Assessing Coping
in Schizophrenia
A Rationally
Devised Scoring Scheme to Assess Coping in Schizophrenia:
Internal Consistency and Associations with Work Performance
Internal Consistency and Associations with Work Performance
Assistant Professor of
Staff Psychologist
Research Assistant. Dept of Psychiatry
Rebecca S. Lancaster, M.S.
Research Assistant. Dept of Psychiatry
Louanne W. Davis, Psy.D
Staff Psychologist
Wayne Zito Psy.D.
Research Psychologist VA
Morris D. Bell Ph.D.
Research Psychologist VA
Lysaker, P.H.,
Johannesen, J., Lancaster. R.S. , Davis, LW., Zito, W., and Bell M.D.
(2004).
Assessing Coping
in Schizophrenia - A Rationally
Devised Scoring Scheme to Assess Coping in
Schizophrenia. International Journal of Psychosocial Rehabilitation. 8, 74-84.
Schizophrenia. International Journal of Psychosocial Rehabilitation. 8, 74-84.
Address correspondence
to:
Paul Lysaker, Ph.D.Day
Hospital W. 10th St Roudebush
VA Medical
Center Indianapolis ,
IN 46202
Email: plysaker@iupui.edu
Fax: (317) 554-0056
Ackowledgements:
Research funded in part by the Department of Veteran Affairs, Rehabilitation, Research, and Development Service,USA
There are no conflicts of interest for any of the authors of this paper. No author has any possible financial gain for the findings presented here.
The authors wish to thank Joseph Ventrua Ph.D. for his helpful comments during the shaping of this research project.
A Rationally Devised Scoring Scheme to Assess Coping in Schizophrenia: Internal Consistency and Associations with Work Performance
Abstract
Most assessments of coping behavior were not developed for the unique needs of persons with schizophrenia. Accordingly, a rationally derived scoring scheme for six aspects of coping implicated in psychosocial function in schizophrenia was constructed using the items of an existent instrument: the Ways of Coping. The revised scales demonstrated higher levels of internal consistency than the original scales across two schizophrenia samples (n=53, n=48). Coping strategies identified using the revised scales were predictive of work performance over a two-month period, while no relationship was found between coping and work function using the original scheme. Implications for research are discussed.
Introduction
Persons with schizophrenia often report chronic difficulty coping effectively with both major and minor stresses (Corrigan, & Toomey, 1995; Frese, 1993; Mueser et. al., 1997). They may possess a relatively limited repertoire of coping strategies (Rollins et al., 1999) and tend to avoid rather than actively attempt to solve problems (Farhall & Gehrke, 1997; Lysaker et al., 2003b; Wilder-Willis, et al., 2002). As a matter of coping style they thus may spend relatively little time thinking or talking about how to resolve a dilemma, and/or be less likely to actively and constructively respond to the stressor. Beyond being intuitively a matter of concern, maladaptive coping patterns in schizophrenia are of larger importance because they have been linked to symptom exacerbation and failure to sustain community tenure (e.g.; Hultman, Wieselgren, & Oehman, 1997; Macdonald et al., 1998; Meyer, 2001; Middleboe & Mortensen, 1997; Takei et al., 1990; Tarrier et al., 1988; Wiedl, 1992).
While the significance of coping deficits in schizophrenia is widely recognized, less is understood about how to measure adaptive coping in schizophrenia, how to detect when key aspects of adaptive coping are absent and how to systematically determine the individualized needs of persons who would benefit from acquiring better coping skills as part of their rehabilitation. One reason for this is that most coping inventories have generally been designed to measure coping among persons in the community with medical but not major psychiatric conditions (Dropkin, 2001; Siegel, Gluhoski, & Karus, 1997) and may, therefore, be less sensitive to the pervasive and possibly idiosyncratic deficits found in schizophrenia. Literature ranging from Bleuler (1911/1950) to contemporary accounts (Carter & Flesher, 1995) suggests that some with schizophrenia may have relatively unique difficulties coping. They may for instance fail to cope adequately because they contemplate action but do not act or because they act decisively but without adequate contemplation. Other authors have pointed out that some with schizophrenia may either consciously choose to not act upon or merely ignore the presence stressors, both due to a lack of any meaningful sense of personal agency (Hoffman et al., 2000; Lysaker et al., 2001; Lysaker et al., in press; Ventura et al., 1999, Young & Ensign, 1999).
In the current study we have, therefore, set out to determine whether the items of an existing coping scale, the Ways of Coping Questionnaire (WCQ, Folkman & Lazarus, 1988) could be scored in a different manner in order to assess the unique difficulties in schizophrenia that might be linked to functional impairments. While this particular instrument has been established as a measure of coping in a community-residing well population, the factor structure of the scale, as with other scales, does not appear to accurately reflect coping behaviors used by individuals with chronic illness (Wineman, Durand, & McCulloch, 1994), which may also be the case for individuals with schizophrenia. We chose the WCQ as a starting point, rather than attempt to invent another instrument, because the WCQ contains multiple items that are face valid for schizophrenia, several of which are not employed in the present scoring instructions. Additionally persons with schizophrenia have been found to validly complete the WCQ and, though perhaps flawed, the profile resulting from the existing scoring scheme has been linked to symptoms and personality in schizophrenia (Lysaker et al., 2003a).
In attempting to create a new scoring method for the WCQ we sought to assign each WCQ item to one of six coping categories, determined a priori based on the literature reviewed above. The first two categories were “considering” which refers to thinking or talking with others about what to do, and “acting,” which refers to taking direct action in regard to a stressor. As suggested by Monti, Abrams, Kadden, & Coney (1989; pg. 83), effective problem solving first requires resisting the temptation to either respond with an initial impulse or to do nothing at all, and then after considering the situation a solution is devised and acted upon. Since either of these elements of problem solving (considering and acting) might be impaired in schizophrenia we reasoned that each should be measured separately. The third category was “ignoring,” which refers to a preference for attempting to put the stressor out of one’s mind, or choosing to “not think” about the stressor. The fourth was “resigning” and that refers to a choice to not act because there is nothing to be done. The fifth category was “positively reappraising.” This dimension is virtually identical in name and description to one of the original 8 categories of the WCQ and refers to the tendency to see “the silver lining,” in a stressor, or to recast a negative stressor in a positive light. The final dimension was “self-soothing.” When this coping strategy predominates, the person’s primary concern is focused on how to regain emotional equilibrium by reducing overwhelming feelings of anxiety or negative affects rather than constructively dealing with the stressor itself.
In the current paper we present the item assignments created according to the new scoring scheme and then compare internal consistency and predictive validity of the revised and original scoring systems. In particular we predicted that if the new scheme were more sensitive to coping deficits in schizophrenia, it would demonstrate an average higher coefficient alpha than the original scale across at least two samples. We secondly predicted that a preference for considering and acting, the two subscales we hypothesize contain essential elements of successful problem solving (Monti, et al., 1989) would prospectively predict better psychosocial function in a work program. In other words, we reasoned that if the new scoring system validly assessed preference for considering and acting when faced with stressors, than participants with a preference for both should be better problem solvers and fare better over time in a work program over time.
Materials and Methods
Participants
Participants were collected from two independent samples: the first was used for the initial comparison of internal consistency between the original and our revised scoring scheme. Informed written consent was obtained for both samples and both studies received full approval from the appropriate hospital and university research review boards. Sample 1 was composed of 51 male and two female participants with a SCID (SCID-I; Spitzer, et al., 1994) confirmed diagnosis of schizophrenia (n = 39) or schizoaffective disorder (n = 14). All were recruited from the outpatient psychiatry clinic of a VA Medical Center for a larger study of the correlates of obsessive-compulsive symptoms in schizophrenia. On average, participants were 44 years old (SD=9), had 13 years of education (SD=2), and 10 lifetime psychiatric hospitalizations (SD=11) with the first occurring at age 26 (SD=6). All participants were in a post-acute phase of illness as defined by having no hospitalizations or changes in medication or housing in the month before entering the study. Participants were excluded who had a diagnosis of mental retardation, or another neurological disorder. WCQ responses from these participants were included in a previously published study of the clinical correlates of personality in schizophrenia that used the original scoring scheme (Lysaker et al., 2003a).
The second sample was used to evaluate the stability of subscales constructed using Sample 1. Participants were recruited from the outpatient psychiatry clinic of a VA Medical Center for a larger study of the clinical effects of psychosocial rehabilitation, which provided outcome data that was later used to evaluate the external validity of the coping scales. Sample 2 was composed of 47 males and one female with a SCID (SCID-I; Spitzer, et al., 1994) confirmed diagnosis of schizophrenia (n = 31) or schizoaffective disorder (n =16). On average, participants were 47 years old (SD=7), had 12 years of education (SD=2), and 12 lifetime psychiatric hospitalizations (SD=13) with the first occurring at age 24 (SD=7). All participants were in a post-acute phase of illness as defined by having no hospitalizations or changes in medication or housing in the month before entering the study. Participants were excluded who had a diagnosis of mental retardation, or another neurological disorder. WCQ responses for these participants have not been used in any studies published elsewhere.
Instruments
Work Behavior Inventory (WBI: Bryson et al, 1997): is a 35 item inventory developed specifically for the purposes of assessing behavior at work for persons with severe and persistent mental illness. A trained rater completes the WBI following direct observation of participants’ work behavior and an interview with the participants’ supervisor. Each WBI item consists of a specific behavior which is rated as a "1" persistent problem area, "2" occasional problem area, "3" average performance, "4" an occasional area of strength, or "5" a frequent area of strength. Items are used to generate a total score, ranging from 35 to 175 possible points, that is the sum of five sub-scales: social skills, cooperativeness, work habits, work quality, and personal presentation. Data supporting the factorial and concurrent validity of the WBI have been reported elsewhere (Bryson et al., 1997). Good to excellent interrater reliability were found in this study with intraclass correlations ranging from .82 to .94. WBI scores were also found to have predictive validity in terms of subsequent vocational outcomes (Bryson et al., 1999)
Ways of Coping Questionnaire (WCQ; Folkman and Lazarus, 1988): is a self-report instrument that asks participants to recall a recent stressor and then rate how often they have used 66 different behaviors to cope with that particular stressor. Scale scores are additively derived from individual items and divided by a total score to provide relative scores for a total of 8 scales: Confrontative Coping, Accepting Responsibility, Distancing, Escape-Avoidance, Planful Problem Solving, Positive Reappraisal, Self controlling and Seeking Social. These eight scales use a total of 50 of the total 66 items available. This instrument has been used in studies of coping in schizophrenia in a variety of samples (Lysaker et al., 2003a; Lysaker et al., 2003b; Lysaker et al., in press). In order to control for response bias, as suggested by the authors, relative scores were calculated. These were obtained for all scales by dividing a scale mean by the mean total such that a score of greater than “1” reflected a greater preference for that coping strategy while a score of less than “1” reflected a lesser preference for that strategy.
Procedures
After all participants provided informed consent, diagnoses were confirmed using the SCID-I (Spitzer, et al., 1994). Participants were then administered the WCQ by a research assistant as part of a battery of two larger studies as noted above. Participants in Sample 2 were offered a paid job placement for up to 20 regularly scheduled hours per week at a VA medical center. Job duties were equivalent to entry-level positions and regular job site supervisors provided supervision. Efforts were made to match participant's interests and skills with work placements in such areas as the escort service, housekeeping, the computer laboratory, and customer service. Job placements were made without reference to WCQ scores. Work behavior was evaluated using the WBI during the first, third, fifth, and seventh week of work by trained bachelor or masters level research assistants. WBI ratings were made blind to WCQ scores. We chose to use multiple assessment of work behavior rather than single points since previous research with other samples has suggested persons with schizophrenia take several weeks to settle into a stable work pattern (e.g. Lysaker et al., 1993).
Results
To derive the rational scales, two raters (PL & RL) separately sorted each of the 66 WCQ items. This yielded an agreement rate of over 90% and the remaining 7 items were sorted according to a consensus between raters. The final solution using the original item numbers was thus: Considering: items # 2,8,18,22,31,42,45,48, 60,63,64; Acting: items # 1,5,6,7,20,25,26,34, 39,46,49,52,62; Ignoring: items # 4,10,13,16,21, 24,32,35,40,41,50,59; Resigning: items # 9,11, 12, 29,43,51,53,55,58,61; Positive reappraisal: items #15,23,27,30,36, 38,65; and Self-soothing: items # 3,14,17,19,28,33, 37,44,47, 54, 56,57, 66. Correlations of the original scales with the new scoring scheme are presented in Table 1.
In order to examine the reliability of the new scoring scheme relative to the original scheme, we assessed the internal consistency of each version of the WQC scale across two independent samples by calculating coefficient alpha for each rational and original scale. Item analysis was conducted at the scale level to identify improvement in internal consistency that could be gained through the removal of select items. A single item (#65) assigned to the Positive Reappraisal scale appeared to detract from the overall coefficient alpha and was therefore dropped from the scale. Maximum inter-item correlations were achieved under the proposed scoring scheme for all other scales. As depicted in Table 2, higher coefficient alphas were consistently observed within both samples at total scale and individual subscale levels for the rational scoring scheme. Total scale reliability for the rational scales was considered adequate to good, with coefficient alpha values averaging .79 and .80 in Sample 1 and Sample 2 respectively. In contrast, the mean coefficient alphas for the original scoring scheme were lower on average for both samples, with four of the eight subscales failing to produce acceptable reliability (below .70) when evaluated independently.
Table 2: Internal Consistency
________________________________________________________________________
Lysaker Scales Sample 1 (n = 53) Sample 2 (n = 48)
________________________________________________________________________
Acting Alpha = .86 Alpha = .87
Considering Alpha = .82 Alpha = .83
Ignoring Alpha = .75 Alpha = .74
Resigning Alpha = .74 Alpha = .75
Positive Reappraisal Alpha = .81 Alpha = .84
Self Soothing Alpha = .77 Alpha = .74
Scale Average Alpha = .79 Alpha = .80
________________________________________________________________________
Original Scales
________________________________________________________________________
Confrontive Coping Alpha = .74 Alpha = .72
Distancing Alpha = .65 Alpha = .65
Self-controlling Alpha = .60 Alpha = .61
Seeking Social Support Alpha = .82 Alpha = .85
Accepting Responsibility Alpha = .66 Alpha = .53
Escape-Avoidance Alpha = .67 Alpha = .73
Planful Problem Solving Alpha = .83 Alpha = .77
Positive Reappraisal Alpha = .80 Alpha = .82
Scale Average Alpha = .72 Alpha = .71
________________________________________________________________________
To examine how preference for the methods of coping involved in successful problem solving impact work performance, participants from Sample 2 who had worked for at least 6 of the first 7 weeks (n=34) were divided into two groups. The first (n = 10) group, the “Considering - Acting” group, was composed of participants with relative scores of greater than “1” on both the “Considering” and “Acting” scales (i.e. participants with a preference for both of the elements essential to problem solving, but which are often impaired in schizophrenia). The second group (n = 24), labeled “Other”, was composed of participants with a relative score of 1 or less on either or both of the “Considering” or “Acting” scales. Thus members of this group could have a preference for only one or neither of those scales. Repeated measures ANOVA was then conducted comparing the WBI Total scores across weeks 1, 3, 5 and 7 for the Considering-Acting group and the Other group. As illustrated in Figure 1, the Considering-Acting group had significantly higher WBI total scores (F = 5.798, p < .05). There was also a significant time effect with both groups improving over time (F = 10.260, p < .01) but no interaction (F = 1.22, p = .28). Post-hoc analyses comparing the individual WBI subscales are reported in Table 3.
Figure 1 - Participants identified as “Considering-Acting” (n = 10) and “Other” (n = 24) based on relative preference for rationally derived coping strategies. Repeated-measures ANOVA across weeks 1, 3, 5 and 7 of work indicated significantly higher WBI total scores (F = 5.798, p < .05) associated with preference for “Considering-Acting”.
Table 3: Repeated Measures ANOVA of WBI Scores Across 7 Weeks of Work
____________________________________________________________________________
WBI Scale Consider-Act Other Time Group
(n = 10) (n = 24) Effect Effect
_____________________________________________________________________________
Mean (SD) Mean (SD)
Social Skills F = 12.33*** F = 3.31 (ns)
Week 1 3.26 (0.31) 3.13 (0.30)
Week 3 3.46 (0.40) 3.15 (0.43)
Week 5 3.48 (0.45) 3.30 (0.38)
Week 7 3.58 (0.48) 3.26 (0.38)
Cooperativeness F = 6.62* F = 7.24**
Week 1 3.41 (0.38) 3.24 (0.30)
Week 3 3.60 (0.35) 3.36 (0.32)
Week 5 3.74 (0.40) 3.44 (0.55)
Week 7 3.89 (0.47) 3.28 (0.50)
Work Habits F = 3.84 (ns) F = 3.56 (ns)
Week 1 3.34 (0.41) 3.00 (0.38)
Week 3 3.43 (0.40) 3.22 (0.55)
Week 5 3.55 (0.54) 3.29 (0.66)
Week 7 3.67 (0.63) 3.16 (0.68)
Work Quality F = 7.40** F = 4.59*
Week 1 3.30 (0.44) 2.92 (0.39)
Week 3 3.53 (0.35) 3.11 (0.46)
Week 5 3.55 (0.54) 3.30 (0.65)
Week 7 3.63 (0.71) 3.26 (0.62)
Personal Presentation F = 10.58** F = 5.02*
Week 1 3.23 (0.28) 3.08 (0.20)
Week 3 3.36 (0.49) 3.11 (0.33)
Week 5 3.43 (0.47) 3.17 (0.38)
Week 7 3.60 (0.46) 3.18 (0.45)
WBI Total F = 10.26** F = 5.80 *
Week 1 3.31 (0.31) 3.08 (0.25)
Week 3 3.47 (0.36) 3.17 (0.37)
Week 5 3.55 (0.47) 3.30 (0.44)
Week 7 3.67 (0.52) 3.23 (0.47)
____________________________________________________________________________
* p<.05; ** p<.01; *** p<.0001
Given the correlations between Considering and Seeking Social Support and between Acting and Planful Problem Solving, as well as roughly equivalent levels of internal consistency between these four scale, we next examined whether participants with preferences for both Planful Problem Solving and Social Support Seeking (n = 10) had better work performance when compared to individuals who use only one or neither of these strategies (n = 24). These analyses failed to find significant differences on the WBI (F = 1.0, p = .30). Of note, the Considering-Acting group and the Planful Problem Solving-Social Support Seeking group were dissimilar with respect to membership, sharing only 7 out of 13 participants in common (54%).
To determine whether a simple preference for Considering (n = 13) or Acting (n =17) from our revised scale prospectively predicted work performance, we conducted two repeated measures ANOVA comparing the WBI total scores. Neither of these analyses produced significant differences: F = 1.94, p = .17, F = 1.33, p = .26, respectively. Lastly, an exploratory analysis was conducted to determine whether preference for Planful Problem Solving (n = 23) or Social Support Seeking (n = 17), according to the original scales, could independently predict better work performance. Again, neither was found to produce significant differences: F = .027, p = .871 and F = .370, p = .548 respectively.
Discussion
In this study a rationally derived scoring scheme was created which reorganized the items of the WOC into six primary according to deficits commonly reported in schizophrenia. Of these scales, five tapped aspects of coping conceptually distinct from the intent of the original scales, and one was virtually identical. Correlations among these scoring schemes suggested that the newly created scale scores were related to the original scores but far from identical. When compared across two samples, the rationally devised scales had better internal consistency than the original scales overall, with several of the original scales but none of the new scales failing to achieve acceptable internal consistency. When used to predict work performance, it was concluded that persons who expressed a relative preference for both acting and thinking prior to starting work achieved consistently better work performance over a period of two months in a work program than individuals who demonstrated preference for either of these two methods independently or, conversely, neither. As might be expected, ad hoc analyses showed that persons who utilized coping strategies involving both acting and thinking performed better at work, particularly in areas of instrumental function relating to work performance and cooperativeness. In addition, personal presentation and sociability were found to be areas of strength for these individuals relative to coworkers. Of note, these findings could not be replicated using a combined preference for the original planful problem solving and social support seeking scales of the WCQ, or when either was present alone.
Taken together these results support the further investigation of a scoring scheme rationally tailored for the study of coping in schizophrenia. The rationally derived conceptualization presented herein appears to provide more reliable and valid assessments of the coping strengths and deficits in schizophrenia, and thus may represent an incremental improvement from the original scoring scheme. With replication, these methods may have practical application as well. In particular, vocational programs may be able to identify persons at risk for poorer outcome on the basis of their coping preferences, and eventually develop or employ interventions aimed at helping these individuals to become more willing and comfortable in using more thoughtful and action-oriented coping strategies.
Of note, there are limitations to this study. Generalization of findings is limited by sample composition. Participants were generally middle-aged males involved in treatment. It may be that a different relationship exists between coping and function among females, younger men with schizophrenia, persons declining treatment or others in different stages of illness. Clearly more research is necessary with broader samples. We also measured coping once as it was reported in response to single recent incident. Future studies may wish to include longitudinal assessments of coping and more qualitative information regarding a broader range of situations with which individuals cope, as it may be that these relationships vary depending upon the stressor and its context. Furthermore, while statistically meaningful, differences in work performance were modest and it cannot ruled out that the relationships between coping and work performance were not mediated by other factors. Replication is necessary with studies designed to control for mediating variables before these findings are used for the basis of intervention. Lastly, assessment of the predictive validity focused on relative preference for acting and considering and the other scales were not examined directly. Future investigation is needed to further validate the remainder of the proposed rational scales.
Paul Lysaker, Ph.D.
Email: plysaker@iupui.edu
Fax: (317) 554-0056
Ackowledgements:
Research funded in part by the Department of Veteran Affairs, Rehabilitation, Research, and Development Service,
There are no conflicts of interest for any of the authors of this paper. No author has any possible financial gain for the findings presented here.
The authors wish to thank Joseph Ventrua Ph.D. for his helpful comments during the shaping of this research project.
A Rationally Devised Scoring Scheme to Assess Coping in Schizophrenia: Internal Consistency and Associations with Work Performance
Abstract
Most assessments of coping behavior were not developed for the unique needs of persons with schizophrenia. Accordingly, a rationally derived scoring scheme for six aspects of coping implicated in psychosocial function in schizophrenia was constructed using the items of an existent instrument: the Ways of Coping. The revised scales demonstrated higher levels of internal consistency than the original scales across two schizophrenia samples (n=53, n=48). Coping strategies identified using the revised scales were predictive of work performance over a two-month period, while no relationship was found between coping and work function using the original scheme. Implications for research are discussed.
Introduction
Persons with schizophrenia often report chronic difficulty coping effectively with both major and minor stresses (Corrigan, & Toomey, 1995; Frese, 1993; Mueser et. al., 1997). They may possess a relatively limited repertoire of coping strategies (Rollins et al., 1999) and tend to avoid rather than actively attempt to solve problems (Farhall & Gehrke, 1997; Lysaker et al., 2003b; Wilder-Willis, et al., 2002). As a matter of coping style they thus may spend relatively little time thinking or talking about how to resolve a dilemma, and/or be less likely to actively and constructively respond to the stressor. Beyond being intuitively a matter of concern, maladaptive coping patterns in schizophrenia are of larger importance because they have been linked to symptom exacerbation and failure to sustain community tenure (e.g.; Hultman, Wieselgren, & Oehman, 1997; Macdonald et al., 1998; Meyer, 2001; Middleboe & Mortensen, 1997; Takei et al., 1990; Tarrier et al., 1988; Wiedl, 1992).
While the significance of coping deficits in schizophrenia is widely recognized, less is understood about how to measure adaptive coping in schizophrenia, how to detect when key aspects of adaptive coping are absent and how to systematically determine the individualized needs of persons who would benefit from acquiring better coping skills as part of their rehabilitation. One reason for this is that most coping inventories have generally been designed to measure coping among persons in the community with medical but not major psychiatric conditions (Dropkin, 2001; Siegel, Gluhoski, & Karus, 1997) and may, therefore, be less sensitive to the pervasive and possibly idiosyncratic deficits found in schizophrenia. Literature ranging from Bleuler (1911/1950) to contemporary accounts (Carter & Flesher, 1995) suggests that some with schizophrenia may have relatively unique difficulties coping. They may for instance fail to cope adequately because they contemplate action but do not act or because they act decisively but without adequate contemplation. Other authors have pointed out that some with schizophrenia may either consciously choose to not act upon or merely ignore the presence stressors, both due to a lack of any meaningful sense of personal agency (Hoffman et al., 2000; Lysaker et al., 2001; Lysaker et al., in press; Ventura et al., 1999, Young & Ensign, 1999).
In the current study we have, therefore, set out to determine whether the items of an existing coping scale, the Ways of Coping Questionnaire (WCQ, Folkman & Lazarus, 1988) could be scored in a different manner in order to assess the unique difficulties in schizophrenia that might be linked to functional impairments. While this particular instrument has been established as a measure of coping in a community-residing well population, the factor structure of the scale, as with other scales, does not appear to accurately reflect coping behaviors used by individuals with chronic illness (Wineman, Durand, & McCulloch, 1994), which may also be the case for individuals with schizophrenia. We chose the WCQ as a starting point, rather than attempt to invent another instrument, because the WCQ contains multiple items that are face valid for schizophrenia, several of which are not employed in the present scoring instructions. Additionally persons with schizophrenia have been found to validly complete the WCQ and, though perhaps flawed, the profile resulting from the existing scoring scheme has been linked to symptoms and personality in schizophrenia (Lysaker et al., 2003a).
In attempting to create a new scoring method for the WCQ we sought to assign each WCQ item to one of six coping categories, determined a priori based on the literature reviewed above. The first two categories were “considering” which refers to thinking or talking with others about what to do, and “acting,” which refers to taking direct action in regard to a stressor. As suggested by Monti, Abrams, Kadden, & Coney (1989; pg. 83), effective problem solving first requires resisting the temptation to either respond with an initial impulse or to do nothing at all, and then after considering the situation a solution is devised and acted upon. Since either of these elements of problem solving (considering and acting) might be impaired in schizophrenia we reasoned that each should be measured separately. The third category was “ignoring,” which refers to a preference for attempting to put the stressor out of one’s mind, or choosing to “not think” about the stressor. The fourth was “resigning” and that refers to a choice to not act because there is nothing to be done. The fifth category was “positively reappraising.” This dimension is virtually identical in name and description to one of the original 8 categories of the WCQ and refers to the tendency to see “the silver lining,” in a stressor, or to recast a negative stressor in a positive light. The final dimension was “self-soothing.” When this coping strategy predominates, the person’s primary concern is focused on how to regain emotional equilibrium by reducing overwhelming feelings of anxiety or negative affects rather than constructively dealing with the stressor itself.
In the current paper we present the item assignments created according to the new scoring scheme and then compare internal consistency and predictive validity of the revised and original scoring systems. In particular we predicted that if the new scheme were more sensitive to coping deficits in schizophrenia, it would demonstrate an average higher coefficient alpha than the original scale across at least two samples. We secondly predicted that a preference for considering and acting, the two subscales we hypothesize contain essential elements of successful problem solving (Monti, et al., 1989) would prospectively predict better psychosocial function in a work program. In other words, we reasoned that if the new scoring system validly assessed preference for considering and acting when faced with stressors, than participants with a preference for both should be better problem solvers and fare better over time in a work program over time.
Materials and Methods
Participants
Participants were collected from two independent samples: the first was used for the initial comparison of internal consistency between the original and our revised scoring scheme. Informed written consent was obtained for both samples and both studies received full approval from the appropriate hospital and university research review boards. Sample 1 was composed of 51 male and two female participants with a SCID (SCID-I; Spitzer, et al., 1994) confirmed diagnosis of schizophrenia (n = 39) or schizoaffective disorder (n = 14). All were recruited from the outpatient psychiatry clinic of a VA Medical Center for a larger study of the correlates of obsessive-compulsive symptoms in schizophrenia. On average, participants were 44 years old (SD=9), had 13 years of education (SD=2), and 10 lifetime psychiatric hospitalizations (SD=11) with the first occurring at age 26 (SD=6). All participants were in a post-acute phase of illness as defined by having no hospitalizations or changes in medication or housing in the month before entering the study. Participants were excluded who had a diagnosis of mental retardation, or another neurological disorder. WCQ responses from these participants were included in a previously published study of the clinical correlates of personality in schizophrenia that used the original scoring scheme (Lysaker et al., 2003a).
The second sample was used to evaluate the stability of subscales constructed using Sample 1. Participants were recruited from the outpatient psychiatry clinic of a VA Medical Center for a larger study of the clinical effects of psychosocial rehabilitation, which provided outcome data that was later used to evaluate the external validity of the coping scales. Sample 2 was composed of 47 males and one female with a SCID (SCID-I; Spitzer, et al., 1994) confirmed diagnosis of schizophrenia (n = 31) or schizoaffective disorder (n =16). On average, participants were 47 years old (SD=7), had 12 years of education (SD=2), and 12 lifetime psychiatric hospitalizations (SD=13) with the first occurring at age 24 (SD=7). All participants were in a post-acute phase of illness as defined by having no hospitalizations or changes in medication or housing in the month before entering the study. Participants were excluded who had a diagnosis of mental retardation, or another neurological disorder. WCQ responses for these participants have not been used in any studies published elsewhere.
Instruments
Work Behavior Inventory (WBI: Bryson et al, 1997): is a 35 item inventory developed specifically for the purposes of assessing behavior at work for persons with severe and persistent mental illness. A trained rater completes the WBI following direct observation of participants’ work behavior and an interview with the participants’ supervisor. Each WBI item consists of a specific behavior which is rated as a "1" persistent problem area, "2" occasional problem area, "3" average performance, "4" an occasional area of strength, or "5" a frequent area of strength. Items are used to generate a total score, ranging from 35 to 175 possible points, that is the sum of five sub-scales: social skills, cooperativeness, work habits, work quality, and personal presentation. Data supporting the factorial and concurrent validity of the WBI have been reported elsewhere (Bryson et al., 1997). Good to excellent interrater reliability were found in this study with intraclass correlations ranging from .82 to .94. WBI scores were also found to have predictive validity in terms of subsequent vocational outcomes (Bryson et al., 1999)
Ways of Coping Questionnaire (WCQ; Folkman and Lazarus, 1988): is a self-report instrument that asks participants to recall a recent stressor and then rate how often they have used 66 different behaviors to cope with that particular stressor. Scale scores are additively derived from individual items and divided by a total score to provide relative scores for a total of 8 scales: Confrontative Coping, Accepting Responsibility, Distancing, Escape-Avoidance, Planful Problem Solving, Positive Reappraisal, Self controlling and Seeking Social. These eight scales use a total of 50 of the total 66 items available. This instrument has been used in studies of coping in schizophrenia in a variety of samples (Lysaker et al., 2003a; Lysaker et al., 2003b; Lysaker et al., in press). In order to control for response bias, as suggested by the authors, relative scores were calculated. These were obtained for all scales by dividing a scale mean by the mean total such that a score of greater than “1” reflected a greater preference for that coping strategy while a score of less than “1” reflected a lesser preference for that strategy.
Procedures
After all participants provided informed consent, diagnoses were confirmed using the SCID-I (Spitzer, et al., 1994). Participants were then administered the WCQ by a research assistant as part of a battery of two larger studies as noted above. Participants in Sample 2 were offered a paid job placement for up to 20 regularly scheduled hours per week at a VA medical center. Job duties were equivalent to entry-level positions and regular job site supervisors provided supervision. Efforts were made to match participant's interests and skills with work placements in such areas as the escort service, housekeeping, the computer laboratory, and customer service. Job placements were made without reference to WCQ scores. Work behavior was evaluated using the WBI during the first, third, fifth, and seventh week of work by trained bachelor or masters level research assistants. WBI ratings were made blind to WCQ scores. We chose to use multiple assessment of work behavior rather than single points since previous research with other samples has suggested persons with schizophrenia take several weeks to settle into a stable work pattern (e.g. Lysaker et al., 1993).
Results
To derive the rational scales, two raters (PL & RL) separately sorted each of the 66 WCQ items. This yielded an agreement rate of over 90% and the remaining 7 items were sorted according to a consensus between raters. The final solution using the original item numbers was thus: Considering: items # 2,8,18,22,31,42,45,48, 60,63,64; Acting: items # 1,5,6,7,20,25,26,34, 39,46,49,52,62; Ignoring: items # 4,10,13,16,21, 24,32,35,40,41,50,59; Resigning: items # 9,11, 12, 29,43,51,53,55,58,61; Positive reappraisal: items #15,23,27,30,36, 38,65; and Self-soothing: items # 3,14,17,19,28,33, 37,44,47, 54, 56,57, 66. Correlations of the original scales with the new scoring scheme are presented in Table 1.
Table
1 Intercorrelations of Original and Rationally Derived Scales
Within Sample 1 (n = 53)
Original Scales
|
|
|||||
|
Acting |
Considering |
Ignoring |
Resigning |
Positive
Reappraial |
Self-
Soothing |
|
|
Confrontive
Coping |
.586*** |
.299* |
-.423** |
-.406** |
-.107 |
.128 |
|
Distancing |
-.377** |
-.393** |
.597*** |
.088 |
.035 |
.162 |
|
Self-Controlling |
-.275* |
-.349* |
.279* |
.385** |
-.237 |
.224 |
|
Social
Support |
-.238 |
.798*** |
-.202 |
-.006 |
-.051 |
-.316* |
|
Accepting
Responsibility
|
-.150 |
-.115 |
-.104 |
.280* |
-.061 |
.128 |
|
Escape-Avoidance |
-.440** |
-.308* |
.529*** |
.624*** |
-.500*** |
.054 |
|
Planful
Problem Solving
|
.644*** |
-.236 |
-.238 |
-.338* |
.189 |
-.152 |
|
Positive
Reappraisal |
.170 |
.081 |
-.297* |
-.553*** |
.736*** |
-.121 |
* p<.05; ** p<.01; *** p<.0001 a Correlations are based on relative scores calculated for each subject.
In order to examine the reliability of the new scoring scheme relative to the original scheme, we assessed the internal consistency of each version of the WQC scale across two independent samples by calculating coefficient alpha for each rational and original scale. Item analysis was conducted at the scale level to identify improvement in internal consistency that could be gained through the removal of select items. A single item (#65) assigned to the Positive Reappraisal scale appeared to detract from the overall coefficient alpha and was therefore dropped from the scale. Maximum inter-item correlations were achieved under the proposed scoring scheme for all other scales. As depicted in Table 2, higher coefficient alphas were consistently observed within both samples at total scale and individual subscale levels for the rational scoring scheme. Total scale reliability for the rational scales was considered adequate to good, with coefficient alpha values averaging .79 and .80 in Sample 1 and Sample 2 respectively. In contrast, the mean coefficient alphas for the original scoring scheme were lower on average for both samples, with four of the eight subscales failing to produce acceptable reliability (below .70) when evaluated independently.
Table 2: Internal Consistency
________________________________________________________________________
Lysaker Scales Sample 1 (n = 53) Sample 2 (n = 48)
________________________________________________________________________
Acting Alpha = .86 Alpha = .87
Considering Alpha = .82 Alpha = .83
Ignoring Alpha = .75 Alpha = .74
Resigning Alpha = .74 Alpha = .75
Positive Reappraisal Alpha = .81 Alpha = .84
Self Soothing Alpha = .77 Alpha = .74
Scale Average Alpha = .79 Alpha = .80
________________________________________________________________________
Original Scales
________________________________________________________________________
Confrontive Coping Alpha = .74 Alpha = .72
Distancing Alpha = .65 Alpha = .65
Self-controlling Alpha = .60 Alpha = .61
Seeking Social Support Alpha = .82 Alpha = .85
Accepting Responsibility Alpha = .66 Alpha = .53
Escape-Avoidance Alpha = .67 Alpha = .73
Planful Problem Solving Alpha = .83 Alpha = .77
Positive Reappraisal Alpha = .80 Alpha = .82
Scale Average Alpha = .72 Alpha = .71
________________________________________________________________________
To examine how preference for the methods of coping involved in successful problem solving impact work performance, participants from Sample 2 who had worked for at least 6 of the first 7 weeks (n=34) were divided into two groups. The first (n = 10) group, the “Considering - Acting” group, was composed of participants with relative scores of greater than “1” on both the “Considering” and “Acting” scales (i.e. participants with a preference for both of the elements essential to problem solving, but which are often impaired in schizophrenia). The second group (n = 24), labeled “Other”, was composed of participants with a relative score of 1 or less on either or both of the “Considering” or “Acting” scales. Thus members of this group could have a preference for only one or neither of those scales. Repeated measures ANOVA was then conducted comparing the WBI Total scores across weeks 1, 3, 5 and 7 for the Considering-Acting group and the Other group. As illustrated in Figure 1, the Considering-Acting group had significantly higher WBI total scores (F = 5.798, p < .05). There was also a significant time effect with both groups improving over time (F = 10.260, p < .01) but no interaction (F = 1.22, p = .28). Post-hoc analyses comparing the individual WBI subscales are reported in Table 3.
Figure 1 - Participants identified as “Considering-Acting” (n = 10) and “Other” (n = 24) based on relative preference for rationally derived coping strategies. Repeated-measures ANOVA across weeks 1, 3, 5 and 7 of work indicated significantly higher WBI total scores (F = 5.798, p < .05) associated with preference for “Considering-Acting”.
Table 3: Repeated Measures ANOVA of WBI Scores Across 7 Weeks of Work
____________________________________________________________________________
WBI Scale Consider-Act Other Time Group
(n = 10) (n = 24) Effect Effect
_____________________________________________________________________________
Mean (SD) Mean (SD)
Social Skills F = 12.33*** F = 3.31 (ns)
Week 1 3.26 (0.31) 3.13 (0.30)
Week 3 3.46 (0.40) 3.15 (0.43)
Week 5 3.48 (0.45) 3.30 (0.38)
Week 7 3.58 (0.48) 3.26 (0.38)
Cooperativeness F = 6.62* F = 7.24**
Week 1 3.41 (0.38) 3.24 (0.30)
Week 3 3.60 (0.35) 3.36 (0.32)
Week 5 3.74 (0.40) 3.44 (0.55)
Week 7 3.89 (0.47) 3.28 (0.50)
Work Habits F = 3.84 (ns) F = 3.56 (ns)
Week 1 3.34 (0.41) 3.00 (0.38)
Week 3 3.43 (0.40) 3.22 (0.55)
Week 5 3.55 (0.54) 3.29 (0.66)
Week 7 3.67 (0.63) 3.16 (0.68)
Work Quality F = 7.40** F = 4.59*
Week 1 3.30 (0.44) 2.92 (0.39)
Week 3 3.53 (0.35) 3.11 (0.46)
Week 5 3.55 (0.54) 3.30 (0.65)
Week 7 3.63 (0.71) 3.26 (0.62)
Personal Presentation F = 10.58** F = 5.02*
Week 1 3.23 (0.28) 3.08 (0.20)
Week 3 3.36 (0.49) 3.11 (0.33)
Week 5 3.43 (0.47) 3.17 (0.38)
Week 7 3.60 (0.46) 3.18 (0.45)
WBI Total F = 10.26** F = 5.80 *
Week 1 3.31 (0.31) 3.08 (0.25)
Week 3 3.47 (0.36) 3.17 (0.37)
Week 5 3.55 (0.47) 3.30 (0.44)
Week 7 3.67 (0.52) 3.23 (0.47)
____________________________________________________________________________
* p<.05; ** p<.01; *** p<.0001
Given the correlations between Considering and Seeking Social Support and between Acting and Planful Problem Solving, as well as roughly equivalent levels of internal consistency between these four scale, we next examined whether participants with preferences for both Planful Problem Solving and Social Support Seeking (n = 10) had better work performance when compared to individuals who use only one or neither of these strategies (n = 24). These analyses failed to find significant differences on the WBI (F = 1.0, p = .30). Of note, the Considering-Acting group and the Planful Problem Solving-Social Support Seeking group were dissimilar with respect to membership, sharing only 7 out of 13 participants in common (54%).
To determine whether a simple preference for Considering (n = 13) or Acting (n =17) from our revised scale prospectively predicted work performance, we conducted two repeated measures ANOVA comparing the WBI total scores. Neither of these analyses produced significant differences: F = 1.94, p = .17, F = 1.33, p = .26, respectively. Lastly, an exploratory analysis was conducted to determine whether preference for Planful Problem Solving (n = 23) or Social Support Seeking (n = 17), according to the original scales, could independently predict better work performance. Again, neither was found to produce significant differences: F = .027, p = .871 and F = .370, p = .548 respectively.
Discussion
In this study a rationally derived scoring scheme was created which reorganized the items of the WOC into six primary according to deficits commonly reported in schizophrenia. Of these scales, five tapped aspects of coping conceptually distinct from the intent of the original scales, and one was virtually identical. Correlations among these scoring schemes suggested that the newly created scale scores were related to the original scores but far from identical. When compared across two samples, the rationally devised scales had better internal consistency than the original scales overall, with several of the original scales but none of the new scales failing to achieve acceptable internal consistency. When used to predict work performance, it was concluded that persons who expressed a relative preference for both acting and thinking prior to starting work achieved consistently better work performance over a period of two months in a work program than individuals who demonstrated preference for either of these two methods independently or, conversely, neither. As might be expected, ad hoc analyses showed that persons who utilized coping strategies involving both acting and thinking performed better at work, particularly in areas of instrumental function relating to work performance and cooperativeness. In addition, personal presentation and sociability were found to be areas of strength for these individuals relative to coworkers. Of note, these findings could not be replicated using a combined preference for the original planful problem solving and social support seeking scales of the WCQ, or when either was present alone.
Taken together these results support the further investigation of a scoring scheme rationally tailored for the study of coping in schizophrenia. The rationally derived conceptualization presented herein appears to provide more reliable and valid assessments of the coping strengths and deficits in schizophrenia, and thus may represent an incremental improvement from the original scoring scheme. With replication, these methods may have practical application as well. In particular, vocational programs may be able to identify persons at risk for poorer outcome on the basis of their coping preferences, and eventually develop or employ interventions aimed at helping these individuals to become more willing and comfortable in using more thoughtful and action-oriented coping strategies.
Of note, there are limitations to this study. Generalization of findings is limited by sample composition. Participants were generally middle-aged males involved in treatment. It may be that a different relationship exists between coping and function among females, younger men with schizophrenia, persons declining treatment or others in different stages of illness. Clearly more research is necessary with broader samples. We also measured coping once as it was reported in response to single recent incident. Future studies may wish to include longitudinal assessments of coping and more qualitative information regarding a broader range of situations with which individuals cope, as it may be that these relationships vary depending upon the stressor and its context. Furthermore, while statistically meaningful, differences in work performance were modest and it cannot ruled out that the relationships between coping and work performance were not mediated by other factors. Replication is necessary with studies designed to control for mediating variables before these findings are used for the basis of intervention. Lastly, assessment of the predictive validity focused on relative preference for acting and considering and the other scales were not examined directly. Future investigation is needed to further validate the remainder of the proposed rational scales.
