Psychoeducation groups for families with a relative with severe mental illness have been associated with a variety of significant gains for the ill family member, including less frequent relapse (1,2,3,4,5,6,7,8,9,10,11,12,13,14), reduced psychiatric symptoms (5,6,7,13), and improved social functioning (3,4,12,13). Research on psychoeducation has also documented gains for family members, including improved social and family adjustments (5,6,7),15,16,17,18) and relief from family burden (19,20). McFarlane and colleagues (14) demonstrated the cost-effectiveness of investment in multifamily psychoeducational groups. A recent federally funded research project, the Schizophrenia Patients Outcomes Research Team, concluded that all families with a relative who has schizophrenia should be offered a psychoeducational intervention (21).
The established benefits of family psychoeducation and family education support further development and dissemination of multifamily psychoeducation models, including curricula that are replicable in a variety of settings. Several such models have been described in books (8,22,23,24,25,26,27,28) and in program manuals (29,30,31,32,33).
Models for psychoeducation generally follow a set curriculum that is usually determined by mental health professionals, although some models have incorporated basic input from families. In addition, existing models generally have no flexibility for differences in group membership. The literature on group dynamics and group therapy suggests, however, that issues of composition and indigenous group leadership strongly affect the process and content of individual groups and hence their outcomes (34,35,36,37).
Therefore, models for psychoeducation that have enough flexibility to take each group's unique composition into consideration should be more effective in achieving outcomes than models lacking this feature. Further, Yalom (38) observed that professionals and their clients may have very different perceptions of what is important in therapy. This observation suggests that professionals may not be well enough informed to modify psychoeducational group curricula to meet a group's particular needs unless they have group members' input about those needs.
North and colleagues (39) have made similar arguments for multifamily psychoeducation groups. Using a model for a multifamily group developed by McFarlane and colleagues (40), they revised the curriculum so that content was selected and ordered based on problems identified by group members themselves. North and colleagues found that family members identified their ill relative's negative symptoms as more problematic than the positive symptoms. Although North and colleagues did not collect data to explore factors that might influence group members' curriculum preferences, they speculated that such factors could include the diagnoses of the ill relative, group members' race or ethnicity, and group members' family roles in relation to the ill family member (for example, parent or sibling).
The purpose of the study reported here was to further explore issues of program content and curriculum for multifamily psychoeducation groups for families with a psychiatrically ill member. Specific aims of the study were to develop a list of categories of program content deemed important by family members, to examine differences in the relative importance of the categories from the perspective of family members, ill family members, and mental health service providers; and to compare the content categories and their relative importance with those of existing curricula for family psychoeducation groups.
Data were gathered at a one-day "family survival" workshop cosponsored by Washington University, the Missouri Institute for Mental Health, and the National Alliance for the Mentally Ill (NAMI) of Metropolitan St. Louis. The workshop was held in October 1996. Participants were recruited through announcements sent to persons on mailing lists developed by the institute and NAMI. The lists included names of persons who had contacted NAMI in the past six months and names of mental health service providers, including psychiatrists, psychiatric nurses, psychologists, social workers, and other clinical staff at agencies and in private practice. The service providers were asked to give information about the workshop to their patients and clients. In addition, all inpatient and outpatient mental health agencies in the St. Louis area were asked to post an announcement for the workshop in their waiting areas.
The workshop was developed to serve families with a relative who had a serious mental illness such as schizophrenia, major depression, or other affective disorders. Workshop registration was not limited to families, and thus some persons with mental illness chose to attend, as did some service providers. The workshop provided descriptive and diagnostic information about schizophrenia and mood disorders; information about the etiology of serious mental illness, including the role of neurochemistry and genetics; and education about medications and community resources. Small-group problem-solving discussions about specific issues of concern were offered.
All family members, including family members with serious mental illness, were asked to fill out a registration form that asked for demographic information. At the start of the program, a form was distributed to participants asking them to identify whether they were a family member, a family member with serious mental illness, or a service provider. Participants were also asked to generate a list of problems they had faced in managing mental illness and rank them by personal importance. (The forms were developed for this workshop and were previously pilot tested with family members at a meeting of NAMI support group leaders.) The forms were collected, and the workshop leaders used the lists of problems to generate topics for discussion groups held later in the day. Examples of discussion topics included ill family members' denial of illness and refusal to take medication and how to get the most out of treatment.
Subsequently the lists of problems were complied into a master list with sections for family members, family members with serious mental illness, and service providers. Based on the aggregated lists, each of the authors independently developed a set of problem categories, then by mutual agreement reconciled them into a single set. Next, the authors assigned each of the items from the aggregated lists of problems to the categories. Reconciliation of differences in assignment of items was accomplished by majority if two authors agreed and by discussion to consensus if all three authors disagreed.
Three numbers were generated for each category: the percentage of times items in the category were mentioned in individual problem lists, the percentage of individual items from problem lists that were associated with the category, and the relative rank of the category, which was derived from the mean rank of the items on the problem lists. Categories whose items were mentioned more than 5 percent of the time or that were associated with more than 5 percent of the items were included in the final analysis.
Differences between responses of the three groups of participants—family members, ill family members, and service providers—were examined using the Mann Whitney U test.
In the final phase of the investigation, content categories from this study were compared with those in curricula from other selected sources. The sources included empirically validated psychoeducation group models, as well as models selected for their national visibility. Of the models included in the comparison, only the model developed by North and colleagues (39) and the curriculum used by NAMI's family psychoeducation program Journey of Hope (32) were generated with the assistance of family members.
The family survival workshop had approximately 150 participants. Registration materials suggested that about 25 or 30 were service providers, and 15 to 20 were persons with serious mental illness.
Demographic surveys were completed by a total of 92 family members and ill family members. Sixty-five percent of the participants who completed the surveys were female, and 99 percent were European American. Eleven percent were age 20 to 29 years; 20 percent, age 30 to 39; 25 percent, age 40 to 49; 20 percent, age 50 to 64; and 20 percent, age 65 or older. Fifteen percent reported having a serious mental illness. Forty-one percent reported being a parent of a person with serious mental illness, 13 percent were siblings, 10 percent were spouses, and 20 percent were sons or daughters.
Problem lists and reliability
Workshop participants completed 86 problem lists. A total of 355 distinct rank-ordered items were mentioned, for an average of 4.1 items per list, with a range from one to ten items. The master list generated from these lists contained 242 distinct items. Overall, family members mentioned a total of 305 items, ill family members mentioned 20 items, and service providers mentioned 39 items.
The authors identified a total of 15 problem categories for analysis: positive symptoms, negative symptoms, stigma, community resources and benefits, denial of illness or noncompliance with treatment, substance use, problem behaviors of the ill family member, education, independence and dependence, medical and psychiatric treatment of the ill family member, relationship with family, harm and violence, interpersonal and social issues of the ill family member, mood symptoms and disorders, and career and employment of the ill family member.
Each of the three raters assigned the problems identified by workshop participants to one or more of the 15 categories. Interrater reliability between all pairs of raters was 90.4 percent, with individual pairs falling between 88.8 percent and 91 percent. The three raters agreed on categorization of 82.6 percent of the items, two agreed on categorization of 14.4 percent, and all three disagreed on categorization of 2.9 percent. Only one consistent relationship of rater disagreement to content was identified. The two raters with social work backgrounds (the first and third authors) included items that were related to psychosocial intervention, such as vocational services and transportation, in the treatment category, while the third rater, a psychiatrist (the second author), used a narrower medical or psychiatric definition of treatment that was limited to medication management, hospitalization, and psychotherapy. About one-third of the paired disagreements were explained by this difference. These specific discrepancies were resolved in favor of the more medical definition of treatment, and the other items were assigned to a different category. The remainder of the disagreements were resolved by majority or through consensus after discussion.
During this procedure, two items that were identified as problems by workshop participants— "everything" and "schizophrenia"—were deleted, because they did not fit within the coding scheme. In addition, 11 items were coded as belonging simultaneously to two separate categories. Therefore, 240 separate items with 364 separate mentions were included in the content analysis.
Of the 15 content categories identified by raters, 11 passed the criteria either of being mentioned in 5 percent of individual problem lists or of being among the 5 percent of items from individual problem lists associated with a category. t1 presents these categories, along with the percentage of times they were mentioned in individual problem lists, the percentage of items associated with the category, and the category's relative rank. The content categories are described below.
Relationship with family.
Problems in this category included family denial of the ill member's illness, family stress and coping, relationship between the ill family member and other family members, and dynamics within the family associated with having an ill relative. This category had the highest frequency of times mentioned and the highest percentage of problem items. Although the category was ranked only fourth of 11, its modal rank—based on the lists of participants who mentioned problems in this category—was first.
Insurance and lack of insurance was the most frequently mentioned concern in this category. Other concerns included entitlements and benefits, financial limitations, and legal issues. A variety of nonmedical treatment issues, such as community services, coordination of care and continuity of care, transportation, and housing, were included in this category. This category had the second highest frequency of times mentioned and problem items. However, the category was deemed of only medium importance by participants, with a rank of seventh among the 11 categories and a modal rank of third.
Denial and noncompliance.
This category, the third most frequently mentioned, consisted of only two items—denial of illness and noncompliance with treatment by the ill family member. Those who mentioned this category rated it with the greatest relative concern, with twice as many individuals ranking items in this category most important, compared with any other single ranking.
Interpersonal and social issues.
This category included problems associated with the ill relative's social isolation and difficulties in friendships and intimate relationships with persons outside the family. Although this category was in the mid-range of frequencies for times mentioned and problem items, it represented more than one-fourth of the responses from ill family members.
Medical and psychiatric treatment.
This category encompassed a variety of items, including medication and physical health concerns, misdiagnosis, and the need for crisis intervention and hospitalization. Although items in this category were mentioned with an average frequency, the category had a modal rank of second and was therefore of great concern for participants who mentioned items in this category.
A variety of negative symptoms were identified as problems, including lethargy, idleness, and excessive sleeping. However, the most frequently mentioned single item was lack of motivation. This category was not ranked as particularly important relative to other categories, nor was it ranked highly within individual problem lists; the modal within-list ranks were third and fourth. However, it is noteworthy that problems in this category were mentioned almost twice as often as problems that were related to positive symptoms.
Multifamily psychoeducation groups have generally focused on schizophrenia. However, that illness was identified as a problem by only one participant. In comparison, more than a dozen participants mentioned problems related to specific mood disorders such as depression, mania, and bipolar disorder. Further, participants who mentioned a problem related to a mood disorder ranked it relatively high, both compared with problems in other categories and within their individual lists. The modal within-list rank was first.
Problem behaviors of the ill family member.
This category included behaviors such as rage and anger, compulsive spending, controlling behavior, manipulative behavior, verbal abuse, self-absorption, and reckless driving. In the discussion group based on this category, the discussion focused on family members' feelings of anger and helplessness in response to the ill family member's problem behaviors.
This category included problems related to gaining information about mental illness. Items mentioned included the need for education of family members, the ill family member, the police, and the general public.
Independence and dependence.
Items in this category focused either on the ill family member's dependence on other family members or on the ill family member's achieving the right amount of independence. One item in this category—achievement of "normality"—was mentioned by more than one respondent. This category was ranked last in importance relative to the other major content categories.
This category generally consisted of problems related to specific symptoms, such as hearing voices, paranoia, delusions, and thought broadcasting. Items in this category were the least frequently mentioned overall, but among ill family members they were the second most frequently mentioned.
Comparisons between participant groups
t1 presents the percentages of time items in each content category were mentioned by family members, ill family members, and service providers. Items associated with the relationship with family category were the most frequently mentioned overall, and the most frequently mentioned among family members. However, this category had no responses from ill family members, a significant difference between this group and family members (z=8.63, p<.01).
Items in the resources and benefits category also were not mentioned by any ill family members but were the primary concern of service providers. One-third of the items mentioned by service providers were associated with this category. They were mentioned by service providers 39 percent of the time, a significantly higher percentage than among family members (z=2.71, p<.05) and ill family members (z=4.80, p<.05). However, this category was the second most frequently mentioned among family members.
Not surprisingly, items grouped in the category of denial and noncompliance were not mentioned by ill family members. They were mentioned statistically more often by both family members (z=5.35, p<.05) and providers (z=2.07, p<.05).
Family members had significantly lower rates of responses in the category of interpersonal and social issues compared with ill family members (z=2.05, p<.05). Service providers did not mention any items in this category—the only category that had no items mentioned by this group. The service providers' response rate for this category was significantly different from that of family members (z=3.58, p<.05). Not surprisingly, the items in the category of problem behaviors of the ill family member were not mentioned by any ill family members, a significant difference compared with the rate of responses in this category by family members (z=4.81, p<.05).
Comparisons with psychoeducation curricula
Of eight books reviewed for their content on family psychoeducation, six were written specifically for families coping with schizophrenia (8,22,23,25,26) and three addressed serious mental illness in general (24,27,28). Of five family psychoeducation programs reviewed in this study, one focused on schizophrenia (33), and four addressed serious mental illness in general (29,30,31,32). t2 summarizes the major content areas of the psychoeducation models described in these sources.
The problem areas identified by families in the psychoeducational group previously developed by North and colleagues (39) were remarkably similar to those identified in the current study, despite the previous group's focus on schizophrenia rather than on serious mental illness in general. Major content categories from the previous group that were not replicated in the current investigation included relapse, recognizing and dealing with feelings, hope and hopelessness, spirituality, and wellness.
Most of the other psychoeducational sources addressed the areas of positive symptoms of psychosis, treatment issues, and relationship with family, which were also identified as problem areas in the current study. Major problem areas that were covered in the other sources but were not included among the major categories identified by the current study participants included severe problem behaviors and crises such as violence, suicide, substance abuse, and risky sexual behaviors; coping skills, including problem solving and stress management; and issues of advocacy and stigma. It is interesting to note that one of the programs designed with input from families—NAMI's Journey of Hope—lacked content on the ill family member's interpersonal and social issues and negative symptoms, issues that were identified as important by families in the current study and in the previous study by North and colleagues (39).
This study empirically derived reliable and valid categories from lists of problems provided by participants in a "family survival" workshop. The frequencies of responses in the various categories provide a view of the relative importance of the concerns expressed by workshop participants. Further, differences in perceptions of relevant problems among family members, ill family members, and service providers yield important implications for program development. Finally, differences between the content categories derived in this study and the content of family psychoeducation curricula developed elsewhere suggest the need for further refinement of curriculum content.
Among the categories specified in this study, issues of most frequent concern to workshop participants were family relationships and community resources and benefits. Based on rank order, however, issues of noncompliance and denial of illness were of paramount importance, and matters of treatment also had a high priority. Consistent with the findings of North and colleagues (39), negative symptoms were of greater concern than positive symptoms. Although most psychoeducation programs have focused on families whose ill relative has schizophrenia, the results of this study suggest that other clinical conditions—particularly mood symptoms and disorders—may be a major source of concern among persons seeking psychoeducational interventions for serious mental illness.
This study found that family members, ill family members, and service providers differed in their desired content focus. Service providers focused on issues of identifying and procuring community resources and benefits. Family members also frequently identified this topic, but they gave it a relatively low priority. Ill family members failed to mention items associated with this category as problems. The three groups of participants also gave different amounts of attention to interpersonal and social issues. This category was of greatest concern for ill family members, of lesser concern for family members, and of no concern for providers. The results suggest that this topic is important enough for persons with serious mental illness and their families to warrant its inclusion in psychoeducation curricula.
Regarding psychotic symptoms, neither positive nor negative symptoms were considered to have great importance. However, the relatively greater focus on positive symptoms by ill family members suggests that attention to this topic may be important in recruiting persons with serious mental illness for psychoeducation groups. The much greater concern with negative symptoms among family members suggests that they would appreciate more information about how to cope with these symptoms.
Programs designed by professionals have often failed to address the problems identified as most important by the participants in this study. In particular, items of concern to families that were often not covered in other curricula included denial and noncompliance, problem behaviors of the ill family member, interpersonal and social issues, issues of dependence and independence, mood disorders, and, especially, negative symptoms. On the other hand, participants in this study did not often mention issues of violence, suicide, and substance abuse, which are clinically important. Professionals who design psychoeducation programs may want to include these topics, which may be considered within the category of problem behaviors.
Family members' ideas about their educational needs may be significantly different from those of professionals who design psychoeducation curricula. The study findings imply that family members would have a significant effect on the content of psychoeducation groups if they were involved in planning their own curricula. Inclusion of families would logically be useful both in development of a general curriculum, as has been done in developing the Journey of Hope, and in choosing specific content for particular groups. Models for psychoeducational groups must be flexible enough to meet the needs of each particular group's members as well as to provide general information relevant to all groups.
Because this study was exploratory, it had significant limitations. The sample was neither random nor representative, and therefore the results may not be generalizable to other situations. However, the sampling bias associated with self-selection of individuals who attended the workshop is inherent in all research on psychoeducational groups. Further bias was associated with the 57 percent rate of completion of the research instrument. Minimally, the findings of this study require replication across different treatment populations.
The sample was also demographically nonrepresentative. In particular, African Americans were underrepresented. Previous research has suggested that the experience of having a family member with serious mental illness may differ among ethnic groups (41). In addition, lack of diagnostic data about the ill family members prevented examination of the effects participants' experience with specific mental illnesses.
Data collection for this study did not allow linkage of demographic data to problem lists. Because of this limitation, we could not explore differences in problems based on the participants' relationship—such as parent or sibling—to the ill family member. The three participant groups—family members, ill family members, and service providers—varied greatly in size; thus the results of the statistical tests should be interpreted with caution.
Further research exploring curriculum content would benefit from comparisons of problems identified by family members with various family roles in relation to the ill family member, family members representing various ethnic groups, and families of persons with a range of diagnoses.