Palliative Care Teams Reduce Hospital Mortality and Utilization: Causal Evidence from a Quasi-Experimental, Nationwide Evaluation in Ireland

Research Objective

To evaluate the causal effect of palliative care consultation teams (PCCTs) on in-hospital mortality and health care utilization in public hospitals in Ireland for the period 2005–2018 using a quasi-experimental research design.

Study Design

Setting: Ireland is a country in north-western Europe with a population of approximately 4.6million. The health system has both public and private provision, but the 40 acute public hospitals account for 95% < of inpatient admissions and 90% < of hospital deaths nationally.

Population: All adults admitted to a public hospital with a primary diagnosis of advanced cancer, organ (heart, liver, kidney, lung) failure, or Alzheimer's disease.

Analysis: We applied decomposed difference-in-differences (Goodman-Bacon method) to exploit the implementation of PCCTs by different hospitals at different times across the study period. For each hospital we identified the year that a PCCT was founded, and so for each year allocated each hospital a treatment variable value of 1 (had a PCCT for the entirety of that year) or 0 (did not). The method then calculates a weighted average of all possible two-group/two-period difference-in-difference estimators in the data. We controlled for hospital size, type, location and socioeconomic catchment as additional predictors.

Sources: We collected PCCT data by hand from each individual hospital service. Outcome data and other predictors were routinely collected administrative data from a national database.

Population Studied

All adults admitted to a public hospital with a primary diagnosis of advanced cancer, organ (heart, liver, kidney, lung) failure, or Alzheimer's disease.

Principal Findings

We had sufficient data for 34 hospitals (85% of sites, 90% < relevant admissions to public hospitals in the study period). There were 600,306 eligible admissions in the data, with an overall mortality rate of 6% (N = 38,595) and a mean LOS of 10.0 days. Sixteen hospitals (47%) had a PCCT at the start of the study period and 34 hospitals (100%) had a PCCT at the end of the study period. Implementation of a PCCT was associated with a mean absolute reduction in in-hospital mortality of 0.9% (95% CI: −1.6% to −0.2%), a proportional reduction of 10%. Implementation of a PCCT resulted in an absolute reduction in LOS of 0.86 days (95% CI: −1.3 days to −0.4 days), a proportional reduction of 8.6%. Secondary analysis of these data using validated unit costs for acute inpatient stays in Ireland suggests a mean estimated saving per admission of €1500 (~$1250) per admission.

Conclusions

PCCTs improve the value of hospital care for people admitted with serious medical illness. These results were derived using quasi-experimental methods, an important advance in a field where observational studies are at high risk of selection bias and high-quality trials are rare. The magnitude of effect estimates is lower than reported by most previous studies. We consider the implications for future research.

Implications for Policy or Practice

PCCTs both reduce costs and improve outcomes for people admitted with serious illness. Services should be expanded to meet current and future needs. The optimal level of this expansion is unknown and requires further development of appropriate methods.

Primary Funding Source

Health Research Board (Ireland).