Occupational gaps 5 years after stroke

Abstract Objectives To investigate the incidence and number of occupational gaps 5 years after stroke and find possible predictors and explanatory factors for increased number of experienced gaps. Material and Methods The participants were diagnosed with first‐time stroke in Gothenburg during 2009–2010. Medical records from their hospital stay were used to obtain baseline data. The Occupational Gaps Questionnaire and the Swedish stroke registers follow‐up questionnaire were sent out. Data from the Occupational Gaps Questionnaire were used as a dependent variable and baseline data and questions from the stroke registry were used as independent variables in logistic regression. Results Five years poststroke, 49.5% experienced a higher number of occupational gaps compared to a healthy reference population. Predictors for an increased number of gaps were higher age at stroke onset and a higher degree of functional dependency. Explanatory factors for an increased number of gaps in the study population were higher age at follow‐up and feelings of depression. Conclusions Older age at the time of stroke and functional dependency can predict an increased number of occupational gaps. Older age and feelings of depression are connected to an increased number of occupational gaps. Individuals at risk should be provided with additional interventions to reduce participation restrictions.


| INTRODUC TI ON
Stroke is a common disorder that can result in a wide array of impairments, both physical and cognitive. These impairments can severely limit an individual's ability to carry out daily activities, resulting in reduced participation.
Participation is often seen as a central concept and a main goal in rehabilitation (Lecture, 2005) and is something that occurs at the intersection of what a person can do, wants to do, has the opportunity to do, and is not prevented from doing (Mallinson & Hammel, 2010).
Participation can be seen from both an objective and a subjective point of view. While the objective aspect focuses on observable behaviors and activities, the subjective targets the individual's own perception of participation (Dijkers, 2010) and is important since it is more connected to the individual's own goals and priorities rather than the actual disability (Witsø, Eide, & Vik, 2012) and thus, can affect the course and result of rehabilitation. This highlights the value of obtaining the individual's own view and thoughts on their situation rather than only observing and quantifying their abilities and limitations.
Approximately two thirds of people experience reduced participation 6 years after stroke compared to prestroke (Singam, Ytterberg, Tham, & von Koch, 2015;Spitzer, Tse, Baum, & Carey, 2011). These high numbers and the connection between participation and life satisfaction (Cooper et al., 2014;Eriksson, Kottorp, Borg, & Tham, 2009) makes reduced participation after stroke an important aspect of rehabilitation. A number of studies have investigated the level of participation after stroke (Desrosiers et al., 2006;Spitzer et al., 2011;van der Zee, Visser-Meily, Lindeman, Jaap Kappelle, & Post, 2013;Vincent-Onabajo & Blasu, 2016) but only a few in a longer time-period than 4 years poststroke (Bergström, Guidetti, Tham, & Eriksson, 2017;Singam et al., 2015). Several factors, such as age, physical dependency, and comorbidity have been identified as predictors for reduced participation at different time points following a stroke (Desrosiers et al., 2006;Singam et al., 2015;van der Zee et al., 2013). Though some factors seem to be reoccurring as predictors for reduced participation, these studies present varying results, leading to an inconclusive view of the situation. Thus, further research is warranted to identify and solidify which factors contribute to reduced participation following a stroke.
The Occupational Gaps Questionnaire assesses participation through restrictions in occupational activities (Eriksson, Tham, & Kottorp, 2013) and has been used in several studies to investigate occupational gaps after stroke and predictors for an increased number of occupational gaps after stroke (Bergström et al., 2017;Bergstrom et al., 2012;Eriksson, Tham, & Borg, 2006). These studies have shown a wide extent of occupational gaps after stroke with up to 90% of a study population experiencing occupational gaps (Bergstrom et al., 2012), making it an interesting and important aspect to further explore.
The aim of the study was to investigate the incidence and number of occupational gaps 5 years after stroke and find possible predictors and explanatory factors for increased number of experienced gaps.

| MATERIAL AND ME THODS
The present study is based on data collected in the SALGOT extended study. Participants were recruited for the SALGOT extended study at Sahlgrenska University Hospital/Sahlgrenska between February 4, 2009 and December 2, 2010. The SALGOT extended study comprises of three different populations treated for first time stroke at a stroke unit, a neurosurgical unit, and an intensive care unit (Persson, Parziali, Danielsson, & Sunnerhagen, 2012;Vikholmen, Persson, & Sunnerhagen, 2015;Wesali, Persson, Cederin, & Sunnerhagen, 2015).
Inclusion criteria for the SALGOT-extended study were the following. Being diagnosed with first time stroke (defined as ischemic stroke I63, intracerebral hemorrhage I61, and subarachnoid hemorrhage I60) and being 18 years or older as well as living within 35 km of Sahlgrenska University Hospital.
As seen in Figure 1 the SALGOT extended study included 725 individuals. A set of questionnaires was sent out 5 years later to the living participants (n = 457). Out of these participants, 281 completed the survey. Participants with incomplete answers in the occupational gaps questionnaire (<27 out of 30 questions answered) were also excluded, resulting in 194 individuals eligible for analysis.

| Questionnaires
The Occupational Gaps Questionnaire (Eriksson, 2012)  The Swedish stroke register's follow-up questionnaire consists of a number of questions regarding patients' health, capacity in different activities, and feeling of support following their stroke. For this study, the questions regarding feelings of depression, living alone, re-stroke, and feeling of support were used.

| Stroke measurements
Clinical data were gathered from patient records. The National Institutes of Health Stroke Scale (NIHSS) (Lyden et al., 1999) was used to assess stroke severity at arrival to hospital of patients with F I G U R E 1 Flow chart of the study population ischemic or hemorrhagic stroke. For patients with subarachnoid hemorrhage, the Hunt & Hess scale (H&H) (Hunt & Hess, 1968) was used at admission instead. The modified Rankin Scale (mRS) (van Swieten, Koudstaal, Visser, Schouten, & van Gijn, 1988) was used to assess functional dependency at discharge for all patients.
The NIHSS is a scale that is used to assess neurological function in patients with stroke, ranging from 0 to 46 points where a lower score equals fewer stroke symptoms. For this study the following cut-off values are used; very mild (0-2), mild (2-4), moderate (5-15), and severe (16-46) stroke.
The H&H scale is used on patients with SAH and the five stages of the scale represent the following states; I = asymptomatic, II = moderate to severe headache without neurological deficits, III = reduced consciousness with or without focal symptoms, IV = unconsciousness with or without focal symptoms, V = deep coma.
Stroke severity: To investigate stroke severity in participants with ischemic or hemorrhagic stroke as well as those with SAH, a combined variable for NIHSS and H&H was created based on clinical experience, and later dichotomized. Very mild to mild stroke according to NIHSS corresponds to grade 1 and 2 in H&H, and moderate to severe stroke according to NIHSS corresponds to grade 3-5 in H&H.
The mRS is a scale used to measure functional independence.
The scale ranges from 0, indicating total functional independence, to 5, indicating severe disability, with 6 indicating death. This scale has a proven validity and reliability in the context of stroke in a clinical setting. A score of ≥3 is considered poor outcome in this study.

| Statistical methods
The SPSS (version 25) was used to carry out statistical analyses.
Logistic regression was used to determine potential predictors and explanatory variables. The study population was divided into two groups according to the number of occupational gaps compared to an age matched reference population of 811 persons living in Sweden (Eriksson, 2012). One group with more than the median number of gaps and one group with the same or lower number of gaps as the median in respective age group. The median number of gaps in the reference population were as follows; age 20-29 = 5 gaps, age 30-49 = 4 gaps, age 50-64 = 2 gaps, age >65 = 1 gap (Eriksson, 2012). Having more gaps than the reference population was used as the dependent variable in logistic regression. Age, functional dependency (mRS), stroke severity, sex, feelings of depression, feeling of support, living alone, and re-stroke were used as independent variables.
Functional dependency (mRS), stroke severity, feelings of depression, and feeling of support were dichotomized based on clinical relevance and to ensure sufficient group sizes. The cut off used for dichotomization of functional dependency (mRS) was a score of 3, where scores <3 were considered a favorable outcome and scores ≥3 were considered a poor outcome. Feelings of depression were dichotomized by creating one group who reported that they never, almost never, or sometimes felt depressed, the other group reported having feelings of depression often or always. Dichotomization of the variable feeling of support was done by creating one group who reported no need for additional support and one group who reported insufficient feelings of support.
Cross tabulations were used to ensure large enough groups (n ≥ 5) for each variable. Correlations between variables were TA B L E 1 Baseline characteristics of the study population Age Mean (SD) 63.0 (14) Median (

| Ethics
The SALGOT study was approved by the Regional Ethics Committee

| RE SULTS
The characteristics of the study population at the time of stroke are shown in  The results from the predictive model showed that a higher age at stroke onset (OR = 1.045 95% C.I. 1.018-1.072) and a higher level of functional dependency at discharge (OR = 3.224 95% C.I. 1.558-6.542) were predictive for having more occupational gaps than the reference population 5 years after stroke (   (Eriksson, Aasnes, Tistad, Guidetti, & von Koch, 2012). These people need to be identified early after stroke in order to prevent reduced participation and isolation .

| Occupational gaps in population
The present study showed that 71.1% of the participants experience occupational gaps 5 years after stroke and half of the study population experienced more gaps than the reference population. Previous studies have shown that approximately 90% experience occupational gaps 1 year (Bergstrom et al., 2012;Eriksson et al., 2012) or 5 years (Bergström et al., 2017) after stroke. An explanation for more experienced gaps could be the older study participants in two of the studies (Bergstrom et al., 2012;Eriksson et al., 2012) since an identified predictor for worse outcome after stroke is older age (Macciocchi, Diamond, Alves, & Mertz, 1998;Nakayama, Jorgensen, Raaschou, & Olsen, 1994) found in the present study as a predictor for an increased number of gaps 5 years after stroke. The activities hobbies, participating in sports, and outdoor life had the highest number of gaps in the present study. Previous studies have shown that traveling for pleasure (Bergstrom et al., 2012;Eriksson et al., 2012) and participating in sports (Bergström et al., 2017) were the activities with most number of gaps. Traveling for pleasure had only the fourth highest number of gaps in the present study with a longer follow-up time and with a younger study population, and only the seventh highest number of gaps in another 5 year follow-up study conducted on a younger population (Bergström et al., 2017).

| Predictors and explanatory factors
Age at the onset of stroke and level of functional dependency at discharge were identified as predictors for an increased number of occupational gaps 5 years after stroke. These variables could explain 16.3%-21.8% of the variance in the outcome, indicating existence of other important predictors. Few studies have investigated predictors for occupational gaps, but one study (Bergstrom et al., 2012) showed that stroke severity, activities of daily living abilities, social participation, and not being born in Sweden as predictors for occupational gaps, whereas age at stroke onset was not found to be a significant predictor. Furthermore, age and physical capability have previously been identified as predictors for a reduced subjective participation 2-4 (Desrosiers et al., 2006) and 6 (Singam et al., 2015) years after stroke, indicating generalizability of the identified predictors in the F I G U R E 3 Illustration of the model building in the two logistic regression models (mRS: modified Rankin Scale) present study since the previous study populations were older and had different sex distribution compared to the present study.
The explanatory model showed that age at time of follow-up and feelings of depression contributed to an increased number of occupational gaps 5 years after stroke. The present study found that 13.9% of included individuals had feelings of depression often or always, while 37.6% sometimes had feelings of depression. 17% reported taking medication for their feelings of depression, which is in line with the reported prevalence of depression which ranges from 19% to 33% (Hackett, Yapa, Parag, & Anderson, 2005;Robinson, 2003).
No studies were found about association between depression and occupational gaps, but depression has been identified as a negative factor for participation after stroke (D'Alisa, Baudo, Mauro, & Miscio, 2005;Micaela Silva et al., 2016), which strengthens the findings of association between depression and occupational gaps after stroke.

| Methodical considerations
There is a risk with questionnaires that the respondent misunderstand the questions and therefore answer incorrectly. Furthermore, a next of kin can have help the respondent to fill out the questionnaire, resulting in a risk of lower accuracy. Since the questions in the OGQ are answered with yes or no, a partial restriction in an activity will not be apparent in the questionnaire.
Ordinal and categorical variables were dichotomized primarily from a clinically relevant perspective and secondarily to attain satisfactory group sizes for logistic regression. The process of dichotomization comes with a loss of information (Fedorov, Mannino, & Zhang, 2009) and consequently a reduction in statistical power (Fedorov et al., 2009). Cut-off levels for NIHSS used in the present study were chosen because of a high prevalence of milder stroke. These levels have been used previously (Muhr, Persson, & Sunnerhagen, 2017) but are not validated. The merging of NIHSS and H&H into a dichotomized variable for stroke severity is not validated and based solely on clinical experience; however, no conclusions were based on the results from this variable.

| Study limitations
The participants were all from Sahlgrenska University Hospital, which is the sole hospital in the region that carries out thrombolysis, thrombec- TA B L E 3 Predictive and explanatory regression models for having more occupational gaps than the reference population more women and more severe cases of stroke which might reduce the generalizability of the present study. Furthermore, the follow-up in this study comprised five different questionnaires, resulting in a time consuming survey that could dissuade persons with lack of endurance or other limitations from participating, which might also reduce the generalizability of the present study.

| Clinical implications
Results from the present study provide knowledge about participation and occupational gaps after stroke, as well as information regarding predictors and factors associated with occupational gaps.
The individuals' feelings of participation are important to obtain a full view of their well-being and to highlight needs for interventions.
The knowledge of predictors is important to be able to identify individuals at risk for reduced participation. Older stroke survivors and stroke survivors with an increased functional dependency could benefit from further evaluation and rehabilitation interventions to enable a more favorable outcome regarding participation in everyday occupations. Health care professionals considering depression in stroke survivors could benefit these individuals and lead to a higher degree of participation.
Further research is needed to evaluate the extent of occupational gaps in different populations and settings as well as to strengthen the knowledge on predictors for occupational gaps.

ACK N OWLED G M ENTS
The study was supported by grants from the Swedish Research Council (VR 2012-70X-22122-01-3, VR2017-00946), the Swedish Heart-Lung Foundation, the Swedish Brain Foundation, Promobilia, and the Swedish Stroke Association.

CO N FLI C T S O F I NTE R E S T S
The authors state that there are no conflicts of interests.