Primary Care Use and Clinical Outcomes Among Patients Initiating Hemodialysis

Abstract

Rationale & Objective

The association between primary care use within 90 days after hemodialysis (HD) initiation and health outcomes is unclear, so evaluating these associations may have implications for population health management. This study examined the association of primary care use after in-center HD initiation and clinical outcomes.

Study Design

Retrospective cohort study.

Setting & Participants

United States Renal Data System registry of incident patients with end-stage kidney disease linked with Medicare claims data between January 2012 and December 2020.

Exposure

Primary care use was defined as ≥1 outpatient visit to a provider with the specialty of family, general, geriatric, or internal medicine within 90 days of HD initiation.

Outcome

The primary study outcome was all-cause mortality, and secondary outcomes were cardiovascular mortality, infectious disease mortality, index hospitalization, emergency department/urgent care, kidney transplantation, intensive care unit care, and a subsequent primary care visit.

Analytical Approach

Inverse probability of treatment weighting to balance patients with primary care use with those without primary care use before fitting an unadjusted Cox proportional hazards regression model and cause-specific hazards models when competing risks were treated as censoring events.

Results

We identified 58,011 (51%) and 56,653 (49%) Medicare beneficiaries (average age 71, 70% White, and 53% men) with and without primary care use within 90 days after HD initiation, respectively. There were 58,982 deaths during a median (interquartile interval) follow-up of 1.79 (0.69-3.38) years. The hazard ratio for all-cause mortality was 0.91 (95% CI, 0.89-0.92), and there were similar reductions in cardiovascular death (HR, 0.94 [95% CI, 0.91-0.96]), infectious disease death (HR, 0.87 [95% CI, 0.81-0.94]), and index hospitalization (HR, 0.98 [95% CI, 0.97-1.00], P = 0.03). There was no difference in the next emergency department/urgent care visit (HR, 1.00 [95% CI, 0.99-1.02]), but there was an increase in receiving subsequent primary care (HR, 3.72 [95% CI, 3.65-3.79]).

Limitations

Patients had pre-existing Medicare coverage before dialysis initiation, limiting generalizability.

Conclusions

Primary care use within 90 days of HD initiation was low at 51% but was associated with lower mortality and hospitalization. Further studies are needed to assess the mechanisms and barriers of primary care use in this vulnerable population.

Plain-Language Summary

Primary care is associated with improved health outcomes for the general population, but whether this also applies to patients on dialysis is unclear. We used the United States Renal Data System database which tracks all US dialysis patients to compare those who saw a primary care provider within 3 months of starting dialysis with those who did not. We found that about 50% of patients saw a primary care provider after starting dialysis, and those who saw a primary care provider had lower rates of mortality and hospitalization. Future studies should examine how primary care use can improve outcomes and identify barriers limiting primary care use.