Access the full text.
Sign up today, get DeepDyve free for 14 days.
P. Loisel, R. Buchbinder, R. Hazard, R. Keller, I. Scheel, M. Tulder, B. Webster (2005)
Prevention of Work Disability Due to Musculoskeletal Disorders: The Challenge of Implementing EvidenceJournal of Occupational Rehabilitation, 15
(2019)
Marginal effects for model objects . Package ‘ margins ‘ . Version 0 . 3 . 26
M. Bethge, Miriam Markus, M. Streibelt, C. Gerlich, M. Schuler (2018)
Implementing the German Model of Work-Related Medical Rehabilitation: Did the Delivered Dose of Work-Related Treatment Components Increase?Archives of physical medicine and rehabilitation, 99 12
Rüdiger Nübling, R. Kaluscha, G. Krischak, D. Kriz, H. Martin, G. Müller, J. Renzland, M. Reuss-Borst, Jürgen Schmidt, U. Kaiser, E. Toepler (2017)
Ergebnisqualität medizinischer Rehabilitation: Zum Zusammenhang zwischen „Patient Reported Outcomes“ (PROs) und geleisteten SozialversicherungsbeiträgenDie Rehabilitation, 56
C. Hetzel, M. Streibelt (2016)
Hängt die berufliche Wiedereingliederung nach beruflichen Bildungsleistungen vom Arbeitsmarkt ab?Rehabilitation Die, 55
(2013)
Integrationserfolge hängen von regionalen Gegebenheiten ab : Neue Typisierung der Agenturbezirke
(2016)
Return to work after inpatient rehabilitation—alternative calculations on the basis of patient data and validation of social security contributions
A. Hayes (2013)
Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach
A. Young, R. Roessler, R. Wasiak, K. McPherson, M. Poppel, J. Anema (2005)
A Developmental Conceptualization of Return to WorkJournal of Occupational Rehabilitation, 15
Sarah Leinberger, C. Hetzel, R. Kaluscha (2023)
Adjustierung des sozialmedizinischen Verlaufs nach medizinischer Rehabilitation: methodische Weiterentwicklung der Reha-Qualitätssicherung der Deutschen RentenversicherungDie Rehabilitation, 62
R. Wedderburn (1974)
Quasi-likelihood functions, generalized linear models, and the Gauss-Newton methodBiometrika, 61
P. McCullagh (1983)
Quasi-Likelihood FunctionsAnnals of Statistics, 11
G. Krischak, L. Tepohl, J. Dannenmaier, Ulrich Hartschuh, R. Auer, R. Kaluscha (2018)
Gesundheitsökonomische Effekte der Rehabilitation bei chronischem Rückenschmerz – Eine Beobachtungsstudie mittels kombinierten Sekundärdaten einer gesetzlichen Krankenkasse und der Deutschen RentenversicherungDie Rehabilitation, 58
(2016)
Rehabilitation and return to work: Analysis report on EU and Member States policies, strategies and programmes. Publications Office of the European Union
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations
Viktor Steiner (2017)
The labor market for older workers in GermanyJournal for Labour Market Research, 50
(2017)
Outcome quality in medical rehabilitation: relationship between „patient-reported outcomes“ (PROs) and social security contributions
G Krischak, L Tepohl, J Dannenmaier, U Hartschuh, R Auer, R Kaluscha (2019)
Health economic effects of rehabilitation for chronic back pain—an observational study with combined secondary data from statutory health insurance and German pension fundDie Rehabil., 58
M. Kolmar (2007)
Beveridge versus Bismarck public-pension systems in integrated marketsRegional Science and Urban Economics, 37
N. Reims, S. Tophoven (2018)
Eine Längsschnittstudie zu Arbeitsmarktübergängen und Beschäftigungsnachhaltigkeit nach beruflichen Weiterbildungsmaßnahmen im Rahmen von Leistungen zur Teilhabe am ArbeitslebenDie Rehabilitation, 57
(2019)
Die medizinische und berufliche Rehabilitation der Rentenversicherung im Licht der Statistik, Reha-Bericht
N. Bülau, Franziska Kessemeier, F. Petermann, M. Bassler, A. Kobelt (2016)
Evaluation von Kontextfaktoren in der psychosomatischen RehabilitationRehabilitation Die, 55
Lene Odgaard, Asger Pedersen, Ingrid Poulsen, S. Johnsen, Jørn Nielsen (2018)
Return to work predictors after traumatic brain injury in a welfare stateActa Neurologica Scandinavica, 137
S. Howard-Wilsher, L. Irvine, H. Fan, T. Shakespeare, M. Suhrcke, S. Horton, F. Poland, L. Hooper, F. Song (2016)
Systematic overview of economic evaluations of health-related rehabilitation.Disability and health journal, 9 1
T. Leinonen, S. Solovieva, K. Husgafvel‐Pursiainen, M. Laaksonen, E. Viikari-Juntura (2019)
Do individual and work-related factors differentiate work participation trajectories before and after vocational rehabilitation?PLoS ONE, 14
A. Fedorets, Franziska Lottmann, Michael Stops (2019)
Job matching in connected regional and occupational labour marketsRegional Studies, 53
C. Hetzel (2015)
Mehrebenenanalysen zu Arbeitsmarkt und Wiedereingliederung nach beruflichen BildungsleistungenDie Rehabilitation, 54
J. Jansen, R. Ooijen, P. Koning, C. Boot, S. Brouwer (2021)
The Role of the Employer in Supporting Work Participation of Workers with Disabilities: A Systematic Literature Review Using an Interdisciplinary ApproachJournal of Occupational Rehabilitation, 31
Barbara Petrongolo, C. Pissarides (2001)
Looking into the black box: a survey of the matching functionLSE Research Online Documents on Economics
(2018)
Rd.: Multilevel analysis: Techniques and applications. Quantitative methodology series
J. Ervasti, M. Joensuu, J. Pentti, T. Oksanen, K. Ahola, J. Vahtera, M. Kivimäki, M. Virtanen (2017)
Prognostic factors for return to work after depression-related work disability: A systematic review and meta-analysis.Journal of psychiatric research, 95
N Reims, S Tophoven (2018)
A longitudinal study on labour market transitions and sustainability of employment after further training measures in the context of vocational rehabilitationDie Rehabil., 57
Leslie Papke, J. Wooldridge (1993)
Econometric Methods for Fractional Response Variables with an Application to 401(K) Plan Participation RatesNBER Working Paper Series
M. Zeisberger, D. Nowik, L. Beck, Stefanie Märtin, U. Beckmann, T. Meyer (2019)
Quality management in medical rehabilitative care by the German statutory pension insurance scheme.European journal of physical and rehabilitation medicine
(2013)
Beeinflusst die Arbeitslosenquote die (Wieder-)Eingliederung in das Erwerbsleben nach medizinischer Rehabilitation
Alison Stewart, Emily Polak, R. Young, I. Schultz (2012)
Injured Workers’ Construction of Expectations of Return to Work with Sub-Acute Back Pain: The Role of Perceived UncertaintyJournal of Occupational Rehabilitation, 22
(2001)
International Social Security Association: Who returns to work and why? A sixcountry study on work incapacity and reintegration
A. Cieza, K. Causey, K. Kamenov, S. Hanson, S. Chatterji, T. Vos (2020)
Global estimates of the need for rehabilitation based on the Global Burden of Disease study 2019: a systematic analysis for the Global Burden of Disease Study 2019Lancet (London, England), 396
N. Echarti, E. Schüring, C. O’Donoghue (2019)
Effects of Vocational Re-training on Employment Outcomes Among Persons with Disabilities in Germany: A Quasi-ExperimentJournal of Occupational Rehabilitation, 30
R. Breen, K. Karlson, Anders Holm (2018)
Interpreting and Understanding Logits, Probits, and Other Nonlinear Probability ModelsAnnual Review of Sociology
Mollie Brooks, K. Kristensen, K. Benthem, A. Magnusson, Casper Berg, Anders Nielsen, H. Skaug, M. Mächler, B. Bolker (2017)
glmmTMB Balances Speed and Flexibility Among Packages for Zero-inflated Generalized Linear Mixed ModelingR J., 9
G. Miyamoto, C. Lin, C. Cabral, J. Dongen, M. Tulder (2018)
Cost-effectiveness of exercise therapy in the treatment of non-specific neck pain and low back pain: a systematic review with meta-analysisBritish Journal of Sports Medicine, 53
C. Hetzel, M. Streibelt (2016)
[Does Return to Work after Vocational Retraining Depend on Labour Market?]Die Rehabilitation, 55 5
C. Hetzel (2015)
[Multilevel analyses of labour market and return to work after vocational retraining].Die Rehabilitation, 54 1
(2018)
Costeffectiveness of cardiac rehabilitation: a systematic review
T. Berg, M. Schuring, M. Avendano, J. Mackenbach, A. Burdorf (2010)
The impact of ill health on exit from paid employment in Europe among older workersOccupational and Environmental Medicine, 67
N. Nevala, I. Pehkonen, A. Teittinen, H. Vesala, Pia Pörtfors, H. Anttila (2019)
The Effectiveness of Rehabilitation Interventions on the Employment and Functioning of People with Intellectual Disabilities: A Systematic ReviewJournal of Occupational Rehabilitation, 29
R. Nübling, R. Kaluscha, G. Krischak, D. Kriz, H. Martin, G. Müller, J. Renzland, M. Reuss-Borst, J. Schmidt, U. Kaiser, E. Toepler (2016)
Return to Work nach stationärer Rehabilitation – Varianten der Berechnung auf der Basis von Patientenangaben und Validierung durch Sozialversicherungs-BeitragszahlungenPhysikalische Medizin, Rehabilitationsmedizin, Kurortmedizin, 26
Anna-Sophia Celsing, K. Svärdsudd, H. Eriksson, K. Björkegren, M. Eriksson, T. Wallman (2012)
Determinants for return to work among sickness certified patients in general practiceBMC Public Health, 12
U. Blien (2001)
Arbeitslosigkeit und Entlohnung auf regionalen Arbeitsmärkten
Background The influence of both individual factors and, in particular, the regional labour market on the return to work after medical rehabilitation is to be analyzed based on comprehensive administrative data from the German Pension Insurance and Employment Agencies. Method For rehabilitation in 2016, pre- and post-rehabilitation employment was determined from German Pen- sion Insurance data for 305,980 patients in 589 orthopaedic rehabilitation departments and 117,386 patients in 202 psychosomatic rehabilitation departments. Labour market data was linked to the district of residence and categorized into 257 labour market regions. RTW was operationalized as the number of employment days in the calendar year after medical rehabilitation. Predictors are individual data (socio-demographics, rehabilitation biography, employment biography) and contextual data (regional unemployment rate, rehabilitation department level: percentage of patients employed before). The estimation method used was fractional logit regression in a cross-classified multilevel model. Results The effect of the regional unemployment rate on RTW is significant yet small. It is even smaller (orthopae - dics) or not significant (psychosomatics) when individual employment biographies (i.e., pre-rehabilitation employ- ment status) are inserted into the model as the most important predictors. The interaction with pre-rehabilitation employment status is not substantial. Conclusions Database and methods are of high quality, however due to the nonexperimental design, omitted variables could lead to bias and limit causal interpretation. The influence of the labour market on RTW is small and proxied to a large extent by individual employment biographies. However, if no (valid) employment biographies are available, the labour market should be included in RTW analyses. Keywords Return to work, Labour market, Rehabilitation, Orthopaedic, Psychosomatic JEL Classification I130, J210, J140 *Correspondence: Christian Hetzel hetzel@iqpr.de Institute for Quality Assurance in Prevention and Rehabilitation at the German Sport University in Cologne, Eupener Str. 70, 50933 Cologne, Germany Institute for Research in Rehabilitation Medicine at Ulm University, Bad Buchau, Germany Federal German Pension Insurance, Berlin, Germany © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. 4 Page 2 of 14 C. Hetzel et al. outpatient or inpatient medical rehabilitation (§ 40 SGB 1 Introduction V) if treatment interventions alone are not adequate to 1.1 Rehabilitative interventions cope with the consequences of illness. To counteract the For people with health restrictions or disabilities, par- effects of a psychological or physical disorder on earn - ticipation in working life is an important individual ing capacity, the GPI provides benefits for medical and and social goal. One means of achieving this is through vocational rehabilitation (§ 9, § 10 SGB VI). For exam- rehabilitation. The WHO defines rehabilitation as “a set ple, if certain formal prerequisites are met, a patient is of interventions designed to optimize functioning and entitled to apply for medical rehabilitation at the GPI. If reduce disability in individuals with health conditions this application is confirmed by the GPI, the patient has in interaction with their environment” (Word Health the legal right to take part in a medical rehabilitation Organization 2020). In 2019, approximately 2.41 billion programme. The fundamental principle applies: reha - individuals worldwide had health conditions that would bilitation has priority over pension. Before someone can benefit from rehabilitation, which counters the common receive a pension owning to reduced earning capacity, view of rehabilitation as a service for the few. This num - the GPI will check if rehabilitation can be carried out. ber increased by 63 percent from 1990 to 2019. The ill - The GPI is the main provider of medical rehabilitation in ness category constituting the largest share of this figure Germany with over 1 million medical rehabilitation pro- was musculoskeletal disorders (approximately 1.71 bil- grammes in 2018 and approximately 5 billion euros spent lion people), with lower back pain being the most preva- in this sector (Deutsche Rentenversicherung Bund 2019). lent condition in 134 of the 204 countries analyzed (Cieza Vocational rehabilitation programmes can be provided et al. 2020). It is evident that many rehabilitative inter- on their own or as a supplement to a completed medi- ventions are cost-effective (Howard-Wilsher et al. 2016; cal rehabilitation programme. There are programmes Shields et al. 2018; Krischak et al. 2019; Miyamoto et al. designed to keep individuals in their job, but there are 2019), because chronification can be counteracted and also education and training interventions designed to the ability to work can be maintained. offer entirely new career prospects. In view of the demographic change and prolonged How much is income replacement during medical working lives, the proportion of older employees is rehabilitation? In principle, employees are entitled to a increasing in most economies, and thus also the number continued payment of their wage or salary for up to six of employees with poor health and functional limitations weeks due to illness. If there is no longer any entitlement (van den Berg et al. 2010). Therefore, one major public to this, the GPI pays transitional benefits (68 percent of health goal should be avoiding premature termination of the net wage) for that time. Individuals are obliged to work due to poor health using primary prevention, reha- pay compulsory contributions into the GPI, in particu- bilitation and RTW strategies. These factors will gain rel - lar if they receive a wage or salary including short-term evance in working life, as for example the aim of medical sickness. People in marginal employment, the long-term rehabilitation is a continuous participation in working unemployed and old-age-pensioners are exempt from life. Understanding which factors favour or slow down paying compulsory contributions. However, voluntary rehabilitation outcomes can help to develop appropri- contributions can be paid. The number of employment ate political concepts to support people return to work days referred to in this article is based on an individual (RTW) after a health shock (Young et al. 2005). From a receiving a wage or salary above the marginal employ- social point of view, social security contribution pay- ment level including short-term sickness and therefore ments are maintained. these individuals pay compulsory contributions into the GPI. 1.2 I nstitutional background in Germany Implemented rehabilitation programmes vary signifi - 1.3 Return to work after medical rehabilitation cantly from country to country (International Social One of the main goals of rehabilitative interventions Security Association 2001; Belin et al. 2016). In contrast financed by the GPI is the partial or complete (re-)inte - to outpatient interventions in other countries, in Ger- gration into working life (Deutsche Rentenversicherung many interventions are mostly conducted as 3-week or Bund 2019). However, the patients’ RTW depends not 4-week inpatient programmes in specialized rehabilita- only on the quality of the rehabilitation programme itself tion departments, but increasingly as an outpatient or but also on personal and contextual factors. There has semi-inpatient, close to the patient’s home. Rehabilitation been surprisingly little research into this area and con- is a part of the social insurance system, mostly provided ceptual and empirical data to substantiate this finding is by the German statutory pension insurance (GPI), Ger- rather inconsistent. This is especially true for the influ - man statutory health insurance or private health insur- ence of the regional labour market on RTW after medical ance. The German statutory health insurances provide Return to work after medical rehabilitation in Germany: influence of individual factors and… Page 3 of 14 4 rehabilitation—in the indication area of “depression”, by systemic conditions such as unemployment benefits there has been an explicit demand for further research to or early retirement (International Social Security Asso- be conducted (Ervasti et al. 2017). ciation 2001; Belin et al. 2016). Individual characteristics The regional labour market is only partially taken into could also obscure existing effects if they are correlated consideration as an influencing factor in RTW concepts, to the labour market (e.g., recent periods of unemploy- e.g., as an environmental contextual factor of the Inter- ment (Celsing et al. 2012)). If the focus is on returning to national Classification of Functioning, Disability and a job that existed before rehabilitation, job-related rather Health, as a facet of the theory of perceived insecurity than labour market-related characteristics are likely to be (Stewart et al. 2012) or as the macro level in system-theo- significant predictors (International Social Security Asso - retical models (Loisel et al. 2005). The matching theory of ciation 2001; Jansen et al. 2021). supply and demand on labour markets (Petrongolo and From a research perspective, controlling the effects Pissaridēs 2001) also implicitly postulates a connection of confounding variables is important to determine between the regional labour market and employment. intervention effects with minimal bias and to enable The theory describes that employers’ wage offers par - fair benchmarking opportunities between regions or tially depend on the wage-lowering effect of high regional providers, for example in the context of GPI quality unemployment (Müller and Blien 2001). management (Zeisberger et al. 2019). To achieve fair Empirical studies take the regional labour market into benchmarking opportunities, the effect of character - account to some extent. Studies from unemployment istics for which the rehabilitation department is not contexts (Hirschenauer 2013), vocational rehabilitation responsible (e.g., the labour market at the patient’s place (Hetzel and Streibelt 2016; Reims and Tophoven 2018; of residence) are to be considered in the assessment of Echarti et al. 2020) and medical rehabilitation (Kalus- treatment outcome (risk adjustment). If RTW is substan- cha et al. 2013) indicate that the regional labour market tially influenced by the labour market, omitting this char - decreases RTW. In these contexts, the labour market acteristic would lead to bias. Another reason to include is operationalized either indirectly via assignment to a the labour market is the limited availability of confound- regional unit (Leinonen et al. 2019; Echarti et al. 2020) ing variables in secondary data. Modeling labour market or directly via characteristics of the regional labour effects on RTW could substitute unobserved character - market, such as the unemployment rate (Kaluscha et al. istics, such as employment biography or job conditions 2013; Hetzel 2015; Hetzel and Streibelt 2016; Reims (Bülau et al. 2016), and thus increase the model quality. and Tophoven 2018). In our opinion, there is evidence All in all, unlike vocational rehabilitation, it is unclear for vocational rehabilitation (Hetzel and Streibelt 2016; whether there are labour market effects on the RTW in Reims and Tophoven 2018; Echarti et al. 2020), but not the context of medical rehabilitation. Evidence in this for medical rehabilitation. To the best of our knowledge, field can help to determine intervention effects with min - only Kaluscha et al. (2013) use labour market conditions imal bias and to enable fair benchmarking opportunities. predicting RTW in medical rehabilitation. Based on ran- domly selected administrative data of the GPI between 2002 and 2009, they explore labour market effects using 1.4 Objective and hypotheses federal states in Germany as regional units and unem- The objective of the present paper is to determine the ployment rates with different standardizations. They influence of the regional unemployment rate on RTW conducted an extensive data-driven selection of 12 pre- after medical rehabilitation for the two largest indica- dictors. The result was that the labour market improved tors—orthopaedics and psychosomatics—using full model fit in some models, but a clear operationalization administrative data from the GPI based on a large num- and estimation of the size of an effect was not possible. ber of observations. The following hypotheses will be Other studies neglect to consider the regional labour tested (Fig. 1). market (International Social Security Association 2001; Celsing et al. 2012; Howard-Wilsher et al. 2016; Ervasti H1. A higher regional unemployment rate lowers et al. 2017; Odgaard et al. 2018; Nevala et al. 2019). Either RTW. there is no labour market effect or an existing effect is not modeled in an appropriate way. This could possibly This expected negative effect is based on the con - be explained by the fact that intervention studies are cepts and empirical findings described above. In con - often realized within a small geographical range, which sistency with matching theory, it has to be examined means that there is little variance in the regional labour whether an exponential correlation exists (Petrongolo market and thus its influence might remain undetected. and Pissaridēs 2001). In addition, the labour market effect could be masked 4 Page 4 of 14 C. Hetzel et al. H1-3 effect of the unemployment rate compared to the labour market models in H1 and H2. H4 H4. The labour market effect on RTW is more evident for patients who were unemployed before the reha- employment biography bilitative intervention than for previously employed RTW patients (interaction effect). socio-demographics and An interaction effect (Hayes 2013) describes that rehabilitaon biography the influence of a predictor on the outcome vari - able depends on other confounding variables. After rehabilitaon department vocational rehabilitation, the regional labour market was found to moderate the influence of the prereha - Fig. 1 Eec ff t relationships and hypotheses (H) bilitation employment status on RTW (Hetzel and Streibelt 2016). There was a clearer dependency of RTW on labour market conditions among prereha- H2. The main effect of the regional unemployment bilitation unemployed patients than among prereha- rate on RTW is independent of socio-demographics bilitation employed patients. and rehabilitation biography. The main effect describes the extent to which the regional labour market influences the outcome vari - 2 Data, variables, and method able directly, without the influences of other vari - 2.1 Data ables. For example, older persons might have more We use the rehabilitation statistics database (RSD) of RTW restrictions than younger persons (Steiner the GPI (Deutsche Rentenversicherung 2015). The RSD 2017), but it is independent of labour market con- contains administratively produced data from all the ditions. Therefore, the main effect of the regional GPI institutions, in this case for medical rehabilitation unemployment rate should be similar to H1, when programmes for the two largest indicators, orthopaedics socio-demographics and rehabilitation biographies and psychosomatics, in 2016. The database also contains are accounted for. all wages or salaries with social insurance contributions H3. The main effect of the regional unemployment until 2017. The individual rehabilitation programme is rate on RTW decreases when the employment biogra- linked to a certain rehabilitation department. We linked phy is controlled for. the regional unemployment rate for the patients’ labour The independency described in H2 should be differ - market region (Bundesinstitut für Bau-, Stadt- und ent for the characteristics of the individual employ- Raumforschung 2020) to their place of residence. These ment biography. This is because they depend on past regional units (n = 257 labour market regions in Ger- labour market conditions, which are also reflected in many) are homogenized based on commuter interde- the current labour market. Therefore, employment pendences (31.12.2015), i.e., there is a lot of commuting biography should already include part of the model’s within a region and less commuting between regions. For labour market effect, resulting in a reduced main the number of cases and exclusion criteria, see Table 1. Table 1 Number of cases and exclusion criteria Orthopaedics Psychosomatics Patients who completed medical rehabilitation in 2016 (the last rehabilitation if more than one) 360,285 131,404 − Exclusion based on theoretical considerations 52,653 14.6% 13,458 10.2% Duration of rehabilitation: shorter than 7 days Special types of rehabilitation: contract performance , aftercare, prevention, cancer, detoxification At the time of application: old-age pensioner, housewife/husband, not in working age (between 18 and 65 years), residence abroad/unknown In the year prior to application: no compulsory contributions to the GPI In the observation period: death − Exclusion because of missing data 1652 0.5% 560 0.4% = Final database 305,980 84.9% 117,386 89.3% a b For example cancer rehabilitation for the statutory health insurance, for example eligible long-term unemployed patients Return to work after medical rehabilitation in Germany: influence of individual factors and… Page 5 of 14 4 Persons with missing data are 0.4 percent in psychoso- • M1: cross-classified multilevel model with regional matics and 0.5 percent in orthopaedics. We made a com- unemployment rate only plete case analysis and did not impute data, especially as • M2: + patients’ personal characteristics and rehabili- some variables, e.g., place of residence, cannot be reliably tation biography estimated. • M3: + patients’ employment biography • M4: + interaction "employment status 3 months 2.2 Variables before" with "regional unemployment rate” The analyzed outcome is the patients’ RTW, operational - ized by the number of employment days in the first cal - In M1, we additionally test nonlinear associations. endar year after medical rehabilitation. Therefore, we use squaring and logarithmising, respec - We use the following predictors: tively. The squaring operationalizes a u-shaped relation - ship between the unemployment rate and the outcome • individual characteristics and rehabilitation biog- variable, and the logarithmising an inverse exponential raphy of the patient: (1) gender, (2) age (categories), relationship. (3) marital status, (4) migration, (5) vocational edu- The outcome variable has a bimodal distribution in cation, (6) place of residence in the area of former the interval between 0 and 365 days (see Fig. 2). Linear West Germany or the newly-formed German states regressions would thus lead to distortions and a classifi - (7), current intervention is post-hospital curative cation or dichotomization of the outcome variable would treatment (AHB) , (8) current intervention is a spe- lead to loss of information. Therefore, we use fractional cial medical programme , (9) additional payment logit regression (FLR) as an estimation method. FLR claim after an individual income check (§ 32 SGB VI), models assume that the outcome variable is a proportion (10) application for or receipt of pension for reduced of the interval from 0 to 1, so we made a linear transfor- earning capacity in temporal relation to the rehabili- mation of the outcome (dividing by 365 days). FLR was tation programme, and (11) number of rehabilitation first described by Wedderburn (1974), generalized by programmes in the 4 years prior to the intervention. McCullagh (1983) and rediscovered by Papke and Wool- • individual factors governing employment biography: dridge (1996). FLR belongs to the family of generalized (1) prerehabilitation employment status (employed/ linear models, is based on quasi-maximum likelihood not employed in the 3rd month before rehabilitation estimators, and is very similar to binary logistic regres- start), (2) employment days in the first calendar year sion. The basis of FLR is a binomial distribution, but with and (3) in the second calendar year prior to the reha- an additional parameter to estimate the deviatoric error bilitation programme (both in categories). variance in the data. This makes the method very flexible • labour market: regional unemployment rate at place and does not require any special distributional assump- of residence (in percent, not centred). tions. At the core is the logit link function, as in a binary • rehabilitation department: percentage of patients logistic regression. with prerehabilitation employment status "employed" We report logits (b-coefficients) where negative logits (in percent). describe a negative association. To quantify the effect size of the labour market, we use average marginal effects (AME) to describe the average effect of the labour mar - ket as the mean of the marginal effects across all obser - 2.3 Method vations. AME are reported in the unit of the outcome The data structure leads to dependencies, as patients are variable (days). For model fit, we use AIC and Pseudo- simultaneously categorized in both departments and in R (the squared correlation of expected and observed labour market regions. Therefore, we use a cross-classi - values). fied multilevel model (Hox et al. 2018). To estimate the main effects (H1-3) and the interaction effect (H4), the 3 Results models (M) are configured in blocks as follows: 3.1 Descriptive results The database contains 257 labour market regions, 589 rehabilitation departments and 305,980 individuals for AHB is a full-day outpatient or inpatient medical rehabilitation program. It is only considered in the case of certain illnesses and immediately follows inpatient hospital treatment (2 weeks after discharge at the latest). 2 3 Special medical program is work-related (medical-vocational oriented We use the R package glmmTMB for the regression modelling (Brooks et al. rehabilitation = MBOR), behavioral (behavioral orthopaedics = VMO, 2017). AME for the regional unemployment rate are reported from the single- behavioral rehabilitation = VOR) or other. level model using the R package margins (Leeper 2021)). 4 Page 6 of 14 C. Hetzel et al. Fig. 2 Number of employment days in the first calendar year after medical rehabilitation. See text for details orthopaedics, and 202 rehabilitation departments and H4: The interaction of the unemployment rate with 117,386 individuals for psychosomatics. For a description prerehabilitation employment status is ambiguous (M4). of the data, see Table 2. The RTW shows a bimodal dis - Arguments for including the interaction in the model are tribution (Fig. 2). This means that the majority of patients the significant interaction effect and the more favourable either work 0 days or 365 days in the first calendar year AIC in M4. Arguments against inclusion are that the AIC after medical rehabilitation. This applies to both indica - is only marginally lower than in M3, that R is the same tion areas. in M3 and M4 and that the effect size of the interaction effect is low. The latter can particularly be seen in Fig. 3. Adding the interaction hardly changes the curve for the 3.2 Orthopaedic rehabilitation two groups, as they are still almost parallel. Apparently, The models for orthopaedics are described in Table 3. the unemployment rate and the employment days are Without control variables (M1), the main effect of the independent. The corresponding coefficients of M4 are unemployment rate is significant and in the expected shown in Table 3: for patients employed prerehabilita- direction, but weak: a 1 percent point higher unem- tion, regional unemployment rate has a minimal effect ployment rate reduces employment days by 3.1 days in (b = − 0.013, p = 0.001, AME = − 0.4 days) and, on the the first calendar year after rehabilitation—averaged contrary, the effect for unemployed patients is significant over all observations in the present sample (H1). Test- (b = 0.011, p = 0.011). It should be emphasized that the ing for nonlinear relationship of the regional unem- interaction effect is very small. ployment rate (UR) with the outcome variable shows The patients’ employment biography dominates the that there is no u-shaped relationship (UR: b = − 0.042, model, as seen from the clear shift of model R from s.e. = 0.016, p = 0.009; UR : b = 0.001, s.e. = 0.001, model 2 to 3. The random effects are close to zero, indi - p = 0.715). Moreover, the relationship is linear rather cating that both rehabilitation departments and the than exponential because the log UR is significant (log labour market are well operationalized. This means that UR: b = − 0.215 s.e. = 0.021, p < 0.001), but the model fit the variance of the outcome variable which is due to the is slightly worse (AIC = 369,391) than for M1. grouping of individuals into departments or labour mar- As expected, by adding the socio-demographics and ket is almost completely modelled. rehabilitation biography (M2), the effect sizes remain very similar (H2). 3.3 Psychosomatic rehabilitation By adding the employment biography (M3), the main The findings for psychosomatic rehabilitation are very effect of the regional unemployment rate still is signifi - similar to those for orthopaedics. In contrast, the find - cant, but weaker (AME only − 0.2 days) (H3). ings for H3 are somewhat clearer, because the labour Return to work after medical rehabilitation in Germany: influence of individual factors and… Page 7 of 14 4 Table 2 Description of patients in orthopaedics and psychosomatics in 2016 Orthopaedics (N = 305,980) Psychosomatics (N = 117,386) Gender Not female 50.7% 36.7% Female 49.3% 63.3% Age, in years < 25 0.9% 1.1% 25–30 2.3% 2.9% 31–35 3.4% 4.7% 36–40 5.0% 6.8% 41–45 7.6% 9.6% 46–50 14.8% 17.5% 51–55 23.0% 23.7% 56–60 26.0% 22.7% 61–65 17.0% 11.1% Marital status Married 67.7% 58.8% Single 15.4% 19.1% Divorced 13.0% 16.3% Widowed 2.6% 3.5% n.a 1.4% 2.4% Migration (country of birth* nationality) Germany*German 87.8% 88.9% Other*other 3.4% 3.2% Other*German 6.2% 5.5% Germany*other 2.5% 2.5% Vocational education No 37.8% 33.2% Yes 62.2% 66.8% Region (German states) Former West Germany 81.6% 85.1% New states in Germany 18.4% 14.9% Post-hospital curative treatment No 64.7% 100.0% Yes 35.3% 0.0% Special medical programmes Normal 87.0% 75.8% Work-related 8.5% 23.6% Behavioral 3.6% 0.0% Other 0.9% 0.6% Additional payment claim No 34.9% 15.5% Yes 65.1% 84.5% Application for reduced earning capacity pension No 97.6% 91.1% Yes 2.4% 8.9% Number of prior rehabilitations in the last 4 years 0 76.7% 81.2% 1 13.4% 12.8% 2 7.8% 4.9% 3 2.1% 1.0% 4 Page 8 of 14 C. Hetzel et al. Table 2 (continued) Orthopaedics (N = 305,980) Psychosomatics (N = 117,386) Prerehabilitation employment status (3rd month before) Employed 77.0% 57.0% Not employed 23.0% 43.0% Employment days one calendar year before < 50 8.4% 13.8% 50–99 1.6% 3.2% 100–149 1.7% 3.4% 150–199 2.6% 4.6% 200–249 3.5% 5.4% 250–299 4.9% 6.0% 300–349 8.1% 9.0% ≥ 350 69.1% 54.7% Employment days two calendar years before < 50 7.7% 7.9% 50–99 1.2% 1.8% 100–149 1.3% 1.9% 150–199 2.0% 2.8% 200–249 2.5% 3.2% 250–299 3.6% 4.0% 300–349 4.9% 5.6% ≥ 350 76.9% 72.7% Prerehabilitation employment status (employed) on department level (%) Mean ± sd 67.73 ± 7.07 52.48 ± 8.81 Regional unemployment rate (%) Mean ± sd 5.93 ± 2.44 6.32 ± 2.38 Employment days one calendar year after Mean ± sd 256.51 ± 149.58 214.80 ± 164.45 market effect is no longer significant when control - duration of the statutory probationary period of employ- ling for employment history. In addition, although ment, and (2) RTW in days in the second calendar year the interaction effect (H4) remains very small, the after intervention. Furthermore, we also monitored find - statistics produce opposing effects for prerehabilita - ings withno change of the outcome variable: (3) subgroup tion unemployed (AME = 0.6 days) and employed indi- analyses based on individual diagnosis (the three largest viduals (AME = − 0.5 days). The models are shown groups of main diagnosis at discharge) and (4) further in Table 4 — the graph for the interaction effect looks predictors of the employment biography, which have similar to Fig. 3 and is not shown here. The findings on proven to be the main predictor of RTW. The results in linearity (H1) are also similar: no u-shaped relationship Table 5 show that the labour market AMEs are robust. (UR: b = 0.019, s.e. = 0.021, p = 0.374; UR : b = − 0.003, s.e. = 0.002, p = 0.070) and a linear rather than expo-4 Discussion nential relationship (log UR: b = − 0.100, s.e. = 0.027, We determined the influence of the regional unemploy - p < 0.001, AIC = 157,166). ment rate on RTW after medical rehabilitation for the two largest indication groups orthopaedics and psycho- 3.4 Robustness tests somatics. To the best of our knowledge, this was the first We did some additional analyses to monitor the robust- time that this was done explicitly using a representative ness of the labour market findings. We used different database, in this case administrative data from the GPI operationalizations of the outcome variable: (1) a dummy for medical rehabilitation programmes in 2016. The core coded RTW with a cut off at 183 days which is the usual We use linear regression coefficients because they are similar to AME based on logit/probit models (Breen et al. (2018). Return to work after medical rehabilitation in Germany: influence of individual factors and… Page 9 of 14 4 Table 3 Fractional logit regression models for RTW in orthopaedics Predictors M1 M2 M3 M4 (M1 + personal and (M2 + employment (M3 + interaction) rehabilitation biography) biography) b s.e b s.e b s.e b s.e {AME} {AME} {AME} {AME} Unemployment rate (UR) − 0.037 *** 0.004 − 0.040 *** 0.005 − 0.010 ** 0.004 − 0.013 *** 0.004 { − 3.1} { − 2,8} { − 0.2} Prerehabilitation employment status − 1.272 *** 0.012 − 1.338 *** 0.029 [not employed, reference: employed] { − 201.2} { − 201.3} UR * [not employed] 0.011 * 0.004 {AME, subgroup “employed”} { − 0.4} {AME, subgroup “not employed”} { 0.2} Random effects τ00, 0.01 0.01 < 0.01 < 0.01 labour market region τ00, 0.05 0.05 0.01 0.01 rehabilitation departments Pseudo-R 0.019 0.144 0.363 0.363 AIC 369,386 339,216 282,902 282,898 Method is cross-classified fractional logit regression with n = 257, n = 589, n = 305,980, * p < 0.05, ** p < 0.01, *** p < 0.001; labour market regions rehabilitation departments patients estimators for intercept, for personal/rehabilitation biography predictors and employment biography predictors in Additional file 1 M model, b coefficients, s.e. standard error, AME average marginal effects (in days), τ variance component of labour market region or rehabilitation department, R square of the correlation between the model’s predicted values and the actual values, AIC Akaike-criterion Fig. 3 Interaction of unemployment rate and prerehabilitation employment status in orthopaedics 4 Page 10 of 14 C. Hetzel et al. Table 4 Fractional logit regression models for RTW in psychosomatics Predictors M1 M2 M3 M4 (M1 + personal and (M2 + employment (M3 + interaction) rehabilitation biography) biography) b s.e b s.e b s.e b s.e {AME} {AME} {AME} {AME} *** *** * Unemployment rate (UR) − 0.019 0.004 − 0.025 0.005 − 0.002 0.004 − 0.011 0.005 { − 2.2} { − 2.2} { 0.0} *** *** Prerehabilitation employment status − 1.229 0.018 − 1.341 0.042 [not employed, reference: employed] { 209.9} { 209.9} ** UR * [not employed] 0.018 0.006 {AME, subgroup „employed “} { − 0.5} {AME, subgroup „not employed “} { 0.6} Random effects τ 0.01 0.01 < 0.01 < 0.01 00, labour market region τ 0.13 0.12 0.04 0.04 00, rehabilitation departments Pseudo-R 0.029 0.206 0.386 0.386 AIC 157,162 138,530 117,661 117,655 Method is cross-classified fractional logit regression with n = 257, n = 202, n = 117,386, * p < 0.05, ** p < 0.01, *** p < 0.001; labour market region rehabilitation departments patients estimators for intercept, for personal / rehabilitation biography predictors and employment biography predictors in Additional file 1 M model, b coefficients, s.e. standard error, AME average marginal effects (in days), τ variance component of labour market region or rehabilitation department, R square of the correlation between the model’s predicted values and the actual values, AIC Akaike-criterion Table 5 Robustness tests Outcome in calendar year Method Sample Orthopaedics Psychosomatics AME for unemployment AME for rate unemployment rate M1 M2 M3 M1 M2 M3 a b RTW no/yes 1stLinear regression All − 2.8 − 2.7 0.1 − 1.6 − 1.9 − 0.2 RTW days 2nd FLR All − 2.7 − 2.4 − 0.4 − 1.9 − 1.8 0.0 RTW days 1st FLR Subgroup back pain (n = 76,289) − 3.9 − 3.4 − 0.5 na na na Subgroup endoprosthesis (n = 46,691) − 3.5 − 1.6 − 0.2 na na na Subgroup depression (n = 66,945) na na na − 2.0 − 2.2 − 0.1 RTW days 1st FLR Added predictors M3: employment status na na − 0.2 na na 0.0 in the 6th, 12th and 24th month before Models (M) 1, 2 and 3 adjusted for the same predictors as before; RTW return to work, AME average marginal effects (in days); dichotomized with cut off 183 days; orthopaedics n = 305,980, psychosomatics n = 117,386 findings reveal that the regional unemployment rate has operationalization. The low effect size also explains the minimal effect on RTW. The effect is even smaller and finding initially reported that most studies on RTW near zero (orthopaedics) or not significant (psychoso - after medical rehabilitation do not even take the labour matics) when individual employment biographies are market into account. In view of the statistical signifi - added to the model as the most important predictors. cance of the labour market, an omission could never- The influence of the regional unemployment rate on theless lead to (presumably small) biases. RTW depending on the prerehabilitation employment The influence of the labour market seems to be status is not substantial. smaller compared to vocational rehabilitation. This is The influence of the labour market is small, but suggested by findings on vocational education inter - still significant depending on the inclusion of fur - ventions that were also based on the RSD using similar ther covariates. These findings are in line with the regression methods but from a different year and with only other study on this topic known to us (Kaluscha a different RTW operationalization (Hetzel and Strei - et al. 2013), even though it was based on a different belt 2016). The RTW range between the regions with Return to work after medical rehabilitation in Germany: influence of individual factors and… Page 11 of 14 4 the lowest and highest unemployment rates is about account for a considerable part of residual confounding, 30 percentage points and 7 percentage points respec- the essential predictors seem to be included in the model. tively. No direct relation can be made to the study by Observational studies are limited. Because of the Reims and Tophoven (2018). They analyzed vocational non-experimental design, self-selection into the treat- rehabilitation between 2009 and 2012 on behalf of the ment might be an issue. Possibly omitted but relevant Federal Employment Agency (FEA) and report hazard control variables could lead to bias and limit causal ratios derived from event history analysis. However, as interpretation. they used types of labour market regions, our findings Self-selection into the treatment is minimized, because are not comparable in this respect. The different labour there is a standardized path from the application to the market dependence on RTW can be explained by the start of rehabilitation with a legal right to access. But fact that in vocational rehabilitation, the focus is usu- we have no information about subjective rehabilita- ally on finding and starting a new job, while in medical tion needs, refused applications and underutilization of rehabilitation, the focus is usually on returning to the rehabilitation. old job. It could also be that the types of treatment used We use the regional unemployment rate. Other labour for medical rehabilitation react appropriately to differ - market characteristics can be omitted in our opinion ing labour market conditions, for example in the con- because the variance components for the labour market text of work-related treatments during rehabilitation are empirically close to zero, indicating a very good oper- (Bethge et al. 2018) or the transition to aftercare. ationalization of the labour market. Other operationali- We operationalized the outcome using the number of zations (including economic structure, commuting links, days in employment with social insurance contributions unemployment structure, and unemployment trend) in the first calendar year. The dataset provides the num - have been tested elsewhere, but have proved to be of low ber of days only by calendar year and not for other time importance (not displayed here). Moreover, this confirms periods. The outcome by calendar year leads to a large the regional unit chosen, which was homogenized in gap between rehabilitation and RTW measures for some terms of commuting links. For alternative regional units, individuals. This is random, and the dataset is large, so such as political-administrative or settlement-structured we see no bias in regression modelling. The social insur - regions, remodeling of commuter links might have been ance contributions from employment in particular are necessary. one advantage of the database, because they are reported Other individual factors would be of importance if they by employers and are therefore without bias. The RTW were associated with both the labour market and the out- can be defined in many ways (Young et al. 2005; Nübling come variable and thus altered the regression estimators. et al. 2016). We have discussed and empirically consid- We excluded diagnoses because the diagnoses mainly ered alternative RTW operationalizations elsewhere affect RTW but should not be associated with the labour (Leinberger et al. 2023). Since these are highly correlated, market. Therefore, they are not important for the unbi - labour market correlations are likely to be largely inde- ased estimation of labour market effects. We demon - pendent of the operationalization choice of RTW. We strated this by additional analyses for subgroups. demonstrated this in additional analyses for a dichoto- In terms of employment biography, we used employ- mous outcome and for days in employment in the second ment days in the first and second year before rehabilita - calendar year. The results may be generalized to other tion, as well as the employment status in the third month outcomes, as patient-reported outcomes are highly corre- before. These predictors have proved to be the main pre - lated with administrative RTW data (Nübling et al. 2017). dictors of RTW. The database offers alternative options For control variables, we included socio-demograph- for operationalizing employment biography: variables ics, rehabilitation characteristics, and employment per calendar year (days in employment, days in receipt of biographies, as well as an aggregated variable on the unemployment benefits, earnings) and status per month department level. The selection is based on theoreti - (yes/no: employment, unemployment, parenthood, etc.). cal considerations, especially the noninfluence of reha - We did not apply all these factors because they are highly bilitation departments, as well as empirical relevance. correlated (multicollinearity) or biased (in the case of The model quality in M3 and M4 can be rated as good. earnings, there is no information about part-/full-time Assuming that unknown treatment quality is likely to employment and about the household income). Even with our parsimonious operationalization, the influence of the labour market was nonexistent (0.0 days in psycho- somatics) or almost nonexistent (-0.4 days in orthopae- For orthopaedics about 8% (10.5 × 2.8 days / 365 days) and for psychosomat- ics about 6% (10.5 × 2.2 days / 365 days), with 10.5% points range between the dics). We have shown in additional analyses that this is regions with the lowest and highest unemployment rates and AME from M2. 4 Page 12 of 14 C. Hetzel et al. robust when we add further employment biography pre- the labour market, such as the COVID-19 pandemic or dictors to model 3. financial crises, limit the generalizability. Occupation, industry, workload, etc., would be relevant to determine labour market effects on RTW because 5 Conclusion they are likely to be correlated with both labour market We conclude from the findings that the influence of the and outcome variables. For example, certain sublabour labour market on RTW is small. It is largely proxied by markets (e.g., for bottleneck occupations such as nurses) individual employment biographies. This finding remains might differ from the general labour market (Fedorets plausible even if the influence of the labour market dif - et al. 2019). However, these characteristics are either fers according to these biographies. However, if no (valid) not present in the database or are insufficiently recorded employment biographies are available, the labour mar- (e.g., occupation key of the last job). Unobserved hetero- ket should be included in RTW analyses. Database and geneity may therefore be present and lead to bias. methods are of high quality, but because of the nonexper- The same applies to characteristics that suggest the imental design, omitted variables could lead to bias and motivation to find a job (e.g., household income, financial limit causal interpretation. obligations, employment status of the partner, domestic care situation). Other interaction effects could be significant. How - Abbreviations AME Average marginal effects ever, nonlinear models such as the FLR used here already GPI German pension insurance implicitly model interactions and are thus less sensitive RSD Rehabilitation statistics database to interactions. This also partially explains the small RTW Return to work interaction effect in M4. Since, to our knowledge, there are no substantial further interaction hypotheses, we did Supplementary Information not test any further interactions empirically. The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12651- 023- 00330-1. Reverse causality is excluded because the direction of effect is predetermined by the data structure and the Additional file 1: Table S1. Fractional logit regression models for RTW in temporal sequence. This means that context effects can orthopaedics (all predictors). Table S2. Fractional logit regression models only affect the individual and not vice versa. Moreover, for RTW in psychosomatics (all predictors). the employment days are timed after the intervention and the labour market is related to the period of the interven- Acknowledgements tion. In this respect, a causal inference is permissible. Not applicable. The FLR estimation method applied is rather uncom - Author contributions monly used. However, given the bimodal distribution of All authors contributed to the study conception and design. Material was the outcome variable and its limited range between 0 and prepared and analyzed by CH, SL and RK. The first draft of the manuscript was written by CH and all authors commented on previous versions of the manu- 365, it is appropriate. Alternatives have been rejected for script. All authors read and approved the final manuscript. reasons presented elsewhere, such as models with zero– one-inflation (Leinberger et al. 2023). Funding Contracted research project "Adjustment of the socio-medical course after The database fully reflects the rehabilitation pro - medical rehabilitation (SMV )" with the German Pension Insurance as client. grammes of the GPI with high data quality. Since only the employment biography and no socio-demographic Availability of data and materials The data used in the present study are available from the German Pen- or rehabilitation characteristics seem to influence the sion Insurance. But restrictions apply to the availability of these data, which estimation of the labour market effect, we consider the were used at our request for this study. Data are however available from the results transferable, at least for welfare states with Bis- authors upon reasonable request and with permission of the German Pension Insurance. marckian systems such as Germany (Kolmar 2007). Social insurance schemes differ according to the relation - Declarations ship between contributions and benefits. Bismarckian systems provide earnings-related benefits, while Bev - Ethics approval and consent to participate eridgean systems offer flat payments. Moreover, the find - Not applicable. ings are relatively stable across the two major indication Consent for publication groups and even across subgroups by diagnosis, which Not applicable. is why transferring them to other indication groups Competing interests seems legitimate. But rehabilitation programmes of only The research leading to these results received funding from the German Pen- one year were analyzed, and period effects relevant to sion Insurance. Return to work after medical rehabilitation in Germany: influence of individual factors and… Page 13 of 14 4 Received: 9 October 2021 Accepted: 1 January 2023 Kaluscha, R., Jankowiak, S., Holstiege J., Krischak G.: Beeinflusst die Arbeit - slosenquote die ( Wieder-)Eingliederung in das Erwerbsleben nach medizinischer Rehabilitation? DRV-Schriften 101, 182–184 (2013) Kolmar, M.: Beveridge versus Bismarck public-pension systems in integrated markets. Reg. Sci. Urban Econ. 37(6), 649–669 (2007) Krischak, G., Tepohl, L., Dannenmaier, J., Hartschuh, U., Auer, R., Kaluscha, R.: References Health economic effects of rehabilitation for chronic back pain—an Belin, A., Dupont, C., Oulès, L., Kuipers, Y., Fries-Tersch, E.: Rehabilitation and observational study with combined secondary data from statutory health return to work: Analysis report on EU and Member States policies, insurance and German pension fund. Die Rehabil. 58(6), 392–397 (2019) strategies and programmes. Publications Office of the European Union, Leeper, T.: Marginal effects for model objects. Package ‘margins‘. Version 0.3.26. Luxembourg (2016) https:// cran.r- proje ct. org/ web/ packa ges/ margi ns/ margi ns. pdf (2021) Bethge, M., Markus, M., Streibelt, M., Gerlich, C., Schuler, M.: Implementing the Leinberger, S., Hetzel, C., Kaluscha, R.: Risk-adjustment of return to work after German model of work-related medical rehabilitation: did the delivered medical rehabilitation: methodical advancements in quality assurance of dose of work-related treatment components increase? Arch. Phys. Med. the German Pension Insurance. Die Rehabil (in press). https:// doi. org/ 10. Rehabil. 99(12), 2465–2471 (2018) 1055/a- 1998- 6574 Breen, R., Karlson, K.B., Holm, A.: Interpreting and understanding logits, probits, Leinonen, T., Solovieva, S., Husgafvel-Pursiainen, K., Laaksonen, M., Viikari- and other nonlinear probability models. Annu. Rev. Sociol. 44(1), 39–54 Juntura, E.: Do individual and work-related factors differentiate work (2018) participation trajectories before and after vocational rehabilitation? PLoS Brooks, M.E., Kristensen, K., van Benthem, K.J., Magnusson, A., Berg, C.W., ONE 14(2), e0212498 (2019) Nielsen, A., Skaug, H.J., Maechler, M., Bolker, B.M.: glmmTMB balances Loisel, P., Buchbinder, R., Hazard, R., Keller, R., Scheel, I., van Tulder, M., Webster, speed and flexibility among packages for zero-inflated generalized linear B.: Prevention of work disability due to musculoskeletal disorders: the mixed modeling. R J. 9(2), 378–400 (2017) challenge of implementing evidence. J. Occup. Rehabil. 15(4), 507–524 Bülau, N.I., Kessemeier, F., Petermann, F., Bassler, M., Kobelt, A.: Evaluation of (2005) contextual factors in psychosomatic rehabilitation. Die Rehabil. 55(6), McCullagh, P.: Quasi-likelihood functions. Ann. Stat. 11(1), 59–67 (1983) 381–387 (2016) Miyamoto, G.C., Lin, C.-W.C., Cabral, C.M.N., van Dongen, J.M., van Tulder, M.W.: Bundesinstitut für Bau-, Stadt- und Raumforschung: INKAR: Indikatoren und Cost-effectiveness of exercise therapy in the treatment of non-specific Karten zur Raum- und Stadtentwicklung. https:// www. inkar. de/ (2020) neck pain and low back pain: a systematic review with meta-analysis. Br. J. Cieza, A., Causey, K., Kamenov, K., Hanson, S.W., Chatterji, S., Vos, T.: Global Sports Med. 53(3), 172–181 (2019) estimates of the need for rehabilitation based on the Global Burden of Müller, W.A., Blien, U.: Arbeitslosigkeit und Entlohnung auf regionalen Arbe- Disease study 2019: a systematic analysis for the Global Burden of Disease itsmärkten, vol. 182. Physica-Verlag HD, Heidelberg (2001) Study 2019. Lancet 396(10267), 2006–2017 (2020) Nevala, N., Pehkonen, I., Teittinen, A., Vesala, H.T., Pörtfors, P., Anttila, H.: The Deutsche Rentenversicherung Bund: Die medizinische und berufliche effectiveness of rehabilitation interventions on the employment and Rehabilitation der Rentenversicherung im Licht der Statistik, Reha-Bericht functioning of people with intellectual disabilities: a systematic review. J. 2019. Deutsche Rentenversicherung Bund, Berlin (2019) Occup. Rehabil. 29(4), 773–802 (2019) Deutsche Rentenversicherung: Datensatz zur Reha-Statistik-Datenbasis (RSD) Nübling, R., Kaluscha, R., Krischak, G., Kriz, D., Martin, H., Müller, G., Renzland, ab Berichtsjahr 2015. http:// forsc hung. deuts che- rente nvers icher ung. J., Reuss-Borst, M., Schmidt, J., Kaiser, U., Toepler, E.: Return to work after de/ FdzPo rtalW eb/ getRe ssour ce. do? key= sk_ 94_ ab_ 2015. pdf (2015). inpatient rehabilitation—alternative calculations on the basis of patient Accessed 31 Nov 2020 data and validation of social security contributions. Phys. Med. Rehab. Echarti, N., Schüring, E., O’Donoghue, C.: Eec ff ts of vocational re-training on Kuror. 26(06), 293–302 (2016) employment outcomes among persons with disabilities in Germany: a Nübling, R., Kaluscha, R., Krischak, G., Kriz, D., Martin, H., Müller, G., Renzland, quasi-experiment. J. Occup. Rehabil. 30(2), 221–234 (2020) J., Reuss-Borst, M., Schmidt, J., Kaiser, U., Toepler, E.: Outcome quality in Ervasti, J., Joensuu, M., Pentti, J., Oksanen, T., Ahola, K., Vahtera, J., Kivimäki, M., medical rehabilitation: relationship between „patient-reported outcomes“ Virtanen, M.: Prognostic factors for return to work after depression-related (PROs) and social security contributions. Die Rehabil. 56(1), 22–30 (2017) work disability: a systematic review and meta-analysis. J. Psychiatr. Res. Odgaard, L., Pedersen, A.R., Poulsen, I., Johnsen, S.P., Nielsen, J.F.: Return to work 95, 28–36 (2017) predictors after traumatic brain injury in a welfare state. Acta Neurol. Fedorets, A., Lottmann, F., Stops, M.: Job matching in connected regional and Scand. 137(1), 44–50 (2018) occupational labour markets. Reg. Stud. 53(8), 1085–1098 (2019) Papke, L.E., Wooldridge, J.M.: Econometric methods for fractional response Hayes, A.F.: Introduction to mediation, moderation, and conditional process variables with an application to 401 (k) plan participation rates. J. Appl. analysis: a regression-based approach. Methodology in the social sci- Econ. 11(6), 619–632 (1996) ences. Guilford Press, New York, NY (2013) Petrongolo, B., Pissaridēs, C.A.: Looking into the black box: a survey of the Hetzel, C.: Multilevel analyses of labour market and return to work after voca- matching function. J. Econ. Lit. 39(2), 390–431 (2001) tional retraining. Die Rehabil. 54(1), 16–21 (2015) Reims, N., Tophoven, S.: A longitudinal study on labour market transitions Hetzel, C., Streibelt, M.: Does return to work after vocational retraining depend and sustainability of employment after further training measures in the on labour market? Die Rehabil. 55(5), 290–298 (2016) context of vocational rehabilitation. Die Rehabil. 57(3), 184–192 (2018) Hirschenauer, F.: Integrationserfolge hängen von regionalen Gegebenheiten Shields, G.E., Wells, A., Doherty, P., Heagerty, A., Buck, D., Davies, L.M.: Cost- ab: Neue Typisierung der Agenturbezirke. IAB-Kurzbericht 5, 1–8 (2013) effectiveness of cardiac rehabilitation: a systematic review. Heart 104(17), Howard-Wilsher, S., Irvine, L., Fan, H., Shakespeare, T., Suhrcke, M., Horton, S., 1403–1410 (2018) Poland, F., Hooper, L., Song, F.: Systematic overview of economic evalua- Steiner, V.: The labour market for older workers in Germany. J. Labour Market tions of health-related rehabilitation. Disabil. Health J. 9(1), 11–25 (2016) Res. 50(1), 1–14 (2017) Hox, J.J., Moerbeek, M., van Schoot, Rd.: Multilevel analysis: Techniques and Stewart, A.M., Polak, E., Young, R., Schultz, I.Z.: Injured workers’ construction applications. Quantitative methodology series. Routledge Taylor & Francis of expectations of return to work with sub-acute back pain: the role of Group, New York, London (2018) perceived uncertainty. J. Occup. Rehabil. 22(1), 1–14 (2012) International Social Security Association: Who returns to work and why? A six- van den Berg, T., Schuring, M., Avendano, M., Mackenbach, J., Burdorf, A.: The country study on work incapacity and reintegration. International Social impact of ill health on exit from paid employment in Europe among Security Association, Geneva (2001) older workers. Occup. Environ. Med. 67(12), 845–852 (2010) Jansen, J., van Ooijen, R., Koning, P.W.C., Boot, C.R.L., Brouwer, S.: The role of the von Celsing, A.-S., Svärdsudd, K., Eriksson, H.-G., Björkegren, K., Eriksson, M., employer in supporting work participation of workers with disabilities: a Wallman, T.: Determinants for return to work among sickness certified systematic literature review using an interdisciplinary approach. J. Occup. patients in general practice. BMC Public Health 12, 1077 (2012) Rehabil. 31(4), 916–949 (2021) Wedderburn, R.W.M.: Quasi-likelihood functions, generalized linear models, and the Gauss-Newton Method. Biometrika 61(3), 439 (1974) 4 Page 14 of 14 C. Hetzel et al. Word Health Organization: Rehabilitation. https:// www. who. int/ news- room/ fact- sheets/ detail/ rehab ilita tion (2020). Accessed 28 May 2021 Young, A.E., Roessler, R.T., Wasiak, R., McPherson, K.M., van Poppel, M.N.M., Anema, J.R.: A developmental conceptualization of return to work. J. Occup. Rehabil. 15(4), 557–568 (2005) Zeisberger, M., Nowik, D., Beck, L., Märtin, S., Beckmann, U., Meyer, T.: Quality management in medical rehabilitative care by the German statutory pen- sion insurance scheme. Eur. J. Phys. Rehabil. Med. 55(6), 845–851 (2019) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations.
Journal for Labour Market Research – Springer Journals
Published: Jan 20, 2023
Keywords: Return to work; Labour market; Rehabilitation; Orthopaedic; Psychosomatic; I130; J210; J140
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.