THE PATHWAYS PROJECT: IMPROVING SUPPORTIVE CARE FOR PATIENTS RECEIVING DIALYSIS

Dialysis patients frequently receive unwanted high intensity end-of-life care. The Pathways Project tested implementation of evidence-based supportive care best practices to identify seriously ill (SI) patients, engage them in goals of care conversations (GOCC), and track outcomes respectful of patient preferences. Teams at 10 dialysis centers participated in a learning collaborative to adopt supportive care best practices. They were trained to screen patients with the surprise question- “Would I be surprised if this patient died the next six months?”- to identify those who are SI;hold GOCC;and document patient preferences. Centers submitted monthly reports to benchmark improvement, including the number of patients screened, percent SI, GOCC, and death. Project processes were interrupted at 9 months by COVID-19, then resumed with adjustments 90 days later. A mortality risk ratio was calculated for the risk of death over the first 9 months for SI versus not SI, and a Cochran-Armitage one-sided test for increasing trend was used to assess potential increases in GOCC. Team members completed a 17-item Kidney Supportive Care Implementation Quotient at baseline and at project completion to assess perceived implementation of each best practice. The average monthly census was 1,529 patients;98.8% were screened, and 18.4% were identified as SI. SI constituted 66% of 114 deaths though only 18.4% of patients. The 9-month mortality risk was 27% for SI versus 3% for not SI (risk ratio: 8.52;95% CI: 5.71 -12.88;p <. 0001). Dialysis center teams implemented site-specific approaches to make GOCC part of usual workflow;centers conducting GOCC within 30 days of hospital discharge increased from 30% to a peak of 80% prior to COVID-19 (p =. 01). Teams reported increased implementation of all 17 best practices (p <. 05) with the largest improvements in screening for SI patients and conducting GOCC (Figure). Through the learning collaborative process, center staff shared successful innovations such as lobby days and chairside GOCC to promote advance care planning. The Pathways intervention empowered dialysis staff to improve supportive care best practices. Future research is needed to determine if the intervention results in outcomes more aligned with patient preferences.

Implementation and effectiveness of a supportive care learning collaborative for hemodialysis

Background: The objective of this study was to determine whether a learning collaborative for hemodialysis providers improved delivery of supportive care best practices. Methods: Ten U.S. hemodialysis centers participated in a hybrid implementationeffectiveness pre-post study targeting seriously ill patients between April 2019 and September 2020. The collaborative educational bundle consisted of learning sessions, communication training and implementation support. The primary outcome was change in proportion of seriously ill patients with complete advance care planning (ACP) documentation. Healthcare utilization was a secondary outcome and implementation was assessed qualitatively. Results: One center dropped out during the COVID-19 pandemic. Among the remaining nine centers, 22.9% (320/1395) of patients were identified as seriously ill in the pre-intervention period and 18.0% (226/1254) were identified in the post-intervention period. From the pre-intervention to post-intervention period, the proportion of patients with complete ACP documentation increased, and hospitalizations and emergency department visits decreased (Table). There was no difference in mortality, palliative dialysis, hospice referral or dialysis discontinuation. Screening for serious illness was widely and sustainably adopted. Goals of care discussions were adopted with variable integration and sustainment. Conclusions: Supportive care best practices were feasible to implement in hemodialysis centers and largely sustained during the COVID-19 pandemic. We observed increased documentation of ACP and lower healthcare utilization after the intervention which could reflect a combination of collaborative and pandemic effects. (Table Presented) .

Understanding dialysis patient COVID-19-related mortality

Background: Reported COVID mortality in dialysis patients is high and ranges from 15-25%. We reviewed data from a prospective 14-month study of seriously ill (SI) dialysis patients pre-COVID (May 2019-January 2020) and during COVID (February 2020-June 2020) to better understand COVID-related mortality in SI and not SI patients. Methods: We recruited 10 dialysis centers (6 in NYC, 3 in Denver, CO, and 1 in Dallas, TX) with 1,507 patients. Dialysis staff screened patients monthly with the surprise question (SQ)-Would I be surprised if this patient died in the next 6 months-and recorded outcomes. Those with a No response were identified as SI. A SQ No response is known to identify older patients with multiple comorbidities and an increased risk of early mortality. In this rolling population, we calculated the monthly mortality risk prior to and during COVID and determined the relative risk of death (RR) for SI compared to not SI during both periods. We also compared the increased mortality risk during COVID between patients dialyzed in NYC vs. Denver and Dallas and used logistic regression to determine whether COVID-19-related mortality differed by geographic region. Results: Over 14 months, dialysis centers screened a monthly average of 1,342/1,507 (89.1%) patients and identified 274 (18.2%) as SI, with more consistent screening pre-COVID than during COVID (98.6% vs. 71.2%). Pre-COVID, the monthly mortality rate for SI patients was 2.8% and for not SI patients 0.4%, (RR 7.02, 95% CI, 4.76-10.44). During COVID, the monthly mortality rate for SI patients increased to 4.8% and for not SI to 1.5% (RR 3.19, 95% CI, 2.28-4.44). The absolute increase in monthly mortality risk from pre-COVID to COVID was greater for SI than for not SI patients, 2.0% vs 1.1%. The excess monthly mortality was higher in NYC (2.3% for SI and 1.2% for not SI) than in Denver and Dallas (1.3% for SI and 0.7% for not SI), but the difference was not significant (p = .12). Conclusions: A No response to the SQ identified SI dialysis patients whose 5-month mortality during COVID increased to 23.9% (annualized rate 57.4%). For not SI, the 5-month mortality rate during COVID increased to 7.5% (annualized rate 18%). These findings underscore the importance of advance care planning not only for SI patients but also for all dialysis patients, who are particularly vulnerable to concurrent infections such as COVID-19.

Results of Intervention to Increase Goals-of-Care Conversations in Dialysis Centers (SCI950)

Objectives: 1. Describe the Pathways Project’s evidence-based change package for kidney supportive care. 2. Describe the Results of implementation of goals of care conversations into dialysis center workflow. Background: Many dialysis patients receive unwanted high intensity end-of-life care. Their families rate the quality of this care lower than families of patients with other chronic diseases. At baseline, seriously ill (SI) patients rated communication with dialysis teams very poorly. Research Objectives: The purpose of this study was to test the feasibility of an evidence-based intervention to identify SI patients, engage them in goals-of-care conversations (GOC), and track outcomes to see if aligned with patient preferences. Methods: Researchers recruited 10 dialysis centers with 1,544 patients. Dialysis staff screened patients monthly with the surprise question—"Would I be surprised if this patient died in the next 6 months?"—to identify SI patients and conduct GOC with them. Recorded patient outcomes included number screened, number SI, GOC, hospitalizations, hospice referral, death, and place of death. An odds ratio of SI versus non-SI dying, and a Cochran-Armitage trend test were used to assess for increasing GOC and deaths at home. The study was interrupted at 9 months due to COVID-19. Results: On average, 98.8% of patients were screened monthly, and 18.4% were identified as SI. Of 114 patients who died, SI constituted 66% of deaths. The mortality for SI was 27% versus 3% for non-SI, and the odds ratio for SI dying was 11.22 (95% CI 7.42 to 16.98, P <.0001). Dialysis interdisciplinary teams implemented site-specific approaches to adding GOC into usual workflow;the number conducting GOC with SI within 30 days of hospital discharge increased from 30% to 80% (P=.02). Patients dying at home increased (32.6% vs 18.8%), but the trend was not yet significant (P=.12). Conclusion: It is feasible and useful to identify SI dialysis patients. After 9 months more dialysis centers were conducting GOC with them. Implications for Research, Policy, or Practice: Future research is needed to determine if the intervention increases outcomes more aligned with patient preferences.

Improving supportive care of seriously ill dialysis patients with goals-of-care conversations

Background: Dialysis patients are frequently known to receive unwanted high intensity end-of-life care. Families rate the quality of this care lower than families of patients with other chronic diseases. The purpose of this study was to test the feasibility of a supportive care intervention-the Pathways Project, an evidence-based change package of best practices-to identify seriously ill patients (SI), engage them in goals of care discussions, and track outcomes for patient goal concordance. Methods: Pathways researchers recruited 10 dialysis centers with 1,546 patients. Dialysis staff screened patients with the surprise question (SQ)-“Would I be surprised if this patient died in the next 6-12 months?”- to identify those who were SI and recorded patient outcomes including the number screened, SI, goals of care conversations, hospitalizations, referred to hospice, death, and place of death. An odds ratio was calculated for the odds of SI dying versus those who were not SI, and one-sided Cochran- Armitage trend tests were used to assess for increasing goals of care conversations and deaths at home. The study was interrupted at 9 months due to COVID-19. Results: On average, 98.8% of patients were screened monthly, and 18.4% were identified as SI. Of 114 patients who died, the SI constituted 66% of deaths though only 18.4% of patients. The mortality for the SI was 27% versus 3% for those who were not, and the odds ratio for SI dying was 11.22 (95% CI 7.42 to 16.98, P < .0001). Dialysis interdisciplinary teams implemented site-specific approaches to adding goals of care conversations into usual workflow;the number conducting conversations with SI within 30 days of hospital discharge increased from 30% to 80% (P=.02). The proportion of the patients who died at home in the last 2 months was higher than baseline (32.6% vs 18.8%), but a trend was not yet evident (P=.12). Conclusions: The Pathways intervention is feasible to implement supportive care best practices into existing workflow of dialysis centers. It takes time for teams to get comfortable with new processes and communication approaches;after 9 months more centers were conducting goals of care conversations and more patients were dying at home. Future research is needed to determine if the Pathways intervention results in outcomes more aligned with patient preferences.