FACTORS WHICH ARE PREDICTIVE OF RETURN TO WORK AFTER STROKE

Correspondence Author: Dr. Veronica Ntsiea Physiotherapy department Faculty of Health Sciences University of the Witwatersrand 7 York Road, Parktown 2193 Email: Veronica.Ntsiea@wits.ac.za ABSTRACT: Stroke impacts on a survivor’s ability to participate in community activities such as return to work (RTW) and affects people who are within the working age. There is a dearth of literature on RTW after stroke in developing countries. This study aimed to bridge this gap in South Africa, and was conducted within the Gauteng province as it comprises the largest share of the South African population. Seventy two stroke survivors participated in this cross-sectional study. A demographic questionnaire; Barthel index; Modified Rivermead mobility index and Montreal cognitive assessment were used to determine the characteristics of study participants. The mean (standard deviation) scores for the Barthel Index (BI), Modified Rivermead mobility index (MRMI) and Montreal cognitive assessment (MoCA) were 19.6 (±0.2), 39.5 (±0.9) and 25.1 (±4.8) respectively. Thirty one (43%) of the stroke survivors returned to work at six months after stroke. Stroke survivors with left hemiplegia had a greater chance of RTW than those with right hemiplegia (odds ratio 7.7). For every unit increase in the BI and MoCA score, the likelihood of RTW increased by 1.6 and 1.3 respectively. Conclusion: Side of hemiplegia, independence in activities of daily living and cognitive ability were found to be predictors of RTW at six months after stroke. It is important to identify people with cognitive impairments after stroke so that efforts can be made to increase awareness of the potential role that cognitive impairments may play in RTW.


INTRODUCTION
Stroke affects people who are still within the working age and the incidence of stroke in people younger than 65 years has increased over the past few years with almost 5% occurring in those younger than 45 years of age (Roding et al 2003).The increasing number of working age stroke survivors was also reported in Wolf et al ( 2009)'s study.They reported that a significant number of people who have had a stroke are less than 55 years old.This implies that with the increasing number of young stroke survivors and improved health services, stroke survivors are likely to live longer and would need to continue working for financial sustainability.Independent income generation would also provide a sense of purpose in life for young stroke survivors.
An independent income is more financially beneficial than relying on a social grant (de Koker et al 2006).Rehabilitation programmes for stroke survivors should focus on resumption of work where possible rather than on encouraging stroke survivors to apply for early retirement or disability grants (McLean 2007).In Duff (2012)'s study on stroke survivors within the Gauteng province, which included 97 stroke survivors up to two years after stroke, only one stroke survivor had vocational rehabilitation.This indicates that reha bilitation of stroke survivors is focused on recovery of activities necessary for daily living, while often overlooking psychosocial recovery which includes RTW (Guise et al 2010).
An important social outcome of rehabilitation for working age stroke survivors is returning to work after stroke (Daniel et al 2009).Duff (2012) found a RTW rate of 34% for stroke survivors within the Gauteng province of South Africa up to two years after a stroke.Peters et al (2012) in their Nigerian study found a RTW rate of 55% for stroke survivors up to eight years after stroke.Most of the people who RTW, do so within three to six months, with a second peak of RTW at 12 to 18 months after stroke (Treger et al 2007).Thus the focus on RTW should be within three to six months period after stroke.For those who do not RTW within six months, there might be a need to follow them up to 18 months to assist them with RTW where possible.
Stroke survivors experience poorer quality of life and worse psychosocial outcomes than healthy controls (Roding et al 2003).This can be due to the significant work and social dysfunction experienced by most stroke survivors (Hommel et al 2009).Return to work can improve these outcomes by contributing to life satisfaction, wellbeing, self worth and social identity, at least partly through independence gained from income generation (Medin et al 2006).It is important to acknowledge that not all stroke survivors will RTW as a variety of factors such as the level of functional ability affect RTW (Vestling et al 2003).Besides level of functional ability, high unemployment rate and poor socio economic status may also affect RTW as disability insurance and routine social security benefits may deter stroke survivors from returning to work (Treger et al 2007) There is a dearth of literature on RTW after stroke in developing countries.Thus, the aim of this study was to esta blish factors that are predictive of RTW.This study was conducted within the Gauteng province as it comprises the largest share of the South African popu lation (Statistics South Africa 2013).

METHODS
A quantitative crosssectional study was carried out to establish factors which are predictive of RTW.Participants were recruited from two government hospi tals and one private hospital.This study was part of a randomised controlled trial (RCT) study which was testing a workplace intervention programme for stroke survivors.The sample for this cross sectional study consisted of stroke survivors who were blue and white collar employees and worked in various employment sectors.The sample size calculation was based on the principle that for every factor that is considered valid and reliable tool that requires mini mal training before use (Johnson and Selfe 2004).The MRMI has responsive ness (effect size =1.15),testretest reli ability (r = 0.731), interrater relia bility (ICC = 0.98) and internal consistency (Cronbach's alpha = 0.93).For there to be a significant change in the patient's mobility level, there has to be a greater than 4.5 change in the MRMI score (Lennon and Johnson 2000).
The Montreal Cognitive Assessment (MoCA) was used to establish the cogni tive ability of the study participants.The MoCA has sensitivity of 90% and 94% when using a cutoff score of 26 (Smith et al 2007).The testretest reliability of the MoCA is 0.92 with an internal con sistency of 0.83 (Nasreddine et al 2005).The total possible score is 30 points and a score of 27.4 (2.2) or above is consi dered normal, 22.1 (±3.11) signifies mild cognitive deficits and that of 16.2 (±4.8) indicates that there is severe cognitive impairment (Nasreddine et al 2005).
A return to work questionnaire was used to establish the rate of RTW.This questionnaire was developed and validated for content by Duff (2012) in the Gauteng province.It is a tick list that indicates whether stroke survi vors returned to work (premorbid job or new job), duration of time they had been back at work, intention to stay at work or resign, which adaptations were made to the working environment, and the current financial role that the stroke survivor holds in the family.If he/she did not RTW, a list of possible reasons was provided for the stroke survivor to select from (for example family did not allow her/him to go back to work; lack of transport; embarrassed about the dis ability; building not accessible).

PROCEDURE
Physiotherapists in charge of the neuro logical rehabilitation sections at the three recruitment hospitals were contacted by the first author on a regular basis to identify new stroke survivors who fitted the inclusion criteria.She then contacted the stroke survivors who met the inclu sion criteria telephonically to make an appointment for interviewer admini stered completion of the demographic questionnaire, MoCA, MRMI, BI and to have a possibility of influencing the results of the study, at least ten partici pants are required (Nunnaly 1978).The previously mentioned seven broad cate gories of factors that were reported to influence RTW were used to calculate the minimum sample size for this study (n=70).Stroke survivors were included if they were employed at the time of having stroke and were independent in activities of daily living (ADL) before the stroke.The diagnosis of stroke was primarily based on the history of the current medical condition and clinical presentation.

INSTRUMENTATION
A demographic questionnaire was deve loped to establish the characteristics of all stroke survivors included in the study.A project development group meeting with eight stroke rehabilitation therapists (two physiotherapists working in the field of neurological rehabilitation, two occupational therapists with vocational rehabilitation experience, three therapists with research experience, and one thera pist working in the field of public health) was held to validate the content of the demographic questionnaire.Activities of daily living were established using the Barthel Index (BI).The coefficient alpha of the BI was found to be 2.0 by Green et al (2001).There are many ver sions of the BI, but the 10 item BI with a total score of 20 (Collin et al 1988) was used in this study.This choice was based on the high reliability coefficient of this version (0.99) when compared with the original BI (0.87), Modified BI (0.95) and the Modified scoring (0.90) (Shah et al 1989).A BI score of at least 60% indicates that a person is independent for vital care, such as moving around unassisted, sphincter control, and eating in addition to personal toileting (Sulter et al 1999).Independence with minimal assistance (e.g.ability to get dressed and move from armchair to bed unassisted) was found to coincide with a score of 85% (Sulter et al 1999).A BI score of 90% (18/20) signifies the turning point whereby help from another person is not required (Uyttenboogaart et al 2005).
Mobility functional level was esta blished using the Modified Rivermead Mobility Index (MRMI).The MRMI is a RTW questionnaire, after consent was obtained.This was done at six months after the stroke at the stroke survivors' homes.

DATA ANALYSIS
The STATA software programme ver sion 12.1 was used to analyse data.Categorical data were summarised as numbers and percentages.Continuous data were summarised as means and standard deviations.The percentage of stroke survivors who returned to work and those who did not RTW was cal culated to obtain the rate of RTW after stroke.To identify factors that are pre dictive of RTW the following methods were used: an initial univariate analysis was done to identify all the variables that had significance in the prediction of RTW.This was followed by a stepwise multivariate analysis, dropping insigni ficant variables in the process.The sig nificance of the study was set at p<0.05.
Permission to conduct the study at the research sites and Gauteng Department of Health was sought and granted.
Ethical clearance was granted by the University of the Witwatersrand com mittee for research on human subjects.

RESULTS
Seventy two stroke survivors partici pated in this study.The average age for the study sample was 45 (±8.7) years, 41 (51%) were male and 41 (51%) had right hemiplegia.The majority (55%) of the stroke survivors were married, breadwinners (63%), had a grade 11 to 12 educational level (64%), an income above R5000 (46%) and had a helper (74%) whom they did not have to pay (81%).The average number of people who were financially dependent on the stroke survivor was 3 (±1.8).Thirty nine percent of the stroke survivors had speech problems with the most common speech problem being dysphasia (45%).Thirty five (49%) of the stroke survivors received a workplace intervention pro gramme within the RCT study.
Figure I illustrates the most common comorbidities found in the study sam ple.All stroke survivors in this study received physiotherapy, followed by occupational therapy (78%), psycho logy/social work (48%) and speech therapy (44%).
The average BI, MRMI and MoCA scores indicating the stroke survivors' ADLs, mobility and cognitive ability six months after stroke are shown in Table 1.The BI mean score was above 18 (90%) indicating that most stroke survivors were independent with ADLs.The mean MoCA score was 25.1 (±4.8) which is less than the score for normal cognitive ability of 27.4 (2.2) (Nasreddine et al 2005) indicating that some of the stroke survivors had cognitive impairments.
Thirty one (43%) of the stroke survi vors returned to work six months after the stroke.The job description was changed for most (55%) of the stroke survivors (light duty or changed from labour intensive to administrative); however very few (23%) had work adaptations made.The majority (97%) of stroke survivors who had returned to work returned to full time employment.The most common occupation was that of administrators (34%) followed by domestic workers/shop packers (23%); however in total there were more blue (55%) than white (45%) collar occupa tions.White collar workers in this study are those that perform work in an office environment and may involve sitting at a computer or desk and blue collar workers are members of the working class who perform manual labour which involve skilled or unskilled, manufactur ing, mining, construction, mechanical, maintenance, technical installation and many other types of physical work.
Table 2 shows results of the occupa tions and key physical job demands of stroke survivors who returned to work and those who did not RTW at six months after the stroke.There was no statistically significant difference between RTW rate and the stroke sur vivors' occupation and physical job demands.The average number of years that stroke survivors had in their current occupation was 10.5 (±8.6) and the ave rage sick leave duration was 7.2 (±3.3) weeks.The reasons for returning or not returning to work after the stroke are presented in Table 3.The most common reason for RTW was financial (61%) and  for not returning to work was that their previous employer would not offer them their job again (29%).Factors which were predictive of RTW are shown in Table 4. Stroke sur vivors who had left hemiplegia had a 7.7 greater odds of RTW than those with right hemiplegia (p < 0.001; CI: 2.6 -23.2). Results of the stepwise multivari ate analysis indicate that for every unit increase in the BI and MoCA score, the likelihood of RTW increased by 1.6 and 1.3 respectively.An analysis to establish if the influence of side of hemiplegia on RTW could be due to speech problems was done.Stroke survivors with right hemiplegia were 7.1 times more likely to have speech problems than those with left hemiplegia (p < 0.001; CI: 2.4 -20.7).Stroke survivors who received a workplace intervention in the RCT study were 6.9 times more likely to RTW than those who did not receive the intervention.

DISCUSSION
The primary goal of this study was to establish factors that are predictive of RTW among stroke survivors in Gauteng at six months after stroke.Factors which were found to have a significant rela tionship with RTW were independence in ADL (BI), cognitive ability (MoCA); and side of hemiplegia.
Improvement in independence with ADL increased the likelihood of RTW for stroke survivors in this This confirms Saeki (2000)'s findings that independence with ADL is an important indicator of whether a stroke survivor would RTW or not.Independence in ADL has also been identified as a sig nificant determinant of RTW by other researchers (Treger et al 2007).Poor performance in ADL is indicative of severity of hemiplegia which is a nega tive predictor of RTW (Peters et al 2012;Treger et al 2007).
Functional ability should not be considered in isolation as a factor that influences RTW because most stroke sur vivors in this study were independent in ADL at six months following the stroke but less than half eventually returned to work.This indicates that a combina tion of factors contribute towards ability to RTW.Most of the stroke survivors would seem to be independent in ADL and functional ability even though they may not be able to leave their homes because of activity limitations such as decreased walking velocity (Schmid et al 2007).A limitation of this study is that walking speed and endurance were not assessed; however Allen et al ( 2011) found that people with hemiparesis generally have low walking speeds with values ranging from 0.23 to 0.73 metres/ second.It is thus possible that inability to RTW in this study population which had good ADL and mobility scores, might be due to poor walking speed and endurance which is required to RTW.Reduced walking speed may add to the travelling time of the stroke survivor if they have to walk to get public transport on their way to work.This may be the case for stroke survivors who indicated that their inability to access transport was the main reason they did not RTW.
Improvement in cognitive ability as measured with the MoCA was related to an increased likelihood to RTW in this study.These findings agree with those reported by Roding et al (2003).Jones et al (2006) also showed that when a person has cognitive impairments, they experience the lowest employment rates as they are less likely to be employed in the competitive labour market.If a less cognitively demanding job is not avail able for stroke survivors with cognitive impairments at their workplace, the employer might not be able to accom modate them.It is important to identify people with cognitive impairments early after stroke so that efforts can be made to increase awareness of the potential role that the cognitive impairments may play in the RTW process (O'Connor et al 2011).Therapists' knowledge of a stroke survivor's cognitive level will also increase her/his chances of receiving cognitive training during the rehabilita tion process and thus increase the like lihood of RTW (McDowd et al 2003).Most stroke survivors in this study did not select cognitive impairments as the main reason for not returning to work even though the mean MoCA scores were below the normal score.This is not an unexpected finding as cognitive impairments are common following stroke (Roding et al 2009).There is a possibility that cognitive impairments were barriers to RTW even though it was not specifically identified by the stroke survivors.
The majority of the stroke survivors with left hemiplegia returned to work and had greater odds of RTW than those with right hemiplegia.It was also found that those with right hemiplegia had a much higher likelihood to suffer speech problems than those with left hemi plegia.Speech problems were more of a limiting factor for RTW than the hemi sphere lateralisation of stroke and this can be seen in that 12% of the stroke survivors indicated that expressive aphasia was their main reason for not returning to work.Bernspang and Fisher (1995) also established that stroke sur vivors with left and right hemiplegia have hemisphericspecific differences in motor impairments, but do not differ significantly in ADLs.Social function ing, including RTW, was also not sig nificantly related to hemisphere latera lisation of stroke in a study by Hommel et al (2009).It would be reasonable to conclude that speech impairments, and not hemisphere lateralisation, influenced RTW in this study.
Since most jobs require speech skills, speech impairments could make some types of work difficult.This may result in some stroke survivors having to change occupations if they cannot be reasonably accommodated for speech impairments in their current occupation.The limiting effect of speech impair ments and the negative association between RTW and aphasia was also reported by Alaszewski et al (2007).The results of this study confirm the findings of these researchers.

LIMITATIONS
Age, educational level, higher income and type of occupation were found to have an influence on RTW in other stu dies (Rollnik and Allman 2011;Trygged et al., 2011;Grammenos 2003;Vestling et al 2003), but did not have an influ ence on RTW in this study.This might be due to the workplace intervention received by 49% of the stroke survivors as the studies referred to were mainly observational with no intervention to facilitate the RTW.Return to work in this study may have been facilitated more by the RTW intervention which may have resulted in employer understanding and willingness to accom modate some of the stroke survivors, irrespective of their age, educational level and type of occupation.Thus, it may be useful to conduct another study with no deliberate intervention group in order to get a true reflection of factors which predict RTW after stroke.

CONCLUSION
Side of hemiplegia, independence in activities of daily living and cognitive ability were found to be predictors of RTW at six months after stroke.Vocational rehabilitation for stroke survivors should include early iden tification of cognitive and speech impairments and should strive for improvement in ADL and mobility functional ability.In addition to improving functional, cognitive and speech function, it is important for therapists to identify stroke survivors with potential to RTW and assist them with the RTW process.This will require involvement in the workability assessment process and will require that therapists explore possibilities of reasonable accommodations.

Figure
Figure I: Co-morbidities among stroke survivors