Feasibility of recruitment and retention
A total of 312 patients admitted with acute stroke were screened for study eligibility and 34% (n=105) met the study inclusion criteria (Figure 1). Consent was provided by 37 individuals. There was an observed recruitment rate of 2.8 participants per month per hospital. At the 7-14 day follow-up call, 17 were randomised to the intervention and 16 to control. Reasons for non-eligibility at randomisation (n=4) comprised: inability to complete the baseline assessment (n=1), inability to locate the participant (n=1), and a surgical intervention that led to prolonged hospitalisation (n=2). Twenty-nine participants (88%) completed their outcome assessment (intervention n=15; active control n=14). The outcome assessments were completed a median of 104 (range 90 to 154) days from randomisation. One outlier who was difficult to contact, had their data collected 154 days post-randomisation. Less than 1% of data were missing at baseline and outcome assessments.
Insert Figure 1 about here
Characteristics of participants
The baseline characteristics were similar between control and intervention groups, although, of note, there were more females and fewer living with a partner in the control than intervention group (Table 3). Twenty-two participants indicated that email was their preferred method for receiving electronic messages, and all participants indicated a preference for follow-up assessments by telephone call, rather than video-conference call.
Table 3. Participant characteristics a
Baseline Characteristics
|
Control
n (%)
N=16
|
Intervention
n (%)
N=17
|
p-value
|
Demographics
|
|
|
|
Age, mean (SD)
|
64 (11)
|
61 (14)
|
0.62
|
Female
|
7 (44)
|
2 (12)
|
0.04
|
Australian
|
14 (88)e
|
13 (81)
|
0.50
|
Married/with partner
|
10 (63)
|
16 (94)
|
0.04
|
Live independently
|
3 (19)
|
1 (6)
|
0.15
|
Own home or unit
|
15 (94)
|
17 (100)
|
0.30
|
University educated
|
5 (31)
|
3 (18)
|
0.51
|
modified Rankin Scale
|
|
|
|
0-No symptoms at all
1-No significant disability
2-Slight disability
|
1 (6)
11 (69)
4 (25)
|
3 (18)
11 (65)
3 (18)
|
0.57
|
Self-reported medical history
|
|
|
|
Hypercholesterolaemia
|
5 (31)
|
7 (41)
|
0.84
|
Heart attack
|
2 (13)
|
1 (6)
|
0.45
|
Atrial fibrillation
|
3 (19)
|
2 (12)
|
0.47
|
Hypertension
|
12 (75)
|
9 (53)
|
0.19
|
Sleep apnoea
|
3 (19)
|
2 (12)
|
0.47
|
Respiratory problems
|
1 (6)
|
5 (29)
|
0.09
|
Diabetes
|
5 (31)
|
2 (12)
|
0.17
|
Arthritis
|
7 (44)
|
7 (41)
|
0.88
|
Depression
|
4 (25)
|
1 (6)
|
0.13
|
Anxiety
|
4 (25)
|
1 (6)
|
0.13
|
Cancer
|
1 (6)
|
1 (6)
|
0.97
|
Other serious illnesses
|
0 (0)
|
4 (24)
|
0.04
|
Lifestyle characteristics
|
|
|
|
Smokingb
|
|
|
|
Current smoker
|
3 (19)
|
1 (6)
|
0.51
|
Past smoker
|
5 (31)
|
7 (41)
|
|
Physically activec
|
8 (50)
|
9 (53)
|
0.87
|
Alcohol consumptiond
|
12 (75)
|
16 (94)
|
0.13
|
Healthy eating
|
|
|
|
Advised to change diet
|
10 (63)
|
6 (35)
|
0.12
|
>5 servings of vegetables daily
|
3 (19)
|
0 (0)
|
0.06
|
>2 servings of fruit daily
|
9 (56)
|
7 (41)
|
0.39
|
a n (%) unless otherwise indicated.
SD - Standard Deviation; b Self-report of current smoking status; c undertaking >20 minutes of vigorous intensity physical activity > 3 times per week; d self-report consumption of alcohol-containing drinks; e 1 case missing;
Characteristics of clinicians
Six hospital clinicians participated in one of three focus groups (with at least one from each participating hospital); two identified as nurse practitioners, one as a physiotherapist, one as an occupational therapist and two as trial nurses. Clinician participants were all involved in the clinical care of adults after stroke. Most (four of the six) reported prior experience with setting person-centred goals. However, the degree of experience varied between clinicians, with two having held responsibility for leading interdisciplinary goal-setting within stroke rehabilitation teams, while others had only contributed to such teams, or had no prior experience.
Quality assessment of SMART goals developed by hospital clinicians
The person-centred goals developed by hospital clinicians were independently assessed as varying in quality, with SMART-GEM scores ranging from two (poor quality), to six out of six (median score 5.3, Q1 4.6; Q3 5.65). One clinician had quality scores below the SMART-GEM threshold of four; this occurred with four of the eleven goals developed with three participants.
Hospital clinician feedback on study processes
Clinicians reported several barriers to recruitment, including competing clinical demands, other trials vying for participants, and time constraints for completing person-centred goal setting with participants before discharge. Clinicians new to clinical trials highlighted being unsure of how to approach potential participants and provide “the right amount of information” about the trial without “overburdening” them (Physiotherapist, female, Hospital 2). Clinicians were responsible for a full clinical caseload in addition to being responsible for study recruitment, limiting their availability to screen and recruit. As highlighted by one clinician: “And I guess what I know is that my role is centrally around acute stroke management so, code stroke calls. This means that there are days where I'm just completely unable to even screen.” (Nurse practitioner, female, Hospital 1).
The inclusion criteria (discharge to home) placed time pressure on recruitment, as discharge destination was not always clear at the time of admission, and the time window for recruitment was often limited.
“Yeah, it's, it's, as I say, it's always towards discharge, it's always sort of towards the end of the admission. And it's typically that I'm providing information and consenting on the same day.” (Trial nurse, female, Hospital 1)
“It's a tricky thing with these sort of discharge support type trials, isn't it because it's not like a drug trial where you find them on admission, and you've got a whole kind of length of time to get through everything you've got to.” (Occupational Therapist, female, Hospital 2)
Hospital clinicians also reported that, baseline measurements were completed over several sessions, due to the time required to complete, and competing clinical demands. Many clinicians left the assessments with the participant for self-completion.
“I am always on the lookout for patients who I think are going to be able to be very independent in filling out their questionnaires, because for me being, unfortunately, a slave to the stroke pager means that there are times where I'm just gonna have to go and interrupt our interview... It [the questionnaire] takes quite a long time. And if I can leave it with a patient, great. And usually I'll leave that with them for a few hours.” (Nurse practitioner, female, Hospital 1)
Nurses, in particular, reported a lack of confidence in developing person-centred goals while for other health disciplines (physiotherapists or occupational therapists), setting person-centred goals was considered to be consistent with their usual practice. One nurse clinician, felt that they had developed new skills during the trial and continued to implement this type of goal setting with all patients at discharge. The tension between spending the required time with the participant for person-centred goal setting whilst meeting health service needs and routine discharge timelines was raised by at least one clinician across each participating hospital, highlighting a universal theme of time constraint in acute hospitalisation.
Feasibility of outcome assessment procedures
The 12-week outcome assessments with participants were completed in one telephone session, and took between 60 and 90 minutes. Some participants reported completing mailed copies of the questionnaires (sent to support telephone completion) but did not return them as there was no return envelope. This highlighted the need for study procedures to be revised to enable self-completion as on option. Revised study processes were introduced in April 2019 to facilitate the self-completion of surveys (return by post or electronically). The assessors subsequently reported the remaining assessments taking between 30 to 45 minutes to complete for those who had completed their surveys prior (electronic or returned paper forms).
There were challenges in meeting the outcome assessment timeframe whereby only 12/29 (41%) were conducted in accordance with study protocol (week 12 post randomisation +14 days). Outcome assessors identified a need to improve participant tracking of follow up visits to meet protocol requirements. Recommended improvements included setting up reminder notifications in REDCap® to be sent to the outcome assessor and project coordinator for when participants’ assessments were due. The tracking database was improved to record the outcome assessor allocated for the call, the participants’ preferred contact days/times, and recording of notes relating to attempted contact calls.
Intervention fidelity
Among the 33 participants, a total of 131 person-centred goals were developed across the five major domains. The proportion of participants who selected goals within each of the major goal domains were similar between the intervention and control groups (Table 4). In total, 63 (47%) person-centred goals were developed from the Health domain (e.g. secondary prevention) with the most frequently associated with exercising (n=13); losing weight (n=12) and management of blood pressure (n=10). Overall, return to driving was the goal most frequently selected by participants (n=17, 52%).
Table 4. Numbers of participants selecting goals within each major goal domain and according to group allocation
Goal menu items
|
Overall
|
Control group n (%) N=16
|
Intervention group n (%) N=17
|
n (%)
|
N=33
|
Major goal domains
|
|
|
|
Your Health
|
32 (97%)
|
16 (100)
|
16 (94)
|
Mind and body
|
18 (55%)
|
9 (56)
|
9 (54)
|
Everyday Activities
|
8 (24%)
|
4 (25)
|
4 (24)
|
Out and About
|
22 (67%)
|
11 (69)
|
11 (65)
|
Healthcare and Support
|
5 (15%)
|
3 (19)
|
2 (12)
|
The time taken to set up the messaging schedule for participants in the intervention group took between 50 and 60 minutes at the start of the trial. This did decrease over time to between 35 and 45 minutes, with the development of a messaging schedule template according to the number of goals developed.
Participants in the intervention group were sent a total of 1115 messages (average 66 per participant, standard deviation [SD] 10) compared with 101 messages sent to the control group (average 6 per participant, SD 1). Two participants in the intervention group sent a ‘stop’ message: participant 1 after receiving 41 of 63 messages, finding them “annoying”; participant 2 after receiving 74 of 76 scheduled messages, and who indicated a particular goal had been achieved. Among the 303 messages sent to participants with an embedded bit.ly hyperlink to additional support resources, the hyperlinks were accessed 109 times (36%). The most frequently accessed hyperlink was the Stroke Foundation Enable Me website (44 times, 40%). Eight messages were received from participants in response to questions such as “would you like to speak with someone for further information?”, or “confirm, by reply SMS or email, that you have received our messages?”.
Participant feedback
Sixteen participants (48%; 10 intervention: 59%; 6 control: 38%) provided feedback via the satisfaction survey. Most respondents (n=14, >80%) reported that the goal setting menu was useful to develop person-centred goals, and the clinicians were helpful during this process. Participants were satisfied with the domains categorised on the goal setting menu and did not provide suggestions for modification. Most participants (n=14, >80%; intervention n=9, control n=8) agreed that it was worthwhile taking part in the study and they would recommend it to others with stroke. Twelve participants responded to the item about the content of the project, and 58% (intervention n=4, control n=4) agreed that they found it relevant to them. Most participants (n=13, >80%; intervention n=8, control n=5) agreed that they trusted the information provided (Figure 2).
Insert Figure 2 about here
All respondents from the intervention group agreed that they led the process of selecting their own goals. Ninety percent reported that the messages received improved their knowledge of stroke and they were comfortable accessing and responding to the messages (via mobile phone or computer). Most respondents (70%) from the intervention group reported the frequency of the electronic messages received suited them and the 12-week period to receive them was sufficient for their needs. One participant reported a reduced length of time would have been sufficient, and one participant felt the messages should be sent over 24 weeks.
Among intervention participants, most of the respondents (>70%) reported the electronic messages helped to increase self-management, in adopting healthy lifestyle behaviours and in working towards achieving goals. One participant reported they had not received any messages from the ReCAPS team. Five responders (50%) in the intervention group reported not accessing web links for additional information from the messages they received as they did not understand how to access the link, or did not notice the weblinks. All responders in the intervention group reported increased confidence in using electronic devices since being involved in the study.
Safety
Five participants (three in the intervention group) reported seven adverse events at the 90-day outcome assessment, all of which were unrelated to the intervention. These included a medication side effect; presentations to the emergency department (n=5) and one planned hospital admission for vascular surgery.