This is a three-arm, multi-site, cluster-randomized controlled trial, comparing three approaches to improve communication between patients and their doctors. Three health systems will participate in the study: UCSD Health, Reliant Medical Group, and Sutter Health.
The study will be carried out in 2 phases.
Phase 1: Phase 1 includes intervention and survey development work to take place prior to the RCT.
OPEN High Touch intervention - The High Touch intervention will be modeled after the Open Communication intervention developed in the pilot (ClinicalTrials.gov Identifier NCT02522286) which contained three components: (a) a one question pre-visit survey delivered through the patient portal of the EHR, asking patients what they most want to discuss with their physician in the upcoming visit; (b) an animated video for patients providing coaching on how to best prepare for their upcoming visits and get the most from the visits; and (c) Standardized Patient Instructor (SPI) providing communication coaching for physicians on how to incorporate what matter most to patients in the visit, with empathy, and clarity.
OPEN High Tech intervention - For the High Tech arm, the patient components of the intervention will be identical to the patient components of the High Touch arm (i.e., the pre-visit survey and patient coaching video). The difference will be in the PCP training: we will replace the in-person SPI with a mobile app with embedded audio and video vignettes demonstrating the communication challenges (e.g., patient with a big list of issues, patients who resist physician recommendations, and patients who disagree with physician) and recommended strategies. A mobile app offers several advantages, including being accessible at a convenient time for busy providers, being easily disseminated, and easily updated. The app will be interactive, posing questions to learners in association with video vignettes and asking learners to answer how they would handle the situation. We will start with the idea of building a set of short mobile modules that mirror the High Touch approach, honing skills on acknowledging patient's agenda, negotiate a joint agenda, invite patient to teachback and incorporate it in the After Visit Summary in the EHR.
ASK intervention - The ASK intervention is intended to activate patients by encouraging them to ask three questions during their primary care visit: (1) What are my options? (2) What are the possible benefits and risks of each option? (3) How likely are each of the benefits and risks to happen to me? These questions will be printed on posters and placed in exam rooms used by providers in clinics randomized to the ASK arm of the trial.
Phase 2: Phase 2 covers the trial recruitment, and three waves of data collection. Prior to the start of the RCT we will collect baseline (T0) data to allow measurement of primary care provider (PCP) performance prior to the trial. Patients participating in the T0 phase will provide only post-visit ratings of their encounters; we will not collect other outcome data or clinical indicators for these patients. For all patients in the intervention phase, we will be collecting information at two time points: 1) immediately post-encounter (T1); and 2) three months post-encounter (T2). We will further sample the top 5% high users of services after the intervention and review their medical records including the indexed visit and subsequent services that had occurred within four weeks after the indexed visit. The chart review will enable us to decipher the reasons for high volume of services after the indexed visit. We will also use existing patient survey data (Press-Ganey and/or NRC) to supplement our data sources.
Analytical approach: We will apply intention to treat analysis, providing descriptive statistics on relevant outcome measures across patients, PCPs, and clinics according to their randomized assignments in the RCT. We will then use generalized estimating equations logistic regression analysis to examine the impact of the assignment into the intervention arms and the outcomes, while controlling for patient and physician covariates.
