[Five Key Questions for Health Services Research: are SHI Claims Data Suitable for Your Research Project?] / 5 Leitfragen für die Versorgungsforschung ­ Eignen sich GKV-Routinedaten für Ihr Forschungsvorhaben?

Health services research examines the structures and processes of health care under everyday conditions. Routine data of the statutory health insurance (SHI) - the so-called routine practice data - represent real health care and are therefore an important data source for health services research. This paper presents 5 key questions that researchers and data-holding institutions can use to assess the suitability of this data source for answering their health services research question. The aim of these guiding questions is to generate a common understanding between researchers and data-holding institutions of the research project, the research objective, and the feasibility of implementation in health services research. The five guiding questions cover the formulation of the research question, the planned method, the target population, the relevant study periods, and the required information from SHI data. These methodologically oriented guiding questions are supplemented by the question of how the results of the research project could improve care. Thus, for researchers, the five guiding questions provide an initial structuring for data requests; for data-holding institutions, they provide a framework for considering possible involvement in or support of a research idea in health services research.

Applying a data-driven population segmentation approach in German claims data.

BACKGROUND: Segmenting the population into homogenous groups according to their healthcare needs may help to understand the population's demand for healthcare services and thus support health systems to properly allocate healthcare resources and plan interventions. It may also help to reduce the fragmented provision of healthcare services. The aim of this study was to apply a data-driven utilisation-based cluster analysis to segment a defined population in the south of Germany. METHODS: Based on claims data of one big German health insurance a two-stage clustering approach was applied to group the population into segments. A hierarchical method (Ward's linkage) was performed to determine the optimal number of clusters, followed by a k-means cluster analysis using age and healthcare utilisation data in 2019. The resulting segments were described in terms of their morbidity, costs and demographic characteristics. RESULTS: The 126,046 patients were divided into six distinct population segments. Healthcare utilisation, morbidity and demographic characteristics differed significantly across the segments. The segment "High overall care use" comprised the smallest share of patients (2.03%) but accounted for 24.04% of total cost. The overall utilisation of services was higher than the population average. In contrast, the segment "Low overall care use" included 42.89% of the study population, accounting for 9.94% of total cost. Utilisation of services by patients in this segment was lower than population average. CONCLUSION: Population segmentation offers the opportunity to identify patient groups with similar healthcare utilisation patterns, patient demographics and morbidity. Thereby, healthcare services could be tailored for groups of patients with similar healthcare needs.

Can minimal clinically important differences in patient reported outcome measures be predicted by machine learning in patients with total knee or hip arthroplasty? A systematic review.

OBJECTIVES: To systematically review studies using machine learning (ML) algorithms to predict whether patients undergoing total knee or total hip arthroplasty achieve an improvement as high or higher than the minimal clinically important differences (MCID) in patient reported outcome measures (PROMs) (classification problem). METHODS: Studies were eligible to be included in the review if they collected PROMs both pre- and postintervention, reported the method of MCID calculation and applied ML. ML was defined as a family of models which automatically learn from data when selecting features, identifying nonlinear relations or interactions. Predictive performance must have been assessed using common metrics. Studies were searched on MEDLINE, PubMed Central, Web of Science Core Collection, Google Scholar and Cochrane Library. Study selection and risk of bias assessment (ROB) was conducted by two independent researchers. RESULTS: 517 studies were eligible for title and abstract screening. After screening title and abstract, 18 studies qualified for full-text screening. Finally, six studies were included. The most commonly applied ML algorithms were random forest and gradient boosting. Overall, eleven different ML algorithms have been applied in all papers. All studies reported at least fair predictive performance, with two reporting excellent performance. Sample size varied widely across studies, with 587 to 34,110 individuals observed. PROMs also varied widely across studies, with sixteen applied to TKA and six applied to THA. There was no single PROM utilized commonly in all studies. All studies calculated MCIDs for PROMs based on anchor-based or distribution-based methods or referred to literature which did so. Five studies reported variable importance for their models. Two studies were at high risk of bias. DISCUSSION: No ML model was identified to perform best at the problem stated, nor can any PROM said to be best predictable. Reporting standards must be improved to reduce risk of bias and improve comparability to other studies.

The detection and prevention of adverse drug events in nursing home and home care patients: Study protocol of a quasi-experimental study.

AIM: To estimate the cost-effectiveness of an intervention facilitating the early detection of adverse drug events through the means of health professional training and the application of a digital screening tool. DESIGN: Multi-centred non-randomized controlled trial from August 2018 to March 2020 including 65 nursing homes or home care providers. METHODS: We aim to estimate the effect of the intervention on the rate of adverse drug events as primary outcome through a quasi-experimental empirical study design. As secondary outcomes, we use hospital admissions and falls. All outcomes will be measured on patient-month level. Once the causal effect of the intervention is estimated, cost-effectiveness will be calculated. For cost-effectiveness, we include all patient costs observed by the German statutory health insurance. RESULTS: The results of this study will inform about the cost-effectiveness of the optimized drug supply intervention and provide evidence for potential reimbursement within the German statutory health insurance system.

[Patients, Partners, Professionals, Payers and Policy Makers Preferences for Integrated Care for Multimorbidity in Germany: A Discrete Choice Experiment]. / Präferenzen von Patienten, pflegenden Angehörigen, Leistungserbringern, Kostenträgern und politischen Entscheidungsträgern bei der integrierten Versorgung in Deutschland: Ein Discrete Choice Experiment.

AIM OF THE WORK: The aim of this study was to measure and compare the relative importance that patients with multimorbidity, partners and other informal caregivers, professionals, payers and policy makers attribute to different outcome measures of integrated care (IC) programmes in Germany. METHODS: A DCE was conducted, asking respondents to choose between two IC programmes for persons with multimorbidity. Each IC programme was presented by means of attributes or outcomes reflecting the Triple Aim. They were divided into the outcomes health/ wellbeing, experience with care and costs with in total eight attributes and three levels of performance. RESULTS: The results of n=676 questionnaires showed that the attributes "enjoyment of life" and "continuity of care" received the highest ratings across all stakeholder groups. The lowest relative scores remained for the attribute "total costs" for all stakeholders. The preferences of professionals and informal caregivers differed most distinctly from the patients' preferences. The differences mostly concerned "physical functioning", which was rated highest by patients, and "person centeredness" and "continuity of care", which received the highest ratings from professionals. CONCLUSIONS: The preference heterogeneities identified in relation to the outcomes of IC programmes between different stakeholders highlight the importance of informing professionals and policy makers about the different perspectives in order to optimise the design of IC programmes. The results also support the relevance of joint decision-making and coordination processes between professionals, informal caregivers and patients.

Capturing modelled and perceived spatial access to ambulatory health care services in rural and urban areas in Germany.

Studies on social and regional inequalities in access to health care often use spatial indicators such as physician density to measure access to health care. However, the concept of access is more complex, comprising, among others, patient perceptions. In this study, we evaluate the association between different spatial measures of access (i.e. physician density, distance to the nearest provider, and measures based on floating catchment area methods) and measures of perceived spatial access to ambulatory health care in rural and urban areas in Germany. Using correlation and regression analysis, we found that the significance and strength of the relation between perceived and modelled spatial access depends on the type of area and the physician group. The distance to the nearest physician is associated with perceived spatial access to GPs only in rural areas but not in urban areas. More sophisticated measures of spatial access seem not to explain perceived access better than the simpler indicators.

The reimbursement of new medical technologies in German inpatient care: What factors explain which hospitals receive innovation payments?

Most hospital payment systems based on diagnosis-related groups (DRGs) provide payments for newly approved technologies. In Germany, they are negotiated between individual hospitals and health insurances. The aim of our study is to assess the functioning of temporary reimbursement mechanisms. We used multilevel logistic regression to examine factors at the hospital and state levels that are associated with agreeing innovation payments. Dependent variable was whether or not a hospital had successfully negotiated innovation payments in 2013 (n = 1532). Using agreement data of the yearly budget negotiations between each German hospital and representatives of the health insurances, the study comprises all German acute hospitals and innovation payments on all diagnoses. In total, 32.9% of the hospitals successfully negotiated innovation payments in 2013. We found that the chance of receiving innovation payments increased if the hospital was located in areas with a high degree of competition and if they were large, had university status and were private for-profit entities. Our study shows an implicit self-controlled selection of hospitals receiving innovation payments. While implicitly encouraging safety of patient care, policy makers should favour a more direct and transparent process of distributing innovation payments in prospective payment systems.

Correction: Effect of corruption on perceived difficulties in healthcare access in sub-Saharan Africa.

[This corrects the article DOI: 10.1371/journal.pone.0220583.].

Effect of corruption on perceived difficulties in healthcare access in sub-Saharan Africa.

BACKGROUND: Achieving Universal Health Coverage (UHC) by improving financial protection and effective service coverage is target 3.8 of the Sustainable Development Goals. Little is known, however, about the extent to which paying bribes within healthcare acts as a financial barrier to access and, thus, UHC. METHODS: Using survey data in adults from 32 sub-Saharan African countries in 2014-2015, we constructed a multilevel model to evaluate the relationship between paying bribes and reported difficulties of obtaining medical care. We controlled for individual-, region-, and country-level variables. RESULTS: Having paid bribes for medical care significantly increased the odds of reporting difficulties in obtaining care by 4.11 (CI: 3.70-4.57) compared to those who never paid bribes, and more than doubled for those who paid bribes often (OR = 9.52; 95% CI: 7.77-11.67). Respondents with higher levels of education and more lived poverty also had increased odds. Those who lived in rural areas or within walking distance to a health clinic had reduced odds of reporting difficulties. Sex, age, living in a capital region, healthcare expenditures per capita, and country Corruption Perception Index were not significant predictors. CONCLUSIONS: We found that bribery in healthcare is a significant barrier to healthcare access, negatively affecting the potential of African countries to make progress toward UHC. Future increases in health expenditures-which are needed in many countries to achieve UHC-should be accompanied by greater efforts to fight corruption in order to avoid wasting money. Measuring and tracking health sector-specific corruption is critical for progress toward UHC.

Strengthening the evidence-base of integrated care for people with multi-morbidity in Europe using Multi-Criteria Decision Analysis (MCDA).

BACKGROUND: Evaluation of integrated care programmes for individuals with multi-morbidity requires a broader evaluation framework and a broader definition of added value than is common in cost-utility analysis. This is possible through the use of Multi-Criteria Decision Analysis (MCDA). METHODS AND RESULTS: This paper presents the seven steps of an MCDA to evaluate 17 different integrated care programmes for individuals with multi-morbidity in 8 European countries participating in the 4-year, EU-funded SELFIE project. In step one, qualitative research was undertaken to better understand the decision-context of these programmes. The programmes faced decisions related to their sustainability in terms of reimbursement, continuation, extension, and/or wider implementation. In step two, a uniform set of decision criteria was defined in terms of outcomes measured across the 17 programmes: physical functioning, psychological well-being, social relationships and participation, enjoyment of life, resilience, person-centeredness, continuity of care, and total health and social care costs. These were supplemented by programme-type specific outcomes. Step three presents the quasi-experimental studies designed to measure the performance of the programmes on the decision criteria. Step four gives details of the methods (Discrete Choice Experiment, Swing Weighting) to determine the relative importance of the decision criteria among five stakeholder groups per country. An example in step five illustrates the value-based method of MCDA by which the performance of the programmes on each decision criterion is combined with the weight of the respective criterion to derive an overall value score. Step six describes how we deal with uncertainty and introduces the Conditional Multi-Attribute Acceptability Curve. Step seven addresses the interpretation of results in stakeholder workshops. DISCUSSION: By discussing our solutions to the challenges involved in creating a uniform MCDA approach for the evaluation of different programmes, this paper provides guidance to future evaluations and stimulates debate on how to evaluate integrated care for multi-morbidity.

Continuity of care and its effect on readmissions for COPD patients: A comparative study of Norway and Germany.

BACKGROUND: This study compares continuity of care between Germany - a social health insurance country, and Norway - a national health service country with gatekeeping and patient lists for COPD patients before and after initial hospitalization. We also investigate how subsequent readmissions are affected. METHODS: Continuity of Care Index (COCI), Usual Provider Index (UPC) and Sequential Continuity Index (SECON) were calculated using insurance claims and national register data (2009-14). These indices were used in negative binomial and logistic regressions to estimate incident rate ratios (IRR) and odds ratios (OR) for comparing readmissions. RESULTS: All continuity indices were significantly lower in Norway. One year readmissions were significantly higher in Germany, whereas 30-day rates were not. All indices measured one year after discharge were negatively associated with one-year readmissions for both countries. Significant associations between indices measured before hospitalization and readmissions were only observed in Norway - all indices for one-year readmissions and SECON for 30-day readmissions. CONCLUSION: Our findings indicate higher continuity is associated with reductions in readmissions following initial COPD admission. This is observed both before and after hospitalization in a system with gatekeeping and patient lists, yet only after for a system lacking such arrangements. These results emphasize the need for policy strategies to further investigate and promote care continuity in order to reduce hospital readmission burden for COPD patients.

Applying sequence clustering techniques to explore practice-based ambulatory care pathways in insurance claims data.

Background: Care pathways are a widely used mean to ensure well-coordinated and high quality care by defining the optimal timing and interval of health services for a specific indication. However, evidence on common sequences of services actually followed by patients has rarely been quantified. This study aims to explore whether sequence clustering techniques can be used to empirically identify typical treatment sequences in ambulatory care for heart failure (HF) patients and compare their effectiveness. Methods: Routine data of HF patients were provided by a large statutory sickness fund in Germany from 2009 until 2011. Events were categorized by either (i) the specialty of the physician, (ii) the type of service/procedure provided and (iii) the medication prescribed. Similarities between sequences were measured using the 'longest common subsequence' (LCS). The k-medoids clustering algorithm was applied to identify distinct subgroups of sequences. We used logistic regression to identify the most effective sequences for avoiding hospitalizations. Results: Treatment data of 982 incident HF patients were analyzed to identify typical treatment sequences. The cluster analysis revealed three distinct clusters of specialty sequences, four clusters of procedure sequences and four clusters of prescription sequences. Clusters differed in terms of timing and interval of physician visits, procedures and drug prescriptions as well as comorbidities and HF hospitalization rates. We found no significant association between cluster membership and HF hospitalization. Conclusions: Sequence clustering techniques can be used as an explorative tool to systematically extract, describe compare and analyze treatment sequences and associated characteristics.

Continuity of care in the ambulatory sector and hospital admissions among patients with heart failure in Germany.

BACKGROUND: Heart failure is one of the most cost-intensive chronic diseases and the most common cause of hospitalization. More than 60% of the treatment costs of heart failure are incurred in the inpatient sector in Germany. However, hospital admissions due to heart failure are considered to be potentially avoidable through effective and continuous ambulatory care. Our aim is to examine whether continuity in ambulatory care is associated with hospitalizations due to heart failure. METHODS: Using insurance claims data from Germany's biggest statutory health insurance company, we defined three measures of continuity of care: Continuity of Care Index (COCI), Usual Provider Index (UPC) and the Sequential Continuity Index (SECON). We analyzed whether these measures are associated with hospitalization due to heart failure using separate logistic regression models. We controlled for a wide range of covariates such as sex, age and the Charlson comorbidity index. RESULTS: Data of 382 118 heart failure patients were included in the analyses. Index values range from 0.77 to 0.89. Results of logistic regression analyses indicate that the continuity indices COCI, UPC and SECON based on visits to general practitioners (GPs), cardiologists and internists are negatively associated with the probability of hospitalization whereas of the continuity indices based on GP visits only SECON is significantly associated with hospitalization. CONCLUSION: The results indicate that the overall continuity in the ambulatory sector is high for heart failure patients in Germany. Public policy should, nevertheless, focus on increasing sequential continuity of specialist and generalist ambulatory care as this was found to be significantly associated with a reduced likelihood of hospitalization.

Developing a composite index of spatial accessibility across different health care sectors: A German example.

The evolving lack of ambulatory care providers especially in rural areas increasingly challenges the strict separation between ambulatory and inpatient care in Germany. Some consider allowing hospitals to treat ambulatory patients to tackle potential shortages of ambulatory care in underserved areas. In this paper, we develop an integrated index of spatial accessibility covering multiple dimensions of health care. This index may contribute to the empirical evidence concerning potential risks and benefits of integrating the currently separated health care sectors. Accessibility is measured separately for each type of care based on official data at the district level. Applying an Improved Gravity Model allows us to factor in potential cross-border utilization. We combine the accessibilities for each type of care into a univariate index by adapting the concept of regional multiple deprivation measurement to allow for a limited substitutability between health care sectors. The results suggest that better health care accessibility in urban areas persists when taking a holistic view. We believe that this new index may provide an empirical basis for an inter-sectoral capacity planning.

The contribution of locational factors to regional variations in office-based physicians in Germany.

There is considerable literature showing that regional variation in the number of office-based physicians is rather explained by factors such as financial attractiveness of a region or employment opportunities for spouses than by health care needs of the population. It remains unclear, however, how much of the variation is explained by each of these determinants. The aim of the present study is to estimate the percentage contribution of a variety of determinants to the measured variation in Germany. Physician density is regressed on a well-defined set of explanatory variables that were identified as determinants of physician location. Regression-based decomposition was applied to decompose the variation in physician density into the percentage contribution of each of the determinants. The results show that varying health care needs of the population explained less than 5.2% of the variation in physician density. Percentage of population with private health insurance explained 14% of the variation in GP density and between 2% and 6% of the variation in specialists' density. For specialists, a higher share of variance was attributable to the variables measuring sociocultural amenities of a region compared to GPs.

Explanatory models concerning the effects of small-area characteristics on individual health.

OBJECTIVES: Material and social living conditions at the small-area level are assumed to have an effect on individual health. We review existing explanatory models concerning the effects of small-area characteristics on health and describe the gaps future research should try to fill. METHODS: Systematic literature search for, and analysis of, studies that propose an explanatory model of the relationship between small-area characteristics and health. RESULTS: Fourteen studies met our inclusion criteria. Using various theoretical approaches, almost all of the models are based on a three-tier structure linking social inequalities (posited at the macro-level), small-area characteristics (posited at the meso-level) and individual health (micro-level). No study explicitly defines the geographical borders of the small-area context. The health impact of the small-area characteristics is explained by specific pathways involving mediating factors (psychological, behavioural, biological). These pathways tend to be seen as uni-directional; often, causality is implied. They may be modified by individual factors. CONCLUSIONS: A number of issues need more attention in research on explanatory models concerning small-area effects on health. Among them are the (geographical) definition of the small-area context; the systematic description of pathways comprising small-area contextual as well as compositional factors; questions of direction of association and causality; and the integration of a time dimension.

Examining regional variation in the use of cancer screening in Germany.

The detection of cancer in its early latent stages can improve a patient's chances of recovery and thereby reduce the overall burden of the disease. Cancer screening services are, however, only used by a small part of the population and utilization rates vary widely amongst the 402 German districts. This study examines to which extent geographic variation in the use of cancer screening can be explained by accessibility of these services and by spillover effects between adjacent areas, while controlling for a wide range of covariates. District level data on cancer screening utilization rates were calculated for breast, cervical, prostate, skin, and colorectal cancers using German data provided by the National Association of Statutory Health Insurance Physicians (Kassenärztliche Bundesvereinigung - KBV) between 2008 and 2011. We estimated the impact of health service variables on cancer screening utilization using spatial and non-spatial regression models. Spatial autocorrelation in the residuals was estimated using Moran's I statistic. After controlling for socioeconomic and other regional covariates, screening rates for breast, prostate, skin, and colorectal cancers are significantly higher in areas with higher physician density. The utilization of Pap-tests, skin cancer screening and colonoscopies is inversely related with average travel time to physicians. The coefficients for the spatial lag are significant and positive in all models. The positive spatial lags indicate that screening utilization rates are determined by knowledge spillovers between neighboring districts. In terms of public policy, our study demonstrates the potential to increase the use of cancer screening services through improving knowledge regarding cancer screening and by ensuring patient access to cancer screening services.

From a conservative to a liberal welfare state: decomposing changes in income-related health inequalities in Germany, 1994-2011.

Individual socio-economic status and the respective socio-economic and political contexts are both important determinants of health. Welfare regimes may be linked with health and health inequalities through two potential pathways: first, they may influence the associations between socio-economic status and health. Second, they may influence the income-related distributions of socio-economic determinants of health within a society. Using the Socio-Economic Panel (SOEP) for the years 1994-2011, we analyze how income-related health inequalities evolved in the context of the transformation from a conservative to a liberal welfare system in Germany. We use the concentration index to measure health inequalities, and the annual concentration indices are decomposed to reveal how the contributions of the explanatory variables age, sex, income, education, and occupation changed over time. The changes in the contributions are further decomposed to distinguish whether changes in health inequalities stem from redistributions of the explanatory variables, from changes in their associations with health, or from changes in their means. Income-related health inequalities to the disadvantage of the economically deprived roughly doubled over time, which can largely be explained by changes in the contributions of individual characteristics representing weaker labor market positions, particularly income and unemployment. The social and labor market reforms coincide with the observed changes in the distributions of these characteristics and, to a lesser extent, with changes of their associations with health.