Analysis of Health Care Systems 2

Einleitung / Introduction

Health system responsiveness - the ability of a health care system to meet the non-medical legitimate expectations of its service users- is an important aspect of health systems performance as defined by the World Health Organization. Responsiveness relates to interpersonal and organizational aspects of health care, e.g., respectfulness and trustworthiness. Although responsiveness is essential to provide equitable and accountable health care while protecting patient rights, little is known about the differences in patient characteristics influencing the perception and assessment of responsiveness.

Methode / Method

Within the study “Integrating the population perspective into health system performance assessment” 20,000 people with substitutive private health insurance (PHI) in Germany were surveyed in 2018. The survey included questions on nine dimensions of health system responsiveness, health literacy (HL), experienced discrimination, health status, subjective socioeconomic status, and sociodemographics. Survey data were linked on patient-level with claims data from one PHI company on service utilization, insurance status, and living area. Multivariate binomial logistic regression was applied to assess the association between responsiveness and service user characteristics.

Ergebnisse / Results

The sample (age 58.9 ± 14.5; 64.9% male) contains 3,109 cases, of whom 95.7% had contact with an ambulatory physician in the past year. Users with high HL are more likely to assess responsiveness as (very) good, e.g., experiences with clear explanations from physicians (OR 3.652). Poor assessment of responsiveness is seen among users who experienced discrimination. Free choice of physicians (OR 2.345) and waiting time for appointments (OR 1.649) are better rated by urban living people.

Zusammenfassung / Conclusion

Differences in the users’ assessment of responsiveness in Germany are associated with several attributes, e.g., HL, experienced discrimination and living area. These results are crucial to improve health services at the organizational level but also health planning and management at the macropolitical level.

AutorInnen

Miriam Blümel, Technische Universität Berlin

Katharina Achstetter, Technische Universität Berlin

Philipp Hengel, Technische Universität Berlin

Julia Köppen, Technische Universität Berlin

Reinhard Busse, Technische Universität Berlin

Einleitung / Introduction

Financial risk protection in case of ill-health is a main function of health systems. It is assessed as the extent of catastrophic and impoverishing health expenditure (CHE/IHE) due to out-of-pocket payments (OOPP), i.e., direct payments at the point of using care net of reimbursements. Unmet medical needs due to costs is used as an additional indicator for how many people refrain from accessing health care (and thereby reduce their OOPP). In Germany, in-detail assessment of financial risk protection is available for statutory health insurance (SHI) but has not yet been analysed for private health insurance (PHI).

Methode / Method

A linked dataset of 2018 survey and claims data (n=3,105) covering adults insured by Debeka, the largest PHI company in Germany, was employed to calculate prevalence of CHE, IHE, and of unmet medical needs due to costs. Household budget survey microdata (n=42,226) were used to calculate nation-wide standardized basic needs expenses (food/rent/utilities) of households in consumption percentiles 25-35 to obtain a household-specific capacity-to-pay (net income minus basic needs) and to serve as a poverty line (5,606€ per annum and person equivalent). OOPP >40% of capacity-to-pay or pushing households (further) below the poverty line were considered catastrophic. The official 2018 60% threshold of median income was used as an alternative, established poverty line. The data was weighted by age, gender, and aid allowance according to the German PHI population.

Ergebnisse / Results

Prevalence of CHE was 1.0% [95% CI: 0.5;1.8], including <0.1% [0.0;0.3] of impoverished and 0.9% [0.5;1.7] of further impoverished households. When using the official poverty line of 60% of median income, 2.8% [2.1;3.9] of households were further impoverished and 0.3% [0.2;0.6] impoverished by OOPP. Those participants with CHE differed significantly from the others in that they reported a threefold prevalence of unmet need due to costs (19.7% [11.8;30.9] vs. 6.7% [5.6;7.9]), higher annual household OOPP (mean: 795€ [546;1,044] vs. 491€ [464;517]), and much lower annual household net income (mean: 19,176€ [16,428;21,924] vs. 63,984€ [62,652;65,328]). Further, all households with CHE were in the lowest person equivalent income quintile.

Zusammenfassung / Conclusion

Prevalence of CHE and IHE is low among the privately insured in Germany, also in international comparison. Despite higher OOPP, it is lower compared to SHI due to differences in income. In contrast to SHI and other solidarity-based systems, only the poorest are affected by CHE. Low prevalence is likely influenced by previous reforms to strengthen financial risk protection in PHI. Future studies may consider premiums next to OOPP, since they are not linked to ability-to-pay and thus might contribute to financial risks.

AutorInnen

Philipp Hengel, Fachgebiet Management im Gesundheitswesen, Technische Universität Berlin

Miriam Blümel, Fachgebiet Management im Gesundheitswesen, Technische Universität Berlin

Martin Siegel, Fachgebiet Empirische Gesundheitsökonomie, Technische Universität Berlin

Katharina Achstetter, Fachgebiet Management im Gesundheitswesen, Technische Universität Berlin

Julia Köppen, Fachgebiet Management im Gesundheitswesen, Technische Universität Berlin

Reinhard Busse, Fachgebiet Management im Gesundheitswesen, Technische Universität Berlin

Einleitung / Introduction

A central aspect of improving quality of care is the diffusion of innovation from research into practice. One way to facilitate this process is system certification and by conceptualization, innovative activities and certification require different competencies and processes. While these disparities may lead to an obstruction of certification, the effect of innovative activity on the diffusion of certification is mostly unknown. In this study, we aim to evaluate this relationship in the context of certified organ cancer centers (OCC) in Germany.

Methode / Method

We perform a retrospective study at district level (NUTS-3) combining scientific publication and certification data from 1998-2018 for eight cancer sites. Scientific publication data was obtained from the National Library of Medicine and mapped to the authors’ geolocation using the MapAffil tool. We matched this data to certification status and timing provided by the German Cancer Society. We consider certification status in years 3, 5 and 10 after initiation of the certification program as outcome measure to evaluate the effect of innovative activity on certification status, estimating linear probability models. We consider cancer incidence, gross domestic product and hospital competition as confounders.

Ergebnisse / Results

Descriptive statistics of the breast cancer certification program suggest high regional variability in scientific publication output and system certification timing by cancer site. We find publication activity in 380 of the 402 districts between 1998 and 2018. Three years after initiation of the breast cancer certificate, 120 centers in 86 of 402 districts (21%) were certified. In 51% of districts patients did not have access to certified cancer care by 2018. Estimates of linear regression models indicate that innovative activity does not have an immediate effect on certification status. However, ten years after initiation, regions with higher innovative activity had on average 0.094 (P<0,001) more certified breast cancer centers if innovative activity increased by one publication per year. Breast cancer incidence had a short term effect on certification status´. Regions had on average 0.0026 (P<0,001) more certified breast cancer centers three years after initiation of the certificate if incidence increased by one case. This effect increased to 0.0035 (P<0,001) ten years after initiation. Competition in the region did not influence certification status.

Zusammenfassung / Conclusion

Certification and innovative activity are no substitutes. Early certification in regions is driven by cancer incidence rather than innovative activity. This might be due to strategic decision making in hospitals aiming at increasing case volume. The positive effect of innovative activity on certification status ten years after initiation might be driven by regions without notable innovative activity that choose not to certify at all.

AutorInnen

Tim Brand, Universität Duisburg-Essen

Katharina Blankart, Universität Duisburg-Essen

Einleitung / Introduction

Hospitals account for 40 % of all healthcare expenditures and play a central role in healthcare provision. Therefor the way hospitals are payed has major implications for the care they provide. Yet reforms of payment schemes are often implemented without unequivocal knowledge about their effects. When Germany introduced the DRG system 20 years ago, its intention was to increase the transparency, efficiency and cost-effectiveness of German hospitals. However, still today, its effects on service provision remain largely unknown. This study aims to provide novel insights into causal effects of the German DRG introduction by using recent methodological advances within international comparisons.

Methode / Method

As a research setting, we focus on German hospital care provision between the years 1994 to 2015. We consider other developed countries, which were not exposed to payment schemes reforms of similar extent, as control group. We take aggregate panel data from OECD, complemented in some cases by data from Eurostat and WHO. Our main outcomes of interest are hospital activity and efficiency. Due to the limited nature of aggregate data, we use hospital discharges and the average length of stay as variables. Additional variables with respect to hospital staffing and population health are considered as secondary outcomes. As baseline model we use a slightly extended Difference-in-Differences (DiD) approach. Since the assumption of parallel trends does not hold for several outcome variables, we make use of other methods that allow estimation under different assumptions. The German DRG introduction is a prime example for the use of the Synthetic Control Method (SCM), one of the most important innovations in policy evaluation literature. The recently proposed Synthetic Difference-in-Differences, which integrates SCM and DiD within a single framework and has attractive estimation properties, complements our approach.

Ergebnisse / Results

Results indicate that the German DRG-Introduction increased hospital activity, measured as discharges, significantly by around 2 % per year. Furthermore, our preliminary findings suggest that the hoped-for efficiency gains did not materialize with respect to the length of stay.

Zusammenfassung / Conclusion

Our results complement two strands of literature. First, it adds to the ongoing policy discussion in Germany about the effects of the DRG introduction. Our findings provide evidence that the DRG reform had long lasting – possibly unintended – consequences on German hospital activity. It increased hospital discharges while failing to shorten the average length of stay. Second, our results add more broadly to the overarching literature regarding the impact of activity based funding. Thus, it informs policy makers about possible consequences and pitfalls of activity based funding.

AutorInnen

Robert Messerle, Universität Hamburg

Jonas Schreyögg, Universität Hamburg