradburn1969) Affect-Balance Scale; (b) Self-Esteem Scale (10 items) (Reference RosenbergRosenberg, 1965); (c) measures of Global Well-Being, including two items: Cantril's Ladder (Reference CantrilCantril, 1965) and a Happiness Scale (Reference Gurin, Veroff and FeldGurin et al, 1960); (d) the Quality of Life Uniscale, which gives an opportunity of evaluating the quality of life of the patient (Reference Spitzer and DobsonSpitzer & Dobson, 1981) independently of the patient's own opinion; (e) the perceived Quality of Life Score, an average of the sum of the subjective items of the first nine domains.
MATERIAL AND METHOD
Study sites
The criteria used to identify study centres, and full details of the general population characteristics of the study sites, are given in Becker et al (Reference Becker, Knapp and Knudsen1999). In brief, the criteria were similar to those used by Dowrick et al (Reference Dowrick, Casey and Dalgard1998). Six partners in five centres joined forces for this collaborative study, and the research teams were located in Amsterdam, Copenhagen, London (Centre for the Economics of Mental Health, and Section of Community Psychiatry, Institute of Psychiatry), Santander and Verona.
Translation and cultural adaptation of the LQoLP-EU
Before the reliability analyses were undertaken, the original English version of the LQoLP was converted into its European versions (Dutch, Danish, Spanish and Italian). The procedure followed was translation, back-translation, focus groups and target checking. Focus groups took place in Amsterdam, Copenhagen, Santander and Verona. The aim of the focus groups was to identify conflicting areas in the wording and cultural applicability of the instrument. As a consequence, a number of changes were made to the local translations of the instruments (for additional details see Reference Knudsen, Vázquez-Barquero and WelcherKnudsen et al, 2000, this supplement). Specifically, as regards the LQoLP-EU, as a result of the experience obtained in the focus group and pilot testing of the instrument, a detailed manual was produced to clarify aspects related to its administration and scoring.
Case identification
Cases included in the study were adults aged 18-65, selected as representative of all people suffering from schizophrenia who were in contact with services in each of the five study sites. Study samples were identified either from psychiatric case registers (in Copenhagen and Verona) or case-loads of local specialist mental health services (in-patient, out-patient and community). Patients included had been in contact with mental health services during the 3-month period before the start of the study (September 1997). Patients with an ICD-10 clinical diagnosis of F20-F25 were considered as candidates for the study. The diagnosis was confirmed using the item group checklist (IGC), which is part of the Schedule for Clinical Assessment in Neuropsychiatry (SCAN) (Reference WingWorld Health Organization, 1992; Reference Vázquez-BarqueroVázquez-Barquero, 1993). Finally, only patients with an ICD-10 F20 research diagnosis were included in the study. In addition, the following exclusion criteria were applied: current residence in prison, secure residential services or hostels for long-term patients; co-existing mental retardation, primary dementia or other severe organic disorder; and periods of in-patient treatment lasting longer than one year.
Study procedure
The numbers of patients finally included in the study varied from 52 to 107 in the five different sites, with a total of 404 participants in the study as a whole. From those patients, a sub-sample of 294 was selected to participate in the test-retest study, and these patients were interviewed twice with the LQoLP-EU. The time interval between the two interviews ranged from 1 to 2 weeks, and the same interviewer performed both interviews with each patient. Patients were given oral information about the purpose and procedures of the study by the interviewer and asked for their consent to participate. Interviewers were previously trained in the use of the LQoLP-EU; in addition, all interviewers received training in the use of the SCAN and all other EPSILON Study instruments. SCAN training was carried out at the Institute of Psychiatry, London, and at the Clinical and Social Psychiatry Research Unit, Santander. There were regular follow-up meetings to ensure the standardised use of instruments, and a series of study coordinating meetings. The coordinating centre (in London) prepared the SPSS templates used at all the participating sites to store the information gathered during the interviews, thus ensuring data consistency and homogeneity.
Reliability testing
Reliability testing in the EPSILON Study is conducted on several levels, depending on the nature of the instruments involved and the way they are administered (interviews v. questionnaires). Two kinds of reliability tests were used: the Cronbach's α statistic to check the internal consistency of scales and sub-scales consisting of more than one item, and intraclass correlation coefficient (ICC), to check the test-retest reliability of scales and sub-scales. These statistics are discussed in Streiner & Norman (Reference Streiner and Norman1995). Each step in the analysis was described in an analysis protocol, which was followed by all sites.
First, differences in sample variances were explored using the Levene test. Cronbach's α was computed for each site and for the pooled sample, and a test for differences in α values between sites was performed as well (Reference Feldt, Woodruff and SalihFeldt et al, 1987). Intra-class correlation coefficients were computed by maximum likelihood estimation of a variance components model with patients entered as random effects, and (in the case of pooled estimates) site entered as a fixed effect.
The ratio of the between-patient to total variance was used for the ICC, and the variance-covariance matrix for the components was used to obtain standard errors based on the delta technique (Reference Dunn, Streiner and NormanDunn, 1989). Fisher's Z transformation was applied to all ICCs to enable approximate comparisons to be made between sites (Reference Donner and BullDonner & Bull, 1983), and differences between sites were tested for significance by the method of weighting (Reference Armitage and BerryArmitage & Berry, 1994) before transforming back to the ICC scale. The standard error of measurement was obtained from the ‘error’ component of variance. Finally, a paired t-test on test-retest data was carried out in order to assess any systematic changes from time 1 to time 2.
Because there were missing items, mean substitution (means over all valid cases) was used before estimating α. This procedure is likely to provide conservative estimates of the true α values (i.e. underestimates), but may magnify the apparent significance of inter-site differences, because it underestimates the standard errors of α. In most cases the number of missing values was small, and they can reasonably be assumed to be random. One exception is the Work sub-scale, where there were many items ‘not applicable’ (such as work satisfaction, for those unemployed). The items from the Work sub-scale were omitted from the estimation of α for the total (ISS) scale.
For reasons of comparability, all sites used the same procedure and the same software for all instruments: SPSS for Windows 7.5 or higher (Reference NorusisNorusis, 1993), the Amsterdam α -testing program ALPHA. EXE (based on Reference Feldt, Woodruff and SalihFeldt et al, 1987), and MicrosoftTM Excel for tests of the homogeneity of ICCs.
RESULTS
The mean time taken for the interview in the first series was 27 ± 12 min (10-90 min). Table 1 shows summary statistics for the first interview, including the results of tests of homogeneity of the means and s.d., for the whole sample of 404 patients (i.e. the main study sample, of which the reliability sample is a sub-sample). The table shows lack of homogeneity in variance for the global quality of life assessment (LSS average score) and mixed values for the subjective sub-scales (significant values for Work, Leisure Activities, Religion, Finances, Family Relations, Social Relations, QoL Uniscale and Positive Self-Esteem were found). A discussion of differences in the mean levels (which in almost all cases differed significantly between sites) will follow in a subsequent paper.
Sub-scales | Pooled n=404 | Amsterdam n=61 | Copenhagen n=52 | London n=84 | Santander n=100 | Verona n=107 | Test equality of means (P-value) | Test equality of s.d. (P-value) | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Items | mean | s.d. | mean | s.d. | mean | s.d. | mean | s.d. | mean | s.d. | mean | s.d. | |||
LSS average score | 27 | 4.67 | 0.76 | 4.61 | 0.72 | 4.98 | 0.70 | 4.24 | 0.60 | 4.83 | 0.69 | 4.67 | 0.84 | <0.01 | 0.01 |
Work situation | 3 | 4.22 | 1.40 | 4.04 | 1.24 | 5.23 | 1.20 | 3.98 | 1.07 | 4.06 | 1.37 | 4.22 | 1.57 | <0.01 | <0.01 |
Leisure activities | 3 | 4.67 | 1.00 | 4.71 | 1.11 | 4.88 | 0.80 | 4.35 | 0.92 | 4.71 | 0.95 | 4.67 | 1.08 | <0.01 | 0.01 |
Religion | 2 | 5.13 | 1.17 | 4.56 | 1.40 | 5.29 | 1.24 | 5.17 | 0.92 | 5.12 | 1.20 | 5.13 | 1.07 | <0.01 | <0.01 |
Finances | 2 | 4.26 | 1.52 | 4.21 | 1.35 | 4.94 | 1.65 | 3.72 | 1.27 | 4.24 | 1.62 | 4.26 | 1.49 | <0.01 | 0.04 |
Living situation | 7 | 4.74 | 1.06 | 4.74 | 0.92 | 5.02 | 0.96 | 4.21 | 1.17 | 4.99 | 0.91 | 4.74 | 1.07 | <0.01 | 0.22 |
Safety | 2 | 4.94 | 1.31 | 4.74 | 1.35 | 5.34 | 1.12 | 4.32 | 1.22 | 5.21 | 1.15 | 4.94 | 1.40 | <0.01 | 0.22 |
Family relations | 3 | 4.96 | 1.35 | 5.01 | 1.11 | 4.81 | 1.50 | 4.35 | 1.33 | 5.55 | 1.22 | 4.96 | 1.32 | <0.01 | 0.02 |
Social relations | 2 | 4.70 | 1.24 | 4.72 | 1.32 | 4.78 | 1.28 | 4.40 | 0.80 | 4.93 | 1.20 | 4.70 | 1.46 | 0.07 | <0.01 |
Health | 3 | 4.61 | 1.15 | 4.82 | 1.10 | 4.64 | 1.07 | 3.98 | 0.98 | 4.67 | 1.13 | 4.61 | 1.17 | <0.01 | 0.27 |
Global well-being | 2 | 4.37 | 1.34 | 4.64 | 1.25 | 4.59 | 1.23 | 3.94 | 1.23 | 4.23 | 1.38 | 4.37 | 1.42 | <0.01 | 0.10 |
Positive affect | 5 | 0.52 | 0.34 | 0.34 | 0.34 | 0.53 | 0.32 | 0.47 | 0.31 | 0.37 | 0.30 | 0.52 | 0.36 | <0.01 | 0.25 |
Negative affect | 5 | 0.45 | 0.33 | 0.42 | 0.29 | 0.38 | 0.32 | 0.60 | 0.33 | 0.38 | 0.31 | 0.45 | 0.33 | <0.01 | 0.34 |
Positive self-esteem | 5 | 0.71 | 0.32 | 0.73 | 0.23 | 0.65 | 0.34 | 0.65 | 0.35 | 0.79 | 0.27 | 0.71 | 0.34 | <0.01 | <0.01 |
Negative self-esteem | 5 | 0.38 | 0.32 | 0.40 | 0.31 | 0.39 | 0.33 | 0.40 | 0.26 | 0.39 | 0.34 | 0.38 | 0.33 | 0.54 | 0.04 |
Cantril's Ladder | 1 | 58.94 | 39.47 | 66.22 | 25.57 | 60.66 | 24.00 | 61.31 | 72.67 | 50.81 | 24.59 | 59.90 | 21.83 | <0.01 | 0.30 |
QoL Uniscale | 1 | 54.28 | 19.74 | 57.46 | 16.41 | 55.27 | 18.67 | 51.08 | 18.42 | 54.60 | 22.65 | 53.20 | 20.43 | <0.01 | 0.01 |
Internal consistency
Internal consistency evaluates the interrelatedness of the items in the LQoLP-EU. It was assessed by the Cronbach's α coefficient with a single administration of the instrument. Three items were excluded from this analysis, since they are administered to certain patients only: those married, retired, or previously admitted to psychiatric hospitals.
The α coefficients were high for LSS average score, as shown in Table 2, with a pooled estimate of 0.87 (95% CI 0.85-0.88). In the remaining nine subjective sub-scales, Cronbach's α ranged from 0.30 (Work) to 0.88 (Finances). The domains with the lowest values were Work (0.30, 0.14-0.42), Leisure Activities (0.56, 0.48-0.52), Religion (0.62, 0.53-0.58) and Social Relations (0.66, 0.58-0.72), and for these four sub-scales the reliabilities were very low for some individual sites (see Table 2).
Sub-scales | Items | Pooled | Amsterdam | Copenhagen | London | Santander | Verona | Test of equality of α |
---|---|---|---|---|---|---|---|---|
n=4041 | n=61 | n=52 | n=84 | n=100 | n=107 | (P-value) | ||
LSS score | 242 | 0.87 (0.85-0.88) | 0.84 (0.77-0.89) | 0.83 (0.75-0.89) | 0.86 (0.81-0.89) | 0.83 (0.77-0.87) | 0.90 (0.87-0.92) | 0.11 |
Work | 2 | 0.30 (0.14-0.42) | 0.12 (-0.50 to 0.45) | 0.64 (0.35-0.78) | 0.76 (0.62-0.84) | 0.18 (-0.23 to 0.44) | 0.47 (0.20-0.63) | <0.01 |
Leisure activities | 3 | 0.56 (0.48-0.62) | 0.67 (0.47-0.78) | 0.29 (0.15-0.55) | 0.66 (0.51-0.76) | 0.42 (0.17-0.58) | 0.62 (0.47-0.72) | 0.07 |
Religion | 2 | 0.62 (0.53-0.68) | 0.33 (-0.15 to 0.58) | 0.48 (0.05-0.69) | 0.92 (0.87-0.94) | 0.71 (0.56-0.80) | 0.62 (0.42-0.74) | <0.01 |
Finances | 2 | 0.88 (0.85-0.90) | 0.72 (0.53-0.83) | 0.93 (0.87-0.95) | 0.98 (0.96-0.98) | 0.88 (0.82-0.91) | 0.85 (0.78-0.89) | <0.01 |
Living situation | 6 | 0.85 (0.82-0.87) | 0.78 (0.68-0.85) | 0.75 (0.62-0.84) | 0.94 (0.91-0.95) | 0.74 (0.65-0.81) | 0.86 (0.81-0.89) | <0.01 |
Safety | 2 | 0.82 (0.78-0.85) | 0.79 (0.64-0.87) | 0.79 (0.63-0.87) | 0.96 (0.93-0.97) | 0.68 (0.52-0.78) | 0.81 (0.72-0.87) | <0.01 |
Family relations | 2 | 0.80 (0.75-0.83) | 0.69 (0.46-0.80) | 0.72 (0.49-0.83) | 0.91 (0.86-0.94) | 0.92 (0.88-0.94) | 0.65 (0.48-0.76) | <0.01 |
Social relations | 2 | 0.66 (0.58-0.72) | 0.65 (0.39-0.78) | 0.58 (0.25-0.75) | 0.56 (0.30-0.70) | 0.49 (0.22-0.65) | 0.83 (0.75-0.88) | 0.02 |
Health | 3 | 0.74 (0.69-0.78) | 0.72 (0.57-0.82) | 0.66 (0.44-0.78) | 0.79 (0.69-0.85) | 0.66 (0.52-0.76) | 0.75 (0.65-0.82) | 0.49 |
Global well-being | 2 | 0.83 (0.79-0.86) | 0.75 (0.56-0.85) | 0.83 (0.69-0.90) | 0.94 (0.91-0.96) | 0.82 (0.72-0.87) | 0.79 (0.69-0.86) | <0.01 |
Positive affect | 5 | 0.74 (0.70-0.78) | 0.71 (0.56-0.81) | 0.71 (0.55-0.81) | 0.64 (0.49-0.74) | 0.70 (0.59-0.78) | 0.83 (0.77-0.88) | <0.01 |
Negative affect | 5 | 0.68 (0.62-0.72) | 0.58 (0.36-0.72) | 0.66 (0.46-0.78) | 0.72 (0.60-0.80) | 0.66 (0.54-0.75) | 0.68 (0.56-0.76) | <0.01 |
Positive self-esteem | 5 | 0.77 (0.73-0.80) | 0.71 (0.56-0.80) | 0.78 (0.66-0.86) | 0.81 (0.73-0.86) | 0.70 (0.59-0.78) | 0.81 (0.74-0.86) | 0.35 |
Negative self-esteem | 5 | 0.67 (0.62-0.72) | 0.64 (0.46-0.76) | 0.74 (0.59-0.83) | 0.45 (0.23-0.61) | 0.75 (0.66-0.82) | 0.74 (0.65-0.81) | 0.04 |
In the Self-Esteem scale the results showed the following coefficients: 0.77 (0.73-0.80) for positive self-esteem and 0.67 (0.62-0.72) for negative self-esteem. For the Affect Balance scale, the results were 0.74 (0.70-0.78) for positive affects and 0.68 (0.62-0.72) for negative affects.
Test-retest reliability
Table 3 shows the intraclass correlation between the test and retest interviews for the patients in the reliability sub-sample (only 294 patients took part in the retesting). The pooled ICC score for global satisfaction (LSS) was 0.82. The nine life domain sub-scale ICCs ranged between 0.61 (Safety) to 0.75 (Living Situation). ICC estimations for the Affect balance scale were 0.72 for positive affect and 0.71 for negative affect. In the Self-Esteem scale the results were 0.71 for positive self-esteem and 0.63 for negative self-esteem. Test-retest reliabilities were 0.65 for Cantril's Ladder, 0.78 for Global Well-Being and 0.81 for QoL Uniscale.
Sub-scale | Pooled n=264 | Amsterdam n=51 | Copenhagen n=46 | London n=51 | Santander n=50 | Verona n=66 | Test of equality of ICCs (P-value) | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ICC | (s.e.)m | ICC | (s.e.)m | ICC | (s.e.)m | ICC | (s.e.)m | ICC | (s.e.)m | ICC | (s.e.)m | ||
LSS score | 0.82 | 0.29 | 0.86 | 0.24 | 0.83 | 0.28 | 0.82 | 0.24 | 0.85 | 0.24 | 0.77 | 0.41 | 0.65 |
Work situation | 0.66 | 0.76 | 0.73 | 0.63 | 0.38 | 0.92 | 0.65 | 0.62 | 0.82 | 0.61 | 0.61 | 0.94 | 0.01 |
Leisure activities | 0.68 | 0.56 | 0.79 | 0.51 | 0.62 | 0.54 | 0.61 | 0.53 | 0.68 | 0.55 | 0.67 | 0.67 | 0.34 |
Religion | 0.62 | 0.67 | 0.51 | 0.88 | 0.71 | 0.64 | 0.58 | 0.66 | 0.74 | 0.51 | 0.54 | 0.69 | 0.24 |
Finances | 0.72 | 0.73 | 0.72 | 0.69 | 0.73 | 0.74 | 0.87 | 0.42 | 0.81 | 0.69 | 0.56 | 1.04 | <0.01 |
Living situation | 0.75 | 0.51 | 0.67 | 0.51 | 0.67 | 0.55 | 0.87 | 0.42 | 0.79 | 0.39 | 0.65 | 0.63 | <0.01 |
Safety | 0.61 | 0.74 | 0.64 | 0.75 | 0.76 | 0.51 | 0.72 | 0.63 | 0.71 | 0.62 | 0.32 | 1.06 | <0.01 |
Family relations | 0.66 | 0.73 | 0.15 | 0.64 | 0.71 | 0.75 | 0.73 | 0.64 | 0.59 | 0.86 | 0.61 | 0.86 | 0.15 |
Social relations | 0.65 | 0.67 | 0.75 | 0.61 | 0.72 | 0.66 | 0.52 | 0.57 | 0.81 | 0.48 | 0.51 | 0.93 | <0.01 |
Health | 0.71 | 0.56 | 0.81 | 0.42 | 0.71 | 0.59 | 0.65 | 0.54 | 0.75 | 0.49 | 0.65 | 0.68 | 0.27 |
Global well-being | 0.78 | 0.58 | 0.73 | 0.64 | 0.83 | 0.45 | 0.72 | 0.59 | 0.81 | 0.56 | 0.81 | 0.62 | 0.36 |
Positive affect | 0.72 | 0.16 | 0.82 | 0.14 | 0.61 | 0.19 | 0.79 | 0.13 | 0.81 | 0.12 | 0.61 | 0.21 | 0.05 |
Negative affect | 0.71 | 0.16 | 0.66 | 0.15 | 0.79 | 0.13 | 0.61 | 0.20 | 0.81 | 0.14 | 0.75 | 0.17 | 0.07 |
Positive self-esteem | 0.71 | 0.16 | 0.68 | 0.14 | 0.83 | 0.13 | 0.73 | 0.17 | 0.41 | 0.19 | 0.77 | 0.14 | <0.01 |
Negative self-esteem | 0.63 | 0.18 | 0.57 | 0.21 | 0.73 | 0.15 | 0.49 | 0.19 | 0.72 | 0.18 | 0.66 | 0.18 | 0.10 |
Cantril's Ladder | 0.65 | 13.49 | 0.59 | 15.09 | 0.85 | 9.36 | 0.81 | 9.64 | 0.64 | 15.03 | 0.38 | 17.54 | <0.01 |
QoL Uniscale | 0.81 | 7.61 | 0.72 | 7.81 | 0.75 | 9.05 | 0.85 | 6.85 | 0.91 | 5.83 | 0.49 | 13.92 | <0.05 |
Test-retest ICCs were good for the LSS (pooled estimate 0.82, 95% CI 0.78-0.85) and also for the individual sub-scales, which ranged from 0.61 to 0.75. The coefficients appear to be higher for the LSS than for the individual sub-scales, perhaps due to the greater stability of the LSS, being the total of many items. There is evidence for differences between sites for the individual sub-scales but not for the LSS. The only centre with relatively low ICCs for sub-scales is Verona, but paired t-tests on the time 2 and time 1 results show this is not due to an overall tendency to higher or lower values at retest in Verona, but can be explained by random variation. There are relatively high values of the standard error of measurement (s.e.)m associated with the low ICCs at Verona, and the higher (s.e.)m in Verona is the explanation for the overall higher standard deviation (0.84 in Table 1).
DISCUSSION
This paper presents the results of the reliability of the LQoLP-EU in relation to its internal consistency and test-retest stability. The internal consistency was assessed in the original instrument by Oliver et al (Reference Oliver, Huxley and Priebe1997). In several studies they obtained a Cronbach's α value ranging between 0.75 and 0.86 for the LSS scale; and other authors (Reference Kaiser, Priebe and BarrKaiser et al, 1997) reported results ranging from 0.53 to 0.79. Our own results corroborate the adequate internal consistency of the total scale (LSS) of the LQoLP-EU, since Cronbach's α value was 0.87.
Additionally, the internal consistency of nine subjective life domains was adequate for most sub-scales, although there were a few that were less satisfactory. The most conflicting domains were: Work (0.30), Leisure Activities (0.56) and Religion (0.62). In the case of Work, values differed widely between centres, the highest being in London (0.76), and the lowest in Amsterdam (0.12) and Santander (0.18). The Pearson correlation between items in this domain was not significantly different from zero in the two latter centres (Amsterdam 0.107, Santander 0.135). It has to be recognised that assessment of the internal consistency of this sub-domain is very difficult, because of items which are missing or ‘not applicable’. It may be that a much larger sample is needed, specifically directed at this particular item, or possibly even reconstruction of the sub-scale from other items. In the domain of Religion, the centre showing the most conflicting results was Amsterdam (0.33), and this may be because religion is a very sensitive topic for people with schizophrenia, as indeed for the general population.
Other authors (Reference Oliver, Huxley and PriebeOliver et al, 1997) had originally found variations between centres in certain domains. In particular, they found most conflict in the results in the areas of Safety, Religion, Living Situation, Leisure and Work. In the five studies quoted, coefficients for the Work domain ranged from 0.53 to 0.80; for Leisure they ranged from 0.59 to 0.8, and for Religion from 0.45 to 0.85. A possible explanation for this (in previous studies as well as in the current study) is that items included in these domains represent discrete concepts that do not compose a single dimension. Thus an individual may achieve a high score on some and a low one on others. For example, a patient could be very satisfied with his/her job, but not with the amount of money he/she earns; or outdoor leisure activities do not require the same abilities as indoor leisure activities.
In our study, the Cronbach's α values for the Affect Balance and Self-Esteem scales were all satisfactory, although α was lower for negative self-esteem (0.67) than for positive self-esteem for (0.77). The values are similar to those detected by previous authors (Reference Hansson, Svensson and BjörkmanHansson et al, 1998).
The test-retest ICCs for the total subjective satisfaction score were good, and they were adequate for each of the nine life domains. Although there was some evidence for differences between the sites in the reliabilities of the sub-domains, they were all above 0.61. The total score showed no evidence of differences, and a pooled estimate is 0.82 (95% CI 0.79-0.86). Unfortunately, in previous studies with this instrument a similar analysis was not carried out, and thus comparison with other data is not possible.
Finally, the paired t-tests analysis between interviews at time 1 and time 2 show that there is no overall tendency to higher or lower values at the retest, thus indicating that the values between test and retest are reasonably stable. Therefore it seems that the time between the test and the retest is not long enough to influence the results. Furthermore, there seems to be no tendency for patients to modify their appraisal of quality of life in the second interview (i.e. because answering the questions make them reassess their quality of life).
CONCLUSIONS
The present study shows that the Lancashire Quality of Life Profile - European Version is a useful instrument, which satisfies the requirements of a good quality-of-life measure in different European settings. The total score has good internal consistency and reliability, and is similar across sites, although some subscales, especially Work, appear to be less satisfactory from the point of view of consistency, as also found in other studies. Furthermore it is user-friendly, takes about half an hour to administer (on average) and is available in different languages. As a consequence of this study, there are versions available in English, Dutch, Danish, Italian and Spanish, each accompanied by a manual clarifying the use of the instrument in the different settings.
Acknowledgements
The following colleagues contributed to the EPSILON Study. Amsterdam: Dr Maarten Koeter, Karin Meijer, Dr Marcel Monden, Professor Aart Schene, Madelon Sijsenaar, Bob van Wijngaarden; Copenhagen: Dr Helle Charlotte Knudsen, Dr Anni Larsen, Dr Klaus Martiny, Dr Carsten Schou, Dr Birgitte Welcher; London : Professor Thomas Becker, Dr Jennier Beecham, Liz Brooks, Daniel Chisholm, Gwyn Griffiths, Julie Grove, Professor Martin Knapp, Dr Morven Leese, Paul McCrone, Sarah Padfield, Professor Graham Thornicroft, Ian R. White; Santander: Andrés Arriaga Arrizabalaga, Sara Herrera Castanedo, Dr Luis Gaite, Andrés Herran, Modesto Perez Retuerto, Professor José Luis Vázquez-Barquero, Elena Vázquez-Bourgon; Verona: Dr Francesco Amaddeo, Dr Giulia Bisoffi, Dr Doriana Cristofalo, Dr Rosa Dall'Agnola, Dr Antonio Lasalvia, Dr Mirella Ruggeri, Professor Michele Tansella.
This study was supported by the European Commission BIOMED-2 Programme (Contract BMH4-CT95-1151). Thanks are due to Dr J. Oliver for his permission to include the LQoLP in the EPSILON Study and for his helpful comments and assistance in its administration and scoring. We would also like to acknowledge the sustained and valuable assistance of the users, carers and the clinical staff of the services in the five study sites. In Amsterdam, the EPSILON Study was partly supported by a grant from the Nationaal Fonds Geestelijke Volksgezondheid and a grant from the Netherland Organization for Scientific Research (940-32-007). In Santander the EPSILON Study was partially supported by the Spanish Institute of Health (FIS) (FIS Exp. No. 97/1240). In Verona additional funding for studying patterns of care and costs of a cohort of patients with schizophrenia were provided by the Regione del Veneto, Giunta Regionale, Ricerca Sanitaria Finalizzata, Venezia, Italia (Grant No. 723/01/96 to Professor M. Tansella).
eLetters
No eLetters have been published for this article.