Grand Rounds August 25, 2023: Pragmatic Trial of an EHR Application to Display Real-time PRO Data: Successes and Challenges (Gabriela Schmajuk, MD, MS)

Posted on by Elizabeth Mccamic

Speaker

Gabriela Schmajuk, MD, MS
Professor of Medicine
UCSF and the San Francisco VA

Keywords

Patient-Reported Outcomes, Rheumatoid Arthritis, EHR

Key Points

  • Clinicians rely on patient-related outcomes (PROs) to track disease and function over time in patients with rheumatoid arthritis (RA). These outcomes include disease activity, function status, and pain score.
  • These PROs have been integrated into RA guidelines and they are recommended by professional societies. The treatment philosophy for RA is to treat-to-target based on PRO disease activity scores. There are 4 components to the treat-to-target approach: record disease activity using a composite measure every 3 months; specify disease activity target; adjust medications to target; and document shared decision making. However, there is a gap between collecting the PROs and communicating about the PROs with patients.
  • To try to bridge the gap between PRO collection and meaningful use during clinical encounters, the RA PRO dashboard was developed. The team collected focus group feedback from patients and clinicians and worked with IT professionals to pull data directly from the electronic health record and design an interface that could be used by all patients.
  • The RA PRO dashboard was tested as part of a stepped wedge, cluster randomized trial at the clinician level. The hypothesis was that the RA PRO dashboard used during clinical encounters will increase patient engagement, foster shared decision making, and improve health outcomes for patients with RA.
  • There were 4 clusters with 4-6 physicians each (a mix of attendings and fellows). There were 1:1 training sessions, regular reminders on how to use the dashboard, clinical research coordinators who collected patient information, conferences for sharing successes and challenges, training for medical assistants on collecting PROs, and information sheet for patients. 552 unique patient participated in the trial.
  • In the 18 months of the trial, there were not see changes in the quantitative outcomes that were measured. Patients completed a dashboard survey, which gave the study a lot of qualitative measures. Despite the negative quantitative outcomes, 80% of patients said they would like to continue to use the dashboard, the dashboard helped them talk to their clinician about their symptoms and medicines, and it helped them understand more about their arthritis.
  • The trial team also did a qualitative analysis with the clinicians and analyzed according to the technology acceptance model. In general, the clinicians perceived a lot of usefulness of the dashboard though they were anxious about discussing the PROs when the disease score came up high and it might not have to do with RA. They found the dashboard easy to use and discussed the lack of patient data that made it less useful.
  • There were technical and non-technical challenges of maintaining the intervention. There was a big gap between when the dashboard was built and launched. There are new medications now that were not showing up that required regular updates to the dashboard. There were also challenges with many software and security updates that impacted to the dashboard. There were also major changes to the clinic workflows during the COVID-19 pandemic and issues with data completeness with clinic turnover.

 

Discussion Themes

-Do you think you would have gotten increased adoption, less technical difficulties if you had built the dashboard within EPIC instead of as a Sidecar app. Building in EPIC as we did several years ago, we would not have been able to get the customization that we wanted for the dashboard.

-Was patient recall about past visits accurate several months later? Could this be a reason for a no effects trial? We struggled with when to collect the surveys. We tried collecting the surveys through a phone call after the visit and found we were not able to reach many of the patients. We tried collecting some of the information at the end of the visit and found people were in a hurry. We had to balance the proportion of people who were going to respond to the survey against the timing of when the survey was delivered.

–For an intervention that did not have demonstrable effect but did have qualitative patient value, how do you think about continuing the dashboard verses de-implementing it or do you think over time it will become more useful in more ways? I am still optimistic that we will see some improvements in the outcomes we specified. I am interested in doing a future survey on some of the outcomes.

Tags

#pctGR, @Collaboratory1