Elsevier

Health Policy

Volume 125, Issue 6, June 2021, Pages 739-750
Health Policy

Assessment of the introduction of DRG-based reimbursement in Switzerland: Evidence on the short-term effects on length of stay compliance in university hospitals

https://doi.org/10.1016/j.healthpol.2021.01.010Get rights and content

Highlights

  • We assess the impact of SwissDRG on hospital performance.

  • Performance is assessed by length of stay relative to a benchmark value.

  • DRG compared to per diems reduced this length of stay compliance in the short-term.

  • Results indicate that hospitals with DRG before adapted more efficiently.

Abstract

The implementation of a nationwide diagnosis-related groups (DRG) reimbursement system in 2012 marked an important step in increasing the transparency and efficiency of hospital services in Switzerland. However, no clear evidence exists to date on the response of hospitals to the introduction of SwissDRG. Using administrative data on inpatient stays in Swiss university hospitals and the length of stay compliance (LOSC) as a measure of hospital performance, we find a significant short-term reduction in LOSC for hospitals that experienced a change from retrospective per diem to prospective DRG reimbursement, compared to hospitals with a prospective payment system already before 2012. LOSC can be interpreted as a performance indicator because it compares the actual length of stay with a benchmark value, taken from the yearly DRG catalogue. The reduction in LOSC implies that hospitals in the treatment group on average had an increase in LOS relative to the benchmark compared to the control hospitals. This may be interpreted as a negative effect of SwissDRG on hospital performance, at least in the short-run, and we provide supporting evidence that hospitals that worked under DRG already before adapted more quickly and efficiently.

Introduction

In 2012, Switzerland implemented a nationwide diagnosis-related groups (DRG) hospital reimbursement system. The main goals of the SwissDRG system were to increase transparency and efficiency in the hospital sector. Policy-makers also aimed at slowing down the growth of health care expenditures, fostering competition between hospitals, increasing the quality of inpatient care, and improving access [1], [2].

The Swiss health system with its decentralized structure offers an interesting setup to study the impacts of a national DRG system. First, Swiss cantons play a multifaceted role, as planners, regulators, payers, and owners of public hospitals, which inevitably creates conflicts of interest and may be at odds with a national payment scheme [3] Second, there is ample evidence for regional disparities in health care provision (e.g., [4]), which, among other things, translates into specific requirements for a payment system. Third, already since 2001 some hospitals used prospective payments based on the all-patient DRG (APDRG) classification. The APDRG system has been identified as a potential driving force behind the reduction in the length of stay (LOS), although limited evidence exists that would establish a causal link (e.g., [5], [6], [7]).

In this study, we explore the coexistence of hospitals with an APDRG payment system and hospitals with a retrospective per diem reimbursement scheme before 2012 to evaluate the introduction of SwissDRG in a before-after comparison using the APDRG hospitals as a comparison group. This quasi-experimental approach allows us to control for factors that potentially confound the regional variation in inpatient services and are difficult to observe otherwise, e.g., differences in patient characteristics or other demand-side factors, the heterogeneity in the medical and non-medical staff, or differences in treatment styles.

We use the length of stay compliance (LOSC) as a main outcome to evaluate the impact of SwissDRG. LOSC relates the actual LOS to a yearly reference value per DRG and has been suggested as a meaningful indicator to compare hospital performance and the efficiency of services [8], [9], [10], [11]. Using administrative data from all Swiss university hospitals and a difference-in-differences methodology, we find a significant reduction in LOSC for hospitals that experienced a change in reimbursement scheme from per diem to DRG, compared to those that had a DRG system in place already before 2012. A reduction in LOSC implies that on average actual LOS values increased relative to the benchmark value, which is indicative of a lower hospital performance or service efficiency in response to SwissDRG. We interpret this as a short-term impact of SwissDRG and provide suggestive evidence for experience effects that may be a potential driving force behind this result, i.e., hospitals that worked with prospective DRG payments already before more quickly and efficiently adapted to the new system.

Our research contributes to an extensive literature on the impacts of DRG (e.g., [12], [13], [14], [15], [16], [17], [18], [19], [20], [21]). This literature studies the consequences of DRG payment systems on hospital operations, with no conclusive evidence yet on the changes in quantity, quality and efficiency of services. In Switzerland, Busato and von Below [5] present a small area analysis and find generally fewer hospitalizations and a higher share of resources in the outpatient sector in areas with APDRG reimbursement. Using a similar approach as in our study, Widmer and Weaver [6] do not find evidence for higher re-hospitalization rates or more early discharges after the introduction of APDRG (see also [22]). Weissenberger et al. [23] combine LOS, patient satisfaction, and quality of life measures and find a positive effect on the quality of hospital services of the APDRG compared to the per diem system (see also [24]). Felder et al. [1] find evidence for a shift from inpatient to outpatient care, but it remains unclear to what extent this result can be attributed to DRG or to general technological progress in the hospital sector. Pellegrini and Roth (2016, 2017) report cost increases from 2011 to 2012 in the inpatient sector, which could be the result of SwissDRG, but they find similar increases in psychiatric hospitalizations, which were not reimbursed by SwissDRG. Moreover, when looking over a longer time horizon, they present descriptive evidence that the cost increases in inpatient acute care after 2012 were mainly driven by a rather constant relative increase in case numbers [25], [26]. Widmer et al. [27] do not find much evidence for an expansion in the quantity of hospital services after the introduction of SwissDRG, but their results indicate that LOS in hospitals reduced due to shifts between the acute and post-acute phase. With our research we aim at supplementing the existing evidence on the impacts of DRG with (i) the use of LOSC as outcome to compare hospital performance, (ii) a quasi-experimental estimation of the short-term effects of SwissDRG on LOSC, and (iii) an empirical evaluation of different channels that might explain these impacts of SwissDRG on LOSC.

Section snippets

Institutional context

The hospital sector in Switzerland can be characterized by a regulated competition market form. There are currently approximately 300 hospitals, 40% of which are classified as general hospitals, the remaining 60% are specialized clinics [44]. The regulatory framework is given by the national health insurance law (KVG/LaMal), in particular articles 49/49a, but the hospital sector in general is rather decentralized and cantons (similar to US states) have the legal authority over the hospital

Data description

Our analysis is based on the medical statistics of the hospitals for the years 2011–2013, provided by the Swiss Federal Statistical Office. Since cantons did not apply payment systems uniformly to all their hospitals before 2012, we accessed the data release with the hospital type and focus on the university hospitals (type K111). Using hospital identifiers and the case numbers per hospital, we were able to uniquely identify these hospitals and assign them to two groups based on their payment

Descriptive findings

Fig. 2 shows the trends in LOSC by type of hospital (APDRG vs. non-APDRG) over the available time frame. The solid lines connect the quarterly means of LOSC, the dotted lines indicate 95% confidence intervals. It can be seen that the mean LOSC values are of similar magnitude for the treatment and the control group in 2011, prior to the introduction of SwissDRG, which confirms our argument of similar pre-treatment trends of both groups in Section 3.3. In 2012, we observe a decline in LOSC for

Discussion

Our results contribute to the literature on the impacts of DRG in at least three dimensions: First, we consider LOSC (as opposed to LOS) as outcome variable in our evaluation, following a recent literature that proposes this measure as suitable indicator for performance comparisons. We think that considering a benchmarked LOS in the evaluation of DRG systems is important because LOS has several limitations. For example, there is a general decline in LOS observed over time that makes it

Conclusion

This study aimed at evaluating the short-term impacts of the introduction of SwissDRG on LOSC, a relative measure that compares the actual LOS of a case to a benchmark value. We find a relative reduction in LOSC for hospitals that changed from per diem to SwissDRG reimbursement, compared to hospitals that already used a DRG system before (based on APDRGs). If interpreted as a proxy for efficiency, our results point to a reduction in service efficiency, at least in the short-run. We explain the

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of interest

Both authors Prof. Stefan Boes and Christoph Napierala do not have any conflict of interest regarding the research presented in “DRG hospital reimbursement and length of stay compliance: Evidence from a system change in Switzerland”. There has not been any financing nor any other incentive been provided to pursue this research. Funding has been solely covered through University of Lucerne.

References (48)

  • P.A. Camenzind

    Explaining regional variations in health care utilization between swiss cantons using panel econometric models

    BMC Health Serv Res

    (2012)
  • A. Busato et al.

    The implementation of DRG-based hospital reimbursement in Switzerland: a population-based perspective

    Health Res Policy Syst

    (2010)
  • M. Widmer et al.

    Der Einfluss von APDRG auf Aufenthaltsdauer und Rehospitalisierungen: Auswirkungen von Fallpauschalen in Schweizer Spitälern zwischen 2001 und 2008

    Obsan Bericht 49

    (2011)
  • A. Tuch et al.

    Qualität der stationären Leistungen unter der neuen Spitalfinanzierung. Monitoring der Qualitätsindikatoren 2009-2016

    Obsan Dossier 65

    (2018)
  • F.M. Accardo et al.

    Diagnosis-related group analysis and strategic healthcare budgeting

    Capsula Eburnea

    (2011)
  • S. Calciolari et al.

    Management Sanitario: principi, esperienze locali e internazionali, prospettive

    (2012)
  • C. Napierala

    Compliance of length of stay a measure of choice in a DRG system. Working paper

    (2017)
  • C. Napierala et al.

    Is the timing of radiological intervention and treatment day associated with economic outcomes in DRG-financed health care systems: a case study

    BMC Health Serv Res

    (2017)
  • R.B. Fetter et al.

    Hospital management by diagnosis related groups: United States and European experience

    Third international conference on system science in health care

    (1984)
  • K.L. Kahn et al.

    Comparing outcomes of care before and after implementation of the DRG-based prospective payment system

    JAMA

    (1990)
  • R.B. Fetter et al.

    DRGs: their design and development

    (1991)
  • W. Böcking et al.

    First results of the introduction of DRGs in Germany and overview of experience from other DRG countries

    J Public Health

    (2005)
  • R. Busse et al.

    Hospital case payment systems in Europe

    Health Care Manag Sci

    (2006)
  • U. Brügger

    Impact of DRGs. Introducing a DRG reimbursement system: a literature review

    (2010)
  • View full text