Web-Based Psychoeducational Intervention for Persons With Schizophrenia and Their Supporters: One-Year Outcomes

Persons with schizophrenia or schizoaffective disorder were recruited from community mental health centers and inpatient units. Enrollment criteria for persons with schizophrenia were age 14 or older, diagnosis of schizophrenia or schizoaffective disorder ( DSM-IV criteria), one or more psychiatric hospitalizations or emergency department visits within the previous two years, the ability to speak and read English, living in the community at the time of study entry, and absence of physical limitations that would preclude using a computer. Enrollment criteria for support persons were age 18 or older, ability to speak and read English, and absence of physical limitations that would preclude using a computer. [Additional enrollment details are provided in an appendix available as an online supplement to this article at ps.psychiatryonline.org .] Informed consent was obtained from participants, and the research protocol was approved by the University of Pittsburgh Institutional Review Board.

Thirty-one persons with schizophrenia or schizoaffective disorder and 24 support persons were randomly assigned to the telehealth or usual care condition.

Participants assigned to the telehealth condition received dial-up Internet access and a computer as needed. Participants were in the study for up to one year (see online appendix for additional details).

Each intervention participant was given a unique login name and password to access SOAR. The SOAR software collected information on each participant's usage of the Web site, including the pages visited, time of day the site was visited, time spent on the site, and user identifier. (See the online appendix for greater detail about the content and design of the site.)

Content. Selection of the content domains and their creation are described elsewhere ( 12, 20 ). The SOAR Web site provides the following components: three therapy forums, one for persons with schizophrenia only, one for support persons only, and one for both groups of users; a capability for asking questions of and receiving answers from the project team within 24–48 hours; a library of previously asked and answered questions; a library of educational reading materials; and a list of community activities, events, and resources.

Facilitation of therapy forums. Each forum was led and moderated by a therapist. In each forum the therapist emphasized discussions that focused on problem solving, alleviating stress, and interacting with peers to develop a supportive forum where members could work together to address problems ( 3, 21 ).

Site design. SOAR was specifically designed to be accessible to persons with cognitive impairments, as detailed elsewhere ( 12 ). The design model is based on five key principles: use of explicit links and labels; a flat, one-page-deep hierarchy; a relatively high number of links on navigation pages; a single constant navigational toolbar; and minimal superfluous and distracting content, such as images.

Before we installed computers in participants' homes and provided access to SOAR via a desktop icon, telehealth participants (persons with schizophrenia and their support persons) attended a joint, four-hour psychoeducation survival skills workshop. The course was modeled on a workshop developed by Dr. Anderson and her colleagues ( 3 ) and is described elsewhere ( 20 ).

Trained, nonblinded interviewers collected data from participants at study entry and at three, six, and 12 months postbaseline. (Additional details about assessment tools are provided in the online appendix.)

Scale for the Assessment of Positive Symptoms. The 34-item Scale for the Assessment of Positive Symptoms was used at all time points to assess for the presence and severity of positive symptoms associated with schizophrenia ( 22 ).

Knowledge About Schizophrenia Instrument. At baseline and at six months, the Knowledge About Schizophrenia Instrument was administered to assess participants' clinical knowledge about the illness ( 23 ).

Usage of the Web site. Server logs from October 2003 to April 2005 were analyzed to determine usage patterns and number of page views or hits. The time between accessing one page and viewing another was considered to be the amount of time that a page was available for viewing. For usage analyses, online time spent on educational activities included viewing reading materials, viewing the library of questions and answers, and submitting questions to be answered and viewing responses.

Intention-to-treat analyses were used to investigate the effects of the intervention versus usual care conditions on outcomes. The three-month data collected from one person with schizophrenia was deemed unreliable by the interviewer and was thus excluded from analyses. Mixed-effects random-intercept models characterized the rates of change over the course of the study, after analyses adjusted for baseline age, gender, and positive symptoms. These models used an autoregressive error structure most appropriate for longitudinal data ( 24 ) and used restricted maximum-likelihood estimate model parameters when data were missing ( 25 ).

The significance of treatment effects was examined by calculating two-tailed t tests of the beta weight for the mixed-model treatment × time interaction term, as is customary. Cohen's d was calculated from predicted means based on these models to estimate the magnitude of treatment effects on rates of changes to symptomatology (schizophrenia group) and knowledge (schizophrenia and support persons). In addition, within-group analyses of the associations between Web site usage and baseline symptomatology and knowledge outcomes were conducted within the intervention group using Pearson correlation coefficients to identify the correlates of the intervention. For symptomatology data collected at multiple follow-up periods, correlation coefficients were pooled across time to obtain an average estimate of the relations between symptomatology and amount of SOAR Web site usage. Positive symptom data were log-transformed because of substantial positive skew. No adjustments for multiple comparisons were made in this initial feasibility study of the intervention.

Demographic characteristics of the study sample are presented in Table 1 . Forty-eight percent of the sample with schizophrenia was Caucasian, and the average age in the group was 38. In the group of support persons, 46% were Caucasian, and their average age was 50. One patient (3%) and four family members (17%) dropped out of the study.

Persons with schizophrenia in the telehealth group. The total amount of time spent on the SOAR Web site by those in the schizophrenia group was 43,789 minutes (730 hours), which involved 47,630 page views. Over the course of the study, the average total time spent on the Web site per person was 46 hours, or 2,737±3,692 minutes (median=971, range=162–11,796), with an average of 2,977±4,546 page views (median=1,170, range=189–14,120) per person. Usage of the main components of the Web site is presented in Table 2 .

Support persons in the telehealth group. The total amount of time that support persons spent on the Web site over the period of observation was 10,901 minutes (182 hours) and involved 8,210 page views. The average total time spent on the Web site per person was 14 hours, or 839±1,125 minutes (median=372, range=30–4,021), with an average of 632±729 page views (median=253, range=19–2,498) per person. Support persons' usage of the main components of the Web site is presented in Table 2 .

Although treatment engagement involves a complex set of processes ( 26 ), an established operational definition of successful initial engagement ( 27, 28, 29 ) is return for treatment after the intake session. To explore engagement in the online treatment environment, we considered attendance at the psychoeducation survival skills workshop as the intake session and defined participants as initially engaged if they subsequently participated at least once in one of the therapy forums and in educational activities. All persons with schizophrenia participated in a therapy forum on at least 13 separate visits, and they used educational resources on at least four visits. These results indicate that 100% of the participants with schizophrenia were engaged in their treatment.

Of the support persons, only one did not fully engage (participating once in each forum but not in educational activities). Of the other 12 support persons in the telehealth condition, all participated in a therapy forum on at least 11 visits and viewed articles on at least five visits, indicating that 12 (92%) were successfully engaged in treatment.

Participants in the schizophrenia group had a total of 4,473 individual sessions or visits to the Web site, with a mean of 280±337 sessions per person (median=128, range=28–1,232). The persons with schizophrenia were active in the therapy forums during 3,200 sessions (72% of total sessions), with an average of 200±318 sessions per person (median=82, range=13–1,173). The average number of months in which each person was active in a therapy forum was 11±4 (median=11, range=4–17), where they spent an average of 3.2 hours per month or 193±316 minutes (median=64, range=14–1,193). Overall, these data indicate strong engagement of most participants in this part of the intervention.

Persons with schizophrenia participated in educational activities during 807 visits (18% of total visits), with an average 50±80 visits per participant (median=30, range=4–340). The average number of months in which each person viewed educational materials was 8±4 (median=8, range=3–16), spending an average of 32±26 minutes per month (median=24, range=7–88). These data suggest that online approaches to education, using materials specifically designed and written for this population, can be well received.

Support persons had a total of 972 sessions on the SOAR site, with an average of 75±59 sessions per person (median=57, range=1–207). They participated in a therapy forum during 539 visits (55% of total), with an average of 42±40 visits per person (median=19, range=1–129). The average number of months in which each support person was active in a therapy forum was 9±6 (median=7, range=1–19), and each spent an average of 55±43 minutes on the site per month (median=41, range=9–161). Support persons participated in educational activities during 339 visits (35% of total), averaging 26±29 visits per person (median=189, range=0–92) over an average of 7±6 months (median=6, range=0–18). They spent on average 29±42 minutes (median=17, range=0–160) during an active month. Support persons used the forum and educational components of SOAR frequently and over a sustained period, indicating that family and friends can become engaged by this approach to psychoeducation.

Analyses using mixed-effects random-intercept models indicated that compared with persons with schizophrenia who received usual care, their counterparts in the telehealth group had significant and large reductions ( 30 ) in positive symptoms during the treatment intervention (t=-2.06, df=97, p=.042, d=-.88) ( Figure 1 ). Using the same analytical methods, we found in the telehealth group a significant and large improvement in knowledge about the diagnosis of schizophrenia (t=-2.34, df=24, p=.028, d=.88) ( Figure 1 ). No effects were found on other domains of knowledge about schizophrenia. Analyses of the association between cumulative Web site usage and symptom severity indicated that individuals with more severe positive symptoms tended to spend more time on the SOAR site (r=.65, p=.005) and to access SOAR more frequently (r=.62, p=.009). Cumulative Web site usage was weakly associated with magnitude of positive symptom reduction (based on within-subject reduction in positive symptoms from baseline to 12 months), but the relationship was not statistically significant. Supporters in the telehealth group showed a large improvement in knowledge about prognosis (t=2.32, df=14, p=.036, d=1.94). No significant effects were observed on other domains of schizophrenia-related knowledge.