Importance of spiritual well-being in assessment of recovery: the Service-user Recovery Evaluation (SeRvE) scale

Abstract

Aims and method

To develop a self-report questionnaire to measure mental health recovery from the service user viewpoint. Literature searches and scoping exercises indicated that psychological, social and spiritual issues should be included. The resultant provisional scale was completed by 107 service users.

Results

The provisional scale was shortened as a result of factor analysis. The finalised version was highly reliable (Cronbach's alpha 0.911) and valid, correlating significantly with an already established recovery scale. It contained nine recognisable subscales, the first two describing existential and religious well-being. Separate well-being and ill-being factors were also identified.

Clinical implications

An inclusive tool for service users' assessment of their own recovery, the Service User Recovery Evaluation (SeRvE) scale, has been validated. This can be used both as a research tool and clinically to monitor interventions. The importance of spiritual care for service users is highlighted.

There is evidence that clinicians and their patients often have a different view of recovery from mental illness. ¹ Traditionally, clinicians tend to focus on specific symptom reduction, ² whereas patients focus more generally on the psychological, the social and the spiritual. ¹ Even clinicians’ rating of symptoms can be very different from that of the service user. ³ There are well-established methods by which clinicians can make assessments of the recovery of their patients. Symptom severity is often recorded using the Health of the Nation Outcome Scales (HoNOS) tool. ⁴ However there is a shortage of scales for service users to rate their own recovery. For service users, recovery can occur even alongside psychosis and does not necessarily mean cure; ⁵ some even see their symptoms as a meaningful creative opportunity. ^6,7 The present study compiles a scale for service users to assess their own view of their recovery. It is concerned with psychological, social and spiritual aspects of recovery without making any attempt to directly measure symptom severity.

The psychological and social dimensions of recovery have been frequently written about ^8-15 and there are already some scales in the literature to measure them, although none inclusive of all the issues raised. ^16-23 To date the spiritual component of recovery has been rather neglected, despite its proven importance to many service users. ²⁴ Spirituality is that which gives meaning, purpose and hope, whether or not it includes a formal religious faith. ²⁵ In his paper of 2004, Cook (p. 548) ²⁶ attempts a working definition:

Spirituality is a distinctive, potentially creative and universal dimension of human experience arising from both within the inner subjective awareness of individuals and within communities, social groups and traditions. It may be experienced as relationship with that which is intimately inner, immanent and personal, within the self or others and/or as relationship with that which is wholly other, transcendent and beyond the self. It is experienced as being of fundamental or ultimate importance and is thus concerned with matters of meaning and purpose in life, truth and values.

Spiritual well-being occurs when this experience of spirituality adds to overall well-being. It is seen as crucial by many service users for their recovery, ²⁷ thus it is important that this dimension is included in our new scale.

Despite the existence of many measures of spiritual well-being. ²⁸ there is not one specifically for mental health service users. The Spiritual Well-Being Scale (SWBS) of Ellison has been used for this purpose in some studies. ^29,30 It has two subscales, religious and existential. In a previous pilot study we found the SWBS to be unsuitable here in Birmingham, UK, due to the religious subscale being expressed in evangelical, Christian language. Understandably, this was largely meaningless in the context of a multicultural, multifaith population. As purely religious issues are important for many service users, ^31,32 it is necessary to include a measure of religious well-being in our scale but the questions need to be more appropriate. The existential subscale of the SWBS is more easily understood by our service users, being concerned with experience and meaning of life in general. This subscale overlaps in subject matter with scales of psychological well-being. ¹⁴ It was decided that it is impossible to separate manifestations of existential well-being and psychological well-being, and that it was suitable to combine the psychosocial and existential dimensions, with a measure of religious well-being, into one inclusive scale of recovery from the service user's viewpoint. It has been named the Service-user Recovery Evaluation (SeRvE) scale.

Method

The team

The team had multidisciplinary input, led by a long-term service user (J.M.B.), supported by a research assistant (M.P.) and a consultant psychiatrist who is also a theologian (C.C.H.C.). A medical statistician (H.P.) was recruited to the team during the project to provide specialist support.

Design of the SeRvE scale

Concerning psychosocial issues, a literature search revealed a wide range of topics valued by service users for recovery. ^8,9,11,33-35 This included hope, self-esteem, empowerment, good relationships, positive and stable affect, stigma and shame, identity, meaning, purpose and satisfaction with life. From this a comprehensive list of questions about psychosocial recovery was compiled. This list had been tested with 37 service users in an unpublished pilot study by the same research team and found to be reliable and valid.

Concerning spiritual issues, a new list of spiritual well-being questions was drawn up, based on broader definitions of spirituality ^25,26,36-42 and existing scales. ^43-46 This included existential as well as purely religious issues. The existential questions overlapped with the psychosocial as described above. The questions on religious matters were designed to be accessible, encouraging participants to insert their own word for God or higher power according to their particular religion or belief systems. Questions were added concerning specific religious problems that some service users report. ⁴⁷ A scoping day at the University of Birmingham was then held, in which 30 mental health service users gave their views on spirituality and its importance for their mental health. Many people had a religious faith of some sort that was important and helpful to them. For some the concept of connectedness was important for their spiritual well-being. Many people derived spiritual inspiration from the arts, music or nature. Some new questions were added accordingly. Each question from the list was then rated by the participants from one to three for both ease of understanding and relevance to spiritual well-being. Questions were discarded if their mean score for either was less than two.

This list of existential and religious questions was then combined with the list of psychosocial questions to make a provisional inclusive scale of service user recovery. Feedback was received from the service users’ in research forum at the University of Birmingham, Suresearch. The resultant scale (provisional SeRvE), contained 67 questions answered on a five-point Likert scale, numbered from one to five, and was set out in subject headings for ease of completion (how you feel about your life, your emotions, your relationships and your religious well-being). Each part contained a mixture of positively and negatively worded questions. For the faith-based questions, the option of answering ‘not applicable’ (n/a) was given, for people to whom these particular questions meant nothing.

Other scales used

Two other scales were used in this study. The first was the Mental Health Recovery Measure (MHRM), ²² an established scale of service user recovery. The second was the Spiritual Well-Being Questionnaire (SWBQ), ⁴⁸ an established scale of spiritual well-being. These scales were chosen to be appropriate to look for correlations with the SeRvE scale for purposes of validation.

Sample

A convenience sample of 107 working age, adult mental health service users was recruited, half from in-patient and half from out-patient units from four wards, four day centres and one out-patient clinic, all in Birmingham and Solihull Mental Health NHS Foundation Trust. Units were visited in turn and clinical staff suggested patients who might be agreeable to participate. At each visit, all available patients were invited. However, we acknowledge that our sample is likely to be biased in favour of the patients who were less acutely unwell, these being more able and willing to participate. Exclusion criteria were individuals under 18, people with known organic brain disease and those with concurrent acute physical illness. People over 65 were also excluded because of the increased possibility of them having early stage organic brain dysfunction.

Study design

This study was approved by the Black Country Research Ethics Committee, on 3 August 2010, REC reference number 10/H1202/52. Potential participants were approached by a member of the clinical staff. If they were agreeable, the research assistant gave them the information sheet to read. If they gave fully informed consent, they were given the three questionnaires (provisional SeRvE scale, MHRM and SWBQ) to complete. Assistance in reading the questions, from a staff member or one of the research team, was available if requested. The completed questionnaires with signed consent forms were then returned to the research assistant. Also recorded were: date of birth, gender, unit they were in, the nature of their religious faith, if applicable, and the importance to them of their spirituality. Participants were asked whether they would consent to their medical notes being accessed by the research team to retrieve diagnosis; only 50% gave such consent.

Data analysis

Results from all the questionnaires were entered manually onto a computer and analysed with SPSS software (version 18 for Windows). All negative questions were scored negatively, for example a high score for agitation (5) would be entered as the appropriate low score (1). Factor analysis was performed to look for meaningful categories within the provisional SeRvE scale. For this, any not applicable responses from faith-based questions were encoded as zeros to distinguish them from random missing values. First, the data were checked for correlations via Bartlett's test of sphericity, which found that the variables were correlated with each other (P<0.01). The Kaiser-Meyer-Olkin measure of sampling adequacy was calculated at 0.739. The data are therefore suitable for principal component analysis (PCA). A PCA using varimax orthogonal rotation was used to maximise the differences between factors. Item communalities were all found to be greater than 0.5, hence all item variances are well represented in the model. Factors were only retained if they had an eigenvalue of greater than one. Items were each checked to see whether they loaded onto a single factor with a correlation of 0.45 or greater. Any items that loaded onto multiple factors, or did not load onto any retained factors, were removed. Hence, a revised, shorter version of the provisional scale was constructed and assumptions re-checked. The variables were still correlated (Bartlett's test of sphericity P<0.001), and the Kaiser-Meyer-Olkin measure of sampling adequacy increased to 0.763. The smallest item communality found was 0.549.

To compensate for faith-based questions that were scored ‘not applicable’ by respondents, the final total SeRvE score for each respondent was calculated as a percentage of the total number of questions answered. This ensures that the total score of each respondent is directly comparable, and is the method of choice for use of the scale both clinically and in research. Pearson's correlations of the SeRvE scale with the MHRM and SWBQ were then calculated, including all participants with a fully completed MHRM (n = 100) and SWBQ (n = 98). The reliability index, Cronbach's alpha, was calculated for the total finalised SeRvE scale and the nine subscales.

Results

Our sample

A wide variety of diagnoses was represented in the sample, however it was only possible to trace exact diagnoses for 42 individuals. Of these, 15 had bipolar disorder, 11 schizophrenia, 12 depression, 2 had an unspecified mental illness and 2 personality disorder. There were 51 in-patients, 24 day patients and 32 out-patients. There was a wide spread of religious or belief affiliation as described by the participants, including Christian, Muslim, Hindu, Sikh, Jewish, Buddhist, Wiccan, Atheist and Humanist. Table 1 shows that a religious belief of some sort is important to the majority of service users, and that most have a significant sense of spirituality. There is also a significant association between the two variables (Pearson's χ² = 48.36, P<0.01), with respondents giving a similar response to both importance of spirituality and belief in a higher power. However, there are a few respondents who place a high importance on spirituality but who have little belief in a higher power.

TABLE 1 Importance of religious faith and spirituality to service users

	Importance of spirituality
	Not at all	A little	Quite a lot	Very much so	Don't know	Total
Belief in a higher power
Not at all	7	2	1	1	1	12
A little	3	6	2	0	0	11
Quite a bit	0	5	4	5	3	17
Very much so	1	12	10	24	5	52
Don't know	3	1	2	1	0	7
Total	14	26	19	31	9	99

Reliability of total finalised SeRvE scale

Cronbach's alpha for the total finalised scale was found to be 0.911, indicating high internal reliability and consistency. For each item, ‘Cronbach's alpha if item deleted’ was computed. No items had an undue influence on the rest of the scale.

Correlations

Despite the difference in subject matter between the three scales, the finalised SeRvE, MHRM and SWBQ correlated highly with each other (Table 2). The correlation between the SeRvE scale and MHRM confirms the validity of the SeRvE as a scale of recovery from mental illness. The high correlation of the SeRvE with the SWBQ indicates that including spiritual well-being in a scale of recovery from mental illness is important. However, the SeRvE scale itself remains unique in including psychosocial and spiritual issues in one scale specifically for mental health service users.

TABLE 2 Correlations between Service-user Recovery Evaluation (SeRvE), Spiritual Well-Being Questionnaire (SWBQ) and Mental Health Recovery Measure (MHRM)

	SeRvE	MHRM	SWBQ
SeRvE	1	0.882 **	0.731 **
MHRM	0.882 **	1	0.739 **
SWBQ	0.731 **	0.739 **	1

^** Pearson's correlation coefficient significant at the 0.001 level, two-tailed.

Factor analysis of the finalised SeRvE scale

Table 3 shows the factor analysis of the finalised scale, forming nine meaningful subscales (see online Table DS1 for a more detailed version with the questions that loaded onto each factor summarised and see the Appendix for the questionnaire). There were no other factors with eigenvalues of more than one. The largest factor, Factor 1, consisted of nine questions regarding the respondents’ existential well-being, and the second largest factor, Factor 2, comprises of seven questions about religious well-being. The reverse-coded questions exploring religious and existential ill-being loaded separately on Factors 7 and 8 respectively. Factor 3 shows the respondent's emotional state and Factor 9 illustrates stigma and shame. Factors 4 and 5 show the social well-being and social ill-being, and Factor 6 measures lack of connectedness, the importance of which had emerged in our scoping day. Cronbach's alpha for each subscale suggests that each one is reliable in its own right.

TABLE 3 Subscales of finalised Service-user Recovery Evaluation scale revealed by factor analysis ^a

Factor	Eigenvalue	% variance	Cumulative variance	Cronbach's α	Mean raw data (1-5)
1. Existential well-being	10.708	26.771	26.771	0.900	3.734
2. Religious well-being	5.007	12.518	39.289	0.882	3.084
3. Emotional state	3.231	8.078	47.366	0.861	2.879
4. Social well-being	2.161	5.403	52.770	0.859	3.944
5. Social ill-being	1.530	3.826	56.596	0.757	3.320
6. Connectedness	1.416	3.540	60.135	0.733	2.969
7. Religious ill-being	1.216	3.039	63.174	0.745	3.446
8. Existential ill-being	1.112	2.780	65.954	0.702	3.431
9. Stigma and shame	1.031	2.578	68.533	0.676	2.972

^a. For a more detailed version that includes the questions summarised by each factor see online Table DS1.

Appendix Service-user Recovery Evaluation (SeRvE) scale

Please answer each question by circling the number, depending on how you have felt this last week
How you have felt about yourself and your life during the last week
Disagree strongly (1), disagree somewhat (2), don't know (3), agree somewhat (4), agree strongly (5)
I feel thankful for my life	1	2	3	4	5
I feel a sense of meaning and purpose in my life	1	2	3	4	5
I am confident I can cope with most things in life	1	2	3	4	5
I feel ashamed of having a mental health problem	1	2	3	4	5
I can find or create something beautiful in life	1	2	3	4	5
I feel my life is pointless	1	2	3	4	5
I have hope for the future	1	2	3	4	5
I can love myself	1	2	3	4	5
I have lost my identity/sense of who I am	1	2	3	4	5
I believe I have the ability to overcome my problems	1	2	3	4	5
I am upset by the stigma of having a mental health problem	1	2	3	4	5
I can do satisfying things despite my problems	1	2	3	4	5
I am positively inspired by the beauty of nature	1	2	3	4	5
I have lost inner motivation	1	2	3	4	5
I am positively inspired by music/the arts or literature	1	2	3	4	5
How you have felt emotionally during the last week
None of the time (1), sometimes (2), don't know (3), quite a bit (4), most of the time (5)
Happy	1	2	3	4	5
Agitated or fearful	1	2	3	4	5
At peace	1	2	3	4	5
Guilty	1	2	3	4	5
Joyful	1	2	3	4	5
Content	1	2	3	4	5
Angry	1	2	3	4	5
How you have related to other people during the last week?
Disagree strongly (1), disagree somewhat (2), don't know (3), agree somewhat (4), agree strongly (5)
I feel other people are against me	1	2	3	4	5
I have some meaningful and close relationships	1	2	3	4	5
I feel loved by some others	1	2	3	4	5
I feel cut off from the rest of the world	1	2	3	4	5
I feel suspicious of most people and find it hard to trust	1	2	3	4	5
My problems make me isolated from other people	1	2	3	4	5
I love some other people	1	2	3	4	5
I feel I need to isolate myself from other people	1	2	3	4	5
I have destructive thoughts towards some other people	1	2	3	4	5
Your personal religious beliefs and practices during the last week
If you believe in a God, higher power, divine spirit, force for good or anything similar, even if only a little, please write your preferred word in here:____
Please substitute your word for X in the following questions, or circle ‘n/a’ (not applicable) if you think the question is not relevant to you
Disagree strongly (1), disagree somewhat (2), don't know (3), agree somewhat (4), agree strongly (5)
I feel I am loved by X	1	2	3	4	5	n/a
I feel that there is a part of X within me	1	2	3	4	5	n/a
My faith/spiritual belief is helpful to me	1	2	3	4	5	n/a
I feel anger towards me from X	1	2	3	4	5	n/a
I find it helpful to pray to X	1	2	3	4	5	n/a
I feel spiritual power/forces are controlling me or others	1	2	3	4	5	n/a
I find it helpful to attend religious services/rituals	1	2	3	4	5	n/a
I feel that X has a purpose for my life	1	2	3	4	5	n/a
My faith/spiritual belief gives me difficult thoughts	1	2	3	4	5	n/a
Thank you for completing this questionnaire

Mean of finalised SeRvE scale

The mean of the complete finalised scale, scored as a percentage of number of questions answered, is 68.7% (s.d. = 13.98). All the subscales have mean raw data scores of around 3.00, which gives scope for sensitivity to change. There was no significant difference in the percentage total means of in-patients and out-patients (Welch's t-test 0.994, P = 0.323). This could be because even our in-patient sample contained few people who were acutely unwell.

Discussion

The SeRvE scale

This scale has been shown to be a reliable and valid measure of holistic recovery from mental illness. It is an inclusive questionnaire for service users to assess their own recovery and the only one to address spiritual and religious issues. The fact that meaningful subscales could be identified added further validity. There are two points of particular interest. First, despite negative scoring for negative questions, the well-being and ill-being factors for existential, social and religious issues do not cluster together, making well-being and ill-being separate concepts for each issue. This means that not only is it crucial to deal with the ill-being but just as important to help people find positive well-being. These may be two quite different tasks. Second, existential well-being was the largest factor in our analysis. Since the questions in this subscale are mainly concerned with meaning and experience of life, this reflects spiritual well-being in its broadest sense. Religious well-being, the specific formal and communal aspect of spiritual well-being, was the second most important factor. The relevance of these two subscales points to the importance of helping people explore their own spirituality/religion in a positive way. More specialised help is also required for the minority of service users who experience religious ill-being. Fulfilling spiritual needs in these ways is the task of spiritual care. The results show the importance of this for all mental health service users.

Use of the SeRvE scale in practice

The SeRvE scale is suitable for mental health service users of all religions and none, and thus could be used in a wide variety of cultural contexts, certainly within the UK, in primary and secondary care.

It can be used as a research tool to evaluate new interventions from the service user viewpoint. Results of the different subscales could help define how an intervention is working and which service users are most likely to be helped by it.

It also has potential in clinical practice:

1. (a) to be used as a new structured approach to taking a complete, person-centred history;

2. (b) to monitor the effectiveness of a particular treatment from the service user viewpoint;

3. (c) comparing scores from the subscales for different service users could assist in identifying interventions specifically targeted to the individual service user.

Limitations of study

The SeRvE scale needs to be tested in its finalised format, in a larger sample of service users including those with acute mental illness. Sensitivity to change and test-retest reliability for the SeRvE need to be established. Comparisons of clinician assessments of recovery from mental illness, for example using the HoNOS, with results using the SeRvE scale would be of further interest. The SeRvE scale may be considered too long to be administered to service users routinely in clinical practice, thus the formation of a shortened form is planned in a further study.

Implications

The SeRvE has wide potential for evaluating interventions in mental health, both in research and in clinical practice. It is a self-report, user-designed scale to monitor recovery from mental illness from the service user's viewpoint. The scale includes measures of spiritual well-being and ill-being, both existential and religious. Factor analysis highlights the importance of these issues for service users. It points to the potential value of increased spiritual care for our service users.