From Documentation to Delivery: Impact of Goals-of-care Discussions on Patient- and Surrogate-Centered Outcomes

Abstract

Older patients commonly receive goal-discordant care during hospitalization and quality of communication is known to be a modifiable factor. While inpatient goals-of-care discussions (GOCDs) can reduce unwanted healthcare utilization, there is a paucity of literature evaluating their impact on patient- and surrogate-centered communication outcomes. This retrospective cohort study evaluated the association between inpatient electronic health record (EHR)-documented GOCDs and three survey-reported outcomes: 1) goal-concordant care, 2) quality of communication, and 3) shared decision-making. Participants were 55 years or older with at least one chronic, life-limiting illness, or 80 years or older regardless of chronic illness status, and were hospitalized at one of three hospitals in the UW Medicine system. For participants lacking decisional capacity during hospitalization, a family or friend involved in their medical care could participate as their surrogate decision-maker and survey respondent. EHR-documented GOCDs within 30 days of randomization were screened using natural language processing methods and confirmed through manual review. Goal-concordant care at 4 to 6 weeks after randomization was the primary outcome measured using patient or surrogate surveys distributed at this time point. Goal-concordant care, quality of communication and shared decision-making at 3 to 5 days after randomization were secondary outcomes measured from surveys distributed at this earlier time point. Linear regression with robust standard errors was used to evaluate the association between the occurrence of an EHR-documented GOCD and these three survey-reported outcomes. Among 617 participants enrolled in the parent trial, 403 had complete exposure and primary outcome data at 4 to 6 weeks and were included in this study. Ninety (22%) participants had an EHR-documented GOCD within 30 days of randomization. The proportion of survey-reported goal-concordant care at 4 to 6 weeks after randomization was higher among participants with an EHR-documented GOCD compared to those without (63.3% vs. 46%; risk difference 17.3%, 95%CI 5.7 – 29%, p = 0.003). Results were similar after adjustment for age, gender, minoritized race/ethnicity, marital/partnership status, education level, severity of acute illness, burden of chronic illness, dementia, life-limiting cancer, and hospitalization site. The association was stronger among surrogate respondents compared to patient respondents. Mean survey-reported quality of communication at 3 to 5 days after randomization was 5.35 (95%CI 1.45 – 9.25, p = 0.007) points higher in participants with an EHR-documented GOCD discussion compared to those without. There was no association between EHR-documented GOCD discussions and reported shared decision-making or goal-concordant care at 3 to 5 days. Survey-reported quality of communication at 3 to 5 days was associated with a higher proportion of goal-concordant care at 4 to 6 weeks. In summary, this study demonstrated an association between the occurrence of an inpatient EHR-documented GOCD and higher survey-reported goal-concordant care and quality of communication in patients with chronic, life-limiting illness; higher quality of communication at 3 to 5 days was also associated with a greater proportion of goal-concordant care at 4 to 6 weeks. Our findings provide important context for future studies evaluating the key domains of EHR-documented GOCDs to understand the mechanisms by which they influence patient-centered outcomes, and ultimately, develop targeted communication interventions in at-risk populations.