Return to 2001 Abstacts List

131. Can Veterans with Posttraumatic Stress Disorder (PTSD) be Safely asked about their Tramatic Experiences Using Mailed Surverys?

K Halek, Minneapolis VAMC Center for Chronic Disease Outcomes Research; L Fortier, Minneapolis VAMC Center for Chronic Disease Outcomes Research; N O'Brien, Minneapolis VAMC; M Murdoch, Minneapolis VAMC Center for Chronic Disease Outcomes Research

Objectives: Mailed surveys may hold advantages over other methods when asking sensitive questions. However, concerns have been raised that asking people with PTSD about traumatic experiences could precipitate psychiatric decompensation. OBJECTIVES: 1) Assess spontaneous comments made by veterans with PTSD about the effects of a mailed survey on their emotional well being and 2) Assess the effect of the survey on their health care utilization relative to unsurveyed controls.

Methods: 1) Content evaluation of veterans' responses to a mailed survey and 2) Cohort study of health care utilization. Cases were 2,466 men and 2,452 women who applied for PTSD disability benefits between 1994 and 1998 and were surveyed as part another study. Among other things, the survey asked about combat exposure; in-service sexual assault; and post-service traumas, such as serious injuries or criminal victimization. Controls were 2,004 male and 1,859 female veterans who applied for PTSD disability benefits at the same time as cases, but were not surveyed. Cases represented 3% of all males and 57% of all females in the parent population. Controls represented an additional 2.4% of men and the remaining 43% of women in the parent population. Male cases and controls were frequency-matched by degree of service connection (SC). Use of VA outpatient clinics was quantified for 8 weeks prior to ("pre-survey") and after ("post-survey") survey mailings.

Results: Male cases' and controls' mean age was 56; their mean degree of SC was 50%. Female cases' mean age was 42, and their controls' mean age was 43. The mean degree of SC was 48% for female cases and 43% for their controls. 1,327 cases contacted us or made comments about the survey. 650 cases made survey-specific comments, of which 30% were positively valenced and 22%, negatively valenced (e.g., "This survey is 'stupid'!"). Despite high rates of trauma, only 131 veterans (2.7% of all cases) reported being upset emotionally by the survey. Three veterans (0.06%) reported mild adverse events requiring, in 2 instances, one-day increases in medications.

Whether they returned a survey or not, cases used fewer clinic services in the post-survey period than in the pre-survey period (p < .001). Controls also used fewer services (p < .001). Comparing pre- and post-survey periods, cases made 0.54 fewer total clinic stops and 0.24 fewer mental health clinic stops. Controls made 0.52 fewer total clinic stops and 0.16 fewer mental health clinic stops. However, the differences in pre- and post-survey clinic use did not differ significantly across cases and controls (p = .92), even after adjusting for gender, %SC, and age. Findings were not affected when the analysis was limited to the 131 cases who reported emotional upset and their respective controls.

Conclusions: Few veterans reported being adversely affected by this survey. To the extent health care utilization serves as a marker of veterans' psychiatric functioning, the survey did not appear to have caused significant psychiatric decompensation.

Impact: Our findings add to a growing literature suggesting that mailed surveys can be safely used to gather sensitive data in subjects with PTSD.