Patient Care Technician Staffing and Outcomes Among US Patients Receiving In-Center Hemodialysis

Key Points Question Is dialysis patient care technician (PCT) staffing associated with outcomes among patients receiving in-center hemodialysis? Findings This cohort study used administrative data for 236 126 US patients initiating hemodialysis and found that the highest vs lowest quartile of facility-level patient-to-PCT ratio was statistically significantly associated with a 7% higher rate of patient mortality, a 5% higher rate of all-cause hospitalization (specifically, an 8% higher rate of sepsis-related hospitalization and a 15% higher rate of vascular access–related hospitalizations), an 8% lower rate of waitlisting, and a 20% lower rate of transplant. Meaning These findings suggest that hemodialysis treatment in facilities with the highest patient-to-PCT ratios may be associated with worse patient outcomes, and further study of the association of PCT staffing with patient safety and quality of US in-center hemodialysis care is warranted.

Using Cox models, without accounting for facility-level shared frailty (models do not converge).c Fine and Gray method used to account for competing risks, which included: mortality=receipt of a kidney transplant, switch to another dialysis modality, or recovery of renal function; transplantation, waitlisting, all-cause hospitalization, and hospital readmission=death, switch to another dialysis modality, or recovery of renal function; and cause-specific hospitalization=death, switch to another dialysis modality, recovery of renal function, or hospitalization due to any other cause.2).From mixed-effects Poisson models including facility as a random effect.b Because patients had to survive and remain on hemodialysis for 90 days to be included, the results for mortality and transplant are the same as for the primary analysis.

d
Statistically significant estimate at P<0.05.

eTable 6 .
Associations of Patient:PCT Ratios With Different Operationalizations of Outcomesrate ratio.a Fully adjusted models (including demographics, clinical characteristics, and facility characteristics; see Table c Statistically significant estimate at P<0.05.Mandated US Interdisciplinary Hemodialysis Care Team APP, advanced practice provider; LVN/LPN, licensed vocational/practical nurses; PCT, patient care technician; RN, registered nurse.*APPs and LVNs/LPNs are used at some facilities but are not required.© 2024 Plantinga LC et al.JAMA Network Open.Selection of Patient Population -81,551 who did not remain on in-center HD for ≥90 days -1991 without a CMS-2728 form -2699 whose incident facility was missing facility data -9966 missing information on patient:PCT ratio -580 with 0 HD stations (n=270) or patients (n=310) -390 patient:PCT ratio <1 (n=183) or >96 (n=207) © 2024 Plantinga LC et al.JAMA Network Open.
LC et al.JAMA Network Open.

eTable 2 .
Characteristics of US Patients Initiating Hemodialysis 1/1/2016-12/31/2018 and Their Initial Facilities, by Inclusion in vs Exclusion From Analytic Models Associations of Patient:PCT Ratio With Outcomes Adjusting For Additional Potential Confounders: Sensitivity Analyses

a effect estimate [IRR (95% CI)] by quartile of patient:PCT ratio, additionally adjusting for:
Area-level poverty and education defined as percentage of households living below the poverty line and percentage of individuals with college degrees, using 2019 American Community Survey data (linked at the zip code level).Associations of Patient:PCT Ratio With Outcomes Using Different Operationalizations of Patient:PCT Ratio: Sensitivity Analyses Fully adjusted models (including demographics, clinical characteristics, and facility characteristics; see Table2).From mixed-effects Poisson models including facility as a random effect.PCT ratios that are at the 1 st percentile (4.8) or lower or 99 th percentile (31) or higher.c Determined as 3-4 patients per shift for 3-4 shifts as the reference range.Associations of Patient:PCT Ratio With Outcomes Using Time-to-Event Analyses: Sensitivity Analyses IRR, incidence rate ratio.aFully adjusted models (including demographics, clinical characteristics, and facility characteristics; see Table 2).From mixed-effects Poisson models including facility as a random effect.bDefined as no vs.any medical coverage at dialysis start.cd Statistically significant estimate at P<0.05.© 2024 Plantinga LC et al.JAMA Network Open.eTable 4. IRR, incidence rate ratio.© 2024 Plantinga LC et al.JAMA Network Open. a b Excluding patient:d Per additional patient per PCT.e Statistically significant estimate at P<0.05.eTable 5. b

Up to 1 year after Annual Facility Survey Assessment of patient outcomes
eFigure 3. Patient Follow-Up for Time-to-Event Outcomes © 2024 Plantinga LC et al.JAMA Network Open.