http://orcid.org/0000-0002-9319-665X
Figures
Abstract
Objectives
We aimed to describe labor market participation (LMP) of persons with spinal cord injury (SCI) in Switzerland, to examine potential determinants of LMP, and to compare LMP between SCI and the general population.
Methods
We analyzed data from 1458 participants of employable age from the cross-sectional community survey of the Swiss Spinal Cord Injury Cohort Study. Data on LMP of the Swiss general population were obtained from the Swiss Federal Statistical Office. Factors associated with employment status as well as the amount of work performed in terms of full-time equivalent (FTE) were examined with regression techniques.
Results
53.4% of the participants were employed at the time of the study. Adjusted odds of being employed were increased for males (OR = 1.73, 95% CI 1.33–2.25) and participants with paraplegia (OR = 1.78, 95% CI 1.40–2.27). The likelihood of being employed showed a significant concave relationship with age, peaking at age 40. The relation of LMP with education was s-shaped, while LMP was linearly related to time since injury. On average, employment rates were 30% lower than in the general population. Males with tetraplegia aged between 40 and 54 showed the greatest difference. From the 771 employed persons, the majority (81.7%) worked part-time with a median of 50% FTE (IRQ: 40%-80%). Men, those with younger age, higher education, incomplete lesions, and non-traumatic etiology showed significantly increased odds of working more hours per week. Significantly more people worked part-time than in the general population with the greatest difference found for males with tetraplegia aged between 40 and 54.
Citation: Reinhardt JD, Post MWM, Fekete C, Trezzini B, Brinkhof MWG, on behalf of SwiSCI Study Group (2016) Labor Market Integration of People with Disabilities: Results from the Swiss Spinal Cord Injury Cohort Study. PLoS ONE 11(11): e0166955. https://doi.org/10.1371/journal.pone.0166955
Editor: Stefano Federici, Universita degli Studi di Perugia, ITALY
Received: April 25, 2016; Accepted: November 7, 2016; Published: November 22, 2016
Copyright: © 2016 Reinhardt et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data and estimates of this study are deposited on Dryad (data set DOI: http://datadryad.org/review?doi=doi:10.5061/dryad.h8p2r). Due to our commitment to SwiSCI study participants, the SwiSCI study center requires on behalf of the SwiSCI Study Group contact prior to any planned data usage (contact@swisci.ch).
Funding: The study was funded by Swiss Paraplegic Research. The funding agency had no influence on design, analysis, or interpretation of the data.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Across the world, persons with disabilities are less likely to be employed or to receive proper assistance and education for employment than the general population [1–3]. This situation entails enormous economic and psychosocial costs, first for the individual in terms of poverty, decreased self-esteem, impaired health, and worse overall social integration and participation [4, 5] and second, for the society in terms of increased social and health care spending, loss of tax revenue and loss of resources such as diversity and creativity [2, 5]. Moreover, work and employment fulfill important social functions for people with disabilities including recognition as full members of society, obtaining income, establishing contacts with other people, contributing to society, and structuring time [6, 7].
Considerable variation regarding labor market participation (LMP) of persons with disabilities has been observed, even across highly developed economies [2, 8]. Against this background, Switzerland provides an interesting example as a high resource-country with low unemployment rates of the general population and an increasing focus of social policies on reintegration of persons with disabilities into the labor market in the last decades [9].
To study LMP of persons with disabilities in-depth, spinal cord injury (SCI) may serve as a case in point. SCI is a major health condition affecting all bodily functions below the lesion level, and may have a substantial impact on activities and social participation against the background of environmental barriers such as inaccessible infrastructure and negative attitudes [10–12]. While SCI is a severe health condition, it usually does not entail major cognitive impairments. Consequently, prospects for employment are good, if sufficient support and accommodations are made available [13]. Employment rates of persons with SCI are nonetheless well below those of the general population [5, 13, 14].
Employment rates and predictors for return to work of people with SCI have been studied extensively over the past decades [4, 5, 13–17]. However, there are several shortcomings in current research. First, most studies rely on highly selective samples or investigate return to work of persons who were employed at time of SCI onset, thus potentially introducing selection bias. Second, past investigations of determinants of LMP after SCI did not consider potential non-linear associations of determinants such as age, time since injury, or years of formal education with the outcome, whereas the likelihood of being employed may, for example, increase up to a certain age and then decrease again. Third, most of the available studies use LMP as binary outcome (employed vs. non-employed) and do not include more detailed information about the amount of work performed (e.g. hours per week). Fourth, comparisons of LMP in SCI with age- and gender-matched general population samples are scarce. However, such comparison is essential in order to account for the overall performance of the labor market of a particular country.
The only study which assessed LMP in persons with SCI in Switzerland before [18] suffered from limited generalizability of its findings since the sampling frame was restricted to a single organization for persons with disabilities (Swiss Paraplegic Association, SPA) and the response rate was low (27%).
By using data from the largest community survey on person with SCI in Europe, we have the opportunity to address the before mentioned shortcomings of current research on LMP in SCI. More specifically, the population-based community survey of the Swiss Spinal Cord Injury Cohort Study (SwiSCI) provides data from almost 2000 persons with SCI and thus enables us to provide more accurate and less biased estimates relying on a large sample of the SCI population in Switzerland [19, 20]. Moreover, we investigate non-linear relationships with selected determinants and consider the amount of work performed, in terms of full-time equivalent (FTE), in our analysis. Finally, by using an age- and gender-matched sample of the Swiss general population, we are able to illustrate differences in LMP between persons with SCI and the general population.
The aims of this study are 1) to report employment rates of people with SCI living in Switzerland and to examine their variation by sociodemographic and lesion characteristics, 2) to compare employment rates between the SCI and the general population providing age- and gender-matched differences, 3) to analyze the amount of work performed by those employed and its association with sociodemographic and lesion characteristics, and 4) to provide a comparison between FTE of people with SCI and the general population.
Materials and Methods
Design
The study draws on cross-sectional data from the first SwiSCI community survey [19–21] performed between late 2011 and early 2013. The community survey was conducted according to the ethical principles formulated in the Declaration of Helsinki and its study protocol was approved by the responsible Medical Ethics Committee of the Canton of Lucerne and subsequently by the Ethics Committees of all other involved Swiss cantons, namely Zurich, Basel, and Valais. The legal age of consent in Switzerland is 16 years and written informed consent for participation in the study was obtained from all subjects prior to data collection.
Data were collected using a mixed-mode approach offering participants the opportunity to respond by postal or online questionnaire. Also, telephone interviews were optional in special cases (e.g., for persons with limited hand function). Following a written invitation, a proactive reminder management including up to two written reminders and one telephone reminder was implemented to improve response rates and minimize response bias. Data were stored in a central database and anonymized before analysis. Further details on the study protocol, the recruitment procedures and the reminder management can be found elsewhere [20, 21].
Sample
Eligible subjects were Swiss residents with a traumatic or non-traumatic SCI aged 16 years or older and living in the community. Persons with congenital conditions leading to SCI, with new SCI in the context of palliative care, with neurodegenerative disorders, and with Guillain-Barré syndrome were excluded from the study. Potential participants were identified from databases of the SPA, three SCI clinical centers, and an SCI home-care service provider (ParaHelp). This approach was chosen since a central registry covering all persons with SCI in Switzerland does not exist to date. As a consequence, the sampling frame did not include all elements of the target population, i.e. all persons aged 16 years or older living with SCI in Switzerland, but the subgroup of those being a member of the SPA, having been treated in one of the above mentioned centers, or being a client of the home-care service provider ParaHelp. In total, 3144 persons were eligible for the study of which 1922 participated in the survey, indicating a response rate of 61.1% [20, 21]. From the 1922 participants, 1458 were of employable age at the time of the study, defined as the age range from 16 to the Swiss statutory retirement age of 65 for men and 64 for women. The response rate in the employable age group was 61.9%.
Measures
The primary outcome of the study was self-reported LMP, defined as employment or self-employment at the time of the survey for at least one hour/week and coded as a dichotomous variable (0 = no LMP; 1 = LMP) [22, 23]. This definition included competitive, sheltered and supported employment. The secondary outcome was the amount of work performed, defined as full-time equivalent (FTE) in percent of the reference level of 42 hours per week in the general Swiss population for those with LMP. To capture the clustered, multimodal distribution in work load (Fig 1) and to facilitate regression modelling, amount of work was recoded as a ordinal categorical variable with the following levels: ≤30%, 31%-60%, 61%-90%, and >90% (full-time).
Independent variables comprised age at the time of the survey in years, gender, years of formal education according to the International Standard Classification of Education combining school and vocational training [24], lesion level (paraplegia vs. tetraplegia), type of lesion (complete vs. incomplete), etiology (traumatic vs. non-traumatic), and time since SCI in years. All independent variables were also derived from self-report. Previous analyses showed that, with the exception of education, all self-reported demographic and lesion characteristics were in substantial agreement with medical record data [20].
General population data for the comparison of employment rates were obtained from the Swiss Federal Statistical Office and represent Swiss labor market statistics for the SwiSCI data collection period (third quarter of 2011) [25].
Data analysis
Basic principles and descriptive analyses.
Statistical analyses were performed using the packages Stata version 13.1 (StataCorp, Texas, USA) and R version 3.1.3 (The R Foundation for Statistical Computing). Reporting of methodology and results follows the STROBE guidelines [26].
Descriptive statistics of study participants’ characteristics were provided for the overall sample as well as stratified by employment status. Unless indicated otherwise, descriptive analysis was based on crude data, not accounting for missing data that resulted from either unit-nonresponse (nonparticipation in the survey) or item-nonresponse (nonexistence of data on a specific survey item). Regression analyses of the primary and secondary outcomes were adjusted for both unit and item nonresponse. Unless reported otherwise, results of adjusted analysis were similar to those of unadjusted analysis.
To account for unit-nonresponse, we used sampling weights which were defined as the inverse of the propensity score of participation in the survey [27, 28]. Propensity scores were derived from a multivariable nonresponse analysis that indicated limited nonresponse bias in relation to membership of the SPA, current age, and time since SCI [20]. Item nonresponse, i.e. missing values in explanatory variables or outcomes, was handled using a random forest imputation technique (R-package missForest [29]).
Primary outcome: employment status.
The primary outcome employment status (employed or self-employed vs. not employed at time of the study) was analyzed with logistic regression models using the following predictor variables: current age (years), gender (reference: female), attained formal education (years), level of lesion (reference: paraplegia), type of lesion (reference: incomplete), etiology (reference: non-traumatic), and time since SCI (years). Predictor variables were used in univariable and multivariable analysis to derive unadjusted and mutually adjusted odds ratios with corresponding 95% confidence intervals (CI), respectively. To evaluate potential effect modification with regards to lesion characteristics, an interaction term between lesion level and lesion type was evaluated separately in the multivariable model. To systematically explore potential nonlinear relationships between employment status and the continuous variables age, years of education, and time since SCI, we used a fractional polynomials algorithm with a predefined set of powers (i.e., −2,−1,−0.5, 0, 0.5, 1, 2, 3; with 0 representing the logarithmic transformation), allowing for repeated powers to further enhance modelling flexibility, and favoring simpler models unless more complicated functions showed superior fit [30–32]. The statistical significance of predictor variables was tested globally (i.e., jointly across all transformations or categorical levels) using adjusted Wald-tests, which adequately account for the usage of robust standard errors in presence of sampling weights. Marginal predictions of employment rates with 95% CI were also derived for graphical display of the relationship of employment rates with continuous predictor variables.
To compare employment rates in SCI with the general population, we used Poisson regression with an offset term to derive age- and gender-standardized employment rate estimates [33, 34]. Count data comprised the number of employed people in the SCI sample stratified by age group (16–24, 25–39, 40–54, 55–64 years of age), gender, and lesion level. The offset term (i.e., the fixed part of the linear predictor) consisted of the log of the expected counts that were calculated by multiplying the number of people in the respective 16 strata with the corresponding age- and gender-specific employment rate in the general population. The average of the individual sampling weights was used as stratum-specific sampling weights. We thus derived employment rate ratios and subsequently rate differences (both with 95% CI) that were standardized to the general population. Modeling assumptions were tested using the deviance and Pearson Goodness-of-Fit (GOF) tests. To gain insight as to whether age and gender differences in employment rates are specific to the SCI population, we also compared effect sizes from models with and without an offset.
Secondary outcome: amount of work performed.
We used stereotype logistic regression [35–37] to evaluate associations of ordinal levels of the amount of work performed with demographic and injury-related characteristics. Brant testing following initial ordinal logistic regression [38] indicated that the proportional odds assumption was not met (χ2 (14) = 42.83, p < 0.001). Stereotype logistic regression provides an alternative method that seeks to minimize the number of parameters, while preserving an adequate model fit by enforcing ordering constraints on adjacent ordinal levels that are statistically indistinguishable. We used the lowest FTE category (≤30%) as base level (reference) and imposed a corner constraint on the highest FTE category (>90%) to make the model identifiable. The distinguishability of adjacent ordinal levels (excluding the base level) was evaluated using Wald-test. From the final multivariable model we derived adjusted odds ratios for being in a given FTE category as compared to being in the lowest FTE category for all predictor variables. Global significance testing using Wald-tests was applied to derive p-values in multivariable modeling. Marginal predictions at means were also derived as to graphically illustrate the predicted distribution of participants over the ordinal levels of FTE in persons with SCI participating in the labor market.
The amount of work (FTE) performed by people with SCI was compared to the general population using basic descriptive statistics that contrasted the distribution in SCI with the average FTE in the general population. Subsequently, the percentage of persons with SCI working less, the same or more than their respective age and gender group in the general population was analyzed using Wilcoxon signed-rank test.
Results
Employment rates
Person characteristics and employment rates in SCI.
Study participants’ demographic and injury characteristics are given in Table 1. The majority of the participants was middle-aged, male, had high levels of education, paraplegia, and a traumatic etiology. On average around 53% were employed at the time of the study. Weighted employment rates were only slightly different from crude rates.
Predictors of employment status in SCI.
The unadjusted analyses of employment status revealed clear associations with gender, age, years of formal education, lesion level and etiology, but not with lesion type and time since SCI. Most of these associations were also found in the adjusted, multivariable analysis with the exception of time since SCI and etiology, which showed slightly higher and lower odds ratios, respectively, than in the univariable analysis (Table 2). Further, there was no indication for an interaction between type and level of lesion in the multivariable model (χ2 (1) = 0.06, p = 0.81). Adjusted odds for being employed or self-employed were 73% higher in males than in females and 78% higher in persons with paraplegia as compared to tetraplegia. In addition, the adjusted odds were elevated by 13% per ten-year increase in time since SCI, while varying non-linearly with age and years of education.
Fig 2 depicts the associations for the continuous variables as marginal predictions of employment rates. Employment rates showed an asymmetrical concave relationship with age peaking at about 60% at around age 40, and an s-shaped relationship with years of education, gradually increasing from low rates to about 45% at around 12 years (vocational) and then gradually leveling off at 70% to 75% at 17 or more years (university level) of formal education. The study population further showed an about 3.2% increase in LMP with each ten years since SCI. Marginal predictions (95% CI) for categorical variables in Table 2 from the adjusted model were: 56.5% (53.2–59.8) and 42.9% (37.6–48.4) for males and females; 57.3% (53.9–60.7) and 43.0% (38.3–47.9) for the paraplegic and tetraplegic group; 54.5% (50.6–58.3) and 50.9% (46.5–55.3) for the incomplete and complete lesion group; and 54.1% (51.0–57.1) and 46.5% (38.8–54.4) for persons with traumatic SCI as compared to non-traumatic SCI.
Predictions are at means for all other variables in the multivariable model with grey areas indicating the 95% confidence interval.
Employment rates of people with SCI in comparison with the general population.
The Poisson regression model for evaluating standardized employment rate ratios demonstrated good statistical fit and revealed associations with gender, age group and lesion level (Table 3).
Adjusted employment rates were estimated as 1.13 times higher in males than in females (a 13% increase); between 1.18 and 1.28 times higher in the persons from the three youngest age groups as compared to the oldest age group (elevations between 18% and 28%); and 1.26 times higher in paraplegia as compared to tetraplegia (a 26% increase). Fig 3 shows the estimated differences of employment rates in people with SCI as compared to the general population, by gender, age and lesion level. On average, the employment rate of people with SCI was around 30% lower than in the general population. Males with tetraplegia aged between 40 and 54 showed the greatest difference, while men with paraplegia aged between 16 and 24 years showed the smallest difference in comparison to the general population.
For stratified estimates of overall employment rate in paraplegia and tetraplegia, see Table 1.
Amount of work performed as FTE
Descriptive analysis of work load in SCI.
From the 771 employed persons in our sample, the vast majority (81.7%) worked part-time (here defined as less than 91% of FTE) with a median of 50% FTE (IRQ: 40%-80%). In total and across all groups, the majority of participants reported 31–60% FTE. Part-time work was particularly more frequent in females, older employees and persons with lower education. Regarding lesion characteristics, persons with complete lesions or having a traumatic etiology less commonly worked full-time (Table 4).
Predictors of amount of work performed by people with SCI.
Stereotype logistic regression with lowest (≤30%) FTE group as base (reference) category and highest (>90%) as constraint (identification) category was fitted. Tests indicated that other levels were distinguishable, implying that the number of ordinal levels in the model could not be further constrained without decrease in model fit. Distinguishability statistics for adjacent levels of FTE were as follows: comparing “>90%” to”61%-90%”, adjusted Wald-test, F(1, 770) = 4.37, P = 0.037; and”61%-90%” to “31%-60%”, adjusted Wald-test, F(1, 770) = 20.22, P < 0.001. Table 5 gives adjusted odds ratios for relevant patient characteristics for being in a given FTE category with reference to the lowest FTE category (i.e., ≤30%). Men, those with younger age and higher education showed significantly increased odds of working more. In addition, persons with an incomplete lesion reported gradually higher FTE as compared to those with a complete lesion, while those with a traumatic lesion worked less than people with a non-traumatic lesion. Time since SCI did not systematically affect the amount of work performed. For a visualization of the marginal predictions of person distributions over FTE categories see Fig 4.
Display is restricted to the two highest level categories with 60%-90% or >90% FTE noting that the two lowest categories predictably showed opposite marginal differences, essentially producing a balancing mirror image. Estimates are at mean values for other parameters using the stereotype logistic regression model in Table 3. Open symbols indicate the reference category for each parameter; solid symbols and error bars the percent difference with 95% confidence interval for other parameter classes as well as for continuous variables. Reading example: Given the overall gender distribution, men are overrepresented by 12.6% (95% CI: 8.3%-16.9%) in the >90% work load category.
Comparison of work load with general population.
The crude median FTE (IQR) in participants who worked was 50% (40%-80%), with higher amounts of work performed by males (50%; 47%-80%) as compared to females (50%; 34%-60%). The majority of employed men and women with SCI worked less than the general population. Further stratification by age group and lesion level (Table 6) revealed a tendency for reduced work in males with older age. Correspondingly, an elevated proportion of persons reported to work less than the general population, ranging from 69.4% to 86.5% across the middle and upper age groups. Females of the intermediate age groups reported similarly high frequencies of reduced work, while trends were indistinct for the youngest and oldest age group due to limited sample size.
Discussion
This study analyzed LMP of persons with SCI living in Switzerland in terms of employment rates and the amount of work performed. About half of the study population participated in the labor market, to a great extent in part-time employment. Employment rates were markedly reduced as compared to the general population. While females with complete tetraplegia had the lowest employment rates in absolute terms, older males with tetraplegia had the greatest disadvantage as compared to men from the general population. Employed men with tetraplegia aged 55 and above also showed the most accentuated reduction in work hours in terms of FTE as compared to the general population.
The found employment rate of 53.4% is around 10% lower than that reported by Marti et al. in 2010 [18]. This is likely due to self-selection bias in the latter study which was flagged as a LMP survey. Employed people may have found this topic more interesting or relevant to their lives than unemployed people. By contrast, the SwiSCI community survey was a survey on functioning in SCI in general, reducing the probability of self-selection due to employment status.
The observed employment rates for Switzerland are comparable to those from Scandinavian countries and the Netherlands averaging around 50% [16, 39], whereas LMP is significantly lower in Southern Europe and globally (37%) [5]. Employment rates of persons with SCI considerably vary by country and continent and range from 11.5% to 74.0% [16], with highest rates in Europe (51%) and lowest in North America (30%) [40] However, the interpretation of these figures is difficult as most studies do not include persons with non-traumatic SCI and use different definitions of employment (e.g., not exclusively paid employment). Reported employment rates may also be considerably affected by the time period under consideration and the sampling frame of respective studies. Also, cited studies do not consider age and gender adjusted risk differences as compared to the general population. It is therefore highly recommended that upcoming between-country comparisons include general population estimates of employment rates in order to determine the relative disadvantage of the SCI population.
Demographic and injury-related factors associated with LMP largely confirm previous research [13, 17]. Our finding that age was non-linearly related with LMP suggests a gradual increase of employment rates between the age of 16 and 40 and then an increasing decline. The later decline might be partially mediated by having pain, which along with increasing age reduced the likelihood of LMP in subsample of participants who reported on their health conditions in a subsequent survey module [41]. Regarding years of formal education, our findings suggest that people with SCI in Switzerland benefit most from an increase between 7 and 15 years of formal education. As education is a modifiable factor, respective interventions are advised. Future research is needed though to compare trajectories with those of the general population, with other countries as well as between different sub-groups of the SCI population. The result of a linear increase in LMP with time since SCI (when adjusted for age) suggests an adjustment process after SCI, potentially involving physical and psychosocial recovery as well as re-training or completion of education in some cases [42]. It is remarkable that this linear relationship continues after the first 5–10 years after SCI.
The literature on the amount of work performed after SCI is sparse and cut-offs to define full-time employment differ, thus limiting comparability. In contrast to our findings indicating a high prevalence of part-time employment, some studies indicated rates of 70.3% [43] and 80.0% [44] full-time employment among the working SCI population. However, methodological issues seem to explain the large differences compared to our findings, as full-time employment was defined as working more than 30 hours a week [44] or was not defined at all [43]. In general, data seem to indicate a shift towards reduced work load after SCI [43], potentially due to physical capacity as well as increased amounts of time needed for other activities. A recent study from Van der Meer and colleagues estimated a median of 13 hours extra time per week needed for self-care, bowel-bladder care, transportation, and handicap-management [45]. Our finding of a higher work load in men mainly reflects patterns in the Swiss general population, while an increased work load of people with incomplete lesions may be related to physical capacity. In line with our findings, higher education was also associated with an increase in the time worked [44], which could be in part explained by less physically demanding jobs. Interestingly, while we found a trend for a lower employment rate of people with non-traumatic SCI, those employed tended to work more, a finding that may be due to differential economic needs which in turn might be mediated by different social insurance benefits depending on whether SCI was due to accident or illness.
Our study has several limitations that warrant discussion. Our study sample may not be representative of the target population. However, we minimized this risk by recruiting participants using a combination of several databases, including rehabilitation hospitals, the Swiss patient association and a home-care institution for SCI [20]. Given our sampling frame, we can also not exclude unit-nonresponse bias related to LMP, because information regarding LMP status was not available at the time of survey invitation and was thus not accounted for in the sampling weights [20]. The cross-sectional data examined further preclude the study of individual trajectories related to age and cohort. Most importantly, other factors than demographics and injury-related variables, such as differences in the environment or psychological resources that have not been accounted for in this study, may explain variation in LMP. In order to identify suitable intervention targets for an enhanced LMP of persons with SCI, future empirical and experimental research needs to address modifiable environmental and personal factors in relation to bodily capacity [15, 17, 46–49]. In the comparison of the amount of work performed by employed people from the SCI and the general population, stratification by gender, age group and lesion level yielded small sample sizes in some cases precluding meaningful interpretation.
Several strengths of our study warrant mentioning as well. We used a large sample of people with SCI based on the combination of several databases. As we had basic demographic information for the sampled population, we were able to correct for unit-nonresponse by employing inverse probability weights. Moreover, we had access to data about LMP of the general population from the Swiss Federal Statistical Office and could therefore provide a detailed comparison between people with SCI and the general population. Furthermore, our analysis considered the possibility of non-linear relationships between LMP and predictors.
Conclusions
LMP of persons with SCI is comparatively high in Switzerland. LMP after SCI is, however, considerably lower than in the general population. Future research needs to show whether the reduced LMP in SCI reflects individual capacity adjustment, contextual constraints on higher LMP, or both.
Acknowledgments
We are grateful to all the participants of the SwiSCI survey for their time and effort spent in responding to the questions. We further thank Albert Marti for fruitful discussions and Binhua Reinhardt for reviewing the manuscript for spelling and readability. The members of the SwiSCI Steering Committee are: Xavier Jordan, Bertrand Léger (Clinique Romande de Réadaptation, Sion); Michael Baumberger, Hans Peter Gmünder (Swiss Paraplegic Center, Nottwil); Armin Curt, Martin Schubert (University Clinic Balgrist, Zürich); Margret Hund-Georgiadis, Kerstin Hug (REHAB Basel, Basel); Hans Georg Koch (Swiss Paraplegic Association, Nottwil); Nadja Münzel (Parahelp, Nottwil); Hardy Landolt (Representative for persons with SCI, Glarus); Mirjam Brach, Gerold Stucki (Swiss Paraplegic Research, Nottwil); Martin Brinkhof, Christine Thyrian (SwiSCI Study Center at Swiss Paraplegic Research, Nottwil).
Author Contributions
- Conceptualization: JDR MWMP CF BT MWGB.
- Data curation: MWGB SwiSCI Study Group.
- Formal analysis: MWGB.
- Investigation: SwiSCI Study Group.
- Methodology: JDR MWGB SwiSCI Study Group.
- Project administration: JDR MWGB.
- Resources: SwiSCI Study Group.
- Software: SwiSCI Study Group.
- Supervision: JDR MWGB.
- Validation: SwiSCI Study Group.
- Visualization: MWGB.
- Writing – original draft: JDR MWGB.
- Writing – review & editing: JDR MWMP CF BT MWGB.
References
- 1.
Commission of the European Communities. Situation of disabled people in the European Union: the European Action Plan 2008–2009 2007 [cited 2012 June 14]. Available from: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2007:0738:FIN:EN:PDF.
- 2.
OECD. Sickness, Disability and Work: Breaking the Barriers: A Synthesis of Findings across OECD Countries. Paris: OECD Publishing; 2010.
- 3. Biering-Sørensen F, Bickenbach JE, El Masry WS, Officer A, von Groote PM. ISCoS-WHO collaboration. International Perspectives of Spinal Cord Injury (IPSCI) report. Spinal Cord. 2011;49(6):679–83. pmid:21423254
- 4. Anderson D, Dumont S, Azzaria L, Le Bourdais M, Noreau L. Determinants of return to work among spinal cord injury patients: a literature review. J Vocat Rehabil. 2007;27(1):57–68.
- 5. Lidal IB, Huynh TK, Biering-Sørensen F. Return to work following spinal cord injury: a review. Disabil Rehabil. 2007;29(17):1341–75. pmid:17729082.
- 6. Meade MA, Reed KS, Saunders LL, Krause JS. It's All of the Above: Benefits of Working for Individuals with Spinal Cord Injury. Top Spinal Cord Inj Rehabil. 2015;21(1):1–9. Epub 2015/03/13. pmid:25762855; PubMed Central PMCID: PMC4349170.
- 7. Schedin Leiulfsrud A, Ruoranen K, Ostermann A, Reinhardt J. The meaning of employment from the perspective of persons with spinal cord injuries in six European countries. Work. 2016;55(1):133–44.
- 8.
WHO, World Bank. World Report on Disability. Geneva: WHO Press; 2011.
- 9. Bollier GE. Die wesentlichen Revisionen. Soziale Sicherheit. 2010:9–14.
- 10. Cieza A, Kirchberger I, Biering-Sorensen F, Baumberger M, Charlifue S, Post MW, et al. ICF Core Sets for individuals with spinal cord injury in the long-term context. Spinal Cord. 2010;48(4):305–12. Epub 2010/01/13. sc2009183 [pii] pmid:20065984.
- 11. Reinhardt JD, Mansmann U, Fellinghauer BA, Strobl R, Grill E, von Elm E, et al. Functioning and disability in people living with spinal cord injury in high- and low-resourced countries: a comparative analysis of 14 countries. Int J Public Health. 2011;56(3):341–52. Epub 2010/12/18. pmid:21165668.
- 12. Reinhardt JD, Ballert C, Brinkhof MW, Post MW. Perceived impact of environmental barriers on participation among people living with spinal cord injury in Switzerland. J Rehabil Med. 2016;48(2):210–8. Epub 2016/03/02. pmid:26926923.
- 13.
WHO, ISCoS. International Perspectives on Spinal Cord Injury. Geneva: WHO Press; 2013.
- 14. Ottomanelli L, Lind L. Review of critical factors related to employment after spinal cord injury: implications for research and vocational services. J Spinal Cord Med. 2009;32(5):503–31. Epub 2009/12/23. pmid:20025147; PubMed Central PMCID: PMC2792457.
- 15. Kent ML, Dorstyn DS. Psychological variables associated with employment following spinal cord injury: a meta-analysis. Spinal Cord. 2014;52(10):722–8. Epub 2014/08/06. pmid:25091108.
- 16. Young AE, Murphy GC. Employment status after spinal cord injury (1992–2005): a review with implications for interpretation, evaluation, further research, and clinical practice. Int J Rehabil Res. 2009;32(1):1–11. Epub 2008/12/06. pmid:19057392.
- 17. Trenaman L, Miller WC, Queree M, Escorpizo R. Modifiable and non-modifiable factors associated with employment outcomes following spinal cord injury: A systematic review. J Spinal Cord Med. 2015;38(4):422–31. Epub 2015/05/21. pmid:25989899; PubMed Central PMCID: PMC4612197.
- 18. Marti A, Reinhardt JD, Graf S, Escorpizo R, Post MWM. To work or not to work: labour market participation of people with spinal cord injury living in Switzerland. Spinal Cord. 2012;50(7):521–6. pmid:22249321
- 19. Post MW, Brinkhof MW, von Elm E, Boldt C, Brach M, Fekete C, et al. Design of the Swiss Spinal Cord Injury Cohort Study. Am J Phys Med Rehabil. 2011;90(11 Suppl 2):S5–16. Epub 2011/11/10. 00002060-201111000-00002 [pii]. pmid:21975676.
- 20. Brinkhof MWG, Fekete C, Chamberlain JD, Post MWM, Gemperli A. Swiss national community survey on functioning after spinal cord injury: Protocol, characteristics of participants and determinants of non-response. J Rehabil Med. 2016;48(2):120–30. Epub 2016/03/02. pmid:26926916.
- 21. Fekete C, Segerer W, Gemperli A, Brinkhof MWG. Participation rates, response bias and response behaviours in the community survey of the Swiss Spinal Cord Injury Cohort Study (SwiSCI). BMC Med Res Methodol. 2015;15(1):80. Epub 2015/10/10. pmid:26450702; PubMed Central PMCID: PMC4599658.
- 22.
ILO. Resolution concerning statistics of work, employment and labour underutilization. 19th International Conference of Labour Statisticians (Geneva, 2–11 October 2013) Geneva: International Labour Organization; 2013.
- 23.
Hussmanns R. Measurement of employment, unemployment and underemployment—Current international standards and issues in their application. Geneva: International Labour Organization; 2007.
- 24.
UNESCO. International standard classification of education (ISCED). UNESCO, editor. Paris1997.
- 25.
Swiss Federal Statistical Office. Die Schweizerische Arbeitskräfteerhebung (SAKE): Konzepte—Methodische Grundlagen—Praktische Ausführung. Neuchâtel, Switserland2004.
- 26. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Bull World Health Organ. 2007;85(11):867–72. pmid:18038077.
- 27.
Little RJA, Rubin DB. Statistical Analysis with Missing Data. New York: John Wiley; 1987.
- 28. Jenkins P, Earle-Richardson G, Burdick P, May J. Handling Nonresponse in Surveys: Analytic Corrections Compared with Converting Nonresponders. Am J Epidemiol. 2008;167(3):369–74. pmid:17998277
- 29. Stekhoven DJ, Bühlmann P. MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics. 2012;28(1):112–8. Epub 2011/11/01. pmid:22039212.
- 30. Sauerbrei W, Meier-Hirmer C, Benner A, Royston P. Multivariable regression model building by using fractional polynomials: Description of SAS, STATA and R programs. Computational Statistics and Data Analysis. 2006;50:3464–85.
- 31.
Royston P, Sauerbrei W. Multivariable model-building: A pragmatic approach to regression anaylsis based on fractional polynomials for modelling continuous variables. Shewhart WA, Wilks SS, editors: John Wiley & Sons; 2008. 322 p.
- 32. Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Stat Med. 2006;25(1):127–41. Epub 2005/10/12. pmid:16217841.
- 33.
Gelman A, Hill J. Data analysis using regression and multilevel/hierarchical models. New York, NY 10013–2473, USA: Cambridge University Press; 2007.
- 34.
Cameron AC, Trivedi PK. Regression analysis of count data. Second ed. Cambridge, U.K.: Cambridge University Press; 2013 May 2013.
- 35. Liu X. Fitting stereotype logistic regression models for ordinal response variables in educational research (Stata). Journal of Modern Applied Statistical Methods. 2014;13(2, Article 31).
- 36. Anderson JA. Regression and ordered categorical variables. Journal of the Royal Statistical Society Series B (Methodological). 1984;46(1):1–30.
- 37. Greenland S. Alternative models for ordinal logistic regression. Stat Med. 1994;13(16):1665–77. Epub 1994/08/30. pmid:7973242.
- 38. Brant R. Assessing proportionality in the proportional odds model for ordinal logistic regression. Biometrics. 1990;46(4):1171–8. pmid:2085632
- 39. Ferdiana A, Post MW, de Groot S, Bultmann U, van der Klink JJ. Predictors of return to work 5 years after discharge for wheelchair-dependent individuals with spinal cord injury. J Rehabil Med. 2014;46(10):984–90. Epub 2014/08/29. pmid:25167341.
- 40.
Anderson C, Ballantyne E, B J.E., Boschen K, Boucher N, Gray D, et al. Chapter 8: Education and empoloyment. International Perspectives on Spinal Cord Injury. 2013. Geneva: WHO Press 2013.
- 41. Marti A, Boes S, Lay V, Reuben Escorpizo PT, Trezzini B. The association between chronological age, age at injury and employment: Is there a mediating effect of secondary health conditions? Spinal Cord. 2016;54(3):239–44. Epub 2015/10/07. pmid:26439844.
- 42. Krause JS, Terza JV, Saunders LL, Dismuke CE. Delayed entry into employment after spinal cord injury: factors related to time to first job. Spinal Cord. 2010;48(6):487–91. Epub 2009/11/26. pmid:19935754.
- 43. Murphy G, Brown D, Athanasou J, Foreman P, Young A. Labour force participation and employment among a sample of Australian patients with a spinal cord injury. Spinal Cord. 1997;35(4):238–44. Epub 1997/04/01. pmid:9143087.
- 44. Meade MA, Barrett K, Ellenbogen PS, Jackson MN. Work intensity and variations in health and personal characteristics of individuals with Spinal Cord Injury (SCI). J Vocat Rehabil. 2006;25 13–9.
- 45. Van der Meer P, Post M, van Leeuwen C, van Kuppevelt H, Smit C, van Asbeck F. Impact of health problems secondary to SCI one and five years after first inpatient rehabilitation. Spinal Cord. 2016. Epub 19 July 2016. pmid:27431661.
- 46. Boyle CL, Nott MT, Baguley IJ, Ranka JL. Contextual influences on employment of people with dual diagnosis: spinal cord injury and traumatic brain injury. Aust Occup Ther J. 2014;61(5):335–43. Epub 2014/05/09. pmid:24810135.
- 47. Thomas FP, Goetz LL, Dixon T, Ho C, Holmes SA, Sandford P, et al. Optimizing medical care to facilitate and sustain employment after spinal cord injury. J Rehabil Res Dev. 2014;51(6):xi–xxii. Epub 2014/12/06. pmid:25479192.
- 48. Frieden L, Winnegar AJ. Opportunities for research to improve employment for people with spinal cord injuries. Spinal Cord. 2012;50(5):379–81. Epub 2012/04/11. pmid:22487953.
- 49. Roels EH, Aertgeerts B, Ramaekers D, Peers K. Hospital- and community-based interventions enhancing (re)employment for people with spinal cord injury: a systematic review. Spinal Cord. 2016;54(1):2–7. Epub 2015/08/26. pmid:26305872.