Adherence to a Long-Term Physical Activity and Exercise Program After Stroke Applied in a Randomized Controlled Trial

Abstract

Background

Persistent physical activity is important to maintain motor function across all stages after stroke.

Objective

The objective of this study was to investigate adherence to an 18-month physical activity and exercise program.

Design

The design was a prospective, longitudinal study including participants who had had a stroke randomly allocated to the intervention arm of a randomized controlled trial.

Methods

The intervention consisted of individualized monthly coaching by a physical therapist who motivated participants to adhere to 30 minutes of daily physical activity and 45 minutes of weekly exercise over an 18-month period. The primary outcome was the combination of participants' self-reported training diaries and adherence, as reported by the physical therapists. Mixed-effect models were used to analyze change in adherence over time. Intensity levels, measured by the Borg scale, were a secondary outcome.

Results

In total, 186 informed, consenting participants who had had mild-to-moderate stroke were included 3 months after stroke onset. Mean age was 71.7 years (SD = 11.9). Thirty-four (18.3%) participants withdrew and 9 (4.8%) died during follow-up. Adherence to physical activity and exercise each month ranged from 51.2% to 73.1%, and from 63.5% to 79.7%, respectively. Adherence to physical activity increased by 2.6% per month (odds ratio = 1.026, 95% CI = 1.014–1.037). Most of the exercise was performed at moderate-to-high intensity levels, ranging from scores of 12 to 16 on the Borg scale, with an increase of 0.018 points each month (95% CI = 0.011–0.024).

Limitations

Limitations included missing information about adherence for participants with missing data and reasons for dropout.

Conclusions

Participants with mild and moderate impairments after stroke who received individualized regular coaching established and maintained moderate-to-good adherence to daily physical activity and weekly exercise over time.

Individuals surviving stroke risk long-term impairments, limitations on activities, and reduced participation.¹ Consequently, the rehabilitation process plays a key role in achieving functional recovery and community reintegration.² Long-term engagement in regular physical activity and exercise is highly recommended to sustain motor functions gained in rehabilitation and to reduce the risk of recurrent stroke.^3,4

Implementing cardiorespiratory training within poststroke rehabilitation programs improves measures of walking performance and reduces dependency.⁵ Guidelines recommend that individuals who have had a stroke and are capable of engaging in physical activity should perform continuous or accumulated exercise of moderate-to-high intensity, defined as vigorous activity for 10 to 60 minutes, once to 5 times per week, sufficient to break a sweat or noticeably raise heart rate.^3,6,7 Nevertheless, several studies demonstrate that many individuals perform little if any regular physical activity or exercise after stroke.^8–11

More knowledge of barriers to regular poststroke physical activity and exercise might improve rehabilitation programs.¹² However, there is little published research that either systematically registers long-term adherence to interventions or that reports on factors predicting adherence and dropout from activity after stroke.¹³

Results from the Life After Stroke (LAST) study were recently published.¹⁴ Despite neutral results concerning the study's primary and secondary outcomes, LAST showed that participants receiving regular individualized coaching on physical activity and exercise were significantly more active than participants receiving standard care.¹⁴ To inform clinicians and increase knowledge about the feasibility of the intervention, a more thorough investigation is required regarding how well the participants adhered to the treatment protocol applied in the LAST study and which factors were associated with good adherence.

The primary aim of the present substudy was to assess the adherence–of participants who had had a stroke and were randomized to the intervention arm of the LAST study--to the prescribed amount of physical activity and exercise required per protocol during 18 consecutive months. The secondary aims were to investigate (1) whether participants’ age, sex, degree of dependency, cognitive function, or goal attainment influenced their adherence; (2) participants’ adherence to the intensity levels of physical activity and exercise required per protocol; and (3) to what extent participants achieved their individual goals related to physical activity and exercise.

Methods

Study Design, Setting, and Participants

This was a prospective, longitudinal study including participants who had had a stroke randomized to the intervention arm of the LAST study.¹⁴ Inclusion criteria were: diagnosis of first-ever or recurrent stroke (infarction or intracerebral hemorrhage); age ≥18 years; discharged from hospital or inpatient rehabilitation at inclusion; community dwelling with a modified Rankin Scale score <5; and cognitive function, as evaluated by the Mini-Mental State Examination, >20 points (>16 points for individuals with aphasia). Exclusion criteria were serious medical comorbidity with short life expectancy, or a condition contraindicating motor training. Following good clinical practice and the current Norwegian guidelines,¹⁵ participants underwent a complete medical history and a physical examination by a medical practitioner during the screening procedures aiming to identify neurological complications or medical comorbidities that could possibly require special considerations or constitute contraindications to high-intensity exercise. Explicitly, participants with uncompensated heart failure and/or unstable coronary function were excluded from the study. Eligible participants were recruited from October 18, 2011 to June 26, 2014 at the outpatient clinics of the stroke units of 2 Norwegian hospitals and consecutively randomized 10 to 16 weeks after the acute stroke. Follow-up assessments were completed by January 15, 2016.

Intervention

In addition to standard care after stroke,¹⁵ participants randomized to the intervention group received a follow-up program delivered by the primary health care services in 3 Norwegian municipalities.¹⁶ The program comprised regular individualized coaching, inspired by motivational interviewing techniques,¹⁷ in physical activity and exercise by a physical therapist for 18 consecutive months after inclusion. Before the intervention, the physical therapists attended a 1-day course in motivational interviewing technique and were recommended to keep updated within this field. Participants were encouraged to perform 30 minutes of physical activity 7 d/wk, and 45 to 60 minutes of exercise once per week. In accordance with the World Health Organization, physical activity was defined as “any bodily movement produced by skeletal muscles that results in energy expenditure.”¹⁸ Exercise was defined as a subcategory of physical activity that was planned, structured, repetitive, and purposeful with the objective of improving or maintaining 1 or more components of physical fitness.¹⁸ Per protocol, exercise should include 2 to 3 bouts of vigorous activity. Hence, participants were encouraged to reach levels of high intensity during exercise corresponding to a score of 15 to 17 on the Borg scale (6–20).¹⁹

The physical therapists’ main purpose was to motivate and encourage the participants to follow an individually adapted training program. Schedules with at least 2 choices for physical activity (eg, housework, walking, gardening) and 2 choices for exercise (eg, hiking, swimming, bicycling) were set for each month based on the participants’ individual preferences and goals. If preferred, individuals were offered participation in existing outpatient, private, and community-based treatment groups, individual physical therapy, or home training. Participants were instructed how to report in training diaries the amount and intensity of each session of physical activity and exercise. During each monthly meeting, the physical therapist conducted a conversation using elements from motivational interviewing and, in collaboration with the participant, reviewed and reassessed the content and progression of the planned training schedule. To enhance adherence, goal setting and regular evaluation of goals were emphasized during follow-up.

Regular meetings between the participant and the physical therapist were arranged once every month. The first 6 meetings were planned face-to-face, preferably at the participant's home. During the next 6 months, every second meeting could be a phone meeting, if preferred; and during the final 6 months, 4 of the 6 meetings could be phone meetings.¹⁶

Baseline Assessment

Before intervention, age, sex, living condition, type of stroke, and medical history were recorded. Stroke severity was assessed by the National Institutes of Health Stroke Scale,²⁰ dependence by the modified Rankin Scale,²¹ and cognitive function by the Mini-Mental State Examination.²²

Outcome Measures

The primary outcome was adherence to the recommended time spent on daily physical activity and weekly exercise, assessed by the combination of 2 outcome measures. The first, and main, outcome measure was participants’ self-reports in standardized training diaries, in which participants were encouraged to report their actual amounts and intensities of physical activity and exercise immediately after each training session. The second was an overall estimation of participants’ adherence assessed by the physical therapists in separate adherence forms. At the regular appointments, the physical therapists reported whether the participants had performed the training program in line with the agreement or not. Reasons for nonadherence, in addition to the setting of the monthly meetings, were also recorded.

Secondary outcome measures in this study were the Borg scale¹⁹ and goal attainment scaling (GAS).²³ The Borg scale was applied to measure participants’ intensity levels during physical activity and exercise. This scale is based on perceived exertion during activity, ranging from 6 (no exertion at all) to 20 (maximal exertion), and appears to be a reasonable indicator of exercise intensity after stroke, at least at moderate intensities.²⁴ GAS was assessed to score the extent to which participants’ individual goals were achieved during the intervention. The validity, reliability, and responsiveness of GAS as an outcome measure for rehabilitation has been supported.²⁵ GAS identifies and quantifies individual goals of treatment, enabling comparison between individuals. In the present study, goal attainment was evaluated at 3, 6, 9, 12, 15, and 18 months after inclusion.

Statistical Analyses

The baseline demographic and clinical characteristics are presented as numbers (percentages) of participants or means (standard deviations [SDs]). Participants who died or were defined as “dropouts” were consecutively excluded. Hence, observations until withdrawal or death were included in the analyses.

Adherence was defined and presented in 2 different ways. Firstly, participants who performed at least 210 minutes of physical activity (ie, 30 minutes, 7 d/wk) and 45 minutes of exercise every week for at least 3 out of 4 weeks within each month, respectively, were defined as adherent to the treatment protocol. Secondly, the weekly amounts of physical activity and exercise, respectively, were accumulated as total sums undertaken during 4 weeks, representing each month. Furthermore, results were categorized into prespecified subgroups and displayed for 18 consecutive months in area plots. These figures were used to illustrate categories of adherence to the amounts and intensities of physical activity and exercise, respectively, in addition to goal achievements over time. Adherence to the amount of physical activity was categorized as follows: (1) ≥30 minutes, 7 d/wk, as recommended per protocol; (2) 30 minutes, 5 to 6 d/wk, within recommendations for adults and older people by the Norwegian Directorate of Health;²⁶ (3) <30 minutes, 5 d/wk, below recommendations; and (4) 0 minutes. Correspondingly, adherence to exercise was categorized as: (1) ≥45 minutes weekly, as recommended per protocol; (2) 20 to 44 minutes weekly, within the Norwegian recommendations of exercise in individuals with stroke;⁷ (3) 1 to 19 minutes weekly, below recommendations; and (4) 0 minutes. Intensity levels were also categorized in accordance with previous research and the clinical application of the Borg scale,^7,27 namely levels of light (6–10 points), moderate (11–14 points), and high (15–20 points) intensity.

The overall GAS score was calculated according to the algorithm presented by Turner-Stokes,²⁸ in which a score of 50 represents the expected level of performance. The proportion of participants who attended at least 2 out of 3 meetings face-to-face within every 3-month period is also reported.

First, we used mixed-effects logistic models with adherence measures (one at a time) as dependent variables, participant as a random effect, and time as the covariate to study change over time. Second, we added the following covariates separately to study their effect: (1) sex; (2) age (in years); (3) modified Rankin Scale;²¹ (4) Mini-Mental State Examination;²² and (5) GAS.²³ Thirdly, change over time in levels of intensity during physical activity and exercise was assessed by linear mixed-effect models with participant as a random effect and time as the covariate.

Two-sided P values ≤.05 were considered statistically significant. Statistical analyses were run in IBM SPSS Statistics (version 23.0; IBM, Armonk, NY, USA), Stata (version 13.1; StatCorp, College Station, TX, USA), and Microsoft Excel 2010 for Windows (Microsoft, Redmond, WA, USA).

Role of the Funding Source

The study was funded by the Liaison Committee for Education, Research and Innovation in Central Norway, the Norwegian Fund for Postgraduate Training in Physiotherapy, and the Research Council of Norway. The funding sources had no influence on the study design, data collection, analysis, interpretation, or manuscript preparation.

Results

A total of 186 participants were randomized to the intervention group of LAST and thus included in the present study. In total, 34 (18.3%) withdrew during follow-up. Among those who withdrew, 4 participants died after withdrawal, but before end of follow-up. In addition, 5 participants from the main sample died, making a total of 9 deaths (4.8%) during the study. Twenty-three (12.4%) participants were excluded from the mixed-effect models because they did not contribute any valid data (Fig. 1).

Figure 1.

Flow chart. LAST = life After stroke; Mors. = deaths.

The total sample, with a mean age of 71.7 years (standard deviation [SD] = 11.9), involved participants affected by mild-to-moderate stroke (97.3% had a score <8 points on the National Institutes of Health Stroke Scale) (Tab. 1). Participants who withdrew were older (mean 75.6 years vs 70.9 years, P = .03), and more men withdrew than women (24/104 = 23% vs 10/82 = 12%, P = .06). The remaining characteristics were similar between the groups.

Table 1.

Baseline Demographic and Clinical Characteristics of 186 Participants in the Intervention Group^a

Characteristic	Valueb
Age, mean (SD)	71.7 (11.9)
<80 y	142 (76.3)
≥80 y	44 (23.7)
Sex
Female	82 (44.1)
Male	104 (55.9)
Domestic circumstances
Living with someone	130 (69.9)
Living alone	56 (30.1)
MMSE score, mean (SD)	27.8 (2.3)
≥25	164 (88.2)
<25	22 (11.8)
Days after stroke, mean (SD)	111.3 (24.5)
Stroke type
Infarction	172 (92.5)
Hemorrhage	14 (7.5)
NIHSS score, mean (SD)	1.5 (2.3)
Mild stroke (<8)	181 (97.3)
Moderate stroke (8–16)	5 (2.7)
Severe stroke (>16)	0
mRS score, mean (SD)	1.45 (1.08)
0	34 (18.3)
1	78 (41.9)
2	36 (19.3)
3	32 (17.3)
4	6 (3.2)
Comorbidity
Stroke	29 (15.6)
TIA	20 (10.8)
Myocardial infarction	19 (10.2)
Heart failure	3 (1.6)
Atrial fibrillation	32 (17.2)
Hypertension	90 (48.4)
Diabetes	25 (13.4)
Lung diseases	19 (10.2)

^aMMSE = Mini-Mental State Examination; mRS = modified Rankin Scale; NIHSS = National Institutes of Health Stroke Scale; SD = standard deviation; TIA = transient ischemic attack.

^bData are reported as numbers (percentages) of participants unless otherwise indicated.

Adherence to Amount of Physical Activity and Exercise

Results showed that the actual number of participants adherent to ≥210 minutes of weekly physical activity decreased from 87 to 79 during follow-up, whereas the proportion they represented of those with available data increased from 55.4% to 73.1% (Tab. 2). Correspondingly, participants who adhered to ≥45 minutes of weekly exercise decreased from 103 to 71, equivalent to a decrease from 66.0% to 65.1% of those with available data (Tab. 2).

Table 2.

Adherence to Physical Activity and Exercise During the 18-month Follow-Up Program^a

	Data for Month:
Parameter	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
Physical activity (PA)b
Total no. of participants	157	148	145	145	140	137	132	131	130	125	126	119	114	115	115	116	113	108
No. (%) of participants adherent to ≥210 min of PA/wkc	87 (55.4)	84 (56.8)	89 (61.4)	95 (65.5)	84 (60.0)	79 (57.7)	86 (65.2)	88 (67.2)	82 (63.1)	64 (51.2)	77 (61.1)	69 (58.0)	71 (62.3)	76 (66.1)	73 (63.5)	75 (64.7)	78 (69.0)	79 (73.1)
Borg scale score (6–20) during PA, mean (SD)	12.2 (1.5)	12.5 (1.5)	12.4 (1.6)	12.5 (1.7)	12.5 (1.8)	12.6 (1.8)	12.6 (1.5)	12.5 (1.5)	12.6 (1.8)	12.4 (1.7)	12.4 (1.6)	12.4 (1.8)	12.4 (1.8)	12.7 (1.8)	12.6 (1.8)	12.9 (1.8)	12.9 (1.8)	12.9 (1.7)
Exerciseb
Total no. of participants	156	148	145	143	140	138	133	131	131	129	128	120	116	118	118	117	113	109
No. (%) of participants adherent to ≥45 min of exercise/wkd	103 (66.0)	94 (63.5)	101 (69.7)	99 (69.2)	101 (72.1)	93 (67.4)	95 (71.4)	92 (70.2)	94 (71.8)	94 (72.9)	93 (72.7)	84 (70.0)	82 (70.7)	94 (79.7)	85 (72.0)	83 (70.9)	76 (67.3)	71 (65.1)
Borg scale score (6–20) during exercise, mean (SD)	13.7 (1.7)	13.9 (1.8)	14.2 (2.1)	14.2 (1.9)	14.2 (1.8)	14.3 (1.8)	14.4 (1.8)	14.4 (1.8)	14.4 (1.8)	14.4 (1.7)	14.3 (1.8)	14.3 (1.7)	14.3 (2.1)	14.2 (2.1)	14.4 (2.0)	14.3 (2.1)	14.1 (2.1)	14.5 (1.9)
Physical activity and exerciseb
Total no. of participants	155	147	144	143	139	136	131	130	129	125	126	118	114	115	115	116	112	107
No. (%) of participants adherent to both ≥210 min of PA/wk and ≥45 min of exercise/wke	58 (37.4)	55 (37.4)	61 (42.4)	64 (44.8)	58 (41.7)	52 (38.2)	58 (44.3)	64 (49.2)	58 (45.0)	46 (36.8)	51 (40.5)	46 (39.0)	49 (43.0)	60 (52.2)	53 (46.1)	54 (46.6)	53 (47.3)	52 (48.6)
Individualized coaching, no. of participants			154			152			147			142			134			125
No. (%) of participants attending ≥2/3 meetings face to face			123 (79.9)			105 (69.1)			59 (40.1)			40 (28.2)			35 (26.1)			32 (25.6)
GASf			138			126			122			109			112			92
GAS score, mean (SD)			44.2 (10.8)			43.3 (9.9)			41.4 (9.6)			42.6 (10.0)			42.9 (10.2)			42.9 (9.1)

^aData show numbers (percentages) of participants who adhered to the program and their reported levels of intensity (ie, scores on the Borg Scale) in 18 consecutive months, each month consisting of 4 wks. GAS = goal attainment scaling; PA = physical activity; SD = standard deviation.

^bParticipants with at least 3/4 wk with no missing data each month.

^cParticipants who performed ≥30 min of PA 7 d/wk (ie, ≥210 min/wk) at least 3/4 wk/mo.

^dParticipants who performed ≥45 min of weekly exercise at least 3/4 wk/mo.

^eParticipants who performed ≥30 min of PA 7 d/wk (ie, ≥210 min/wk) and ≥45 min of weekly exercise at least 3/4 wk/mo.

^fNumber of participants with goals set and evaluated.

Further, Figure 2A and B illustrate the increasing deaths, withdrawals, and missing data over time. On average, 6.8% and 2.2% reported zero minutes of exercise and physical activity, respectively. The monthly proportion of participants reporting amounts of exercise required per protocol ranged from 42.5% to 64.0%, whereas 41.9% to 57.0% of participants reported physical activity required per protocol. Moreover, the proportion of participants reporting monthly amounts of physical activity corresponding to ≥30 minutes, 5 d/wk, ranged from 48.9% to 66.7%.

Figure 2.

Area plots illustrating adherence to the amounts (A, B) and intensities (C, D) of physical activity and exercise, in addition to goal achievements (E), during 18-month follow-up. Adherence to amounts of physical activity and exercise (A, B) were categorized as follows: blue shading indicates amounts required per protocol; red shading indicates amounts within the levels of activity for adults and older people recommended by the Norwegian Directorate of Health²⁶ (A) or amounts within the Norwegian recommended levels of exercise in participants with stroke⁷ (B); green shading indicates amounts below recommendations; and purple shading indicates no activity or exercise reported. Adherence to intensity levels during physical activity and exercise (C, D) were categorized into light, moderate, and high. Goal achievements (E) were dichotomized into achieved (≥50 points) and not achieved (<50 point) goals. Mors = deaths.

Data from 163 participants (87.6%) were included in the mixed-effect models. Results showed a monthly increase in adherence to physical activity (odds ratio [OR] = 1.026, 95% CI = 1.014–1.037, P < .001), but not for exercise (OR = 1.003, 95% CI = 0.992–1.012, P = .62). Adherence to the combination of physical activity and exercise increased significantly during follow-up (OR = 1.018, 95% CI = 1.008–1.028, P < .001) (Tab. 3). Adjusted for time, the modified Rankin Scale was significantly associated with adherence to physical activity (OR = 0.60, 95% CI = 0.44–0.82, P < .001) as well as adherence to the combination of physical activity and exercise (OR = 0.78, 95% CI = 0.60–0.99, P = .05). Further, male sex (OR = 1.44, 95% CI = 0.92–2.26, P = 0.11) and age (OR = 0.977, 95% CI = 0.957–0.998, P = .03) were associated with adherence to exercise, although only the latter was statistically significant. There was no statistically significant interaction effect between the 2 covariates.

Table 3.

Mixed-Effects Logistic Regression With Adherence as the Dependent Variable and Participant as a Random Effect^a

	Adherence to Physical Activity			Adherence to Exercise			Adherence to Physical Activity and Exercise
Covariate	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Monthb	1.026	1.014–1.037	<.001	1.003	0.992–1.012	.624	1.018	1.008–1.028	<.001
Men vs womenc	1.044 vs 1.0	0.532–2.049	.900	1.444 vs 1.0	0.923–2.260	.108	1.037 vs 1.0	0.603–1.785	.895
Age (y)c	0.990	0.960–1.021	.507	0.977	0.957–0.998	.030	0.984	0.959–1.008	.190
mRS scorec	0.602	0.443–0.818	<.001	0.972	0.787–1.200	.791	0.777	0.604–0.999	.049
MMSE scorec	0.992	0.862–1.143	.916	0.920	0.837–1.010	.081	0.926	0.827–1.037	.181
GAS scorec	1.021	0.996–1.047	.098	1.010	0.990–1.032	.327	1.015	0.994–1.036	.152

^aGAS = goal attainment scaling; MMSE = Mini-Mental State Examination; mRS = modified Rankin Scale; OR = odds ratio.

^bUnadjusted.

^cAdjusted for time.

Intensity Levels of Physical Activity and Exercise

Among the participants who were defined as adherent to the amounts of physical activity and exercise, the mean (SD) ratings on the Borg scale ranged from 12.2 (1.5) to 12.9 (1.8) for physical activity and from 13.7 (1.7) to 14.5 (1.9) for exercise (Tab. 2).

Figure 2C and D illustrate that, on average, approximately half the amount of physical activity and one-third of the exercise were performed with moderate intensity (ie, 11–14 on the Borg scale). On average, almost one-fourth of the exercise was performed with levels of high intensity (ie, Borg scale ≥15) as recommended per protocol. Results from the linear mixed models (Tab. 4) showed highly significant increases in reported Borg scale for both physical activity and exercise over time (P < .001).

Table 4.

Mixed-Effects Linear Regression With the Borg Scale Score as the Dependent Variable, Participant as a Random Effect, and Month as the Covariate

Evaluation of Borg Scale Score	Nonstandardized Regression Coefficient	95% CI	P
During physical activity	0.020	0.015–0.025	<.001
During exercise	0.018	0.011–0.024	<.001

Goal Achievements

The mean GAS scores were below the expected level (ie, score of <50) at every evaluation point, ranging from 41.4 (SD = 9.6) to 44.2 (SD = 10.8) (Tab. 2). Overall, goal achievements over time were low (Fig. 2E), with the proportion of participants achieving their goals ranging from approximately one-fifth to one-third.

Setting of Monthly Meetings

Table 2 shows that among participants with valid data, 79.9% attended ≥2 out of 3 of the individual face-to-face coaching sessions within the first 3-month period, and 69.1% within the next 3-month period. Face-to-face meetings declined to 25.6% within the final 3-month period.

Discussion

The main results showed that participants who received individual coaching over 18 months managed to establish and maintain moderate-to-good adherence to daily physical activity and weekly exercise over time poststroke. There were slightly increasing amounts of physical activity among those who completed the long-term follow-up. However, discrepancies between participants were large. Individuals with a higher degree of dependency were less adherent to physical activity, and both increasing age and female sex were associated with lower adherence to exercise. Both physical activity and exercise primarily corresponded to scores on the Borg scale equivalent to moderate intensities. Only one-fourth of the reported amount of exercise was performed at levels of high intensity as required per protocol. Goals related to the individuals’ training programs were poorly achieved over time.

This study sample was limited to voluntary participants, possibly leading to selection bias. This might reflect greater adherence to physical activity and exercise than could be expected among the eligible participants who declined to participate. In addition, the sample consisted of participants only mildly-to-moderately affected by stroke. Hence, the results should be interpreted with caution when considering individuals more severely affected in both motor function and cognition. To ensure safety, the intervention was conducted in line with good clinical practice and according to the current Norwegian national guidelines of treatment and rehabilitation after stroke.¹⁵ However, it should be noted that other organizations recommend that people who have had a stroke undergo graded exercise testing with ECG monitoring before beginning an exercise program.²⁹ Nevertheless, previous analysis showed 39% fewer hospital admissions due to vascular events in the intervention group compared with the control group (17 vs 28 events, P = .110).¹⁴ Further, there is no established consensus on how to measure and rate adherence of home-based rehabilitation interventions.³⁰ The use of training diaries might be considered a weakness, as self-reported measurements are limited by recall bias, social desirability bias, and vulnerability to inaccuracies.^31,32 Because stroke survivors often suffer cognitive deficits, this might limit their ability to report activity levels accurately. The use of self-reports could possibly under- or overestimate activity levels compared with objective measures.^30,33 However, inflated self-reports of physical activity seem to be the largest challenge in people who have had a stroke.³⁴ This might have resulted in overestimation of adherence in the current study, although daily reviewing of training diaries probably helped to ensure adherence in accordance with the protocol.¹⁶ Because training diaries are among the simplest and least expensive methods to register adherence in the long term, they were considered valid.^32,35 Further, the Borg scale appears to be a reliable and valid indicator of exercise intensity after stroke at levels of moderate, but perhaps not high, intensity.²⁴ GAS enables important changes in function to be identified, which many current scales might not address.²³ However, GAS depends on the experience and ability of the clinicians to predict outcome.²⁸ Work is needed to assess construct validity and generalizability of findings if GAS is to become a routine outcome measurement in stroke rehabilitation. Lastly, it should be noted that the participant-therapist relationship might also have influenced the individuals’ degree of adherence and could be of interest for future research.³⁶

Results of the present study showed that between 51.2% and 79.7% of participants reached the monthly amounts of physical activity and exercise, respectively, as recommended per protocol (Tab. 2). Compared with previous studies, these findings suggest enhanced adherence rates. A large cross-sectional survey assessing levels of self-reported exercise among individuals in the chronic phase after stroke found 31% reporting regular exercise and 27% reporting infrequent or no exercise.³⁷ Furthermore, a Norwegian survey assessing self-reported activity levels in the general population suggested that, on average, 32% of people aged ≥65 years achieve 150 minutes of weekly physical activity.³⁸ Moreover, in studies measuring physical activity using accelerometers or other electronic devices among individuals with chronic stroke, ambulatory activity levels were very low.^8–10 A significantly smaller proportion of individuals seemed to meet the recommendations of 30 minutes of daily physical activity compared with the findings of the present study. The lack of established methods for assessing adherence to activity means that various methods have been used, which limits the possibilities of comparing adherence rates between studies.^35,39 Nevertheless, considering the long-term perspective and the strict requirements for adherence in the present study, the adherence rates achieved are evaluated as moderate to good.

The stable and slightly increasing amounts of daily physical activity and weekly exercise observed across the 18-month follow-up are also in contrast to earlier research indicating that most individuals poststroke can plateau, reduce, or cease their activity levels over time.^11,40 Several longitudinal studies, which demonstrated that intensity, frequency, and duration of physical activity increased within the first 3 months poststroke, but then plateaued, suggested that behavior patterns established within the first 3 months poststroke might predict long-term physical activity habits.^9,41,42 In contrast, the findings of the present study suggest the possibility of increasing amounts of physical activity over time. This indicates that several elements of the follow-up program, such as self-monitoring one's behavior by the use of training diaries and participant-centered coaching techniques, might have promoted adherence to physical activity and exercise. The good adherence might also explain the possible benefit of the regular coaching reported in the LAST study, assessed by the International Physical Activity Questionnaire⁴³ during follow-up.¹⁴

There were large discrepancies in degrees of adherence between participants. Regarding physical activity, dependency levels were the only covariate contributing significantly to explain these differences. Hence, future work should investigate whether the intervention could adopt a more differentiated approach, in that those with a higher degree of dependency might benefit from closer follow-up compared with their more independent counterparts. Noticeably, the individual face-to-face coaching was performed to a lesser extent than required per protocol. Therefore, increasing the number of face-to-face meetings might also increase adherence for these participants. Nevertheless, there is a subtle balance between independent activity behavior and dependency on coaching. Further, increasing age was negatively associated with adherence to exercise; additionally, although not statistically significant, men were more likely than women to adhere to exercise. Additional research is needed to determine the causes of these associations. Despite the lack of an interaction effect between age and sex, future interventions might need to specifically target exercise after stroke with regards to these aspects.

Adherence to high intensity levels during exercise proved challenging in the present study. Despite a statistically significant increase equivalent to an average of 0.32 points on the Borg scale over 18 months, this is not necessarily of clinical importance. Common impairments, such as hemiparetic gait, reduced balance, poststroke fatigue, and depression might be barriers to conducting high-intensity training for individuals after stroke.⁶ Also, some participants might have been afraid to push themselves beyond comfortable limits, fearing acute illness or discomfort.⁴⁴ Facing the challenge of achieving high intensities in clinical practice, and particularly as part of a home training program,^45,46 teaching people who have had a stroke how high intensity actually is experienced might be included in the intervention. Further, developing interventions to enhance adherence to high intensities in the long term after stroke should be prioritized in future research.

The low degrees of goal achievement throughout the follow-up might be explained by several factors. Turner-Stokes emphasizes that goal achievement depends on both the participant's ability to achieve his or her goals and the clinician's ability to predict outcome, the latter requiring knowledge and experience.²⁸ The present results suggest that the physical therapists, in collaboration with the participants, had a tendency to incorporate overambitious goals. In addition, several participants chose somewhat vague goals that remained unchanged over time, struggling to define new and more specific goals. It is reasonable to assume that these long-term goals (eg, “be able to catch the bus independently”) made it difficult to see the direct link to the actual physical activity and exercise program (eg, balance and walking training), which might possibly have decreased adherence. Hence, greater emphasis should be given to focusing more on intermediate goals. Turner-Stokes underlines how the involvement of individuals with acquired brain injury presents particular challenges for GAS, as cognitive impairments, including executive dysfunction, and communication problems can limit their ability to remember and articulate goals.⁴⁷ These were common challenges for some of the participants in the present study, affecting their ability and motivation to achieve goals.

Overall, the long-term follow-up program appears feasible for individuals poststroke with mild-to-moderate impairments. Regular coaching by a physical therapist might contribute to enhanced adherence to activity levels over time, and the findings help in understanding which factors influence adherence to persistent physical activity and exercise over time after stroke. The intervention might also contribute to a continuation of well-established habits, perhaps beyond the duration of the long-term follow-up program. However, these suggestions are yet to be explored.

In conclusion, this study has shown that individuals who had had mild-to-moderate stroke and were receiving individualized regular coaching by a physical therapist managed to establish and maintain moderate-to-good adherence to physical activity and exercise in the long term. However, dependent participants were less adherent to physical activity than those who were independent, and female sex and advanced age were associated with less adherence to exercise. Future research should investigate how to improve adherence for these subgroups of individuals.

Author Contributions and Acknowledgments

Concept/idea/research design: M. Gunnes, B. Langhammer, H. Ihle-Hansen, B. Indredavik, T. Askim

Writing: M. Gunnes, B. Langhammer, I.L. Aamot, T. Askim

Data collection: M. Gunnes, B. Langhammer, H. Ihle-Hansen, K.H. Reneflot, W. Schroeter

Data analysis: M. Gunnes, S. Lydersen, B. Indredavik, T. Askim

Project management: B. Langhammer, B. Indredavik, T. Askim

Fund procurement: T. Askim

Providing participants: B. Langhammer, B. Indredavik

Providing facilities/equipment: H. Ihle-Hansen, B. Indredavik

Providing institutional liaisons: B. Indredavik, T. Askim

Consultation (including review of manuscript before submitting): M. Gunnes, B. Langhammer, I.L. Aamot, S. Lydersen, H. Ihle-Hansen, B. Indredavik, T. Askim

The authors would like to thank Christian Sesseng for assistance with the analyses of the adherence data.

Ethics Approval

The study was approved by the Regional Committee of Medical and Health Research Ethics (REC no. 2011/1427).

Funding

The study was funded by the Liaison Committee for Education, Research and Innovation in Central Norway, the Norwegian Fund for Postgraduate Training in Physiotherapy, and the Research Council of Norway.

Disclosures

The authors completed the ICJME Form for Disclosure of Potential Conflicts of Interest and no conflicts were reported.