Re-evaluation of re-hospitalization and rehabilitation in renal research

A Comparison of US Medicare Expenditures for Hemodialysis and Peritoneal Dialysis

BACKGROUND

Observations that peritoneal dialysis (PD) may be an effective, lower-cost alternative to hemodialysis for the treatment of ESKD have led to policies encouraging PD and subsequent increases in its use in the United States.

METHODS

In a retrospective cohort analysis of Medicare beneficiaries who started dialysis between 2008 and 2015, we ascertained average annual expenditures (for up to 3 years after initiation of dialysis) for patients ≥67 years receiving in-center hemodialysis or PD. We also determined whether differences in Medicare expenditures across dialysis modalities persisted as more patients were placed on PD. We used propensity scores to match 8305 patients initiating PD with 8305 similar patients initiating hemodialysis.

RESULTS

Overall average expenditures were US$108,656 (2017) for hemodialysis and US$91,716 for PD (proportionate difference, 1.11; 95% confidence interval [CI], 1.09 to 1.13). This difference did not change over time (P for time interaction term=0.14). Hemodialysis had higher estimated intravenous (iv) dialysis drug costs (1.69; 95% CI, 1.64 to 1.73), rehabilitation expenditures (1.35; 95% CI, 1.26 to 1.45), and other nondialysis expenditures (1.34; 95% CI, 1.30 to 1.37). Over time, initial differences in total dialysis expenditures disappeared and differences in iv dialysis drug utilization narrowed as nondialysis expenditures diverged. Estimated iv drug costs declined by US$2900 per patient-year in hemodialysis between 2008 and 2014 versus US$900 per patient-year in PD.

CONCLUSIONS

From the perspective of the Medicare program, savings associated with PD in patients ≥67 years have remained unchanged, despite rapid growth in the use of this dialysis modality. Total dialysis expenditures for the two modalities converged over time, whereas nondialysis expenditures diverged.

Challenging Assumptions of Outcomes and Costs Comparing Peritoneal and Hemodialysis

OBJECTIVES

Policy makers have suggested increasing peritoneal dialysis (PD) would improve end-stage kidney disease (ESKD) outcomes and reduce Medicare spending compared with hemodialysis (HD). We compared mortality, hospitalizations, and Medicare spending between PD and HD among uninsured adults with incident ESKD.

METHODS

Using an instrumental variable design, we exploited a natural experiment encouraging PD among the uninsured. Uninsured patients usually receive Medicare at dialysis month 4. For those initiating PD, Medicare covers the first 3 dialysis months, including predialysis services in the calendar month when dialysis started. Starting dialysis later in a calendar month increases predialysis coverage that is essential for PD catheter placements. The policy encourages PD incrementally when ESKD develops later in the month. Dialysis start day appears to be unrelated to patient characteristics and effectively "randomizes patients" to dialysis modality, mitigating selection bias.

RESULTS

Starting dialysis later in the month was associated with an increased PD uptake: every week later in the month was associated with an absolute increase of 0.8% (95% confidence interval [CI] 0.6%-0.9%) at dialysis day 1 and 0.5% (95% CI 0.3%-0.7%) at dialysis month 12. We observed no significant absolute difference between PD and HD for 12-month mortality (-0.9%, 95% CI -3.3% to 0.8%), hospitalizations during months 7 to 12 (-0.05, 95% CI -0.20 to 0.07), and Medicare spending during months 7 to 12 (-$702, 95% CI -$4004 to $2909).

CONCLUSIONS

In an instrumental variable analysis, PD did not result in improved outcomes or lower costs than HD.

Prior Hospitalization Burden and the Relatedness of 30-Day Readmissions in Patients Receiving Hemodialysis

BACKGROUND

Thirty-day readmissions are common in patients receiving hemodialysis and costly to Medicare. Because patients on hemodialysis have a high background hospitalization rate, 30-day readmissions might be less likely related to the index hospitalization than in patients with other conditions.

METHODS

In adults with Medicare receiving hemodialysis in the United States, we used multinomial logistic regression to evaluate whether prior hospitalization burden was associated with increased 30-day readmissions unrelated to index hospitalizations with a discharge date from January 1, 2013 to December 31, 2014. We categorized a hospitalization, 30-day readmission pair as "related" if the principal diagnoses came from the same organ system.

RESULTS

The adjusted probability of unrelated 30-day readmission after any index hospitalization was 19.1% (95% confidence interval [95% CI] 18.9% to 19.3%), 22.6% (95% CI, 22.4% to 22.8%), and 31.2% (95% CI, 30.8% to 31.5%) in patients with 0-1, 2-4, and ≥5 hospitalizations, respectively. Cardiovascular index hospitalizations had the highest adjusted probability of related 30-day readmission: 10.4% (95% CI, 10.2% to 10.7%), 13.6% (95% CI, 13.4% to 13.9%), and 20.8% (95% CI, 20.2% to 21.4%), respectively. Renal index hospitalizations had the lowest adjusted probability of related 30-day readmission: 2.0% (95% CI, 1.8% to 2.3%), 3.9% (95% CI, 3.4% to 4.4%), and 5.1% (95% CI, 4.3% to 5.9%), respectively.

CONCLUSIONS

High prior hospitalization burden increases the likelihood that patients receiving hemodialysis experience a 30-day readmission unrelated to the index hospitalization. Health care payers such as Medicare should consider incorporating clinical relatedness into 30-day readmission quality measures.

Health Insurance in the First 3 Months of Hemodialysis and Early Vascular Access

BACKGROUND AND OBJECTIVES

Patients without Medicare who develop ESKD in the United States become Medicare eligible by their fourth dialysis month. Patients without insurance may experience delays in obtaining arteriovenous fistulas or grafts before obtaining Medicare coverage.

DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS

In this retrospective cohort study, we used a national registry to compare uninsured patients starting in-center hemodialysis with a central venous catheter between 2010 and 2013 with similar patients with Medicare or Medicaid. We evaluated whether insurance status at dialysis start influenced the likelihoods of switching to dialysis through an arteriovenous fistula or graft and hospitalizations involving a vascular access infection. We used multivariable logistic and Cox regression models and transformed odds ratios to relative risks using marginal effects.

RESULTS

Patients with Medicare or Medicaid were more likely to switch to an arteriovenous fistula or graft by their fourth dialysis month versus uninsured patients (Medicare hazard ratio, 1.63; 95% confidence interval, 1.14 to 2.43; Medicaid hazard ratio, 1.23; 95% confidence interval, 1.12 to 1.38). There were no differences in rates of switching to arteriovenous fistulas or grafts after all patients obtained Medicare in their fourth dialysis month (Medicare hazard ratio, 1.17; 95% confidence interval, 0.97 to 1.42; Medicaid hazard ratio, 1.01; 95% confidence interval, 0.96 to 1.06). Patients with Medicare at dialysis start had fewer hospitalizations involving vascular access infection in dialysis months 4-12 (hazard ratio, 0.60; 95% confidence interval, 0.37 to 0.97).

CONCLUSIONS

Insurance-related disparities in the use of arteriovenous fistulas and grafts persist through the fourth month of dialysis, may not fully correct after all patients obtain Medicare coverage, and may lead to more frequent vascular access infections.