The Impact of Pre-visit Contextual Data Collection on Patient Activation: Results from a Randomized Control Trial

Abstract

Background The majority of health indicators are outside of the healthcare system, and current electronic health records (EHR) do not capture those indicators. There is a compelling opportunity to test consumer informatics tools that integrate patient's life circumstances, goals, supports, risks, and care preferences into their EHR for point-of-care discussions. Purpose To determine whether the use of a patient-generated contextual data (PCD) tool designed to enhance the capture and sharing of PCD influenced patient activation. Design, Setting, and Participants A two-armed, non-blind, randomized control trial was conducted between May 2019, and October 2019 at two urban, academically affiliated primary care clinics. 301 patients were enrolled, randomized with stratification by race to study arms. Nearly equal percentages of control and intervention group participants (60.5% vs. 62.4%) and two-thirds of White vs one-third of Black participants completed both assessments (67% vs. 33%). Main outcomes and Measures The main outcome was the pre-/post-visit change in Patient Activation Measure (PAM) score, evaluated using intention-to-treat principles. Analysis was also conducted to determine if patient factors mediated racial differences in baseline PAM scores. Results Using intention-to-treat analysis, there were no significant differences in pre-/post-visit change in PAM scores by arm (p=.079). When allowing for an interaction between race and treatment arm, all interaction terms were not significant (p>0.05). Mediation analysis results indicate income (p=0.025) and difficulty paying monthly bills (p=0.04) when treated as continuous variables, mediated the relationship between race and baseline PAM score. Conclusions and Relevance The findings indicate the PCD Tool did not affect patient activation. Socioeconomic status (SES) mediated baseline racial differences in the PAM score. These data indicate that further study of the relationships among SES and patient activation are needed. The findings also indicate that interventions targeting patient activation need to account for and be sensitive to patient's SES.