HSR&D Citation Abstract
Search | Search by Center | Search by Source | Keywords in Title
Mapping the Evidence: Sex Effects in High-Impact Conditions for Women Veterans - Depression, Diabetes, and Chronic Pain
Duan-Porter W, Goldstein K, McDuffie J, Clowse M, Hughes JM, Klap R, Masilamani V, Allen LaPointe NM, Williams JW. Mapping the Evidence: Sex Effects in High-Impact Conditions for Women Veterans - Depression, Diabetes, and Chronic Pain. Washington (DC): Department of Veterans Affairs (US); 2015 Sep 1. 93 p.
Women are entering the military at unprecedented rates and comprise a rapidly increasing segment of Veterans Health Administration (VHA) enrollees. In response, the VHA Women's Health Service requested an evidence map to (1) identify effective interventions in women, (2) better understand sex differences in intervention effects for high-impact medical conditions, and (3) identify gaps in evidence about the efficacy of interventions in women.
We used a stakeholder-driven approach to identify high-priority conditions and interventions. From an initial list of 36 conditions, we used a forced-rank methodology to identify 3 conditions for evaluation: depressive disorders, type 2 diabetes mellitus, and chronic pain conditions (chronic low back pain [CLBP], chronic knee osteoarthritis [OA], and fibromyalgia [FM]). We evaluated treatments in broad categories, including medications, behavioral interventions, supervised exercise, and quality improvement interventions, along with certain condition-specific interventions. For each condition, we searched MEDLINE and the Cochrane Database of Systematic Reviews (CDSR) to identify relevant systematic reviews published from January 1, 2009, through October 31, 2014. Data abstracted from eligible systematic reviews included study design, outcomes, the number and design of primary studies, proportion of men and women in included studies, and whether sex effects were part of study aims, analysis plan, or results. For studies containing sex-specific results, we also abstracted the method used for evaluating sex effects (eg, meta-regression) and the outcomes that differed due to sex effects. When information on sex effects was absent from eligible reviews, we selected high-priority interventions for further evaluation. For these interventions, we examined the largest recent systematic review to identify primary randomized controlled trials (RCTs) as candidates for review. We examined RCTs that randomized at least 75 patients per treatment arm to determine whether they reported sex effects. We chose this sample size criterion in order to limit evaluation to RCTs that had the potential for adequate statistical power to detect interaction effects (intervention * sex).
A combined search of PubMed and CDSR yielded 2531 unique citations, of which 582 full-text articles were retrieved; 313 systematic reviews were eligible, and 268 were fully abstracted. Of these, 86 addressed interventions for depression, 114 addressed interventions for diabetes, and 68 addressed interventions for 3 types of chronic pain: CLBP (n = 26), FM (n = 34), and knee OA (n = 8). Most reviews limited eligibility to RCTs, and the number of primary studies included in the systematic reviews ranged from 0 to 347. Only half (48%) of the reviews summarized the gender distribution of the populations of the included studies, but when summarized, women were well represented. Sex effects were reported in only 30 of the 313 (10%) eligible reviews: 14 (16%) for depressive disorders, 13 (8%) for diabetes, and 3 (4%) for chronic pain. Individual patient data (IPD) meta-analysis-the analysis method best suited to evaluating sex effects-was rarely used (n = 16 of 268 abstracted reviews, 6%). Overall, we found only a minority of RCTs had sample sizes large enough to examine moderator effects, and only 14% of these (9 of 66) examined interactions between sex and the main comparison, intervention type versus control group. When sex effects analyses were identified, most commonly no effect was found. Those with evidence of a sex effect often showed greater benefit in women (Table), but differential effects were typically small. There were important gaps in evidence on sex effects for multiple interventions in all conditions examined, either because no reviews evaluating sex effects were identified or because no IPD meta-analyses were identified.
There is a large body of evidence for many of the examined interventions, particularly medications, psychotherapy, and exercise. However, systematic reviews and primary RCTs examined sex effects infrequently. When examined, sex effects generally favored greater benefits in women, but the differential effects were small and the analysis approaches were often suboptimal. All RCTs and systematic reviews should report the proportion of men and women enrolled, and sex effects should be examined in adequately powered RCTs or IPD meta-analyses.