More than 50,000 individuals in Canada experience a stroke every year, and an estimated 315,000 Canadians are currently living with the prolonged consequences of stroke. 1 More than 60% of stroke survivors have physical deficits requiring some degree of physical therapy, 2 and approximately 40% have persisting cognitive deficits.Reference Patel, Coshall, Rudd and Wolfe 3 Following acute care, stroke survivors are often discharged to inpatient rehabilitation where the aim is to maximize physical, psychosocial, and cognitive recovery.Reference Good, Betterman and Reichwein 4 Despite the progress made during inpatient rehabilitation, approximately 33% of stroke survivors will still have deficits requiring additional rehabilitation in the community.Reference Mayo, Wood-Dauphinee, Cote, Durcan and Carelton 5
The inability to access health care services may impede a patient’s recovery and consequently result in an increased risk of medical complications, depression, cognitive difficulties, and decreased quality of life.Reference Good, Betterman and Reichwein 4 , Reference Teasell, Foley, Salter, Richardson, Allen and Hussein 6 It has been shown that individuals residing in rural settings often have poorer access to health care, especially rehabilitation services, poststroke.Reference Jia, Cowper, Tang, Litt and Wilson 7 - Reference Rodriguez, Cox, Zimmer, Olson, Goldstein and Drew 9 Given these inequalities, it is presumed that recovery poststroke may differ between rural and urban residing individuals; however, there is inadequate research examining such a relationship. What it means to live in a rural area also varies according to geographical region.Reference Jia, Cowper, Tang, Litt and Wilson 7 In Canada, the criteria for an urban area includes a population of at least 1000 and a density of at least 400 people per square kilometre, with the remaining area defined as rural. 10 By this definition, approximately 19% to 30% of Canadians live in rural areas. 11
Ontario, Canada’s most densely populated province, is divided into 14 Local Health Integration Networks (LHINs). The southwest LHIN (Figure 1) serves nearly 1 million people in an area of 21,639 square kilometers. It consists of eight counties; the southern four counties are the most densely populated and are home to half the region’s population, many of whom reside in London, a city with 350,000 residents. The northern and middle four counties are almost entirely rural with many small towns and only two small cities. 12
Map of the southwest local health integration network. 25
A potential solution for improving access to rehabilitation services for rural stroke survivors is home-based rehabilitation. The Community Stroke Rehabilitation Teams (CSRTs) in the southwest LHIN are one such example of a home-based rehabilitation program that provides coordinated personalized support and therapeutic services to clients recovering from stroke in both urban and rural locations. The objective of this study was to compare the gains between urban and rural residents who received specialized multidisciplinary stroke rehabilitation services from CSRTs.
Methods
The CSRTs provide interdisciplinary, individualized services to any stroke survivor with ongoing rehabilitation needs and consist of a physiotherapist, occupational therapist, speech language pathologist, social worker, registered nurse, therapeutic recreation specialist, and rehabilitation therapist. Clients accessing the CSRT program are individuals for whom traditional outpatient stroke rehabilitation following hospital discharge is either unavailable or inaccessible. Individuals may enter the program at any time poststroke and receive visits based on individual need (both therapy type and frequency). Individual therapists travel to a client’s home to provide therapy. Team rounds are held weekly to discuss client rehabilitation goals.
This study was granted ethics approval by the Western University Research Ethics Board.
Dataset
Data were obtained from the CSRT central administrative database. Data were collected on CSRT clients between January 2009 and June 2013 (N=1222). Client demographic information collected on admission to the CSRT program included age, gender, postal code, marital status, and date and type of stroke. CSRT service information included date of client referral, date of first CSRT team visit, and date of discharge from active services. All visits where therapy was provided by a physiotherapist, occupational therapist, speech language pathologist, social worker, registered nurse, therapeutic recreation specialist, or rehabilitation therapist were recorded by date.
Outcomes
Outcomes were assessed on admission to, and discharge from, the CSRT program with a follow-up assessment completed within 6 months postdischarge. All outcome measures were administered by a member of the CSRTs. Assessments captured functional and psychosocial outcomes. The Functional Independence Measure (FIM)Reference Keith, Granger, Hamilton and Sherwin 13 and either the Stroke Impact ScaleReference Duncan, Wallace, Studenski, Lai and Johnson 14 or the Reintegration to Normal Living Index (RNLI)Reference Wood-Dauphinee, Opzoomer, Williams, Marchand and Spitzer 15 were used to assess functional outcomes. The Stroke Impact Scale was assessed by CSRT staff from January 2009 to October 2012, but was replaced with the RNLI in May 2012. Client’s psychosocial status was assessed using the Hospital Anxiety and Depression Scale (HADS)Reference Zigmond and Snaith 16 until January 2011, when it was replaced by the Patient Health Questionnaire-9. For further detail regarding methodology and outcome measures, please refer to Allen et al.Reference Allen, Richardson, McIntyre, Janzen, Meyer and Ure 17
Inclusion Criteria and Rurality
To be included in this study, clients in the dataset had to have satisfied four inclusion criteria: (1) demographic information available; (2) at least one baseline outcome measure completed; (3) active therapy received; and (4) at least one discharge or follow-up outcome measure completed. Active therapy was defined as a client receiving a minimum of four visits from any one discipline. Individuals were assigned a Rurality Index for Ontario score generated using the client’s six-digit postal code. The Rurality Index for Ontario score provides an indication of the degree of rurality of residence on a continuous scale (0 to 100), with scores greater than or equal to 40 indicating a rural residence. 18 Each postal code was individually entered into the Ontario Medical Association’s Rurality Index for Ontario Postal Code Look Up tool bar. 18 Rurality of the client’s residence was the main variable of interest in this study.
Data Analysis
Rehabilitation intensity was measured by the number of visits a client received from each health care discipline as well as the average number of visits received per week in the program. Descriptive analysis, using means with standard deviations and frequencies where appropriate, were completed to describe both rural- and urban-dwelling cohorts. Independent t-tests and chi-squared tests were conducted to examine similarities in baseline demographic (age, sex, stroke type, marital status, weeks since stroke) and CSRT service delivery characteristics (i.e. wait time, number of visits, length of stay in the program) between the two populations.
Mean changes in outcome measures scores between admission and discharge, and discharge and follow-up, were also calculated for both rural and urban populations. Significant between-group differences were tested using independent t-tests in an unadjusted analysis.
A series of multiple linear regression analyses was completed to assess the relationship between a client’s rurality of residence and change in functional and psychosocial outcomes over time. The dependent variable for each regression analysis was the mean change in outcome score from admission to discharge, and discharge to follow-up. To isolate the effects of rurality of residence on client outcomes, the independent variables of interest were entered into the analysis in two steps. The first step included age and admission FIM score as well as service and demographic characteristics considered conceptually appropriate to the particular outcome of interest. The second step included the binary rural/urban variable. The change in the amount of variance explained by the model (R 2) after the addition of the rurality variable was assessed for significance (p<0.05). To account for multiple comparisons, a Bonferroni corrected level of statistical significance was defined as p<0.003 (95% confidence interval [CI]). All analyses were completed using SPSS, v.21.0.
Results
A total of 786 clients (64%) of a possible 1222 met the inclusion criteria for this study. Table 1 describes the demographic and program characteristics of both the urban and rural cohorts. No statistically significant differences were observed between the two groups for any of the baseline demographic characteristics. There were significant differences between the groups with respect to amount of services provided by the CSRT. Rural residing individuals received significantly more visits overall as well as significantly more visits from the therapeutic recreation specialist, physiotherapist, occupational therapist, and rehabilitation therapist. The total length of stay in the program was also significantly longer for the rural group (p<0.001). However, there were no statistically significant differences between the groups according to intensity of rehabilitation received.
Demographic and CSRT service-related variables
CSRT=Community Stroke Rehabilitation Team; OT=occupational therapist; PT= physiotherapist; RN=registered nurse; RT=rehabilitation therapist; SD=standard deviation; TRS=therapeutic recreation specialist; SLP=speech language pathologist; SW=social worker.
Baseline outcome scores and change in outcome scores between time points (unadjusted analysis)
ADL=activities of daily living; FIM=Functional Independence Measure; HAPS=; RNLI=Reintegration to Normal Living Index; SIS=Stroke Impact Scale.
* Physical indicates the composite score of strength, mobility, hand strength, and social participation.
Results of multivariate linear regression analyses (adjusted analysis)
FIM=Functional Independence Measure; HAPS=; RNLI=Reintegration to Normal Living Index; SIS=Stroke Impact Scale.
* Physical indicates composite score of strength, mobility, hand strength, and social participation.
When examining baseline scores of all outcome measures, the only significant between-group differences observed were on the FIM (p=0.004; 95% CI, −6.4 to −1.3) as well as the communication domain of the SIS (p=0.028; 95% CI, −8.7 to −0.52). In both instances, rural residing individuals had higher scores, on average, at baseline than urban clients. There were no significant differences between groups in the ability to make gains on any functional or psychosocial outcome measure between admission to and discharge from the program or between discharge and follow-up.
Rurality of residence did not significantly improve the amount of variance explained in any of the regression analyses. However, before conducting a Bonferroni adjustment, the HADS depression subscale between admission and discharge was found to be significant (p=0.040, R 2 change: 0.026). In this case, the rurality variable increased the amount of variance explained by 2.6%. Urban residents were able to demonstrate a greater ability to reduce their depression subscale score (mean difference: urban −1.2±3.7 vs rural −0.5±3.6), suggesting that there was a greater improvement in depressive symptoms in this cohort during service provision.
Discussion
CSRTs are an intensive and multidisciplinary stroke rehabilitation service offered in southwestern Ontario for individuals who are unable to access outpatient rehabilitation services or who are believed to benefit more from home-based rehabilitation. Clients living in both rural and urban settings are eligible to receive rehabilitation services via CSRTs. The objective of this study was to compare differences in improvement of poststroke outcomes between urban and rural residing individuals who received services from CSRTs. Findings indicated that both sets of clients experienced improvements in functional and psychosocial outcomes. Encouragingly, there was no evidence to suggest a disparity between the gains made between rural and urban stroke survivors. Our findings provide preliminary evidence that the CSRTs are “leveling the playing field” by providing stroke care in the community regardless of a clients’ rurality of residence.
In the literature, there is limited and conflicting evidence regarding differences in stroke recovery between these geographically distinct populations. In a large review comparing urban and rural settings around the world, Joubert et alReference Joubert, Prentice, Moulin, Liaw and Jourbert 8 reported that stroke demographics and information on medical care (e.g. stroke incidence, prevalence, mortality, disability, provision of services, management) were extremely variable regardless of a country’s level of economic development (i.e. high, medium, or low). Much of the research that does exist has been completed in the United Kingdom, Australia, and United States. Comer and MuellerReference Comer and Mueller 19 compared urban-rural differences on several measures of access to health care in Nebraska in the United States. In contrast, O’Neil et alReference O’Neill and Godden 20 examined provision of stroke resources and outcomes in Scotland in accessible-remote communities (population <3000 individuals 21 ). This research is not particularly applicable to a Canadian context because of large variations in health care system structures and provision of services as well as geographical inconsistencies and varying taxonomies, resulting in diverse definitions of rurality.Reference Jia, Cowper, Tang, Litt and Wilson 7 Despite this, the current study’s findings are supported by O’Neil et al,Reference O’Neill and Godden 20 who also found no differences in recovery patterns between the two populations. Overall, there are currently too few studies to offer direct comparisons between our findings from Canada and other countries.
Disparities in health and rehabilitation between rural and urban dwelling individuals remain an important issue throughout Canada and the world. There are a large number of individuals living in population centres with fewer health care supports and services than is typically found in large urban areas. It has been shown that the geographic location where one resides has an association with their health status, personal health behaviours, and health service utilization.Reference Sibley and Weiner 22 According to a survey of the Canadian population, individuals living in rural communities were less likely to have a general practitioner and had reduced access to specialized medical services than those in urban areas.Reference DesMeules, Pong, Lagace, Heng, Manuel and Pitblado 23 Sibley et alReference Sibley and Weiner 22 noted that rural residing individuals tend to delay seeking health care; this is of particular concern given the proportion of rural residents in Canada. The CSRTs aim to bridge this gap in service availability and to facilitate referrals to additional health care resources.
Although similar improvements in outcomes were demonstrated between the two groups, rural individuals had slightly higher baseline FIM scores, received significantly more total visits from therapists, and were enrolled in the program for a longer period than their urban counterparts. Although reasons for these findings are speculative, the differences may be due to differential referral practices between urban and rural clients. For example, the CSRT program may have acted as a replacement to inpatient rehabilitation for rural patients who do not live in close proximity to specialized stroke services. Although one may suspect that urban clients could have accessed outpatient based rehabilitation programs to supplement CSRT visits, this is not likely the case because the CSRT program is intended to act as a replacement for traditional outpatient services for those with limited access. Unfortunately, evolving program characteristics make the reason for these differences in baseline characteristics difficult to assess.
It is interesting to note that before correcting for multiple comparisons using the Bonferroni adjustment, urban residents improved more in terms of depressive symptoms than their rural counterparts (p=0.04) from admission to discharge from the program. Although this finding may be due to the higher probability of obtaining false positives with multiple comparisons, the initial significance of the relationship may be important to consider. For example, it has been shown that positive social role functioning (e.g. spending time with friends and family; engaging in social, community, and leisure activities) is associated with a reduction in depressive symptoms poststroke.Reference Schmid, Damush, Tu, Bakas, Kroenke and Hendrie 24 It is possible that many of the urban residents in this study had greater access to additional community programs and were less limited by geographic distance to family, friends, and community-based social activities, resulting in greater opportunities for social participation and inclusion. Future studies should explore the relationship between social participation, depression, and rurality.
Limitations
Although most of the outcome measures were sufficiently powered to detect a medium effect size, the RNLI outcome measure was not. This was the result of clinicians changing the outcome measure from the SIS midway through the data collection period, resulting in an inadequate sample size. Second, many clients did not meet inclusion criteria due to a number of factors including: (1) this was an administrative dataset and was not collected as rigorously as would be for research purposes; (2) many more clients were assessed for the program than admitted to it; and (3) some clients were lost to the final follow-up. An additional challenge with this analysis is that this is an evolving program. Changes in target lengths of stay, wait times for the program, and client referral practices have occurred over time and are difficult to account for. Finally, the results should be generalized with caution because the data are from a program covering a single geographical area that may differ from other rural areas where distances to health care services may be greater.
Conclusions
This research aimed to assess the differences in recovery between urban and rural residing stroke survivors, particularly in a Canadian context. Previous research on the CSRT program indicates that this intervention approach has significant positive effects on both physical and psychosocial outcomes of their clients.Reference Allen, Richardson, McIntyre, Janzen, Meyer and Ure 17 The findings of the current analysis further suggest that home-based stroke rehabilitation services benefitted equally both rural- and urban-residing individuals, with no significant differences in improvement between the two groups. This type of poststroke rehabilitation program should be considered in other rural areas where health services are not available or are limited.
Disclosures
LA, AM, SJ, MR, MM, DU, and RT have nothing to disclose.
Author Contributions
LA undertook data analysis and interpretation, methods, and an overall review. AM provided discussion and an overall review. SJ wrote the introduction and provided an overall review. MR provided data interpretation, analysis approach, and overall review. MM provided data interpretation and overall review. DU provided clinical interpretation and overall review. RT provided clinical interpretation and overall review.
