Abraham Mukolo, PhD, MS
Instructor
Vanderbilt Institute for Global Health, Vanderbilt University

Craig Anne Heflinger, PhD
Professor and Associate Dean for Graduate Education
Peabody College of Education and Human Development, Vanderbilt University

Jane Baxter, B
Director
Middle Tennessee Mental Health & Substance Abuse Coalition

 

Abstract
The correlation between the Recovery Assessment Scale (RAS) and the Rosenberg Self-Esteem Scale (RES) was examined among adult participants in consumer-directed services. Hierarchical regression analysis was conducted with RES subscales as dependent variables predicted by RAS variables. Controls were race, age, gender, diagnosis and involvement in peer-directed activities. Regression models for global and positive self-esteem demonstrated moderate fit (explaining 48 to 51% variance, p<.001). As expected negative self-esteem models demonstrated poor fit. ‘Personal confidence and hope’ RAS subscale scores were significantly associated with positive self-esteem. Non-dominance by symptoms was associated with positive and global self-esteem in unadjusted analysis, but attenuated in regression analyses. The RAS demonstrated concurrent validity with the RES in this sample. RAS domains reflective of self-concept correlated with global and positive self-esteem. Reliance on others, non-dominance by symptoms and goal and success orientation might not to be related to self-esteem among participants in consumer-directed mental health services.

Key words: Recovery assessment scale; Self-esteem; Concurrent validity; Consumer-directed services

Introduction:
Recovery is a complex yet increasingly appreciated and insightful concept, particularly in the fields of mental health and substance abuse (Davidson  & White, 2007; Deegan, 1988; Essock & Sederer, 2009; Harvey & Bellack, 2009; Holloway, 2008; Mountain & Shah, 2008).  Literature suggests two potentially conflicting concepts of recovery, one derived from predominantly biomedical models of illness and the second is grounded in the experiences and aspirations of people directly impacted by illness (Frese, Knight, & Saks, 2009; Holloway, 2008; Lieberman, et al., 2008; SAMHSA, 2006). The former conceptualizes recovery as cure and thus prioritizes clinical remission (Essock & Sederer, 2009; Lieberman, et al., 2008). Literature indicates that the absence of symptoms or illness is the most desirable treatment and rehabilitation outcome (Andreasen, et al., 2005; Boden, Sundstrom, Lindstrom, & Lindstrom, 2009; Corrigan, 2006b; Drake, et al., 2006; Harvey & Bellack, 2009; Robinson, Woerner, McMeniman, Mendelowitz, & Bilder, 2004; Xie, McHugo, Helmstetter, & Drake, 2005). In the case of schizophrenia, for example, clinical remission is widely recognized as one of the many goals of treatment (Harvey & Bellack, 2009). The second concept of recovery, generally attributed to Deegan (Anthony, 1993; Barbic, Krupa, & Armstrong, 2009; Deegan, 1988) and popularized by the Substance Abuse and Mental Health Services Administration (SAMHSA) consensus statement (2006), largely focuses on ability to claim/reclaim full and meaningful life in the community, and so emphasizes functional remission (Corrigan, 2006b; Davidson, Schmutte, Dinzeo, & Andres-Hyman, 2008; Frese, et al., 2009; Harvey & Bellack, 2009; Roberts & Wolfson, 2004).
Consumer-driven conceptualizations of recovery emphasize self-agency (Abbott, 2008; Mancini, 2008; Rogers, Chamberlin, Ellison, & Crean, 1997). Self-agency oriented recovery scales (such as the Recovery Assessment Scale: Giffort, Schmook, Woody, Vollendorf, & Gervain, 1995), constitute dimensions indicative of perceived capability to deal with disease assault and its consequences, hope and optimism about the future (independent of disease severity), internal locus of responsibility for one’s own life, and confidence to lead a full and satisfying life (Giffort, et al., 1995; Jacobson & Greenley, 2001; Mancini, 2008; Resnick, Fontana, Lehman, & Rosenheck, 2005; Rogers, et al., 1997; SAMHSA, 2006). Such attributes and capabilities likely engender and sustain the motivation to engage in recovery enhancing activities (Harvey & Bellack, 2009; Mancini, 2008).

SAMHSA has acknowledged the recovery facilitation role of consumer-directed services (SAMHSA, 2006). Participation in consumer-operated services has been shown to correlate with scores on scales assessing recovery and empowerment among a sample of adults consumers (Corrigan, 2006a). Cook and colleagues (2009) observed that participating in peer-led mental illness self-management programs had a positive effect on recovery, self-advocacy and physical health among a sample of adults with serious mental illness.  However, theory and empirical assessments of consumer-defined recovery from mental illness are emergent (Anthony, 1993; Corrigan, 2006a; Davidson, et al., 2007; Lieberman, et al., 2008; Mausbach, Moore, Bowie, Cardenas, & Patterson, 2009; McNaught, Caputi, Oades, & Deane, 2007; Ramon, Healy, & Renouf, 2007; Resnick, et al., 2005; Robinson, et al., 2004; Schennach-Wolff, et al., 2009; Yanos, Roe, Markus, & Lysaker, 2008). The few available assessment scales require testing in diverse samples (McNaught, et al., 2007).

Self-esteem, the feeling of satisfaction with the magnitude of congruence between ones’ self-image and his/her ideal self-image, is central to the emergence and maintenance of a recovery orientation(Jacobson & Greenley, 2001); hence its popularity as an outcome variable in psychiatric rehabilitation studies (Cook, et al., 2009; Corrigan, 2006a; Karatzias, Gumley, Power, & O'Grady, 2007; Lysaker, Davis, & Tsai, 2009; Mancini, 2007; Provencher, Gregg, Mead, & Mueser, 2002; Robson, 1988; Rogers, et al., 1997; Torrey, Mueser, McHugo, & Drake, 2000; Van Dongen, 1996; Yanos, et al., 2008).  Jacobson and Greenley (2001, p. 483) have hypothesized that the sense of self-esteem (and self-respect) emerges from the recovery process of ‘reconnecting with the self’, particularly when one gains insight into (an begins to apply) the distinction between the illness and the self.  Although most of the recovery literature is drawn from experiences of persons with schizophrenia spectrum disorders (Frese, et al., 2009; Harvey & Bellack, 2009; Ho, Chiu, Lo, & Yiu, 2010; Lieberman, et al., 2008; Mausbach, et al., 2009; Novick, et al., 2007), self-esteem studies involving diverse samples have also been reported (Lysaker, et al., 2009; Robson, 1988; Rogers, et al., 1997; Silverstone & Salsali, 2003; Torrey, et al., 2000; Van Dongen, 1996).  For example, the self-esteem oriented study by Torrey and colleagues (2000) involved adults with chronic psychotic illness (46.9% of their sample), severe affective disorders (42.7%) and predominantly severe personality disorders (10.4%). Silverstone and Salsali (2003) found that compared to non-psychiatric patients, psychiatric patients generally had low self-esteem. However, the magnitude of self-esteem has been shown to differ by psychiatric diagnosis. In Silverstone’s study, for example, patients with major depressive, eating and substance use disorders had lower self-esteem than patients with anxiety, conduct and bipolar disorders. In addition, co-morbidity with major depressive disorder was associated with even lower self-esteem than having only one of these disorders (Silverstone & Salsali, 2003). 

In this study, we examined the extent to which the Recovery Assessment Scale (RAS)(Giffort, et al., 1995) fares against the Rosenberg Self-Esteem Scale (RES)(Rosenberg, 1965) among a diverse sample of adults from consumer-operated mental health services. We intend to add to the growing evidence base for the Recovery Assessment Scale since similar work has been reported (Barbic, et al., 2009; Corrigan, Salver, Ralph, Songster, & Keck, 2004; McNaught, et al., 2007). 

Recently, Liberman and colleagues (2008) have advocated for domain-specific conceptualizations of recovery, particularly in relation to schizophrenia disorders. According to Liberman and colleagues (2008), this operationalization of recovery allows investigators and therapists to better circumscribe specific areas in which interventions effectively improve patients’ health and well-being. In this study we used the recovery assessment scale that focuses on improvements in self-agency, hope and perceived mastery of symptoms (Corrigan, et al., 2004; McNaught, et al., 2007). These are some of the domains of recovery that our intervention (the Creative Arts Project) also intends to engender (see methods section).  

We hypothesized that scores on the Recovery Assessment Scale (RAS) would predict scores on the Rosenberg Self-Esteem scale (RES), in particular, subscales of the RAS that are reflective of self-concept (such as optimism and self-agency) would be associated with high global and positive self-esteem and low negative self-esteem. We anticipated a non-direct relationship between perceived capacity to manage symptoms and self-esteem, particularly given the impersonal nature of symptoms, their specificity to disease, and variance in the clinical prognosis of mental illnesses. Given differences in the causation and implications of different disorders on cognitive function (e.g., self-schema in schizophrenia vs. bipolar disorders), we expected psychiatric diagnosis to be independently associated with self-esteem and to moderate the relationship between RAS and RES scores.
 
Methods
Participant Description
Data were collected in 2008 from 110 consenting adults (61% females, Mean age =44.05, SD=12.22) recruited at five peer support centers in one southern state. This was part of a larger study to assess the effectiveness of one art training project implemented by the Middle Tennessee Mental Health and Substance Abuse Coalition (MTMHSAC). The Creative Arts Project started in 2005. It facilitates adults with mental health and substance use disorders to develop artistic talent and showcase their works to the general public via public exhibits at local arts galleries, civic and educational establishments. Participation is voluntary and open to all adults at institutions affiliated with MTMHSAC. The project aims to increase the self-esteem, respect and self-worth of artists (Cook, et al., 2009), overcome stigma about mental illness and substance use disorders through public recognition of artistic talent, and to enhance community functioning by connecting participants to local networks of professional artists. All these are aspects of self-concept and recovery tapped into by the RES and the RAS.

Data Collection
Data for this study were collected prior to the commencement of the 2008 arts classes and from all peer support center members who were willing and consented to the study.
Informed consent. The study team administered oral and written informed consent gathering procedures at each peer support center following detailed protocol approved by two institutional review boards. 

Survey administration. Participants completed a self-completion questionnaire. Assistance with questionnaire completion was provided as requested by participants.
Measures. The questionnaire constituted the Rosenberg Self-Esteem Scale (RES) and the Recovery Assessment Scale (RAS). Both scales have been validated among comparable samples (Corrigan, et al., 2004; McNaught, et al., 2007). 

The Rosenberg Self-Esteem Scale (RES)(Rosenberg, 1965) scale is a recognized self-esteem measure, and has been used to assess the efficacy of psychiatric rehabilitation programs (Rogers, et al., 1997; Torrey, et al., 2000; Van Dongen, 1996; Yanos, et al., 2008). Torrey and colleagues (2000) reported coefficient alphas ranging from 0.88 to 0.90. The scale consists of 10 items: 5 items measure positive self-concept and 5 items assess negative self-concept. The actual number of sub-scales derivable from the RES scale is debatable (Corwyn, 2000; Dunbar, Ford, Hunt, & Der, 2000; Tomas & Oliver, 1999).  We used the three conventional categories: (a) Global self-esteem, i.e., the sum of scores for the 10 items (the 5 negatively worded items were reverse coded first before the summation so that higher scores reflect greater self-esteem), b) Positive self-esteem, i.e., the sum of the scores for the 5 positively worded items and, c) Negative self-esteem, i.e., the sum of the 5 negatively worded items (higher scores reflect greater negative self-esteem). See Table 1 for results.

The Recovery Assessment Scale (RAS)(Corrigan, et al., 2004) is a 41-item measure that originally emerged from research conducted by Giffort and colleagues (Giffort, et al., 1995) to derive a consumer oriented measure of recovery. We used the RAS version modified by Corrigan and colleagues (2004). The scale has been used to assess consumers’ personal confidence and hope for the future, their goal and success orientation, perceived mastery of symptoms of mental illness, willingness to ask for help, and reliance on other people for help (Barbic, et al., 2009; Cook, et al., 2009; Corrigan, 2006a; Corrigan, et al., 2004). Two studies have re-examined the robustness of the five factor structure of the RAS (Corrigan, et al., 2004; McNaught, et al., 2007). Corrigan and colleagues (2004) reported internal consistency estimates (Chronbach’s α) ranging between .74 and .87 among a US sample. McNaught and colleagues (2007) tested the scale among an Australian sample and reported Chronbach’s α ranging between .73 and .91.   They also found the RAS to have good convergent validity with the stages of recovery measure (SRM)(Andresen, Oades, & Caputi, 2003) and the Mental Health Recovery Measure (MHRM)(Young & Ensing, 1999). Our Chronbach’s alphas ranged from .75 to .92 for the five factors described in table 5. 

Participants were also asked to self-report their age, race (White, Black, Other), gender, mental health diagnosis, and the peer support center activities in which they take part based on the known list of activities the participating centers generally offer to members, i.e., vocational training & rehabilitation programs, center maintenance activities such as general cleaning and food preparation, health and wellness enhancing programs, life-skills training to facilitate community functioning, and leisure and play oriented  activities.   Participants were free to include other more generalized activities that did not fit into these categories: all who responded to this portion of the questionnaire mentioned leisure and play type activities or one-off social events that are not part of the formal program of activities. There were no clinical assessments of psychiatric or psychological status or independent verification of self-reports. 

Schizophrenia, major depression and bipolar disorder were the most common primary diagnoses reported (See Table 1). A number of participants reported co-morbidity. As would be expected, certain disorders coexisted with some clarity but some did not. We consulted with a generous colleague who is a practicing psychiatrist regarding the classification of co-morbidities.  Any bipolar disorder type was then grouped as bipolar, same with schizophrenia. With either of those diagnoses a coexisting depression or anxiety was placed under the primary diagnosis (bipolar or schizophrenia). If they reported all three primary diagnoses, then they were placed under schizophrenia.  Those reporting depression and bipolar were grouped under bipolar.  Co-morbid schizophrenia and bipolar was treated as a separate diagnosis (9.1% of sample, Table 1).  About 42% reported what amounts to symptoms rather diagnosis such as anger, other mood problems, hallucinations or that they are mentally unwell, and 5% either did not disclose any diagnosis/ symptoms or stated they did not know their diagnosis.

Data Analysis
Data were analyzed to assess the extent to which RAS scores would predict RES scores controlling for variance in participant demographics, self-reported mental health diagnosis and involvement in peer support center activities.   In regression analyses, each of the self-esteem subscales was entered as a dependent variable. Predictors included the five RAS sub-scales, participant race, age, gender, diagnosis, and participation in peer center activities associated with self-esteem, i.e. leisure and play and generalized activities. Apart from age which was used as a continuous variable, all these data were converted into dummy variables. First, descriptive analyses compared the dependent variables and predictors (Table 1). With the exception of age (in years) demographic, diagnosis and activity involvement variables were categorical.  Next, hierarchical multiple regression analysis was conducted. RAS variables were entered first in each of the regression models (model 1), followed by demographic and activity involvement variables (model 2).  The diagnosis variable was entered last (Tables 2, 3 & 4, model 3). Data analysis was conducted using SPSS Statistics 17.0 software. 

Results
Table 1 provides the unadjusted means, standard deviation, and correlation coefficients of global, positive and negative self-esteem. The self-esteem scale was scored so that scores would range from 1 (low) to 5 (high). Table 1 indicates that the mean scores of self-esteem were above the mid-score of 3.0.  The scores generally ranged between 2.7 and 4.15.  These self-esteem scores did not differ statistically by demographic, diagnosis or services use variables (Table 1). Self-esteem was related only to leisure and play (statistically significant, p<.05) and general activities (marginally significant, p<.10).  These are unstructured non-programmatic activities and the association observed was with positive self-esteem and not global or negative self-esteem.  Post-hoc analysis (results not shown) of the unadjusted relationships between self-esteem and diagnosis indicated that, in the case of global and positive self-esteem, the mean differences between bipolar and undisclosed diagnosis were statistically significant (p<.05). The difference in mean global self-esteem between participants self-reporting schizophrenia and bipolar reporters were marginally significant (Mean difference = 0.436, p<.10).  In the case of negative self-esteem, the mean difference between participants disclosing depression and those with undisclosed diagnosis was statistically significant (Mean difference = 0.344, p<.05). 

Table 1: Participant characteristics and means, standard deviation and correlations of self-esteem by predictor variables

		Global Self-Esteem		Positive Self-Esteem		Negative Self-Esteem
	% (N=100)*	M (SD)	p	M(SD)	p	M(SD)	p
Total Sub-Scale Scores		3.348 (.753)		3.825 (.881)		3.113 (.997)
Gender Female Male	61 39	3.359 (.756) 3.331 (.757)	.861	3.886 (.893) 3.732 (.867)	.403	3.159 (1.021) 3.039 (.967)	.574
Race White Caucasian	72	3.344 (.722)	.654	3.861 (.835)	.810	3.167 (.947)	.240
Black (African American)	16	3.250 (.810)		3.713 (1.090)		3.213 (1.153)
Other	12	3.530 (.906)		3.764 (.907)		2.636 (1.019)
Residence Rural Urban	55.5 45.5	3.301 (.740) 3.400 (.774)	.554	3.913 (.878) 3.716 (.884)	.066	3.286 (1.044) 2.907 (.907)	.279
Past 3 months Mental Health Treatment Case Management No Yes	40.4 59.6	3.337 (.637) 3.356 (.830)	.907	3.897 (.796) 3.775 (.939)	.508	3.221 (.873) 3.039 (1.074)	.389
Medication Management  No Yes	54 46	3.326 (.667) 3.371 (.840)	.773	3.736 (.911) 3.922 (.847)	.305	3.053 (.884) 3.178 (1.113)	.547
Individualized Counseling No Yes	77 33	3.340 (.721) 3.363 (.821)	.892	3.840 (.936) 3.794 (.773)	.807	3.136 (1.013) 3.069 (.979)	.760
Group counseling No Yes	76 24	3.396 (.719) 3.213 (.843)	.308	3.892 (.887) 3.625 (.851)	.201	3.082 (1.001) 3.200 (1.003)	.622
Inpatient hospital No Yes	90 10	3.352 (.709) 3.311 (1.136)	.879	3.855 (.844) 3.533 (1.208)	.299	3.143 (.951) 2.911 (1.411)	526
Mental Health Diagnosis			.238		.889		.128
Schizophrenia	15.5	3.486 (.910)		3.743 (1.080)		2.771 (1.155)
Bi-polar disorder	14.5	3.050 (.671)		3.667 (.922)		3.440 (.882)
Co-morbid Bipolar & Schizophrenia	9.1	3.580 (.839)		4.040 (.793)		2.880 (.944)
Depression	13.6	3.193 (.753)		3.840 (.836)		3.457 (.920)
Other (unspecified)	42.7	3.339 (.642)		3.821 (.847)		3.135 (.958)
Undisclosed	4.5	3.950 (.998)		4.150 (1.063)		3.113 (.997)
Peer Center Activity Involvement
Vocational rehabilitation program:    No Yes	80.9 19.1	3.350 (.771) 3.353 (.719)	.989	3.857 (.886) 3.788 (1.006)	.780	3.125 (1.002) 3.082 (.993)	.875
General cleaning/work No Yes	71.9 28.1	3.336 (.774) 3.384 (.727)	.793	3.887 (.883) 3.736 (.967)	.484	3.180 (.994) 2.968 (.999)	.376
Food service program No Yes	85.4 14.6	3.313 (.734) 3.554 (.868)	.294	3.814 (.925) 4.015 (.794)	.461	3.155 (.989) 2.908 (1.041)	.413
Health and wellness program No Yes	78.7 21.3	3.361 (.730) 3.316 (.860)	.821	3.850 (.920) 3.821 (.874)	.903	3.095 (1.002) 3.190 (.992)	.719
Life skills training program No Yes	58.4	3.371 (.807) 3.337 (.727)	.842	3.689 (1.041) 3.953 (.787)	.182	2.936 (.997) 3.240 (.984)	.170
Leisure and play No Yes	41.6 58.4	3.222 (.844) 3.449 (.674)	.177	3.578 (1.086) 4.040 (.691)	.018	3.128 (1.040) 3.108 (.970)	.930
Other generalized activities No Yes	43.3 56.7	3.229 (.779) 3.440 (.734)	.211	3.642 (.999) 4.000 (.800)	.067	3.161 (.995) 3.098 (.998)	.774
Creative Arts Project Exposure Never participated Past participant	82.2 17.8	3.427 (.687) 3.279 (.805)	.370	3.932 (.821) 3.677 (.913)	.176	3.055 (.972) 3.111 (1.034)	.801
Over 12 months at Peer Center No Yes	42.5 57.5	3.423 (.880) 3.297 (.724)	.517	3.773 (.843) 3.737 (.955)	.867	2.927 (1.033) 3.102 (.970)	.465
	M(SD)	r	p	r	p	r	p
Age (years)	44.05 (12.22)	-.012	.914	.058	.579	.065	.538
Recovery Subscales Reliance on others Personal confidence & hope Willingness to ask for help Non-dominance by symptoms Goal & success orientation Total RAS score	4.01 (.853) 3.35 (.803) 4.07 (.852) 3.61 (.976) 4.11 (.800) 85.36 (16.36)	.328 .432 .653 .390 .511 .624	.002 .000 .000 .000 .000 .000	.595 .848 .634 .763 .646 .877	.000 .000 .000 .000 .000 .000	.038 -.284 -.028 -.147 -.077 -.145	.727 .007 .789 .170 .485 .207
*  Data from 10 of the original 110 participants was incomplete. M=mean, SD=standard deviation.  r= Pearson's product-moment coefficient, p= P-value.  Sample N was reduced by participants who did not provide reliable responses (e.g., not completing more than 90% of the questions) and/or requested that information be not used even if they did not withdraw their completed questionnaires. The latter might be participants’ own test of researchers’ commitment to informed consent protocols.

 

As expected, both global and positive self-esteem were positively correlated (p<.05) with the five sub-scales of RAS and the total RAS score (Table 1).  The coefficients for positive self-esteem were much higher than those for global self-esteem, with marked differences in the case of personal confidence and hope and non-dominance by symptoms.  Apart from the positive relationship between negative self-esteem and reliance on others, RAS subscales (and total RAS score) were negatively correlated with negative self-esteem as hypothesized.  However, only the inverse association between negative self-esteem and personal confidence and hope was statistically significant (p<.01). 

In regression analyses, the global and positive self-esteem models demonstrated moderate fit, explaining 48 to 51% of self-esteem variance (Tables 2 and 3), and were statistically significant (p<.001). Based on the F-test, the models for negative self-esteem demonstrated poor fit. That is, the RAS and demographic variables considered did not adequately account for observed negative self-esteem.

Unadjusted relationships between RES and RAS variables were not substantially changed by adjusting for demographics, diagnosis, and activity involvement (Tables 2, 3 & 4). In the case of positive self-esteem (Table 3), personal confidence and hope and willingness to ask for help were directly associated with self-esteem (p<.001). The correlations (unadjusted β) were robust at all levels of adjustment (Table 3). Personal confidence and hope was the only RAS variable marginally associated with negative self-esteem (p<.10), see Table 4. However, this negative association was attenuated by demographic variables.   In Table 2, personal confidence and hope was also the only RAS variable associated with global self-esteem (p<.05) and remained robust at all levels of adjusting for participant characteristics (Table 2). 

Table 2 Regressing global self-esteem against RAS variables

	Model 1 (R2= .475, df= 5 , F=9.942, P>.001)					Model 2 (R2= .524 , df=12, F=4.399, p<.001)					Model 3 (R2= .524, df=13, F=3.983, P<.001)
	B	SE	95%CI for B	p		B	SE	95%CI for B	p		B	SE	95%CI for B	p
Recovery subscales
Reliance on others	-.108	.124	-.358 - .141	.387		-.118	.133	-.386 - .150	.381		-.111	.138	-.390 - .167	.425
Personal confidence & hope	.575	.177	.221 - .929	.002		.521	.184	.150 - .891	.007		.5190	.186	.145 - .894	.008
Willingness to ask for help	.150	.113	-.077 - .376	.191		.139	.122	-.107 - .385	.261		.140	.124	-.109 - .386	.264
Non-dominance by symptoms	.055	.139	-.124 - .334	.693		.123	.149	-.176 - .422	.411		.118	.152	-.187 - .424	.439
Goal & success orientation	-.074	.133	-.341 - .193	.581		-.067	.140	-.347 - .214	.635		-.066	.141	-.350 - .217	.641
Demographics
Gender (ref=male)						-.204	.177	-.560 - .153	.256		-.195	.184	-.565 - .176	296
Race (Ref=white)						-.125	.121	-.368 - .118	.305		-.128	.123	-.374 - .119	.303
Residence (Ref=Rural)						.214	.169	-.125 - .554	.210		.231	.188	-.147 - .608	.225
Age (yrs)						.000	.007	-.015 -.014	.922		-.001	.008	-.016 - .014	.880
Art experience (ref=nil)						.061	.118	-.175 - .298	.604		.057	.121	-.185 - .300	.638
Activity Involvement
Leisure & Play						.063	.208	-.355 - .482	.763		.066	.211	-.358 - .490	.755
General activities						.096	.208	-.320 - .511	.646		.087	.213	-.343 - .515	.686
Mental Health Diagnosis											.013	.061	-.109 - .135	.835

Table 3 Regression positive self-esteem against RAS variables

	Model 1 (R2= .822 , df=5, F=53.600, P<.001)					Model 2 (R2 = .849, df=12, F= 23.876, p<.001)					Model 3 (R2= .851, df=13, F=22.026, P<.001)
	B	SE	95%CI for B	p		B	SE	95%CI for B	p		B	SE	95%CI for B	p
Recovery subscales
Reliance on others	-.067	.089	-.244 - .110	.450		-.081	.091	-.263 - .100	.373		-.062	.093	-.249 - .125	.507
Personal confidence & hope	.620	.125	.370 - .870	.000		.598	.125	.346 - .849	.000		.594	.126	.342 - .847	.000
Willingness to ask for help	.345	.081	.183 - .507	.000		.371	.084	.202 - .539	.000		.372	.084	.203 - . 541	.000
non dominance by symptoms	.095	.097	-.100 - .290	.332		.128	.101	-.074 - .330	.211		.113	.102	-.092 - .318	.273
Goal & success orientation	.087	.095	-.103 - .277	.362		.063	.096	-.129 - .256	.512		.066	.096	-.127 - .258	.498
Demographics
Gender (ref=male)						-.163	.121	-.407 - .081	.186		-.135	.125	-.387 - .117	.286
Race (Ref=white)						-.149	.082	-.314 - .017	.077		-.155	.083	-.322 - .011	.066
Residence (Ref=Rural)						-.118	.114	-.346 - .110	.303		-.068	.127	-.322 - .187	.595
Age (yrs)						-.000	.005	-.009 - .010	.943		-.000	.005	-.010 - .009	.890
Art experience (ref=nil)						.054	.080	-.106 - .215	.500		.042	.082	-.122 - .205	.612
Activity Involvement
Leisure & Play						.036	.138	-.241 - .312	.797		.043	.138	-.234 - .320	.756
General activities						.068	.138	-.211 - .343	.633		.042	.141	-.240 - .350	.764
Mental Health Diagnosis											.037	.041	-.045 - .120	.369

Table 4 Regression negative self-esteem against RAS variables

	Model 1 (R2=.080, df= 5, F=.954, P>.10)					Model 2 (R2 =.164, df=12, F=.784, p<.10 )					Model 3 (R2= .164, df=13, F=.713, P>.10)
	B	SE	95%CI for B	p		B	SE	95%CI for B	p		B	SE	95%CI for B	p
Recovery subscales
Reliance on others	.152	.212	-.273 - .577	.477		.166	.228	-.292 - .624	.469		.169	.236	-.306 - .644	.477
Personal confidence & hope	-.531	.301	-1.135 - .072	.083		-.446	.314	-1.077 - .186	.163		-.446	.318	-1.085 - .193	.167
Willingness to ask for help	.041	.193	-.345 - .427	.832		.076	.209	-.3446 - .496	. 718		.076	.211	-.349 - .501	.720
non dominance by symptoms	-.016	.237	-.491 - .460	.947		-.118	.254	-.628 - . 392	.644		-.120	.259	-.642 - .401	.645
Goal & success orientation	.250	.227	-.205 - .704	.276		.211	.238	-.268 - .690	.381		.211	.241	-.273 - .695	.385
Demographics
Gender (ref=male)						.251	.302	-.357 - .859	.410		.255	.314	-.377 - .887	.421
Race (Ref=white)						.093	.206	-.321 - .508	.652		.092	.209	-.329 - .513	.661
Residence (Ref=Rural)						-.509	.288	-1.088 - .070	.083		-.502	.320	-1.146 - .142	.124
Age (yrs)						.003	.012	-.022 - .029	.786		.003	.013	-.023 - .029	.803
Art experience (ref=nil)						-.066	. 201	-.469 - .338	.744		-.068	.206	-.482 - .346	.743
Activity Involvement
Leisure & Play						-.073	.355	-.787 - .641	.839		-.071	.360	-.795 - .652	.844
General activities						-.111	.352	-.819 - .598	.755		-.115	.363	-.846 - .616	.754
Mental Health Diagnosis											.006	.103	-.202 - .214	.955

Table 5 RAS subscales and their reliability estimates

RAS Subscale (and items)	Cronbach’s  Alpha
Personal confidence and hope	0.92
Fear doesn't stop me from living the way I want to
I can handle what happens in my life
I like myself
If people really knew me, they would like me
Something good will eventually happen
I am hopeful about my future
I continue to have new interests
I can handle stress
Willingness to ask for help	0.86
I know when to ask for help
I am willing to ask for help
I ask for help, when I need it
Goal and success orientation	0.83
I have a desire to succeed
I have my own plan for how to stay or become well
I have goals in life that I want to reach
I believe I can meet my current personal goals
I have a purpose in life
Non-dominance by symptoms	0.75
My symptoms interfere less and less with my life
Coping with my mental illness is no longer the main focus of my life
My symptoms seem to be a problem for shorter periods of time each time they occur
Reliance on others	0.77
Even when I don't care about myself, other people do
I have people I can count on
Even when I don't believe myself, other people do
It is important to have a variety of friends

 

Discussion
In this study we used the recovery assessment scale that focuses on improvements in self-agency, hope and perceived mastery of symptoms. We examined the association between these domains of recovery and self-esteem among a sample with several different psychiatric diagnoses.  The RAS was associated with global and positive self-esteem as hypothesized. As anticipated the robust association was the one observed between positive self-esteem and RAS subscales that are reflective of positive self-concept. Not all domains captured by the RAS were found to be related to self-esteem. This does not mean that participants in our study did not achieve recovery in those contexts, but rather that the domains of the RAS that are reflective of self-concept were correlated with an established measure of recovery, i.e., the RES. 

Scores on the personal confidence and hope subscales of the RAS were significantly associated with positive self-esteem in our study. Participants had above average scores on perceived mastery of symptoms (Table 1, non-dominance by symptoms). Given that the absence of symptoms or illness is generally considered to be the most desirable treatment and rehabilitation outcome (Andreasen, et al., 2005; Boden, et al., 2009; Corrigan, 2006b; Drake, et al., 2006; Harvey & Bellack, 2009; Robinson, et al., 2004; Xie, et al., 2005), it is interesting to note that non-dominance by symptoms was associated with positive and global self-esteem in unadjusted analysis. However, in the regression analyses that controlled for demographic characteristics and participants’ scores on personal confidence and hope, this relationship was attenuated. These observations suggest a complex relationship between non-dominance by symptoms and self-esteem. Resnick and colleagues examined correlates of recovery and found a significant association between recovery orientation and lower severity of depressive symptoms. 

In our study, negative self-esteem was not related to scores on the RAS in either descriptive or multivariate analysis. The negative correlation between negative self-esteem and personal confidence and hope was not robust.  The model fit estimates for negative self-esteem suggest that RAS variables did not explain negative self-esteem in this sample. Adjusting for differences in demographics and mental health diagnosis did not improve the explanatory power of RAS variables, including those RAS variables that are theoretically reflective of self-concept.  

However, we observed some relationship between mental health diagnosis and self-esteem only in unadjusted analysis.  Most notably between participants self-reporting bipolar disorders vs. those with undisclosed diagnosis (including those who stated they had no diagnosis). The differences in mean global self-esteem between participants self-reporting schizophrenia and bipolar reporters were marginally significant, most likely because of the co-morbidity of schizophrenia and bipolar.  It is possible that self-reports might not have accurately distinguished co-morbid bipolar and schizophrenia.  In a study reported by Karatzias and colleagues (2007), adults with schizophrenia who had a co-morbid anxiety or affective disorder were observed to have low self-esteem (based on the Rosenberg Self-Esteem Scale) as well as poor functioning and higher negative appraisal of their mental health status. Overall, and contrary to our hypothesis, self-reported mental health diagnosis did not moderate the relationship between RAS and RES variables among our sample.

While other studies have indicated that self-esteem might be influenced by demographic variables of age, gender, educational achievement, income, and employment among mental health samples (Salsali & Silverstone, 2003), we considered age, gender and race variables and found no relationship with self-esteem. 

Consumer-operated services have been hypothesized to facilitate recovery. In our study, we observed some relationship between self-esteem and the non-programmatic consumer-directed activities. Programmatic activities such as health and wellness programs, job preparation support, and life skills training were not related to self-esteem in our sample.  However, the positive association between positive self-esteem and either leisure and play or involvement in generalized social activities was attenuated by recovery variables. It is likely that personal confidence and hope and willingness to ask for help moderated these relationships with positive self-esteem. 

Study Limitations
Data analyzed were obtained just before art classes commenced but after peer center members had had the chance to register for classes. It is likely that the anticipated commencement of art classes might have primed some participants to respond to questions about goal and success orientation more favorably than those for whom current art-class attendance was not a novelty. The mental health statuses considered did not cater for severity of illness/symptoms and how it might moderate the relationship between RAS and RES variables. In other samples, the RAS total score has been shown to correlate negatively with severity of psychiatric symptoms (McNaught, et al., 2007). 

The RES is structured to indicate an inverse relationship between negative self-esteem and positive self-esteem. However, we observed some counterintuitive relationships, particularly among female participants. This could have been influenced by the fact that positive self-esteem items preceded negative self-esteem items on our questionnaire. Therefore, some participants might have paid less attention to items differences and so maintained the high scores they were allocating to positive self-esteem items. We had anticipated this error at the time of the study and had informed participants about the differences in the items.  It could also be indicative of the fact that the two subscales tap into two different aspects of self-concept (Corwyn, 2000) and that most participants were aware of these and their responses reflect that degree of insight. 

Although the RAS has demonstrated convergent validity with other recovery scales, it has been noted that the scale does not consider all the aspects of recovery that have been reported by mental health consumers (McNaught, et al., 2007). It also does not adequately capture objective aspects of clinical remission, captured by established clinician-completed scales/protocols.  However, McNaught and colleagues (2007) have advocated for the potential of the RAS to inform consumer-led interventions.

Conclusions
We have attempted to examine the extent to which scores on the Recovery Assessment Scale (RAS) might be related to how adults with mental illness score on the Rosenberg Self-Esteem scale (RES). Our findings collaborated findings that the RAS demonstrates concurrent validity with self-esteem among samples of mental health consumers (Corrigan, 2006a). Based on our findings, reliance on others, non-dominance by symptoms and goal and success orientation might not to be related to the construct of self-esteem among participants in consumer-directed mental health services. Probably the RAS is biased towards positive self-concept. Participants in this study evidently maintained positive self-esteem irrespective of their symptom management capabilities or goal and success orientation. Jacobson and Greenley (2001:483) have theorized that recovery from mental illness entails “recovering the self by reconceptualizing illness as only part of the self, not as a definition of the whole…As consumers reconnect with their selves, they begin to experience a sense of self-esteem and self-respect.”  Perhaps the RAS is best considered within a stages-of-recovery framework, given its convergent validity with the stages of recovery measure (Andresen, Oades, & Caputi, 2003). 

References

Abbott, P. (2008). Another step towards understanding recovery?: INVITED COMMENTARY ON... SELF-DETERMINATION THEORY. Advances in Psychiatric Treatment, 14(5), 366-368.

Andreasen, N. C., Carpenter, W. T., Jr., Kane, J. M., Lasser, R. A., Marder, S. R., & Weinberger, D. R. (2005). Remission in schizophrenia: Proposed criteria and rationale for consensus. American Journal of Psychiatry, 162(3), 441-449.

Andresen, R., Oades, L., & Caputi, P. (2003). The experience of recovery from schizophrenia: Towards an empirically validated stage model. The Australian and New Zealand Journal of Psychiatry, 37(5), 586-594.

Anthony, W. A. (1993). Recovery from mental iIllness: The guiding vision of the mental health service system in the 1990s. [Reprint]. Psychosocial Rehabilitation Journal, 16(4), 11-23.

Barbic, S., Krupa, T., & Armstrong, I. (2009). A randomized controlled trial of the effectiveness of a modified recovery workbook program: Preliminary findings. Psychiatric Services, 60(4), 491-497.

Boden, R., Sundstrom, J., Lindstrom, E., & Lindstrom, L. (2009). Association between symptomatic remission and functional outcome in first-episode schizophrenia. Schizophrenia Research, 107(2-3), 232-237.

Cook, J. A., Copeland, M. E., Hamilton, M. M., Jonikas, J. A., Razzano, L. A., Floyd, C. B., et al. (2009). Initial outcomes of a mental illness self-management program based on wellness recovery action planning. Psychiatric Services, 60(2), 246-249.

Corrigan, P. W. (2006a). Impact of consumer-operated services on empowerment and recovery of people with psychiatric disabilities. Psychiatric Services, 57(10), 1493-1496.

Corrigan, P. W. (2006b). Recovery from schizophrenia and the role of evidence-based psychosocial interventions. Expert Review of Neurotherapeutics, 6(7), 993-1004.

Corrigan, P. W., Salver, M., Ralph, R. O., Songster, Y., & Keck, L. (2004). Examining the factor structure of the Recovery Assessment Scale. Schizophrenia Bulletin, 30(4), 1035-1041.

Corwyn, R. F. (2000). The factor structure of global self-esteem among adolescents and adults. Journal of Research in Personality, 34, 357-379.

Davidson, L., Schmutte, T., Dinzeo, T., & Andres-Hyman, R. (2008). Remission and recovery in schizophrenia: Practitioner and patient perspectives. Schizophrenia Bulletin, 34(1), 5-8.

Davidson, L., Tondora, J., O'Connell, M. J., Kirk, T., Jr., Rockholz, P., & Evans, A. C. (2007). Creating a recovery-oriented system of behavioral health care: Moving from concept to reality. Psychiatric Rehabilitation Journal, 31(1), 23-31.

Davidson, L., & White, W. (2007). The concept of recovery as an organizing principle for integrating mental health and addiction services. The Journal of Behavioral Health Services & Researchs, 34(2), 109-120.

Deegan, P. E. (1988). Recovery: The lived experience of rehabilitation. Psychosocial Rehabilitation Journal, 11, 11-19.

Drake, R. E., McHugo, G. J., Xie, H., Fox, M., Packard, J., & Helmstetter, B. (2006). Ten-year recovery outcomes for clients with co-occurring schizophrenia and substance use disorders. Schizophrenia Bulletin, 32(3), 464-473.

Dunbar, M., Ford, G., Hunt, K., & Der, G. (2000). Question wording effects in the assessment of global self-esteem. European Journal of Psychological Assessment, 16(1), 13-19.

Essock, S., & Sederer, L. (2009). Understanding and measuring recovery. Schizophrenia Bulletin, 35(2), 279-281.

Frese, F. J., 3rd, Knight, E. L., & Saks, E. (2009). Recovery from schizophrenia: With views of psychiatrists, psychologists, and others diagnosed with this disorder. Schizophrenia Bulletin, 35(2), 370-380.

Giffort, D., Schmook, A., Woody, C., Vollendorf, C., & Gervain, M. (1995). Construction of a scale to measure consumer recovery. Springfield, IL.: Illinois Office of Mental Health.

Harvey, P. D., & Bellack, A. S. (2009). Toward a terminology for functional recovery in schizophrenia: Is functional remission a viable concept? Schizophrenia Bulletin, 35(2), 300-306.

Ho, W. W., Chiu, M. Y., Lo, W. T., & Yiu, M. G. (2010). Recovery components as determinants of the health-related quality of life among patients with schizophrenia: Structural equation modelling analysis. The Australian and New Zealand Journal of Psychiatry, 44(1), 71-84.

Holloway, F. (2008). Is there a science of recovery and does it matter?: INVITED COMMENTARY ON... RECOVERY AND THE MEDICAL MODEL. Advances in Psychiatric Treatment, 14(4), 245-247.

Jacobson, N., & Greenley, D. (2001). What Is Recovery? A Conceptual Model and Explication. Psychiatric Services, 52(4), 482-485.

Karatzias, T., Gumley, A., Power, K., & O'Grady, M. (2007). Illness appraisals and self-esteem as correlates of anxiety and affective comorbid disorders in schizophrenia. Comprehensive Psychiatry, 48(4), 371-375.

Lieberman, J. A., Drake, R. E., Sederer, L. I., Belger, A., Keefe, R., Perkins, D., et al. (2008). Science and recovery in schizophrenia. Psychiatric Services, 59(5), 487-496.

Lysaker, P. H., Davis, L. W., & Tsai, J. (2009). Suspiciousness and low self-esteem as predictors of misattributions of anger in schizophrenia spectrum disorders. Psychiatry Research, 166(2-3), 125-131.

Mancini, A. D. (2008). Self-determination theory: a framework for the recovery paradigm. Advances in Psychiatric Treatment, 14(5), 358-365.

Mancini, M. A. (2007). The role of self-efficacy in recovery from serious psychiatric disabilities: A qualitative study with fifteen psychiatric survivors. Qualitative Social Work, 6, 49-74.

Mausbach, B. T., Moore, R., Bowie, C., Cardenas, V., & Patterson, T. L. (2009). A review of instruments for measuring functional recovery in those diagnosed with psychosis. Schizophrenia Bulletin, 35(2), 307-318.

McNaught, M., Caputi, P., Oades, L. G., & Deane, F. P. (2007). Testing the validity of the Recovery Assessment Scale using an Australian sample. The Australian and New Zealand Journal of Psychiatry, 41(5), 450-457.

Mountain, D., & Shah, P. J. (2008). Recovery and the medical model. Advances in Psychiatric Treatment, 14(4), 241-244.

Novick, D., Haro, J. M., Suarez, D., Lambert, M., Lepine, J. P., & Naber, D. (2007). Symptomatic remission in previously untreated patients with schizophrenia: 2-year results from the SOHO study. Psychopharmacology (Berl), 191(4), 1015-1022.

Provencher, H. L., Gregg, R., Mead, S., & Mueser, K. T. (2002). The role of work in the recovery of persons with psychiatric disabilities. Journal of Psychiatric Rehabilitation 26, 132-144.

Ramon, S., Healy, B., & Renouf, N. (2007). Recovery from mental illness as an emergent concept and practice in Australia and the UK. International Journal of Social Psychiatry, 53(2), 108-122.

Resnick, S. G., Fontana, A., Lehman, A. F., & Rosenheck, R. A. (2005). An emprirical conceptualization of the recovery orientation. Schizophrenia Research, 75, 119-128.

Roberts, G., & Wolfson, P. (2004). The rediscovery of recovery: open to all. Advances in Psychiatric Treatment, 10(1), 37-48.

Robinson, D. G., Woerner, M. G., McMeniman, M., Mendelowitz, A., & Bilder, R. M. (2004). Symptomatic and functional recovery from a first episode of schizophrenia or schizoaffective disorder. American Journal of Psychiatry, 161(3), 473-479.

Robson, P. J. (1988). Self-esteem: A psychiatric view. British Journal of Psychiatry, 153, 6-15.

Rogers, E. S., Chamberlin, J., Ellison, M. L., & Crean, T. (1997). A consumer-constructed scale to measure empowerment among users of mental health services. Psychiatric Services, 48(8), 1042-1047.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

Salsali, M., & Silverstone, P. H. (2003). Low self-esteem and psychiatric patients: Part II - The relationship between self-esteem and demographic factors and psychosocial stressors in psychiatric patients. Annals of General Hospital Psychiatry, 2(1), 3.

SAMHSA. (2006). National census statement on mental health recovery. Rockville, MD: National Mental Health Information Center, US Government.

Schennach-Wolff, R., Jager, M., Seemuller, F., Obermeier, M., Messer, T., Laux, G., et al. (2009). Defining and predicting functional outcome in schizophrenia and schizophrenia spectrum disorders. Schizophrenia Research, 113(2-3), 210-217.

Silverstone, P. H., & Salsali, M. (2003). Low self-esteem and psychiatric patients: Part I - The relationship between low self-esteem and psychiatric diagnosis. Annals of General Hospital Psychiatry, 2(1), 2.

Tomas, J. M., & Oliver, A. (1999). Rosenberg’s self-esteem scale: Two factors or method effects. Structural Equation Modeling, 6, 84-98.

Torrey, W. C., Mueser, K. T., McHugo, G. H., & Drake, R. E. (2000). Self-esteem as an outcome measure in studies of vocational rehabilitation for adults with severe mental illness. Psychiatric Services, 51(2), 229-233.

Van Dongen, C. J. (1996). Quality of life and self-esteem in working and nonworking persons with mental illness. Community Mental Health Journal, 32(6), 535-548.

Xie, H., McHugo, G. J., Helmstetter, B. S., & Drake, R. E. (2005). Three-year recovery outcomes for long-term patients with co-occurring schizophrenic and substance use disorders. Schizophrenia Research, 75(2-3), 337-348.

Yanos, P. T., Roe, D., Markus, K., & Lysaker, P. H. (2008). Pathways between internalized stigma and outcomes related to recovery in schizophrenia spectrum disorders. Psychiatric Services, 59(12), 1437-1442.

Young, S. L., & Ensing, D. S. (1999). Exploring recovery from the perspective of people with psychiatric disabilities. Psychiatric Rehabilitation Journal, 22(3), 219-232.