Hospital Economics

Einleitung / Introduction

Acute myocardial infarction and stroke make up for around 25% of all death in Germany. However, these diseases are not only deadly but go along with high follow-up costs for those who do not die and therefore for the whole economy. In2015, treatment of stroke and myocardial infarction cost the German economy around 11.5 billion Euro. In this paper, we want to understand the impact of structural quality on result quality and gain insights on regional differences in over- and undersupply of high-quality hospital treatment of the two indications.

Methode / Method

Based on a delphi approach and administrative data, we use a matching algorithm to build a structural indicator for hospital treatment of acute myocardial infarction and stroke. The algorithm examines whether patients were treated at the ideal department given there diagnosis and further characteristics. We then examine if treatment at a not ideal department is linked to poorer outcomes suchs a mortality.

Ergebnisse / Results

With the matching indicator, we show that more than a third of patients experiencing one of these indications are treated in a unfitting hospital department. Applying risk-adjusted regression analysis, we find a significant negative impact of a treatment in a wrong department on in-hospital and one-year mortality. We differentiate between hospitals that are equipped for offering ideal treatment and hospital that do not have the prerequisites to offer FirstBest-Treatment for the indications. Building on that, we can estimate how many unfitting treatments could be avoided within the hospital or by changing to a hospital that is equipped to offer ideal treatment. Consequently we also identify regions that show a undersupply of hospitals that are able to offer high-quality treatment for the indications of heart attack and stroke.

Zusammenfassung / Conclusion

In conclusion, we offer a new approach to measure structural hospital quality and show its linkage to outcome quality. We show that the structural quality of stroke and heart attack treatment has improved but is still far from ideal.

AutorInnen

Anna Werbeck, RWI - Leibniz-Institut für Wirtschaftsforschung

Alexander Haering, RWI - Leibniz-Institut für Wirtschaftsforschung

Christiane Wuckel, RWI - Leibniz-Institut für Wirtschaftsforschung

Einleitung / Introduction

It has been documented that a hospital's volume of a specific procedure is positively correlated to the health outcomes of treated patients (Birkmeyer et al. 2002, Halm et al. 2002 ). As a consequence, many countries have implemented governmental or institutional minimum volume requirements (MVR) for hospitals (Morche et al. 2018). These MVRs apply to different medical diagnoses or procedures and set a respective lower bound which ought to be met by a hospital. One way to institutionalize (enforce) these regulations is to link the reimbursement system to these minimum volume requirements. This may create (unintended) consequences on treatment choices due to changed financial incentives. In this paper, we investigate (i) if and to which extent the introduction of MVRs, that are linked to the reimbursement schedule, leads to changes in treatment choices and (ii) whether MVRs affect patients’ health outcomes. We exploit the introduction of MVRs for complex surgical procedures in Germany that mandated a minimum number of procedures in a calendar year. More specifically, we use individual-level patient billing data from all German hospitals between 2008 to 2018 to analyze how hospitals respond to these MVRs dynamically in a given calendar year and how these affect patients’ health.

Methode / Method

We create a hospital-level panel of procedure counts over the course of a year. Based on the time in a year and the number of completed procedures, we create a “pressure” variable that captures the level of necessity to do more procedures to reach the MVR threshold. We use an adapted „Difference in differences” framework that exploits the fact that pressure equalizes at all hospitals at the end of the year as the count reverts to zero. We use this framework across hospitals – comparing hospitals under high pressure to hospitals under low pressure in a given year – and within hospitals – comparing years with low pressure to years with high pressure for single hospitals. Furthermore, we analyze the years 2011-2014 in which the MVRs for knee replacement were suspended. In addition, we use hip replacement surgery (which carries no MVRs) as a counterfactual to knee replacement to detect MVR-specific trends.

Ergebnisse / Results

Preliminary results show that hospitals under “high pressure” based on their procedure count on October 1st indeed increase the number of procedures significantly before the end of the year. This is the all types of complex surgery that we analyze. Next, we will explore the mechanisms of this by evaluating patient characteristics and health outcomes.

Zusammenfassung / Conclusion

We evaluate how hospitals react to minimum volume requirements (Mindestmengen). We find that the number of procedures increases at the end of a calendar year when a hospital is under pressure to reach a minimum threshold.

AutorInnen

Corinn Hartung, LMU München

Peter Redler, LMU München

Einleitung / Introduction

The effect of hospital ownership (private-for-profit (FP), private-not-for-profit (NFP), public (PUB)) on quality outcomes is inconclusive. Theory suggests that FP hospitals intend to maximise profits. To achieve this, they may aim to provide higher quality to increase their reputation, especially if the quality is observable to the patient. If quality is unobservable, they may have incentives to exploit information asymmetries and provide a lower quality. In this paper, we want to contribute to the literature by studying differences in quality outcomes between three different ownership types. We analyse hip replacement surgery, in (i) emergency care for patients with a femur fracture where quality is potentially less observable and (ii) elective care for patients with hip arthrosis where quality is potentially observable. This allows us to study hospital behaviour and outcomes in two different care and information settings, and compare outcomes in PUB against FP and NFP hospitals.

Methode / Method

We analyse hospital discharge records of all German hospital cases, merged with data on hospital and regional level for the years 2012-2016 for two case studies using an instrumental variable (IV) approach. To overcome endogeneity issues in hospital choice, we use differential distances between the patient residence and the locations of different hospital types to instrument hospital choice. We analyse the outcomes for different post-surgery adverse events (e.g. thrombosis/embolism), surgery-related adverse events, in-hospital mortality as well as hospital reimbursement.

Ergebnisse / Results

After specifying several in- and exclusion criteria, we identified 211,466 patients with the emergency indication hip fracture who underwent total or partial hip replacement surgery. For the elective indication of patients with hip arthrosis who underwent a total hip replacement surgery, we identified 465,714 patients. For the elective case study, the first results of our IV-approach indicate fewer post-surgery adverse events (e.g. thrombosis/embolism) in FP hospitals compared to PUB and NFP hospitals. For the emergency case study, we do not find statistically significant differences in adverse events between different ownership types in our preliminary results, but FP hospitals seem to receive higher reimbursement.

Zusammenfassung / Conclusion

The preliminary results suggest that FP hospitals can provide in some studied dimensions higher quality of care for elective surgery, but differences in quality are insignificant for the emergency interventions compared to PUB hospitals and NFP hospitals. We plan to apply different methods, such as matching/weighting on patient and hospital characteristics, to ensure that our IV-approach accounts properly for differences in patient morbidity or hospital characteristics.

AutorInnen

Stefan Rabbe, Universität Hamburg - Hamburg Center for Health Economics

Ricarda Milstein, Universität Hamburg - Hamburg Center for Health Economics

Jonas Schreyögg, Universität Hamburg - Hamburg Center for Health Economics

Einleitung / Introduction

Many countries with prospective hospital payment have increased the number of reimbursement groups that are defined by specific medical interventions. While prospective payment was often started with diagnoses related case-budgets (so called diagnoses related groups, DRGs), many prospective payment systems today feature case-budgets that differentiate between surgical treatment and the non-surgical alternatives. Although having such more fine-grained group definitions is better for reimbursing the actual expected costs, such changes can also incentivize hospitals to perform more surgeries. In this paper, we analyze whether prospective payment with surgeries as a new grouping criteria incentivizes hospitals to perform more surgeries.

Methode / Method

We use claims data from the universe of German hospitals from 2005-2008 and analyze how higher reimbursement for surgeries in neonatology affected surgery rates among low birth weight newborns. A change in the definition of DRGs increased reimbursement for surgeries by 30 to 80 percent for a specific group of newborns while the majority was not affected. This change in reimbursement dependent on the type of care provided allows us to identify the effects of extra reimbursement on mode of treatment using regression discontinuity and difference-in-differences estimations.

Ergebnisse / Results

Our results suggest an increase in surgeries when their performance gains more financial relevance in reimbursement. We find a substantial increase of 12 to 14 percentage points in the probability for surgical treatment for newborns being affected by the refined reimbursement groups. There is neither evidence that surgeries generally increase over the analyzed period nor for a change in treatment of newborns unaffected by the new definitions of DRGs. We can therefore show that hospitals adjust their treatment behavior according to the height of financial rewards in the German prospective payment scheme.

Zusammenfassung / Conclusion

Our findings are in line with the results of the general literature on hospital responsiveness to financial incentives in reimbursement systems. Whereas often upcoding is the addressed issue, we see changes in treatment choices. As a policy implication, we stress the importance of not only considering the criteria of precise mapping of costs but possible altered treatment behavior when modifying prospective payment schemes. Further research is needed on the consequences of increased surgeries on the patient health outcomes.

AutorInnen

Simon Reif, ZEW Mannheim

Sabrina Schubert, ZEW Mannheim