Systematic evidence review and meta-analysis of outcomes associated with cancer genetic counseling

Abstract

Purpose: Genetic counseling (GC) is standard of care in genetic cancer risk assessment (GCRA). A rigorous assessment of the data reported from published studies is crucial to ensure the evidence-based implementation of GC.

Methods: We conducted a systematic review and meta-analysis of 17 patient-reported and health-services-related outcomes associated with pre- and post-test GC in GCRA in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology.

Results: Twenty-five of 5393 screened articles met inclusion criteria. No articles reporting post-test GC outcomes met inclusion criteria. For patient-reported outcomes, pre-test GC significantly decreased worry, increased knowledge, and decreased perceived risk but did not significantly affect patient anxiety, depression, decisional conflict, satisfaction, or intent to pursue genetic testing. For health-services outcomes, pre-test GC increased correct genetic test ordering, reduced inappropriate services, increased spousal support for genetic testing, and expedited care delivery but did not consistently improve cancer prevention behaviors nor lead to accurate risk assessment. The GRADE certainty in the evidence was very low or low. No included studies elucidated GC effect on mortality, cascade testing, cost-effectiveness, care coordination, shared decision making, or patient time burden.

Conclusion: The true impact of GC on relevant outcomes is not known low quality or absent evidence. Although a meta-analysis found that pre-test GC had beneficial effects on knowledge, worry, and risk perception, the certainty of this evidence was low according to GRADE methodology. Further studies are needed to support the evidence-based application of GC in GCRA.

Keywords: Cancer genetics; GRADE; Genetic counseling; Health services; Outcomes; PRISMA.

Figure 1

Identification of studies included in systematic review.

Figure 2. Anxiety.

Forest plot of random effects meta-analysis of studies reporting on anxiety. Outcome measures included the state scale of the State-Trait Anxiety Index,^,, anxiety subscale of the Brief Symptom Inventory, or anxiety subscale of the Hospital Anxiety and Depression Scale.

Figure 3. Depression.

Forest plot of random effects meta-analysis of studies reporting on depression. Outcome measures included the depression subscale of the Brief Symptom Inventory or Hospital Anxiety and Depression Scale (Quinn et al [0-21; 11 considered “abnormal”]).

Figure 4. Worry.

Forest plot of random effects meta-analysis of studies reporting on worry. Outcome measures included the 4-item Cancer Worry Scale,^, 6-item Cancer Worry Scale (Brain et al^, [6-24; cut-off: 12.5]), Impact of Event Scale (Quinn et al [0-75; clinically significant: 40]), or revised Impact of Event Scale (Esplen et al [cut-off: 24]).

Figure 5. Decisional conflict.

Forest plot of random effects meta-analysis of studies reporting on decisional conflict. Outcome measures included 2 different versions of a Decisional Conflict Scale (Matloff et al [1-5; those who make decisions: ≤2] measured at 1 month and 6 months (1 month included in Figure for comparability to Quinn et al); Quinn 2017 [0-100]) measured immediately after intervention, Matloff et al reported data from the Decisional Conflict Score but not in aggregate; only subscale data are reported. Given that subscales in the Decisional Conflict scale are summed and evenly weighted, we plotted the individual subscale data in the Forest Plot to better illustrate the overall effect.

Figure 6. Knowledge.

Forest plot of random effects meta-analysis of studies reporting on knowledge. Outcome measures included a 30-item true/false questionnaire was developed to assess knowledge of menopause, an 11-item true-false scale (Lerman), 10 true/false items (Quinn), the Colorectal Cancer Risk Factor and Screening Knowledge Questionnaire (Esplen), 4 true/false items (Brain), or a questionnaire comprising 14 “true” or “false” statements (Torr). GC outcomes were combined for the telephone and in-person GC intervention (Esplen).

Figure 7. Perceived risk.

Forest plot of random effects meta-analysis of studies reporting on perceived risk. Perceived risk was measured by single-item measure, a 4 items scaled designed by the research team, a 2 item scale,^, and a “counseling feedback questionnaire.” Perceptions of CRC risk were assessed in various formats, including participants’ risk perception on a scale from 0 to 100, 1 item risk perception (scale 0-100) and scale detailing risk to develop breast cancer, heart disease, and osteoporosis sometime during their life on a scale from 0% (not at all likely) to 100% (extremely likely). GC outcomes were combined for the telephone and in-person GC intervention (Esplen). For Bowen et al,^, the comparison and psychosocial groups were lumped and compared with the GC group.

Figure 8. Intent to pursue genetic testing.

Forest plot of random effects meta-analysis of studies reporting on intent/interest to pursue genetic testing. Outcomes were measured with a 4-question scale or a single question (4-point Likert Scale^,-). For Helmes, Culver, and Bowen, both GC subgroups were combined into 1 group. For Bowen et al,^, psychosocial counseling and no counseling were combined into a single subgroup. The exact “n” for each subgroup in Bowen et al for who completed the survey could not be determined from the paper; an overall follow-up rate of 96% was noted, such that we assumed 96% of the original enrollment (Genetic 77, Psychosocial 68, and control/comparison 75) for each study subgroup. For Burke et al, a Likert scale was used to measure intent/interest, although the paper combined responses for a score of 1 or 2 combined or 3 and 4. We assumed equal numbers of responses for 1 and 2, and 3 and 4, respectively, to calculate means and standard deviations. For Lerman, the education and counseling groups were combined for the analysis.