This is a summary of a chapter to be published in "Distance Writing and Computer-Assisted Interventions in Psychiatry and Mental Health," edited by Luciano L'Abate. The chapter will be published in Fall of 2000 by Ablex Corporation, Stamford, CT.

To supplement or as alternatives to current f2f psychotherapeutic practices, computer-assisted interventions like workbooks may be used on their own, without any additional intervention, in preventive or para-preventive activities. Proponents of these supplements argue that f2f verbal treatment is inherently inefficient. It dramatically limits the number of respondents (individuals, couples, and families) that can receive treatment from mental health professionals. Inclusion of these supplements and alternatives, hopefully will lead to more cost-effective treatments. Through Distance Writing (DW) and Computer-Assisted Interventions (CAI), it should be possible to treat greater numbers of respondents.

A critical assumption of this argument, however, is whether DW/CAI, and workbooks are effective. Will DW/ CAI or workbooks as supplements to, or replacements for, more traditional therapies actually lead to improvement? The goal of this chapter was to examine this assumption within the limited domain of workbooks. It attempted to determine to what degree workbooks lead to improvements in mental and/or physical health.

Writing about emotionally traumatic experiences has been shown to be effective in promoting health and well-being in both healthy and chronically ill individuals (Esterling, L'Abate, Murray, & Pennebaker, Clinical Psychology Review, 1999; Smyth, JCCP, 1998). An annotated bibliography of mental health workbooks (http://www.mentalhealthhelp.com) reported over 60. Despite this prevalence, very few workbooks have established an empirical basis for their preventive or clinical use. Such support would need to include demonstration of both efficacy and efficiency.

Current changes and trends in the health-care system (for both mental and physical health) suggest that treatments which fail to meet an increasingly stringent cost-effectiveness criterion will not be made available, or will not receive reimbursements from insurance and managed care companies. Conversely, those treatments that are demonstrated to be cost-effective will become commonplace in the arsenal of managed-care. Cost-effectiveness, therefore, becomes a critical concept as the combination of both effectiveness and efficiency. Although the nature of workbooks make them inherently low-cost and mass-produced, we remain confronted with the bugaboo of effectiveness. To justify the continued use of workbooks in clinical settings, as well as to promote more stringent empirical examination of this issue, this chapter examine if workbooks show any effectiveness.

Meta-analytical methods for the cumulation and examination of research studies can thus provide an alternative, and in many ways preferable, approach to evaluate a research literature. Meta-analysis is a statistically based approach for generating effect size estimates by aggregating effect size information from all available sources. There are, however, some unique difficulties posed by the literature on workbooks. Vast differences exist between studies in overall design and quality of methods and statistics used. There are, however, a relatively small number of studies that contrast workbooks to a control condition (often a delayed treatment condition).

All studies had to include the use of a workbook in conjunction with, or solely as, therapy (with the explicit aim of producing positive change on some outcome measure). Studies were also required to contain some outcome measure of health, although, for the purposes of this chapter, we defined health in the broadest possible terms. Such health outcomes could be in the domain of either mental or physical health, or more general measures of performance (e.g., student grades, cognitive performance). Studies also had to contain statistical information necessary to calculate an effect size.

Following these criteria, 18 studies were included in this analysis. Magnitude and significance of the overall mean weighted effect size was computed for all outcomes and all studies. One effect size was computed for each study by generating effect sizes for each outcome in the study, and then averaging across all outcomes. These study effect sizes were then cumulated across all studies (corrected for bias) for an overall effect size of workbooks.

The overall effect size across all outcomes and all studies was d = .30 (r = .15; 95% CI +0.21 / +0.39, p<.0001). There was considerable variability in effect sizes between studies, ranging from -.22 to +1.16. Effect size for workbooks alone was d = .36 (r = .15; 95% CI +0.22 / +0.50), whereas the effect size for workbooks used in conjunction with other treatment was d = .26 (r = .13; 95% CI +0.15 / +0.37). The difference between these two effect sizes was not significant (Qb(1) = 1.27, p = .26).

Overall effect size for mental health outcomes was d = .44 (r = .21, 95% CI +0.29 / +0.59). Overal effect size of workbooks on physical health outcomes was d = .25 (r = .12, 95 %CI +0.14 / +0.35). Unlike mental health outcomes, the test for homogeneity of effect sizes was significant (Qw(5) = 10.79, p = .05).

Results of this analysis suggest that workbooks may produce a medium effect size in mental health and a somewhat lower effect size for physical health (although there was great variability for such outcomes). This analysis supports the use of workbooks as additions or as alternatives in preventive and psychotherapeutic practices, making them tools for secondary prevention. Furthermore, as long as they are external to traditional psychotherapeutic practices, they are not yet part of the mental or physical health mainstream. It behooves practitioners using them to include informed consent and pre-post objective evaluation as standard operating procedures in their preventive and clinical practices. Finally, we explicitly encourage future research on the use of workbooks.