gms | German Medical Science

GMS Medizinische Informatik, Biometrie und Epidemiologie

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e.V. (GMDS)

ISSN 1860-9171

Assessing the socioeconomic status in the German Pharmacoepidemiological Research Database (GePaRD): Description and exemplary application using the association with obesity

Research Article

  • Marieke Asendorf - Department of Clinical Epidemiology, Leibniz Institute for Prevention Research and Epidemiology – BIPS, Bremen, Germany
  • Jonas Reinold - Department of Clinical Epidemiology, Leibniz Institute for Prevention Research and Epidemiology – BIPS, Bremen, Germany
  • Tania Schink - Department of Clinical Epidemiology, Leibniz Institute for Prevention Research and Epidemiology – BIPS Bremen, Germany
  • Bianca Kollhorst - Department of Biometry and Data Management, Leibniz Institute for Prevention Research and Epidemiology – BIPS, Bremen, Germany
  • corresponding author Ulrike Haug - Department of Clinical Epidemiology, Leibniz Institute for Prevention Research and Epidemiology – BIPS, Bremen, Germany; Department of Human and Health Sciences, University of Bremen, Germany

GMS Med Inform Biom Epidemiol 2022;18(1):Doc02

doi: 10.3205/mibe000235, urn:nbn:de:0183-mibe0002350

This is the English version of the article.
The German version can be found at:

Published: May 12, 2022

© 2022 Asendorf et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at


Background: The socioeconomic status is often an important confounder in the context of epidemiological studies. However, this information is typically not included in studies based on claims data, possibly due to a lack of knowledge regarding the completeness and validity of the available information.

Objectives: Our aim was to assess and evaluate the socioeconomic status in the German Pharmacoepidemiological Research Database (GePaRD).

Methodology: First, an algorithm was developed to assign an educational status to individuals insured in 2017 in order to estimate the socioeconomic status at the individual level. For comparison, the socioeconomic status was estimated based on the place of residence. For plausibility, we examined whether the known association between the presence of obesity and the socioeconomic status could be reproduced.

Results: Based on different age groups it was possible to dichotomously estimate the socioeconomic status for 86–93% of those under 60 years of age and for 67% of the 60- to 69-year-olds. For individuals 70 years and older, the proportion of missing values was very high. In all subgroups, the prevalence of obesity was higher among persons of lower socioeconomic status than among persons of higher socioeconomic status. The estimation based on the place of residence produced plausible results, but with smaller differences by socioeconomic status.

Discussion and conclusion: Overall, the study results suggest that in terms of plausibility and completeness the information available in GePaRD is well suited for estimating the individual socioeconomic status for age groups up to 69 years.

Keywords: socioeconomic status, educational status, obesity, index of social deprivation, electronic health care data


For decades, scientific studies have shown a social imbalance in the distribution of the burden of various diseases. In most cases, socially disadvantaged groups bear the greater burden of disease [1], [2], [3]. For example, there is a pronounced association between the socioeconomic status and the presence of obesity [4], [5], [6], [7], [8]. Based on survey data, Großschädl/Stronegger report a 13-percentage point higher obesity prevalence for women with primary/vocational school education compared to those with university entrance qualification. For men, a difference of about 9 percentage points was observed in the study [9].

Consequently, it is important to integrate information on the socioeconomic status into epidemiological studies as a possible influencing or confounding variable. Studies based on primary data usually collect information on school education and vocational qualification, occupation or income for this purpose [6], [10]. In addition to primary data, health insurance claims data are an increasingly important data source for health services and epidemiological research. For example, the German Pharmacoepidemiological Research Database (GePaRD) contains claims data from four statutory health insurance providers, covers approximately 20% of the German population, and currently includes the data years 2004–2017. GePaRD is utilized for numerous studies in the areas of drug utilization and drug safety research and cancer screening. In many of these studies, the incorporation of the socioeconomic status is also of great relevance or interest. However, it has not yet been systematically investigated to what extent variables available in GePaRD are suitable to characterize the socioeconomic status and to what degree of completeness they are available.

There are several ways of estimating the socioeconomic status in data from statutory health insurance providers (see Table 1 [Tab. 1]). In GePaRD, two variables could provide information on the socioeconomic status. First, the data include the county code of the residence of the insured, which – linked to the regional deprivation index of the respective county – may be used to estimate the socioeconomic status based on aggregated information, i.e., the same socioeconomic status is assigned to all persons in a county. Second, there is a variable containing information on occupation, occupational status, and education of the insured according to the key list of the Federal Employment Agency, thus providing a way to estimate the socioeconomic status on an individual level. The aim of this study was to investigate the completeness of these two types of information in GePaRD and to indirectly assess their informative value and plausibility with regard to the socioeconomic status by determining the association with the presence of obesity.


Data source and study population

We used the German Pharmacoepidemiological Research Database (GePaRD) for this study. GePaRD is based on claims data from four statutory health insurance providers in Germany and currently includes information on approximately 25 million persons who have been insured with one of the participating providers since 2004 or later. In addition to demographic data, GePaRD contains information on drug dispensings as well as outpatient and inpatient services and diagnoses. Per data year, there is information on approximately 20% of the general German population and all geographical regions of Germany are represented. For this study, we used data from the years 2004 to 2017. We included all persons who were insured for at least one day in 2017 and were alive on that day, for whom information on sex, year of birth, and place of residence was available and whose place of residence was in Germany.

Estimation of the socioeconomic status on the individual level

The starting point was a variable included in the master data of insured persons containing information on occupation, occupational status, and education. This information, summarized by the so-called occupation key (German: Tätigkeitsschlüssel) is transmitted at least annually by employers to the social insurance institutions for all persons covered by statutory health insurance and for primary insurance holders covered voluntarily. The transfer occurs in accordance with Section 28a of the German Social Code (German: Sozialgesetzbuch – SGB) IV. This information is thus not available for persons who have retired in the respective year or for “other” insured persons and co-insured family members [11].

For the operationalization of this variable, the following principles were applied:

A) The focus was to estimate the socioeconomic status based on information regarding school education.

B) Where unambiguous, information on occupational degree or status was included to draw conclusions about school education. For example, individuals with a university degree were assumed to have a (technical) degree allowing university entrance regardless of the information on their school education.

C) In order to assign an educational status to persons insured in 2017, earlier data years were also used. For example, in the case of some retired persons, this made it possible to identify past periods of employment, i.e., years in which an occupation key had still been transmitted.

D) Individuals covered by the insurance of family members were assumed to have the same socioeconomic status as the associated primary insurance holders. Based on this assumption, we attempted to link individuals for whom an occupation code was not available in any of the data years to a family member with joint insurance in GePaRD to whom a socioeconomic status had already been assigned. The family-ID contained in the master data of the insured was used for this linkage.

E) If data from different calendar years yielded discrepant information on education, the highest level of education was always used. This is based on the assumption that the level of education acquired does not decrease, but may at best increase as a result of further school education.

Based on this operationalization, insured persons were classified into the following three categories:

(technical) university entrance qualification (German: (fachgebundene) Hochschulreife),
basic secondary degree or secondary degree (German: Hauptschulabschluss oder mittlere Reife),
school education unknown.

In the last category, insured persons with “school education unknown” or “no degree” and those without data on school education were grouped together, since it was expected that the group with no degree would be small. This classification is based on the options provided by the occupation key before it was changed to a new classification system in 2011 [11], [12], [13]. The new classification system offers the additional option of distinguishing between the basic secondary degree and the secondary degree (see Figure 1 [Fig. 1]). However, for the most part, the occupation key corresponding to the new classification was transmitted to GePaRD only starting with the 2017 data set. To investigate the potential of the new occupation key as well, we conducted sensitivity analyses. For this purpose, the population of the main analysis was restricted to persons with an educational status based on information on vocational or school education of the classification system after 2011.

Estimation of the socioeconomic status based on information on the place of residence

The most recent information on residence at the county level available in GePaRD was linked to the German Index of Socioeconomic Deprivation (GISD) developed by the Robert Koch Institute (RKI) in its revised version based on 2014 data. The index reflects the extent of deprivation based on the three dimensions income, education, and occupation [14]. In line with the smallest spatial unit in GePaRD, the GISD was used at the county level (see appendix (Attachment 1 [Attach. 1]) for further explanation). Analogously to the procedure of Kroll et al., the deprivation scores assigned to counties were divided into quintiles (first quintile: “low deprivation”; second to fourth quintile: “medium deprivation”; fifth quintile: “high deprivation” [14].

Collection of information on the presence of obesity

The presence of obesity was determined on the basis of the relevant ICD-10 codes (inpatient or confirmed outpatient) and codes for procedures indicating obesity therapy (e.g., bariatric surgery). Persons with at least one of the relevant codes in GePaRD were classified as obese. For these codes, we considered not only the data year 2017, but also all data years prior to 2017 available in GePaRD for each person given that obesity is generally a chronic and therefore permanent health condition. Since it is presumably often only coded if relevant as a reason for billing, which may not be the case every year, reverting to earlier data years increases the sensitivity of detecting obesity.

Analysis of the data in GePaRD

First, to assess the completeness of individual schooling information in GePaRD, we determined the proportion of the insured in 2017 to whom school education could be assigned based on the algorithm described above. To determine the prevalence of obesity, the persons insured in 2017 who were classified as obese as described above were considered in the numerator and all persons from the described cohort of the insured were considered in the denominator. The obesity prevalence calculated in this way was determined in the main analyses stratified by (technical) university entrance qualification vs. basic secondary degree or secondary degree (dichotomous classification). In the sensitivity analysis, prevalence was stratified according to the three categories (technical) university entrance qualification, secondary degree, and basic secondary degree. For the analyses based on the GISD, we followed an analogous procedure, i.e., obesity prevalence was determined stratified according to the categories low, medium, and high socioeconomic deprivation. All analyses were stratified by age and sex of the insured.

Comparative analysis of primary data

In addition to the analyses based on GePaRD, primary data collected as part of the GEDA 2014/2015 survey of the RKI were comparatively analyzed [15], i.e., obesity prevalence was also determined according to age, sex, and school education (in the categories (technical) university entrance qualification, secondary degree, and basic secondary degree). The survey, conducted by means of electronic questionnaires or telephone interviews, involved 24,016 persons aged 18 years and older with main residence in Germany [16].


After applying the inclusion and exclusion criteria, 17,317,559 individuals were included in the study (proportion of women: 53%). 258,013 individuals were excluded due to missing information (in 84% of these individuals, information on the place of residence in Germany was missing or not valid). Overall, the socioeconomic status could be approximated at the individual level according to the algorithm described for 73% of women and 78% of men.

For about half of the men and women, there was information on their educational level. Additionally, for about 20%, there was no information on their education, but it was possible to match them to family members with information on their education and thus indirectly estimate their socioeconomic status (see Figure 2 [Fig. 2]). The proportion of individuals to whom a socioeconomic status could be assigned with the algorithm was consistently above 85% for men and women in the age groups under 60, it was 67% for 60- to 69-year-olds (women: 63%, men: 72%), 16% for 70- to 79-year-olds (women: 10%, men: 23%), and for persons aged 80 and over below 1% (women: 0.3%, men: 1%) (see Table 2 [Tab. 2]). The proportion of individuals without a formal degree (no school completed) among all individuals was less than 1% across all subgroups.

Figure 3 [Fig. 3] shows the prevalence of obesity observed in GePaRD stratified by age group, sex, and assigned educational level (basic secondary degree/secondary degree and (technical) university entrance qualification) at the individual level. In all age groups and in both sexes, the prevalence of obesity, which increased overall with age, was higher in individuals with lower educational level. For example, the point estimates of prevalence among women aged 30–39 years were approximately 9 percentage points higher than of those with higher education (22% vs. 13%). In the same age group, the difference for men was about 5 percentage points (12% vs. 7%). Fig. 7 (Attachment 1 [Attach. 1]) illustrates the comparative obesity prevalence for all categories including education unknown, no formal education, and persons without any information on educational level. Fig. 8 (Attachment 1 [Attach. 1]) also compares obesity prevalence for the two groups with and without information on education. This shows a largely consistent prevalence in both groups for men, except for 18- to 29-year-olds (lower obesity prevalence in the group without information on education). For women up to the age of 50, a pattern similar to that of men emerges. Older women show a higher obesity prevalence in the group with information on education than in women without such information.

In the sensitivity analyses classifying the educational status into three categories (basic secondary degree, secondary degree, and (technical) university entrance qualification), 27% of the insured could be included (24% of men, 29% of women). There was an inverse relationship between obesity prevalence and educational level in all subgroups under 80, with a smaller difference between the categories secondary degree vs. basic secondary degree (e.g., 5% percentage points among 30–39-year-old women) than between the categories secondary degree vs. (technical) university entrance qualification (8 percentage points among 30- to 39-year-old women) (Figure 4 [Fig. 4]). The maximum difference in obesity prevalence between persons with a basic secondary school diploma vs. persons with a (technical) university entrance qualification was 13 percentage points and was observed in the subgroup of 30- to 39-year-old women.

The results on the association between the obesity prevalence observed in GePaRD and the socioeconomic status estimated by means of the spatial deprivation index of the place of residence (county level) are shown in Figure 5 [Fig. 5]. Again, there is an increase in obesity prevalence with increasing deprivation in all age groups and for both sexes. The differences were less pronounced compared to the analysis based on the individually estimated socioeconomic status. The difference between the subgroups with the highest and lowest deprivation, respectively, is about 8 percentage points among 30–39-year-old women (21% vs. 13%) and about 4 percentage points among men in this age group (11% vs. 7%).

Figure 6 [Fig. 6] illustrates the results of the comparative analysis based on the GEDA 2014/2015 dataset. This dataset contains information on educational attainment at the individual level, i.e., the results can be viewed in comparison with the GePaRD results shown in Figure 3 [Fig. 3]. Compared with GePaRD, some subgroups – particularly those in older age groups – showed somewhat different, often lower, obesity prevalences. However, overall, the patterns were similar in terms of differences between educational groups. In tendency, the absolute difference in obesity prevalence between education groups was more pronounced for GEDA and showed fewer sex differences than for GePaRD. For example, among women aged 40–49, the difference between the higher and lower education groups was 7 percentage points (GePaRD: 7 percentage points) and among men aged 40–49, the difference was 8 percentage points (GePaRD: 4 percentage points).


The results of this study show that based on individual information available in the Pharmacoepidemiological Research Database GePaRD, it is possible to classify about three quarters of the insured into a higher or lower educational category as an approximation of the socioeconomic status. The plausibility of this classification was confirmed by the reproduction of the known association between obesity and socioeconomic status based on this classification. Overall, the estimation of the socioeconomic status based on the county of residence also rendered plausible results. However, the difference observed in these analyses between 20% of the insured with the lowest and the highest deprivation score, respectively (i.e., 60% of the insured were not considered in the comparison), was similar in magnitude to the difference observed in the individual estimation of the socioeconomic status based solely on a dichotomous classification of all insured with available information on education. The differences were even more pronounced when the educational level was divided into three categories, which also indicates that the discriminatory power of the estimated socioeconomic status based on individual information is greater than in the case of an estimate based on the county of residence.

The magnitude of the association between the presence of obesity and the socioeconomic status, as well as the differences by age and sex, were overall consistent with our results of the comparative analyses based on GEDA data. For some subgroups, the difference by socioeconomic status was slightly more pronounced in the GEDA than in the GePaRD analyses. For example, among those aged 50–59 years, the GEDA analyses showed a 2-fold (women) and 1.5-fold (men) increase in obesity prevalence for those with a basic secondary degree or secondary degree compared with the higher education group, whereas the GePaRD analyses showed a 1.4-fold increase in prevalence for men and women. However, when comparing the results one should bear in mind that some of the estimators in the GEDA analyses have wide confidence intervals. It should also be noted that full agreement was not to be expected simply because of the different ways in which obesity was assessed. In insurance data, diagnoses appear primarily if relevant for billing purposes, i.e., if services were provided in connection with the diagnosis. A sensitive case definition was applied in the GePaRD analyses, which may have led to an overestimation of prevalence in adulthood. In the GEDA definition of obesity, which is based on self-reported height and weight, there may have been misclassification for other reasons. For example, it is known that body weight and height are often misreported in surveys [17]. According to Maukonen et al., this misclassification leads to an underestimation of obesity prevalence by 0.6 to 8.4 percentage points in European studies [17]. Thus, none of the data sources contains gold standard information on the presence of obesity, and differences in the strength of the association should therefore not be over-interpreted. This generally also applies to the comparison with other studies, which, however, show similar overall associations between the individually assessed socioeconomic status and the presence of obesity [6], [7], [9].

The estimation of the individual socioeconomic status in our analyses was based on information related to the occupation key available in GePaRD. In principle, insurance data contain information that would facilitate an even more detailed estimate of the socioeconomic status, such as the premium level, which is indicative of income, or the exact place of residence. However, since this is very sensitive information, it is generally not available for secondary data use outside the health insurance provider. Most of the previous studies based on health insurance data taking the socioeconomic status into account therefore used (rough) information on the place of residence. An older study by Geyer based on data of 416,000 insured persons from the insurance provider AOK Mettmann from 1987 to 1996 already illustrated the potential of using information on education and occupation to estimate the individual socioeconomic status in insurance data [1], but this was given little consideration in more recent studies. On the one hand, this could be because the data analysts did not have access to the occupation key. On the other hand, it is also widely believed that the information is not available with the necessary completeness or unambiguity [18]. We were able to significantly increase completeness in the GePaRD analyses by exploiting the long lookback period in combination with the linkage of family members to primary insurance holders. Furthermore, we made assumptions, for example, that the socioeconomic status may be transferred from the primary insurance holder to family members. While this assumption is controversial in the literature [19], [20], [21], we believe it is justified and similar in terms of underlying assumptions to the use of family income in studies based on primary data.

With the GePaRD data source in combination with the methodology applied here, it was possible to roughly estimate the individual socioeconomic status (basic secondary degree/secondary degree and (technical) university entrance qualification) for 86–93% of those under 60. Even among 60- to 69-year-olds, this proportion was still 67% (men: 72% women: 63%). Undoubtedly and for several reasons, this is not a perfect classification of the socioeconomic status. Apart from the underlying assumptions, one cannot generally presume that the occupation keys reported by employers are completely free of errors [11]. Moreover, the educational level does not cover all dimensions of the socioeconomic status [22]. However, this rough estimate of the socioeconomic status is still of great value for numerous descriptive or analytical questions based on insurance data. This is particularly relevant since it is also an important surrogate parameter for other factors poorly represented in insurance data (e.g., lifestyle factors). For example, the socioeconomic status estimated in this way can be used for stratification and adjustment in order to identify or account for its relevance as an influencing or confounding factor, at least to some extent. When considering the limitations mentioned above, it should be noted that estimating the socioeconomic status based on primary data is often also suboptimal due to missing information, among other reasons.

When using the GISD to estimate the socioeconomic status, the known association between low socioeconomic status (or high deprivation) and obesity was weaker than when the socioeconomic status was estimated individually in three categories, which was not surprising [23], [24]. Since the place of residence in GePaRD is available at the county level, we also used the GISD with this spatial reference. This results in 401 different regional units for Germany. This classification is very rough and does not take into account the differences in socioeconomic status between individuals in a region, i.e., considerable misclassification must be assumed. Accordingly, the estimation of the individual socioeconomic status based on GePaRD is preferable to the GISD, both dichotomously and in three categories. The GISD could be considered as a second-best option if the individual socioeconomic status cannot be estimated, which is currently often the case for individuals aged 70 and older, as it is often impossible to look back at the period of employment for them. However, in the long term, i.e., with an even longer lookback period in GePaRD, the proportion with missing occupation keys will likely continue to decrease among the elderly as well. Similarly, it is to be expected that educational and occupational information according to the new classification system will be available for ever more individuals, i.e., in the future, it will be possible to classify young people in particular into three educational categories.

The comparison of obesity prevalence among individuals with vs. without information on education (Fig. 8 in Attachment 1 [Attach. 1]) shows differences between the two groups, particularly among 18- to 29-year-old men and women and among women aged 50 and older. In these groups, there appears to be a correlation between the presence of information on education and social status. In the case of 18- to 29-year-olds, this seems plausible as this group presumably includes students (higher educational level) who are still insured through their parents. Since some of the parents have already reached retirement age by that time resulting in missing occupation keys, there are also missing values for the co-insured family members. In contrast, persons undergoing vocational training (generally lower educational level) already earn a salary at this age and are self-insured, i.e., an occupation key is available. The pattern among older women may be explained by the fact that women with lower educational level are more often not employed and therefore co-insured with their husband, who is already retired (i.e., often lack an occupation key). Due to these correlations, the category “missing” should be considered separately in future analyses, especially with regard to 18- to 29-year-olds and women over 50.


Overall, the study results presented here indicate that the information available in GePaRD is well suited for the dichotomous estimation of the individual socioeconomic status for age groups up to 69 years, both in terms of completeness and plausibility. For individuals aged 70 and older, an estimation of the socioeconomic status based on the GISD could be used as a second-best option until a higher completeness can be achieved for these age groups in GePaRD as well – provided an even longer lookback period becomes available.



The authors would like to thank all statutory health insurance providers, which provided data for this study, namely AOK Bremen/Bremerhaven, DAK-Gesundheit, Die Techniker (TK), and hkk Krankenkasse.

Competing interests

Marieke Asendorf, Jonas Reinold, Tania Schink, Bianca Kollhorst, and Ulrike Haug are working at an independent, non-profit research institute, the Leibniz Institute for Prevention Research and Epidemiology – BIPS. Unrelated to this study, BIPS occasionally conducts studies financed by the pharmaceutical industry. Almost exclusively, these are post-authorization safety studies (PASS) requested by health authorities. The design and conduct of these studies as well as the interpretation and publication are not influenced by the pharmaceutical industry. The study presented was not funded by the pharmaceutical industry and was carried out in compliance with the ENCePP Code of Conduct.

The authors declare that they have no conflicts of interest related to this article.

ORCIDs of the authors


Geyer S. Sozialstruktur und Krankheit. Analysen mit Daten der Gesetzlichen Krankenversicherung [Social inequalities in health. Analysis using data from statutory health insurance companies]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2008 Oct;51(10):1164-72. DOI: 10.1007/s00103-008-0651-1 External link
Richter M, Hurrelmann K. Gesundheitliche Ungleichheit: Ausgangsfragen und Herausforderungen. In: Richter M, Hurrelmann K, editors. Gesundheitliche Ungleichheit. Wiesbaden: VS Verlag für Sozialwissenschaften; 2009. p. 11-31.
Wilkinson R, Marmot M. Soziale Determinanten von Gesundheit. Die Fakten. 2. Ausgabe. Kopenhagen: World Health Organization Europe; 2004.
Greiner W, Batram M, Damm O, Scholz S, Witte J. Kinder- und Jugendreport 2018 – Gesundheitsvorsorge von Kindern und Jugendlichen in Deutschland. Schwerpunkt: Familiengesundheit. Bielefeld & Hamburg; 2018.
Lamerz A, Kuepper-Nybelen J, Wehle C, Bruning N, Trost-Brinkhues G, Brenner H, Hebebrand J, Herpertz-Dahlmann B. Social class, parental education, and obesity prevalence in a study of six-year-old children in Germany. Int J Obes (Lond). 2005 Apr;29(4):373-80. DOI: 10.1038/sj.ijo.0802914 External link
Lampert T, Kroll LE, von der Lippe E, Müters S, Stolzenberg H. Sozioökonomischer Status und Gesundheit: Ergebnisse der Studie zur Gesundheit Erwachsener in Deutschland (DEGS1) [Socioeconomic status and health: results of the German Health Interview and Examination Survey for Adults (DEGS1)]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2013 May;56(5-6):814-21. DOI: 10.1007/s00103-013-1695-4 External link
Mensink GB, Schienkiewitz A, Haftenberger M, Lampert T, Ziese T, Scheidt-Nave C. Übergewicht und Adipositas in Deutschland: Ergebnisse der Studie zur Gesundheit Erwachsener in Deutschland (DEGS1) [Overweight and obesity in Germany: results of the German Health Interview and Examination Survey for Adults (DEGS1)]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2013 May;56(5-6):786-94. DOI: 10.1007/s00103-012-1656-3 External link
OECD. Health at a Glance. Paris: OECD Publishing; 2019.
Großschädl F, Stronegger WJ. Long-term trends (1973-14) for obesity and educational inequalities among Austrian adults: men in the fast lane. Eur J Public Health. 2019 Aug;29(4):790-6. DOI: 10.1093/eurpub/cky280 External link
Hajek A, Lehnert T, Ernst A, Lange C, Wiese B, Prokein J, Weyerer S, Werle J, Pentzek M, Fuchs A, Luck T, Bickel H, Mösch E, Heser K, Wagner M, Maier W, Scherer M, Riedel-Heller SG, König HH; AgeCoDe Study Group. Prevalence and determinants of overweight and obesity in old age in Germany. BMC Geriatr. 2015 Jul;15:83. DOI: 10.1186/s12877-015-0081-5 External link
Damm K, Lange A, Zeidler J, Braun S, Graf von der Schulenburg JM. Einführung des neuen Tätigkeitsschlüssels und seine Anwendung in GKV-Routinedatenauswertungen. Möglichkeiten und Limitationen [Implementation of the new German job role code and its application in claims data analysis. Possibilities and limitations]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2012 Feb;55(2):238-44. DOI: 10.1007/s00103-011-1418-7 External link
Bundesagentur für Arbeit. Schlüsselverzeichnis für die Angaben zur Tätigkeit in den Meldungen zur Sozialversicherung. Nürnberg: Bundesagentur für Arbeit; 2007.
Bundesagentur für Arbeit. Schlüsselverzeichnis für die Angaben zur Tätigkeit in den Meldungen zur Sozialversicherung. Ausgabe 2010. Stand: Mai 2017. Nürnberg: Bundesagentur für Arbeit; 2010.
Kroll LE, Schumann M, Hoebel J, Lampert T. Regionale Unterschiede in der Gesundheit – Entwicklung eines sozioökonomischen Deprivationsindex für Deutschland. J Health Monit. 2017;2(2):103-20. DOI: 10.17886/RKI-GBE-2017-035 External link
Robert-Koch-Insitut, Abteilung für Epidemiologie und Gesundheitsmonitoring. Gesundheit in Deutschland aktuell (GEDA 2014/2015-EHIS). Scientific Use File. 1. Version. Berlin: Robert-Koch-Insitut; 2015.
Saß AC, Lange C, Finger JD, Allen J, Born S, Hoebel J, Kuhnert R, Müters S, Thelen J, Schmich P, Varga M, von der Lippe E, Wetzstein M, Ziese T. „Gesundheit in Deutschland aktuell“ – Neue Daten für Deutschland und Europa Hintergrund und Studienmethodik von GEDA 2014/2015-EHIS. J Health Monit. 2017;2(1):83-90. DOI: 10.17886/RKI-GBE-2017-012 External link
Maukonen M, Männistö S, Tolonen H. A comparison of measured versus self-reported anthropometrics for assessing obesity in adults: a literature review. Scand J Public Health. 2018 Jul;46(5):565-79. DOI: 10.1177/1403494818761971 External link
Schubert I, Köster I, Küpper-Nybelen J, Ihle P. Versorgungsforschung mit GKV-Routinedaten. Nutzungsmöglichkeiten versichertenbezogener Krankenkassendaten für Fragestellungen der Versorgungsforschung [Health services research based on routine data generated by the SHI. Potential uses of health insurance fund data in health services research]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2008 Oct;51(10):1095-105. DOI: 10.1007/s00103-008-0644-0 External link
Geyer S. Die Bestimmung der sozioökonomischen Position in Prozessdaten und ihre Verwendung in Sekundärdatenanalysen. In: Swart E, Ihle P, Gothe H, Matusiewicz D, editors. Routinedaten im Gesundheitswesen Handbuch Sekundärdatenanalyse: Grundlagen, Methoden und Perspektiven. Bern: Verlag Hans Huber; 2005. p. 203-7.
Lidfeldt J, Li TY, Hu FB, Manson JE, Kawachi I. A prospective study of childhood and adult socioeconomic status and incidence of type 2 diabetes in women. Am J Epidemiol. 2007 Apr;165(8):882-9. DOI: 10.1093/aje/kwk078 External link
Muschik D, Jaunzeme J, Geyer S. Are spouses' socio-economic classifications interchangeable? Examining the consequences of a commonly used practice in studies on social inequalities in health. Int J Public Health. 2015 Dec;60(8):953-60. DOI: 10.1007/s00038-015-0744-1 External link
Lampert T, Kroll LE. Die Messung des sozioökonomischen Status in sozialepidemiologischen Studien. In: Richter M, Hurrelmann K, editors. Gesundheitliche Ungleichheit: Grundlagen, Probleme, Perspektiven. Wiesbaden: VS Verlag für Sozialwissenschaften; 2009. p. 309-34.
Jacobs E, Tönnies T, Rathmann W, Brinks R, Hoyer A. Association between regional deprivation and type 2 diabetes incidence in Germany. BMJ Open Diabetes Res Care. 2019;7(1):e000857. DOI: 10.1136/bmjdrc-2019-000857 External link
Kauhl B, Maier W, Schweikart J, Keste A, Moskwyn M. Exploring the small-scale spatial distribution of hypertension and its association to area deprivation based on health insurance claims in Northeastern Germany. BMC Public Health. 2018 Jan;18(1):121. DOI: 10.1186/s12889-017-5017-x External link
Grobe T, Ihle P. Versichertenstammdaten und sektorübergreifende Analyse. In: Swart E, Ihle P, Gothe H, Matusiewicz D, editors. Routinedaten im Gesundheitswesen. Handbuch Sekundärdatenanalyse: Grundlagen, Methoden und Perspektiven. Bern: Verlag Hans Huber; 2005. p. 17-34.
Wahrendorf M, Rupprecht CJ, Dortmann O, Scheider M, Dragano N. Erhöhtes Risiko eines COVID-19-bedingten Krankenhausaufenthaltes für Arbeitslose: Eine Analyse von Krankenkassendaten von 1,28 Mio. Versicherten in Deutschland [Higher risk of COVID-19 hospitalization for unemployed: an analysis of health insurance data from 1.28 million insured individuals in Germany]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2021 Mar;64(3):314-21. DOI: 10.1007/s00103-021-03280-6 External link