Call It a Monster for Lack of Anything Else

Abstract

Individuals diagnosed with schizophrenia often appear to be unaware of having an illness or actively reject their diagnostic label. It is unclear, however, how this lack of awareness relates to important outcomes. Broadening the definition of awareness to include “narrative insight” may clarify this issue. The objective of this study was to identify profiles of narrative insight and test how these relate to standardized measure of insight. Sixty-five individuals with schizophrenia spectrum disorders participated in an assessment that included the Scale of Unawareness of Mental Disorder (SUMD) and an in-depth semi-structured interview. Qualitative analysis revealed 5 central themes related to insight on the basis of which each interview was then rated. Cluster analysis of these ratings resulted in 4 profiles of narrative insight: (1) accepts illness/rejects label, (2) rejects illness/searches for a name (3) passive insight of illness and label, and (4) integrative insight. The SUMD differentiated between individuals assigned to profile 2 who showed low insight to their illness and those assigned to the other profiles of narrative insight, but could not differentiae between them. Results support the claim that illness narratives are multifaceted and that traditional measures of insight may not be sensitive to different ways in which people understand their illness.

One of the most important advances of modern descriptive psychiatry is improved reliability of psychiatric diagnoses. But, even when offered by clinicians with the greatest of care and concern, reliable diagnoses are often rejected by the individuals for whom they are intended, leaving clinicians feeling frustrated and helpless. These mental health consumers, in turn, are often left feeling just as frustrated and helpless, as they often find it impossible to relate their experiences to the carefully rendered reliable diagnostic label. They may not agree with the label and reject it, feeling that it objectifies personal experiences in a manner that robs them of personal meaning.

In traditional psychiatric literature, clients’ disagreement with clinicians is often considered a reflection of poor insight. Labeling the dispute between clinicians and clients about the meaning of the clients’ experience as lack of insight implies that clinicians, because of their clinical knowledge and experience, are more able to produce a valid interpretation of the clients’ experiences than the clients themselves. It also suggests that the clients’ difficulties would be considerably reduced if they genuinely accepted the clinicians’ interpretation of their experiences. Individuals with schizophrenia are especially considered to suffer from the consequences of lack of insight. Research has consistently shown that approximately 50% to 80% of individuals with schizophrenia exhibit varying degrees of lack of insight into their illness (Amador et al., 1991; 1994; Lincoln et al., 2007). Nevertheless, considerable theoretical and empirical controversy exists about the consequences of lack of insight. On one hand, lack of insight has been shown to be associated with negative outcomes such as lower levels of treatment adherence (Cuffel et al., 1996), less favorable clinical outcome (Schwartz, 1998), poorer social function (Francis and Penn, 2001; Lysaker et al., 1998; Olfson et al., 2001), and vocational dysfunction (Lysaker and Lysaker, 2002). On the other hand, elevated insight has also been associated with different undesirable outcomes such as higher levels of dysphoria (Amador et al., 1994; Dixon et al., 1998; Drake et al., 2004; Mintz et al., 2003), lowered self-esteem (Warner et al., 1989), and decreased quality of life (Hasson-Ohayon et al., 2006; Pyne et al., 2001; Schwartz, 2001). Therefore, the dispute between clinicians and clients about the latter's diagnosis generates a difficult professional dilemma with regard to naming and classifying the clients’ internal experiences.

Roe and Kravetz (2003) argued that differentiating between descriptive and narrative insights may be one way of dealing with the dilemma. The introduction of the term narrative insight broadens the definition of awareness of illness. It suggests that insight can be studied as a story people tell about their illness rather than as a set of descriptive beliefs that are consistent or inconsistent with what are officially considered to be the symptoms of an illness. The current study tested this presumed differentiation between descriptive and narrative insights. Specifically, it attempted to identify profiles of narrative insight and examine their relationship to a measure of traditional descriptive insight so that information provided by the narrative but not conveyed by the Unawareness of Mental Disorder (SUMD) might be identified.

The general trend in assessing insight has been to shift from categorical or unidimensional approaches to multidimensional continuous approaches (Jovanovski et al., 2007). Several new instruments have been developed over the last 2 decades to assess insight (Amador and Kronengold, 2004), most of them based on clinical ratings (ie, Amador et al., 1993) or self-report (i.e., David, 1990). Although these instruments vary by method and emphasis, the assumption underlying their assessment of insight is that their final scores reflect the match between the individual's account of her or his disorder and the normative description of the disorder attributed to that individual by a mental health professional.

In addition to these professional and descriptive approaches to insight, there have been attempts to adopt a narrative perspective and, thus, to broaden the horizons of insight. Lysaker and coworkers (2002; 2003) emphasized the importance of understanding the awareness of illness as a narrative produced by complex and ongoing conversations, not only between individuals but also between the diverse aspects that make up an individual. This view emphasizes the belief that any genuine awareness of an illness, aimed at guiding an individual's recovery must refer to a story that has been and is being meaningfully coauthored by the individual experiencing it.

According to Roe and Kravetz's (2003) multifunctional approach to insight, descriptive facts are essential to providing a reliable account of states of affairs, whereas narrative facts represent attempts to communicate the emotional experiences associated with a series of events. In keeping with this approach, a model of narrative insight was proposed. This model consists of 3 components: the illness narrative; the themes of narrative insight conveyed by this narrative; and consequences of the communication of these themes such as empathy, control over the illness, and quality of life. In a recent study, Ko et al. (2006) analyzed the narratives of 50 individuals with schizophrenia and identified 4 stages of insight formation: (1) the feeling that symptoms were unbearable, (2) comparisons of experiences with references, (3) perception that medication was not working, and (4) awareness of illness after medication relieved symptoms. According to the conclusions of the latter study, the empathic understanding of an individual's progress along the stages generated by this narrative can serve as a means of providing feedback about the source of a client's dissatisfactions and of the relationship of these dissatisfactions to that individual's illness. Consistent with this line of thought, the current study had 3 aims. First, we sought to identify the different themes that are embedded in an individual's story of their mental illness. Beyond simply accepting or rejecting the label offered by a mental health professional, we sought to identify the themes that arise when individuals are engaged in describing whether they have a mental illness. Second, we sought to determine whether these themes form different insight profiles that show how individuals have distinctly different ways of being aware of their condition, their condition vary along identifiable axes. Finally, we used a semi-structured interview measure of traditional insight, the Scale of Unawareness of Mental Disorder (Amador et al., 1993) to compare individuals’ levels of traditional insight to their different narrative insight profiles. The purpose of this comparison was to determine ways in which insight as characterized by a multifunctional perspective and measured by the analysis of illness narratives corresponds to and differs from insight as traditionally defined and measured.

METHODS

Participants

Sixty-four men and 1 woman whose diagnoses of schizophrenia (n = 32) or schizoaffective disorder (n = 33) was confirmed by means of the Structured Clinical Interview for the DSM-IV (SCID) (Spitzer et al., 1994) were recruited from a comprehensive day hospital at a VA Medical Center. All participants were receiving ongoing outpatient treatment and were in a postacute or stable phase of their disorder, defined as no hospitalizations or changes in medication or housing in the last month. Participants with a documented history of mental retardation, verified in a chart review, were excluded. participants had a mean age of 47.23 years (SD = 6.16), a mean educational level of 12.60 (SD = 1.22), and a mean of 10.50 lifetime hospitalizations, the first one occurring on average at the age of 24.14 years (SD = 8.03). Thirty-one participants were white; 33, African American; and 1, Latino.

Instruments

Indiana Psychiatric Illness Interview

The Indiana Psychiatric Illness Interview (IPII; Lysaker and Lysaker, 2002) is a semi-structured interview developed to elicit illness narratives. This interview is designed to elicit individuals’ narrative understanding of their psychiatric difficulties. The narratives elicited by this interview can be analyzed qualitatively. To develop this interview, a narrative theory of self (Gallagher, 2000), which stresses that the meaning of a key life event depends largely on the story within which it is embedded was adopted (Lysaker and Lysaker, 2002). In this view, what has and has not gone wrong for a individual with mental illness is not so much a collection of statements of facts by the individual who is living with the illness, but it is a story with a past and present, which points to a future toward which the individual is evolving.

Research assistants conducted the interviews that typically lasted between 30 and 60 minutes. Responses were audio-taped and later transcribed. The interview was divided conceptually into 4 sections. First, a rapport was established and participants were asked to tell the story of their lives in as much detail as they were able. Second, the participants were asked whether they thought they had a mental illness and how they understood it. This was followed with a question about what has and has not been affected by their condition. In the third section, participants were asked whether and, if so, how their condition controlled their lives and how they controlled their condition. Fourth, the participants were asked what they expect to remain the same and what would be different in the future with respect to interpersonal and psychological functioning. Thus, IPII produces a narrative of the self and the illness that can be analyzed as part of a larger story being told and not merely as the presence or absence of specific beliefs.

The Abbreviated SUMD (Amador et al., 1993) is a rating scale completed by clinically trained research staff after a semi-structured interview and chart review. For the purposes of this study, we used the total score, which is the sum of the 3 central items of SUMD (awareness of mental disorder, the consequences of mental disorder, and effects of medication). Each item was rated on a 5-point scale ranging from “1” (complete awareness) to “5” (severe unawareness). Interrater reliability for this study was found to be in the excellent range, with an intraclass correlation of 0.89.

Procedure

After informed consent, diagnosis was confirmed using the SCID (Spitzer et al., 1994). Next, participants completed a testing and interview battery that included the SUMD and IPII as part of the baseline assessment for a larger study of the effects of cognitive therapy on rehabilitation outcome. The participants’ responses to the IPII were qualitatively analyzed by raters blind to the SUMD ratings.

ANALYSIS AND RESULTS

All the IPIIs were transcribed and then analyzed using the following 4 steps. First, a qualitative analysis of the narratives elicited by the IPII was performed and major themes relating to the different ways in which people perceived and understood their experiences and condition were articulated. Second, these themes were transformed into quantitative rating scales. These were then used to rate each of the transcribed narratives, and interrater reliability for these ratings was estimated. Third, a cluster analysis of the quantitative ratings was performed to identify the distinct patterns of awareness of illness, which were reflected by the IPII interviews. Fourth, the groups produced by the cluster analyses were compared on demographics and on a traditional measure of insight, the SUMD. These steps will be described in greater detail in the following 2 sections.

Qualitative Analysis: Identifying Narrative Themes of Insight

The qualitative analysis of the transcripts of 65 interviews was based on grounded theory (Strauss and Corbin, 1990) and included open-coding case analysis to identify central themes related to insight into the mental illness. These analyses were performed by 2 coauthors who had not conducted the interviews and who are trained and experienced qualitative researchers. The 2 investigators independently reviewed all of the transcribed interviews. They examined and compared the narratives for similarities and differences, trying to understand the thematic phenomena reflected by each interview and assigning conceptual labels to instances that best captured the essence of these themes. The conceptual labels were then grounded by more abstract categories, resulting in the following 5 analyst-constructed categories (Patton, 1990): (1) belief in having a mental illness, (2) belief in having symptoms, (3) acceptance of the diagnostic label, (4) active involvement in searching for a name and explanation for the illness, and (5) attributing experiences to the label. Rating anchors and a scoring system were then developed by the 2 authors conducting the qualitative analysis so that each of the narratives could be rated on each of the themes on a 5-point scale. Interrater reliability for the ratings was high and ranged from 0.74 to 0.97. Discrepancies in ratings were resolved through discussion until a consensus was reached. Mean scores for each scale are presented in Table 1. Intercorrelations between the subscales are presented in Table 2.

TABLE 1.

Mean and Standard Deviations of Scores for Narrative Dimensions of Awareness of Illness (n = 65)

	Mean	Standard Deviation	Range
Belief in having a mental illness	4.26	1.31	1–5
Belief in having symptoms	3.67	1.67	1–5
Acceptance of the diagnostic label	4.23	1.54	1–5
Attributing experiences to the label	3.18	1.63	1–5
Active search for name or explanation	3.35	1.80	1–5

TABLE 2.

Intercorrelation of Narrative Dimensions of Awareness of Illness (n = 65)

	1	2	3	4	5
Belief in having a mental illness	—	0.25*	0.25*	0.28*	0.00
Belief in having symptoms		—	0.11	0.83**	0.46**
Acceptance of the diagnostic label		—	0.38*	0.08
Attributing experiences to the label			—	0.49*
Active search for name or explanation				—

^*p < 0.05

^**p < 0.01.

Quantitative Analysis 2: Identifying Profiles of Narrative Insight

In the second step of analysis, a k-means cluster analysis was used to determine whether there were distinct clusters of participants with different kinds of insight. Because the fifth theme (attribution of experiences to illness) was highly correlated with the other 4 themes (r = 0.25–0.83), the theme was considered redundant and was not included in the cluster analyses. K-means cluster analysis is a nonhierarchical form of cluster analysis that produces an optimal number of clusters by minimizing variability within clusters and maximizing variability between clusters. Based on ratings of the clients’ narratives on the 4 themes, 4 distinct clusters of clients were uncovered. These clusters were labeled according to the mean ratings of the clients assigned to each cluster on the 4 themes that emerged from the qualitative analysis.

Table 3 presents the 4 profiles that were produced by the cluster analysis. As shown in the table, the individuals characterized by the first profile had significantly higher ratings on the belief in having a mental illness and symptoms, and lower ratings on the acceptance of label and on searching for a label or explanation. Participants with this profile admitted both to having an illness and to having symptoms but rejected the terms used officially to characterize their illness and symptoms. We therefore labeled this group as “Accepts illness/rejects label” (n = 9). The following is an example of a narrative assigned to this profile.

TABLE 3.

Cluster Analysis of Narrative Dimensions of Insight and Demographics

Narrative Dimension	Cluster 1 Accepts Illness/Rejects Label (n = 9)	Cluster 2 Rejects Illness/Searches for Name (n = 9)	Cluster 3 Passive Acceptance (n = 18)	Cluster 4 Integrative Insight (n = 29)	ANOVA p < 0.05	Fishers LSD
Belief in having a mental illness	4.78 (0.66)	1.67 (0.87)	5.00 (0.00)	4.44 (0.99)	40.87**	2 < 1,3,4
Belief in having symptoms	4.67 (0.70)	1.56 (1.33)	2.38 (1.69)	4.83 (0.76)	28.93**	2,3 < 1,4
Acceptance of the diagnostic label	2.33 (2.00)	2.89 (2.02)	4.78 (0.94)	4.89 (0.41)	16.05**	1,2 < 3,4
Active search for name/explanation	2.22 (0.67)	3.00 (1.80)	1.56 (0.86)	4.55 (0.83)	36.67**	1,2,3 < 4; 3 < 2
Demographics
Age	46.22 (4.15)	52.67 (5.57)	47.72 (6.43)	45.55 (5.93)	3.54*	1,2,4 < 3
Education	12.33 (1.58)	12.78 (0.97)	12.78 (1.06)	12.52 (1.30)	0.37
Lifetime hospitalizations	9.11 (12.57)	14.22 (14.03)	10.56 (9.64)	9.71 (9.21)	0.47
SUMD total score	7.56 (2.60)	10.00 (2.45)	7.83 (1.58)	7.10 (2.66)	3.86a*	1,3,4 < 2

^aThis analysis was conducted as an ANCOVA controlling for age.

^*p < 0.05

^**p < 0.001.

Interviewer: do you think you have a mental illness?

Participant: Yeah, I uh, I haven't heard any voices lately, and I have hallucinations; I see little things running around on the floor. I mean, I knew it wasn't right. (Laughs). Sometimes I can hear breathing other than my own. I don't know what that is, call it a monster for lack of anything else.

The participant in the above dialogue readily admitted to having a mental illness and even differentially attributed symptoms to it, but the participant's terse reference to the label of the illness was obviously ambivalent, and describes being haunted by something evil that lacks a name.

By contrast, the individuals assigned to the second profile did not believe that they had an illness. However, they seemed to be searching for a label or explanation for their condition. We therefore called this group as “Rejects illness/searches for name.” (n = 9). The following is an excerpt of a dialogue with an individual with this profile:

Interviewer: Do you think you have a mental illness?

Subject: No.

Interviewer: Have you ever had a mental illness?

Subject: No.

Interviewer: Have you ever experienced anything, or has anybody talked with you about having one?

Subject: Yes. When I first . . . I used to hear voices, and I still hear them sometimes, but I don't pay any attention to them . . . I believe that I'm controlled by a demon. Otherwise I wouldn't have done all these bad things.

The participant in the above exchange clearly did not believe he had a mental illness. It is evident though, that he had been searching for a way of explaining and naming his experiencing the symptoms.

In contrast to the first 2 groups, members of the third group believed they had a mental illness and accepted the diagnostic label but rejected having had symptoms and were not searching for a name or explanation. We therefore named this group as “Passive insight of the illness and label” (n = 18). These individuals used the traditional psychiatric diagnostic label to denote their illness but were unable to identify the symptoms associated with the diagnostic label and expressed no interest in naming or understanding the illness. They passively repeated information they had learned about their illness, but did not show any signs of understanding it or being actively involved in applying what they had learned within the context of their lives.

Interviewer: Do you think you have a mental illness?

Subject: Yes, I know I have an illness.

Interviewer: What do you think it is?

Subject: Manic depressive.

Interviewer: What does that mean, what are the symptoms?

Subject: I don't know. They've told me many times.

Finally, individuals assigned to the fourth group (n=29) believed they had an illness, identified its symptoms, and accepted the label. However, they also remained curious about it and active in understanding it. The illness narratives assigned to this profile seem to express an open ended and exploratory experience of the mental illness that seems to match the concept of integration (Levy et al., 1975; McGlashan, 1975; Thompson, McGorry, and Harrigan, 2002), which connotes flexible attempt to incorporate the illness into a comprehensive experience of life and was termed. Therefore, this profile of narrative insight was termed integrative insight. Following is a brief section from a narrative of a person with this narrative profile.

It's labeled schizoaffective, and what happens is I have audio and visual hallucinations. Um, I draw erroneous conclusions to some of these audio and visual hallucinations that I have. I've even felt like I was being beaten when there was no one there to do the beating. Uh . . . I have had strange thoughts about being attacked by spirits from the spirit world, and I've had a difficult time living with the thoughts and the experiences that I have had.

This participant recognized his symptoms and attributed them to a diagnostic label with which he seemed familiar and believed to be appropriate. His description of the symptoms was also clear. He realized that the tension between the attribution of his symptoms to his illness and the vividness of these symptoms were a source of distress with which he needed to cope.

Quantitative Analysis 2: Relating the Profiles of Narrative Insight to Demographic Variables and to Traditional Insight

Analyses of variance were conducted to determine whether the groups produced by the cluster differed in age, education, or history of hospitalization. As shown in Table 3, these analyses showed that the second group, which rejected the illness and searched for a label, was significantly older than the other groups. No other differences were found. Chi square analyses similarly did not show that there were higher proportions of individuals with schizoaffective disorder in any of the groups, produced by the cluster analysis.

Finally, an analysis of covariance controlling for age was used to compare the profiles on summary scores for the SUMD. As shown in Table 3, the individuals assigned to the “rejects illness/searches for name” cluster received significantly lower ratings of awareness of illness than any of those assigned to other profiles. No other differences were found.

DISCUSSION

Narratives of mental illness may be understood as the stories individuals tell themselves and others to account for their illness. These stories may vary from one another in a number of different ways. The results of this study's qualitative analysis of such stories and then cluster analysis revealed 4 different profiles of the ways in which individuals have insight into their illness: integrative insight (cluster 4 in Table 3) involves believing that one has an illness with identifiable symptoms and a label, which one actively tries to understand. A second profile, which we labeled as “accepts illness/rejects label” (cluster 1 in Table 3), involves acceptance of a cognitive or emotionally based problem with symptoms, but rejection of a general label. A third profile, which we labeled as “passive insight” (cluster 3 in Table 3), involves accepting a mental illness and its official label with minimal curiosity, involvement, or understanding. A fourth profile, which we labeled as “rejects illness” (cluster 2 in Table 3), involves both the denial of symptoms and label of mental illness. Importantly, only this group had significantly lower scores on the SUMD, a traditional measure of insight.

The above findings suggest that measures of traditional insight may fail to distinguish between individuals who accept their label but have only a shallow understanding of its implications and those who accept the label and are able to integrate it into their understanding of themselves. This interpretation is consistent with equivocal findings regarding the association between insight and the personal and social consequences of the mental illness (Lincoln et al., 2007). It is also consistent with literature indicating that people must make sense of their problems in terms of their life stories (Lysaker and Lysaker 2002; Lysaker et al., 2003).

This study constitutes an initial effort at identifying and exploring narrative insight. Nevertheless, its results raise several important related clinical and research issues. One issue concerns the importance of distinguishing between compliance with the normative demands of the professional and clinical situation and genuine awareness of the illness. The differences between the passive insight and integrative narrative profiles point to the possibility that the client's desire to comply with the clinician's authoritative pronouncements regarding their experience of mental illness may be mistaken for genuine awareness of the illness. The passive insight and integrative insight profiles are similar in that they are characterized by high levels of belief in the illness and acceptance of the illness label. However, they differ with regard to the narrative themes of belief in symptoms and of active engagement in searching for an explanation or name for one's condition. Although the narratives of individuals with a passive insight profile exhibited less belief in symptoms and almost no interest in searching for a name or explanation for their illness, the narratives of individuals with an integrative insight profile expressed higher levels of both these narrative themes. Possibly, because of compliance, individuals with a passive insight profile may feel that it is not appropriate to continue to explore the sources of their symptoms and illness, whereas those with an integrative insight profile may experience the acceptance of their illness as a preliminary step toward integrating it as part of their life history.

This aspect of narrative insight incorporates the distinction between pliant and tracking rule-governed behavior as described by Hayes et al. (1999). According to this distinction, pliant rule-governed behavior is supported by an individual's history of the general socially mediated consequences of following rules, whereas tracking is rule-governed behavior undertaken because of a history of experiencing the correspondence between a rule and the natural social or nonsocial contingencies implied by that rule. Clients who are pliant are more concerned with pleasing the professional care provider than they are in exploring the experience of their interaction with their environment outside of therapy or the institution. Therefore, narratives that reflect compliance can be counterproductive and an indication of resistance (Hayes et al., 1999).

A second issue is concerned with whether the client continues to search for an account of the experience of the illness despite a lack of awareness or rejection of prevalent psychiatric accounts of the mental illness (Kravetz and Roe, 2007). Thus, individuals with profile 1 exhibit high levels of narrative themes that are highly congruent with the classic definition of psychiatric insight, believing in both the illness and symptoms, whereas those with profile 2 exhibit low levels of these themes. In addition, the narratives of individuals assigned to profile 1 are characterized by low levels of acceptance of the label and a search for a label or explanation for the mental illness, whereas the narratives of individuals assigned profile 2 evince moderate levels of acceptance of the mental illness label and of the search for a label or explanation for it. Despite relative acceptance of the illness label by the latter, they seem to be searching for a more adequate explanation for their condition. Individuals who displayed this relatively anomalous profile received a statistically significant high rating of lack of awareness on SUMD (higher scores on SUMD are indicative of lower levels of awareness of mental illness). This finding suggests that the professional who administered the SUMD interview was sensitive to these individuals’ lack of awareness of their illness. But, SUMD is not designed to detect the client's search for a label or explanation of the illness, which is conveyed by the narrative. Additional research is required to assess the reliability and implications of the present findings concerning the profiles of narrative insight.

Both presumed facets of narrative insight described above are examples of how the story one tells about his or her mental illness emerges within an interpersonal dynamic context. Consequently, the meaning and development of this story may be affected by the nature of the audience to whom the story is addressed and by one's relationship with this audience (Lysaker et al., 2006). As noted above, findings about the relations between insight into mental illness and the personal and social consequences of SMI are inconsistent. Roe and Kravetz (2003) argued that a narrative approach to insight could contribute to resolving these inconsistencies. Considering the strong influence of context in assessing insight, which has been supported by recent research (Jovanovski et al., 2007; Young et al., 2003), narrative insight may provide a method for increasing the usefulness of insight.

From a methodological perspective, the findings of the current study draw attention to how conceptualizing insight as the degree to which an individual agrees with a description of a set of objective phenomena runs the risk of overlooking potentially important narrative accounts. By contrast, narrative insight may provide a more valid account of insight as a story one tells about one's illness, although it may be more challenging to measure it reliably. But, as the interpretation of the mental illness narrative analyzed in this study implies, the meaningfulness of measures of narrative insight should be evaluated by their relation to the goals that motivated the narration and to the extent that these goals promote recovery.

The findings of this study, consistent with those of other studies (Greenfeld et al., 1989; Kleinman, 1988; Lobban et al., 2003), reveal that individuals hold a range of beliefs and attitudes about their illness. The clinical implication is that any intervention directed at exploring or altering perceptions of illness must take into account the existing ones. Reliable profiles of insight may provide the necessary first step in this direction. Shifting away from a singular descriptive approach to one of narrative process requires not only scientifically based information about the illness but also helping people make sense of their anomalous experiences in a way that is consistent with the context of their lives and social environments. Similarly, facets of the profiles can help better tailor interventions according to their focus on motivation, (Kemp et al., 1995), psychoeducation (Azrin and Teichner, 1998), self-management (Mueser et al., 2006), cognitive behavioral therapy (Kingdon et al., 2006), social cognition (Penn et al., 2007), or narrative (Lysaker et al., 2007). Beyond actual technique, and perhaps most important, because of its inherently constructivist foundation, a narrative insight approach offers individuals the opportunity and authority to actively negotiate the meaning of their disorder, construct their own narrative of it in which they are the protagonists, and be the authors of the story of their illness (Roe and Davidson, 2005). Accordingly, cluster analysis and other forms of profile analysis can be used in identifying meaningful patterns of subjective experiences of other mental illnesses that raise the possibility that these experiences may be misinterpreted as lack of insight.

The present study has several important limitations. Because of its cross-sectional nature, no conclusions can be drawn about causality, and we cannot rule out alternative explanations of the findings. Generalization of findings is also limited because of sample composition. Participants were mostly men in their 40s, all of whom were involved in treatment. It may be that different themes and different relationships exist between narrative themes of women with schizophrenia, individuals in an early stage of illness, and among those who decline treatment. Thus, more research is necessary, with broader samples and based on data collected at multiple time points.