Abstract A10: Multimodal training of primary care physicians in cancer risk communication

Abstract

Abstract Background: Limited health literacy (HL) is associated with misunderstandings about cancer susceptibility and low adherence to cancer screening. Training primary care physicians (PCPs) to engage in cancer risk communication with limited HL patients may improve information exchange and thus improve limited HL patient adherence to screening guidelines. Methods: This is a 4-year randomized controlled trial of PCPs who practice in safety-net clinics in New Orleans, LA and their limited HL patients (men [age 50-75]; women [age 40-75]); Rapid Estimate of Adult Literacy in Medicine score < 60; overdue for cancer screening). PCPs in both study arms undergo 3 unannounced standardized patient (SP) encounters (new patients with family history of colon and breast cancer and overdue for screening) at 0-, 6-and 12-months. SPs use checklists to rate PCPs general cancer risk communication (GenRisk) and shared decision making (SDM) about colon cancer screening (SDMcrc) [1 =poor; 5=excellent]. Intervention PCPs undergo academic detailing after SP visit 1 and receive SP verbal feedback after each visit. All physicians receive annual reports of their patient panel's cancer screening status. We examined group differences in SP ratings using Student's t-test and multi-linear linear regression adjusting for PCP characteristics (practice organization, specialty, age, PCP-SP race/gender concordance). Results: 18 PCPs enrolled in the study. Of 746 clinic patients screened thus far, 133 met eligibility criteria (target sample size: 180). Most patients are African American (91 %), female (80%), and overdue for mammography (51.9%) or colorectal cancer screening (67.5%). Baseline assessments show PCPs self-rated “very good” their proficiency in discussing cancer risks (mean [SD]: 4.1[0.7]) and eliciting patient preference for treatment (3.9[0.3]) and decision-making (4.0[1.0]). Clinic patients rated “good” doctor facilitation of involvement in care (4.2[1.2]) and information exchange (3.5[1.0]) but rated participation in decision-making “poor” (1.5[1.3]). In unadjusted analysis, there was no difference in SP ratings of control vs. intervention PCPs’ communication at baseline (GenRisk: 2.7[1.2] vs. 3.3[1.1]; SDMcrc: 2.5[1.2] vs. 3.0[1.1], all p >0.05); however, there were significant group differences in SP ratings at 6-months (GenRisk: 3.1[1.3] vs. 4.1[1.1]; SDMcrc: 2.7[1.1] vs. 3.9[1.0]) which were sustained at 12-months (GenRisk: 2.3[0.8] vs. 4.1[0.7]; SDMcrc: 2.1[0.7] vs. 3.9[0.8], all p<0.05). Among PCP characteristics, only practice organization predicted SP ratings at 12 months. Nonetheless, group differences in GenRisk and SDMcrc remained at 12-months after adjusting for practice organization. Conclusions: Training in cancer risk communication and SDM improves SP ratings of PCP communication behaviors. Practice organizations may influence communication styles. Next steps include using generalized estimating equations to examine whether physician behavior and practice location are associated with changes in patient cancer screening behaviors. Citation Information: Cancer Epidemiol Biomarkers Prev 2010;19(10 Suppl):A10.