• Sessions
• Presentations
• Posters
• Rooms

Presentation

Session: Hospital Costs
Room: Meeting Rooms 22+23
Time: Fri 11:45-13:00

Presenter: Tom Stargardt (Helmholtz Zentrum Munich. Institute of Health Economics and Health Care Management)

Abstract

Objective: This paper analyses the relationship between hospital costs and health outcomes for patients with acute myocardial infarction (AMI) that were treated in German hospitals using individual-level data for costs and outcomes.

Methods: We used mortality and readmission following AMI after index hospitalization as health outcome measures. As the German DRG system is very explicit in differentiating reimbursement according to type of AMI treatment, i.e. by using multiple DRGs and supplementary fees, we used provider reimbursement for index hospitalization as a proxy for treatment costs. We also assumed that costs were endogenous to health outcomes, which we confirmed with the Hausman test. Thus, we subsequently used two stage residual inclusion. The method requires the use of instrumental variables, i.e. variables that are highly correlated with the endogenous variable (costs), but not with unobserved determinants of the main outcome variables (time to readmission, time to death). We believe that average hourly wages and price per square meter in the hospital’s county meet these criteria. In the first stage, we estimated a Generalized Linear Mixed model with cost as the dependent variable. We assumed a gamma distribution for the variable “cost”, a normal distribution for differences in average hospitals costs, and employed a log-link function. In the second stage, we estimated two random-intercept-cox-proportional-hazard models that accounted for competing risk between mortality and readmission. To control for patient co-morbidities, we relied on the Ontario Acute Myocardial Infarction Mortality Prediction Rules. Data were collected from the Techniker Krankenkasse, a sickness fund with more than 6 million insured in 2006.

Results: We obtained data from 18,043 patients with a primary diagnosis in the area of ischemic heart disease and a treatment for AMI in 2005 or 2006. Ultimately, 14,085 patients were included into our study. For both equations, the majority of coefficients had the expected signs. We found that costs in thousand Euros were negatively associated with mortality (p=0.0009; OR=0.957; 95%CI [0.932-0.982]), but positively associated with readmissions (p=0.0002, OR=1.072; 95%CI [1.034-1.112]).

Discussion/Conclusions: Our results suggest that there is no clear trade-off between costs and outcomes. A possible explanation for the positive association between costs and readmissions may be found in the nature of readmissions as an outcome measure. In contrast to mortality, a readmission always requires that certain individuals make a decision to readmit a patient. It is conceivable that patients with multiple comorbidities that we could not control for in our model – and for which acute hospital treatment was not appropriate – were less likely to be readmitted. If these patients incurred low costs during their index hospitalization for treatment of AMI (i.e. because interventions were not deemed appropriate due, for example, to a high mortality risk), this might explain the positive coefficient for costs in the readmission equation. The negative association between costs and mortality confirms the often-stated hypothesis that increased resource input for patients should clearly lead to higher outcomes. However, the negative association between costs and mortality also suggests that outcomes should be monitored closely when introducing cost-containment programs.

Key Terms

hospital costs, acute myocardial infarction, readmission, mortality, Germany

Authors:

Tom Stargardt (Helmholtz Zentrum Munich. Institute of Health Economics and Health Care Management) and Jonas Schreyoegg (Munich University. Department for Health Services Management)

Search with Google