Top

The European Journal of Health Economics

Published in:

01-09-2019 | Care | Original Paper

Spatial risk adjustment between health insurances: using GWR in risk adjustment models to conserve incentives for service optimisation and reduce MAUP

Author: Danny Wende

Published in: The European Journal of Health Economics | Issue 7/2019

Abstract

This paper presents a new approach to deal with spatial inequalities in risk adjustment between health insurances. The shortcomings of non-spatial and spatial fixed effects in risk adjustment models are analysed and opposed against spatial kernel estimators. Theoretical and empirical evidence suggests that a reasonable choice of the spatial kernel could limit the spatial uncertainty of the modifiable area unit problem under heavy-tailed claims data, leading to more precise predictions and economically positive incentives on the healthcare market. A case study of the German risk adjustment shows a spatial risk spread of 86 Euro p.c., leading to incentives for spatial risk selection. The proposed estimator eliminates this issue and conserves incentives for services optimisation.

In Germany, the term “solidarity” is explicitly anchored in the Basic Law (§ 72 (2) GG). It means "the creation of equivalent living conditions throughout the federal territory".

The German scientific advisory council discusses a similar approach [10].

van de Ven and Ellis labelled the risks with “s”-type (solidarity desired) and “n”-type (solidarity not desired) [2]. The c/r term (compensation/responsibility) is frequently used in literature [20].

In Germany, health insurances have incentives to optimise or even manipulate medical diagnoses to fulfil a rash of technical requirements to get more reimbursement. Bauhoff, et al. [38] showed that the share of diagnoses which are relevant for reimbursement grew by 3–4% in the last 5 years which could only be explained by an optimisation of the coding practice.

In Germany, the full sample consists of 70 million observations. Additionally, all \(\hat{\beta }\) are required to be positive and ranged according to a predefined cost hierarchy. In the case of negative \(\hat{\beta }\), corresponding variables are eliminated. In the case of an offended hierarchy, the corresponding variables are regrouped. The process happens iteratively until all criteria are reached [66]. The Leapfrog algorithm limits this extra effort to the WLS-step.

Note that a similar method allows multiple (spatial) varying coefficients, for example an additional kernel on age.

Health insurance can add diagnoses which would increase their measured portion of morbidity but reduce the corresponding estimator. Furthermore, the insurance can avoid reporting diagnoses or costs and increase the corresponding estimator, which generates profits if the insured people with the manipulated diagnoses are relatively cheap.

In Germany, there is a premium boundary of 53.100 € p.a. (§ 6 (7) SGB V - §§ 3 and 4 (2) Sozialversicherungs-Rechengrößenverordnung 2018) defining a wage ceiling. Employees who earn less are compulsorily insured. The income dependent premium rates above the premium boundary are zero.

The “HMG” are equivalent to the American “Hierarchical Condition Categories” (HCC), which are a hierarchical extension of the disease groups “DxCG” building on ICD10 codes.

The actual reimbursement also covers additional costs for sick-pay, administration expenses, disease management programs and other activities. The proposed formulation covers the major part of the reimbursement.

The dummy for individuals receiving a pension for reduced earning capacity is a weak social indicator, since the pension could only be received if the individual was at least 5 years employed before the causing illness.

Data specifications are defined by the German federal health care authority (BVA) and revaluated for every year. Historical information can be received from http://www.bundesversicherungsamt.de/risikostrukturausgleich/verfahrensbestimmung.html 16.06.2018.

Note, the German privacy protection does not allow the use of full addresses. Information on zip-codes and municipalities are the slenderest alternatives. Zip-code areas do not fit into municipalities. They are smaller in cities and bigger in rural areas.

There is no meaning in a t test against the hypotheses of zero coefficients as the mean regression error of the model without spatial effects and thus the mean of the spatial coefficient itself is zero. The F-test against the hypotheses that all spatial coefficients are zero could be rejected for every model with less than 1% type-I error.

The coefficients of model 2 are an aggregate of model 3. Coefficients of model 6 are like model 3 in big cities and like model 5 in rural areas.

The lower risk stems from guest workers who are insured while working in Germany (§ 118 (4) SGB V) but remain (illegally) in the data for much longer and generate therefore additional reimbursements (see first draft of the Ministry of Health (GKV-VEG); [72]).

BMG: Statistiken über die Finanzergebnisse des Gesundheitsfonds. Bundesministerium für Gesundheit. https://www.bundesversicherungsamt.de/fileadmin/redaktion/Gesundheitsfonds/20180515UEbersicht_Endgueltige_RE2016-2017.pdf (2018). Accessed 26 June 2018

van de Ven, W., Ellis, R.: Risk adjustment in competitive health plan markets. In: Culyer, A., Newhouse, J. (eds.) Handbook of Health Economics, pp. 755–845. North Holland, Amsterdam (2000)

Ash, A., Ellis, R., Pope, G., Ayanian, J., Bates, D., Burstin, H., Iezzoni, L., Mckay, E., Qu, W.: Using diagnoses to describe populations and predicts costs. Health Care Financ. Rev. 21, 7–28 (2000)PubMedPubMedCentral

Kautter, J., Pope, G., Keenan, P.: Affordable Care Act risk adjustment: overview, context, and challenges. Medicare Medicaid Res. Rev. 4, E1–E11 (2014). https://doi.org/10.5600/mmrr.004.03.a02 CrossRef

van Veen, S.H.C.M., van Kleef, R.C., van de Ven, W.P.M.M., van Vliet, R.: Is there one measure-of-fit that fits all? A taxonomy and review of measures-of-fit for risk-equalization models. Med. Care Res. Rev. 72, 220–243 (2015). https://doi.org/10.1177/1077558715572900 CrossRefPubMed

Ellis, R., Martins, B., Rose, S.: Risk adjustment for health plan payment. In: McGuire, T.G., van Kleef, R. (eds.) Risk Adjustment, Risk Sharing and Premium Regulation in Health Insurance Markets: Theory and Practice, pp. 1–73. Elsevier Press, Amsterdam (2017)

van de Ven, W., Beck, K., Buchner, F., Chernichovsky, D., Gardiol, L., Holly, A., Lamers, L., Schokkaert, E., Shmueli, A., Spycher, S., Van de Voorde, C., van Vliet, R., Wasem, J., Zmora, I.: Risk adjustment and risk selection on the sickness fund insurance market in five European countries. Health Policy 65, 75–98 (2003)CrossRefPubMed

van de Ven, W., Beck, K., Van de Voorde, C., Wasem, J., Zmora, I.: Risk adjustment and risk selection in Europe: 6 years later. Health Policy 83, 162–179 (2007). https://doi.org/10.1016/j.healthpol.2006.12.004 CrossRefPubMed

McGuire, T., Van Kleef, R.: Risk Adjustment, Risk Sharing and Premium Regulation in Health Insurance Markets, 1st edn. Academic Press, Cambridge (2018)

10.

Drösler, S., Garbe, E., Hasford, J., Schubert, I., Ulrich, V., van de Ven, W., Wambach, A., Wasem, J., Wille, E.: Sondergutachten zu den Wirkungen des morbiditätsorientierten Risikostrukturausgleichs. Bonn (2018)

11.

Drösler, S., Garbe, E., Hasford, J., Schubert, I., Ulrich, V., van de Ven, W., Wambach, A., Wasem, J., Wille, E.: Gutachten zu den regionalen Verteilungswirkungen des morbiditätsorientierten Risikostrukturausgleichs. Bonn (2018)

12.

Ulrich, V., Wille, E.: Zur Berücksichtigung einer regionalen Komponente im morbiditätsorientierten Risikostrukturausgleich (Morbi-RSA). Bayreuth und Mannheim (2014)

13.

Ulrich, V., Wille, E., Thüsing, G.: Die Notwendigkeit einer regionalen Komponente im morbiditätsorientierten Risikostrukturausgleich unter wettbewerbspolitischen und regionalen Aspekten. Bonn (2016)

14.

Bieri, O., Köchli, H.: Regionale Unterschiede bei der Belastung durch die obligatorischen Gesundheitsausgaben. Neuchatel (2013)

15.

Schokkaert, E., Van de Voorde, C.: Belgium’s health care system: should the communities/regions take it over? Or the sickness funds? Re-Bel initiative. E-book 10 http://www.rethinkingbelgium.eu/rebel-initiative-ebooks/ebook-10-belgium-healthcare-system-communities-regions-take-over-sickness-funds (2011). Accessed 26 June 2018

16.

Bauhoff, S.: Do health plans risk-select? An audit study on Germany’s Social Health Insurance. J. Public Econ. 96, 750–759 (2012)CrossRef

17.

Göpffarth, D., Kopetsch, T., Schmitz, H.: Determinants of regional variation in health expenditures in Germany. Health Econ. 25, 801–815 (2016). https://doi.org/10.1002/hec.3183 CrossRefPubMed

18.

Reschke, P., Lauterbach, K., Wasem, J.: Klassifikationsmodelle für Versicherte im morbiditätsorientierten Risikostrukturausgleich. IGES 7, 7–15 (2005)

19.

van Kleef, R.C., van Vliet, R., van de Ven, W.: Risk equalization in The Netherlands: an empirical evaluation. Expert Rev. Pharmacoecon. Outcomes Res. 13, 829–839 (2013). https://doi.org/10.1586/14737167.2013.842127 CrossRefPubMed

20.

Schokkaert, E., Van de Voorde, C.: Risk selection and the specification of the conventional risk adjustment formula. J. Health Econ. 23, 1237–1259 (2004)CrossRefPubMed

21.

Arbia, G., Petrarca, F.: Effects of MAUP on spatial econometric models. Lett. Spat. Resour. Sci. 4, 173–185 (2011)CrossRef

22.

Diehr, D., Yanez, A., Ash, M., Nornbrook, D.: Methods for analyzing health care utilization and costs. Annu. Rev. Public Health 20, 125–144 (1999)CrossRefPubMed

23.

Schokkaert, E., Van de Voorde, C.: Incentives for risk selection and omitted variables in the risk adjustment formula. Ann. Écon. Stat. (2006). https://doi.org/10.2307/20079173 CrossRef

24.

Skinner, J.: Causes and consequences of regional variations in health care. In: Pauly, M.V., Mcguire, T.G., Barros, P.P. (eds.) Handbook of Health Economics, pp. 45–93. Elsevier, Amsterdam (2012). https://doi.org/10.1016/b978-0-444-53592-4.00002-5 CrossRef

25.

Finkelstein, A., Gentzkow, M., Williams, H.: Sources of geographic variation in health care: evidence from patient migration. Q. J. Econ. 131, 1681–1726 (2016). https://doi.org/10.3386/w20789 CrossRefPubMedPubMedCentral

26.

Ozegowski, S., Sundmacher, L.: Understanding the gap between need and utilization in outpatient care—the effect of supply-side determinants on regional inequities. Health Policy 114, 54–63 (2014). https://doi.org/10.1016/j.healthpol.2013.08.005 CrossRefPubMed

27.

Busato, A., Matter, P., Künzi, B., Goodman, D.: Geographic variation in the cost of ambulatory care in Switzerland. J. Health Serv. Res. Policy 17(1), 15–23 (2012). https://doi.org/10.1258/jhsrp.2011.010056 CrossRef

28.

Eibich, P., Ziebarth, N.: Analyzing regional variation in health care utilization using (rich) household microdata. Health Policy 114, 41–53 (2014). https://doi.org/10.1016/j.healthpol.2013.04.015 CrossRefPubMed

29.

Charlton, J., Rudisill, C., Bhattarai, N., Gulliford, M.: Impact of deprivation on occurrence, outcomes and health care costs of people with multiple morbidity. J. Health Serv. Res. Policy 18, 215–223 (2013). https://doi.org/10.1177/1355819613493772 CrossRefPubMedPubMedCentral

30.

Siegel, M., Mielck, A., Maier, W.: Individual income, area deprivation, and health: do income-related health inequalities vary by small area deprivation? Health Econ. 24, 1523–1530 (2015). https://doi.org/10.1002/hec.3102 CrossRefPubMed

31.

Shohaimi, S., Welch, A., Bingham, S., Luben, R., Day, N., Wareham, N., Khaw, K.: Residential area deprivation predicts fruit and vegetable consumption independently of individual educational level and occupational social class: a cross sectional population study in the Norfolk cohort of the European Prospective Investigation into Cancer (EPIC-Norfolk). J. Epidemiol. Community Health 58, 686–691 (2004). https://doi.org/10.1136/jech.2003.008490 CrossRefPubMedPubMedCentral

32.

Felder, S., Tauchmann, H.: Federal state differentials in the efficiency of health production in Germany: an artefact of spatial dependence? Eur. J. Health Econ. 14(1), 21–39 (2013). https://doi.org/10.1007/s10198-011-0345-8 CrossRefPubMed

33.

Anthony, D., Herndon, M., Gallagher, P., Barnato, A.E., Bynum, J., Gottlieb, D., Fisher, E., Skinner, J.: How much do patients’ preferences contribute to resource use? Health Aff. 28, 864–873 (2009). https://doi.org/10.1377/hlthaff.28.3.864 CrossRef

34.

Laurence, B., Kate, B., Daniel, K.: Patients’ preferences explain a small but significant share of regional variation in medicare spending. Health Aff. 33, 957–963 (2014). https://doi.org/10.1377/hlthaff.2013.1184.Patients CrossRef

35.

Cutler, D., Skinner, J., Stern, A., Wennberg, D.: Physician beliefs and patient preferences : a new look at regional variation in health care spending. Nber working paper 19320 (2013). https://doi.org/10.3386/w19320

36.

Schönfelder, T., Balázs, S., Klewer, J.: Costs of DRG-upcoding due to the introduction of the diagnosis related groups in Germany. Heilberufe 61, 77–81 (2009). https://doi.org/10.1007/s00058-009-1516-1 CrossRef

37.

Helmchen, L., Lo Sasso, A.: How sensitive is physician performance to alternative compensation schedules? Evidence from a large network of primary care clinics. Health Econ. 19, 1300–1317 (2010). https://doi.org/10.1002/hec CrossRefPubMed

38.

Bauhoff, S., Fischer, L., Göpffarth, D., Wuppermann, A.: Plan responses to diagnosis-based payment: evidence from Germany’s morbidity-based risk adjustment. J. Health Econ. 56, 397–413 (2017). https://doi.org/10.1016/j.jhealeco.2017.03.001 CrossRefPubMed

39.

Laffont, J.-J., Tirole, J.: A Theory of Incentives in Procurement and Regulation, 1st edn. The MIT Press, Cambridge (1993)

40.

Rothschild, M., Stiglitz, J.: Equilibrium in competitive insurance markets: an essay on the economics of imperfect information. Q. J. Econ. 90(4), 629–649 (1979)CrossRef

41.

Glazer, J., Mcguire, T.: Optimal risk adjustments in markets with adverse selection: an application to managed care. Am. Econ. Rev. 90(4), 1055–1071 (2000)CrossRef

42.

Schramm, A.: Diagnosekodierung in der Praxis. Dissertation, TU Dresden (2016)

43.

Flowerdew, R., Manley, D., Sabel, C.: Neighbourhood effects on health: does it matter where you draw the boundaries? Soc. Sci. Med. 66, 1241–1255 (2008). https://doi.org/10.1016/j.socscimed.2007.11.042 CrossRefPubMed

44.

Singh, G.: Area deprivation and widening inequalities in US mortality, 1969–1998. Am. J. Public Health 93, 1137–1143 (2003). https://doi.org/10.2105/AJPH.93.7.1137 CrossRefPubMedPubMedCentral

45.

Fleurbaey, M., Schokkaert, E.: Unfair inequalities in health and health care. J. Health Econ. 28, 73–90 (2009). https://doi.org/10.1016/j.jhealeco.2008.07.016 CrossRefPubMed

46.

Fotheringham, S., Brunsdon, C., Martin, C.: Geographically Weighted Regression: The Analysis of Spatially Varying Relationships. Wiley, New York (2002)

47.

Kautter, J., Pope, G., Ingber, M., Freeman, S., Patterson, L., Cohen, M., Keenan, P.: The HHS-HCC risk adjustment model for individual and small group markets under the Affordable Care Act. Medicare Medicaid Res. Rev. 4, E1–E11 (2014)

48.

Melorose, J., Perroy, R., Careas, S.: Verfeinerung des Risikoausgleichs auf Basis von Daten zur stationären Abrechnung, Bundesamt für Gesundheit (BAG) (2015). https://doi.org/10.1017/cbo9781107415324.004

49.

ESPON: The modifiable areas unit problem. UMS 2414 RIATE (French ECP) and UMR 8504 GEOGRAPHIE-CITES. Luxembourg (2006)

50.

Brunsdon, C., Fotheringham, S., Charlton, M.: Geographically weighted regression-modelling spatial non-stationarity. J. R. Stat. Soc. 47, 431–443 (2012)CrossRef

51.

Nadaraya, E.: On estimating regression. Theory Probab. Appl. 9, 141–142 (1964). https://doi.org/10.1137/1109020 CrossRef

52.

Watson, G.: Smooth regression analyses. Indian J. Stat. 26, 359–372 (1964)

53.

Priestley, M., Chao, M.: Non-parametric function fitting. J. R. Stat. Soc. 34, 385–392 (1972)

54.

Gasser, T., Müller, H.: Estimating regression functions and their derivatives by the kernel method. Scand. J. Stat. 11, 171–185 (1984)

55.

Tobler, W.: A computer movie simulating urban growth in the Detroit region. Econ. Geogr. 46, 234–240 (1970)CrossRef

56.

Nakaya, T., Fotheringham, A., Brunsdon, C., Charlton, M.: Geographically weighted Poisson regression for disease association mapping. Stat. Med. 24, 2695–2717 (2005). https://doi.org/10.1002/sim.2129 CrossRefPubMed

57.

Guagliardo, M.: Spatial accessibility of primary care: concepts, methods and challenges. Int. J. Health Geogr. 3, 3 (2004). https://doi.org/10.1186/1476-072X-3-3 CrossRefPubMedPubMedCentral

58.

Altman, N.: Kernel smoothing of data with correlated errors. J. Am. Stat. Assoc. 85, 749–759 (1990). https://doi.org/10.1080/01621459.1990.10474936 CrossRef

59.

Picard, R., Cook, R.: Cross-validation of regression models. J. Am. Stat. Assoc. 79, 575–583 (1984). https://doi.org/10.1080/01621459.1984.10478083 CrossRef

60.

Mei, C., Wang, N., Zhang, W.: A note on the mixed geographically weighted regression model. J. Reg. Sci. 44, 143–157 (2004)CrossRef

61.

Smyth, G.: Partitioned algorithms for maximum likelihood and other non-linear estimation. Stat. Comput. 6, 201–216 (1996). https://doi.org/10.1007/BF00140865 CrossRef

62.

Wende, D.: Anreize für regionale Risikoselektion unter dem Morbiditätsorientierten Risikostrukturausgleich. Gesundheitswesen 79, 816–820 (2017). https://doi.org/10.1055/s-0042-123851 CrossRefPubMed

63.

Göpffarth, D.: Regionalmerkmale im Risikostrukturausgleich. Ein Beitrag zum funktionalen Wettbewerb und zu bedarfsgerechter Versorgung? BARMER GEK Gesundheitswessen aktuell 2011, pp. 16–127 (2011)

64.

Vogt, V.: The contribution of locational factors to regional variations in office-based physicians in Germany. Health Policy 120, 198–204 (2016). https://doi.org/10.1016/j.healthpol.2016.01.006 CrossRefPubMed

65.

GKVSV: Grundprinzipien der gesetzlichen Krankenversicherung. https://www.gkv-spitzenverband.de/krankenversicherung/kv_grundprinzipien/gkv_grundprinzipien.jsp (2018). Accessed 10 June 2018

66.

BVA: Verfahrensbestimmung nach § 39a Abs. 3 Satz 6 RSAV und § 41 Absatz 5 Satz 2 RSAV für die Zuweisung im Ausgleichsjahr 2018. Bundesversicherungsamt Referat 312 - Risikostrukturausgleich, Bonn. https://www.bundesversicherungsamt.de/fileadmin/redaktion/Risikostrukturausgleich/Verfahrensbestimmung/20171110_Verfahrensbestimmung_AJ2018_Versand.pdf (2018). Accessed 28 June 2018

67.

BVA: Datenzusammenstellungen und Auswertungen. https://www.bundesversicherungsamt.de/risikostrukturausgleich/datenzusammenstellungen-und-auswertungen.html (2018). Accessed 28 June 2018

68.

Heinmueller, J.: Entropy balancing for causal effects: a multivariate reweighting method to produce balanced samples in observational studies. Polit. Anal. 46, 20–25 (2012)

69.

Anselin, L.: Lagrange multiplier test diagnostics for spatial dependence and spatial heterogeneity. Geogr. Anal. 20, 1–17 (1988). https://doi.org/10.1111/j.1538-4632.1988.tb00159.x CrossRef

70.

Keeler, E., Carter, G., Newhouse, J.: A model of the impact of reimbursement schemes on health plan choice. J. Health Econ. 17, 297–320 (1998). https://doi.org/10.1016/S0167-6296(97)00029-5 CrossRefPubMed

71.

Moran, A.: A test for the serial independence of residuals miscellanea. Biometrika 37, 178–181 (1950)CrossRefPubMed

72.

BMG: Entwurf eines Gesetzes zur Beitragsentlastung der Versicherten in der Gesetzlichen Krankenversicherung (GKV-Versichertenentlastungsgesetz – GKV-VEG). https://www.vgsd.de/wp-content/uploads/2018/05/GKVVEG_RefE_Versichertenentlastungsgesetz_190418.pdf (2018). Accessed 19 May 2018

Title: Spatial risk adjustment between health insurances: using GWR in risk adjustment models to conserve incentives for service optimisation and reduce MAUP
Author: Danny Wende
Publication date: 01-09-2019
Publisher: Springer Berlin Heidelberg
Keyword: Care
Published in: The European Journal of Health Economics / Issue 7/2019
Print ISSN: 1618-7598
Electronic ISSN: 1618-7601
DOI: https://doi.org/10.1007/s10198-019-01079-6

Keynote webinar | Spotlight on medication adherence

Springer Medicine

Spatial risk adjustment between health insurances: using GWR in risk adjustment models to conserve incentives for service optimisation and reduce MAUP

Abstract

Keynote webinar | Spotlight on medication adherence

Springer Medicine

Abstract

Please log in to get access to this content

Other articles of this Issue 7/2019

Is rate of return pricing a useful approach when value-based pricing is not appropriate?

Assessing progressivity and catastrophic effect of out-of-pocket payments for healthcare in Canada: 2010–2015

Economic evaluations of pharmacist-led medication review in outpatients with hypertension, type 2 diabetes mellitus, and dyslipidaemia: a systematic review

Value of a QALY and VSI estimated with the chained approach

How to value safety in economic evaluations in health care? A review of applications in different sectors

Changing payment instruments and the utilisation of new medical technologies