Vocational rehabilitation following traumatic brain injury: A quantitative synthesis of outcome studies

Authors: Khan, F., Ng, L., & Turner-Stokes, L.
Year Published 2008
Publication Journal of Vocational Rehabilitation
Volume 25
Number 3
Pages 149-160
Publisher IOS Press

Return to work following traumatic brain injury can range from 20% to 90%. Rehabilitation professionals need to sound research to guide best practices in service delivery. For years there has been a call for using evidenced based approaches to medical care and rehabilitation. Despite this fact, evidence related to the efficacy of vocational rehabilitation remains inconclusive. To further complicate the issue the wide range of outcomes and interventions makes it difficult to determine employment outcomes for individuals with TBI. A synthesis of the existing literature should help identify effective vocational rehabilitation practices to assist individuals with TBI with employment.


The purpose of the study was to consolidate findings from existing studies in order to classify return to work outcomes for individuals with traumatic brain injury.


This study is a systematic review. The included studies were undertaken in various locations and settings. The quantitative synthesis took place at the Centre of National Research on Disability and Rehabilitation Medicine and Griffith University in Australia.


The sample included 26 studies involving individuals with traumatic brain injury. Among these the number of participants ranged from 9 to 1170 with a total of 3792.

Data Collection

A multimethod search strategy was used to find all possible studies related to vocational rehabilitation and traumatic brain injury between the years of 1993 and 2000. Twenty six studies were included. Data extracted from each of the studies included baseline employment percentages for the sample where reported (i.e., pre-injury employment rate), post-injury employment percentages, definitions of return-to-work (i.e., restricted definition pertaining to full-time competitive employment only or inclusive definition pertaining to any employment or productive activity including any competitive work, study, or other functional activity), and follow-up time(s) post-injury. The traditional meta-analytic approach was not applied in this study, as the assumptions of parametric analysis were violated. Specifically, both return-to-work and intervention are intrinsically linked to time, producing non-linear effects and non-independent observations. Rather than employing a single aggregate method as in meta-analysis, quantitative synthesis was conducted using non-linear exponential modeling through the SAS PROC NLIN1 module. To confirm the validity of the non-linear analysis, the results were checked using an approximate permutation test (also through SAS PROC NLIN). The analysis was conducted using follow-up time, return-to-work patterns, and the return-to-work definition (restricted or inclusive) to develop an overall model of unemployment following injury. Non-linear models were constructed to predict unemployment at a given time from all available data points. Comparisons were made by comparing the parameters obtained against parameters computed by random permutation of the relevant covariates thus producing an approximate permutation test of the original estimates. Degrees of freedom were based on the number of studies included in the analysis rather than the number of participants variance estimates could not be derived as a result of biased means and standard deviations (caused by possible right censoring in the data).


There were no comparison or control conditions.


The selected studies (n = 26) were combined to produce 37 separate observations or data points. In terms of return-to-work, the 17 intervention studies generally appeared to produce higher percentages of successful adjusted return-to-work using an inclusive definition across all follow-up time points than did the 20 non-intervention studies. Using an inclusive definition increased the apparent rate of return to work by an average of 17% (SD = 12.5%). When the rate using the restricted definition for these 11 data points was aggregated and compared against the rate using the inclusive definition for the remaining 26 data points a similar finding emerged, with average return to work at 68% for the inclusive definition values and 48% for the restricted definition values. This finding approached significance. There was a negative correlation between unemployment and follow up time where the longer duration following injury led to a return to workplace or productive activity.


The research seems to indicate that rehabilitation interventions may be a major determinant in return to work regardless of severity of injury or other injury variables. Standardized terminology and methods for reporting return to work outcomes over the long term are needed. Synthesis of the literature that fail to do so will likely reach wrong conclusions.

URL http://content.iospress.com/articles/journal-of-vocational-rehabilitation/jvr00352
Disabilities Traumatic brain injury (TBI)
Populations Male & Female
Outcomes Employment acquisition | Return to work
Research Design Systematic reviews and meta-analysis
Peer Reviewed Yes