Psychological impact of genetic counseling for familial cancer: a systematic review and meta-analysis.

backgroundIdentification of a genetic basis underlying certain types of cancer has led to an increase in demand for genetic counseling about individual risks of the disease. We conducted a systematic review of the literature to determine the quality and strength of evidence relating to psychological outcomes of genetic counseling for familial cancer.

methodsSix electronic databases were searched to identify controlled trials and prospective studies that examined the effect of genetic counseling on risk perception, knowledge, anxiety, cancer-specific worry, depression, and cancer surveillance. Twenty-one studies from 25 papers met inclusion criteria, including five controlled trials and 16 prospective studies. Analysis of each outcome was stratified by short-term (</=1 month) and long-term (>/=3 months) follow-up. Trial evidence was assessed with standardized differences of the means at follow-up between intervention and comparison groups, and these data were pooled by use of random-effects meta-analysis.

resultsMeta-analysis of controlled trials showed that genetic counseling improved knowledge of cancer genetics (pooled short-term difference=0.70 U, 95% confidence interval (CI)=0.15 to 1.26 U) but did not alter the level of perceived risk (pooled short-term difference=-0.10 U, 95% CI=-0.23-0.04 U). Prospective studies reported improvements in the accuracy of perceived risk. No effect was observed in controlled trials on general anxiety (pooled long-term effect=0.05 U, 95% CI=-0.21-0.31 U) or cancer-specific worry (pooled long-term difference=-0.14 U, 95% CI=-0.35-0.06 U), although several prospective studies demonstrated short-term reductions in these outcomes. Few studies examined cancer surveillance behaviors, and no studies attempted to measure informed choice.

conclusionsGenetic counseling for familial cancer is associated with improvement in knowledge but does not have an adverse effect on affective outcomes. We urge further investigation of these findings through well-designed, well-reported, randomized controlled trials with suitable comparison groups and additional outcome measures [J Natl Cancer Inst 2004; 96:122-33].