Stroke Riskometer Application (SRA™) influence on lifestyle changes of home bound familial Malaysian stroke caregivers: a randomised controlled trial in a primary care based longer term stroke care facility

As this study was a pilot study with numerical outcomes, based on Billingham et al.’s [17] recommendation, a minimum of 30 participants were required for each group. This decision was also made based on practical factors such as willingness of the caregivers to participate in the study, attendance rate of patients, recruitment and monitoring to be done at same follow up visit for the stroke patient.

Computer generated number was used for simple randomisation to identify potential participants from the KLS appointment list, and those eligible were contacted via telephone by researcher. Familial or blood-related caregivers of stroke patients aged ≥ 18 years old, owns a smartphone and literate in either Malay or English were recruited. Pregnant caregivers or caregivers with possible depression (i.e. respondents screened positive using the Two Questions With Help Questionnaire, TQWHQ) were excluded. Shortlisted participants were contacted by the researcher and briefed about the study. Participants who agreed to participate signed the informed consent document.

Participants were asked to provide details of their sociodemographic data (proforma), stroke risk (SRA™, 20 questions) and lifestyle practices (LS7 questionnaires), and were then interviewed by the researcher to verify their risk factors for stroke (i.e. hypertension, diabetes, dyslipidaemia, previous history of stroke), including physical examination (i.e. blood pressure, body mass index) and reviewing blood tests for total cholesterol and fasting blood sugar levels within the last 6 months. The researcher then used the SRA™ to calculate stroke risk and informed study participants about the findings. All participants in the usual care and SRA™ intervention group received standard counselling and explanation about their stroke risk. Once briefed about their stroke risk, they were randomly assigned to either intervention (SRA™) or conventional care group (i.e., education pamphlet from Ministry of Health, MOH).

An assistant researcher, who was not part of the study, managed allocation concealment. This involved enrolling participants and assigning them to different treatments using a randomly generated code from sealed envelopes. The main researcher created the code before the study and kept it concealed until participants were recruited after baseline measurements.

Due to the nature of the intervention, masking was not possible. Caregivers in the intervention group received the SRA™ which was downloaded onto their smartphones by the researcher during intervention assignment, in addition to the standard health education (pamphlet). The pamphlet, which is authored by the MOH contains general healthy lifestyle advice and did not contain any information or endorsement on usage of app to reduce risk of stroke or NCD (MOH 2012).

To prevent contamination of data among participants during the study, caregivers were asked if they were aware or have downloaded any healthcare application on stroke risk reduction prior to the study and during exit visit assessment. Study participants were then assessed three months later (exit visit assessment), during a face-to-face evaluation visit or via telephone by the researcher based on participants’ availability for assessment.

The researcher assisted caregivers in the intervention group to install the SRA™ app, and provided individualised training on how to use the app. They had free access to the app including the health education sections during the intervention period and also received the pamphlet. Caregivers could communicate with the care management team about the app through WhatsApp or email.

The care management team included four members—one medical officer, two staff nurses, and one research assistant. The team received training from the researcher prior to commencement of the trial. The team performed the following functions: provided guidance when required to ensure that the app functioned well, conducted interviews with participants regarding lifestyle according to LS7, tracing blood investigations and informing participants. At baseline, all participants answered the questions in the application, with minimal assistance from researcher. Their absolute and relative risk of getting stroke for the next 5- and 10-years’ time was automatically estimated and standard medical care (motivational education) according to the app-based information was given for those at risk. Standard medical care was defined as treatment provided at primary care level according to Malaysian Clinical Practice Guidelines from Academy of Medicine Malaysia and Ministry of Health (MOH), Malaysia. Standard medical care was given at primary care clinic by registered doctors to all study participants. In addition, participants assigned to the app were able to calculate their own stroke risk and had access to self-help advice embedded in the app and how to act on high-risk domains. The main benefit for interventional group was the ability to learn more about their individual stroke risk with additional opportunity to save their stroke risk profile and monitor changes in their stroke risk over time. Reminders on managing their stroke risk such as exercise and adhering to medication was accessible via the app. Participants in the control group were expected to act on reducing their known risk of getting stroke according to the standard medical care only. Both control and interventional groups was assessed again after 12 weeks by filling up the questionnaire either by email or phone call. During exit assessment only BMI, BP, lifestyle components such as self-reported physical activities were assessed. Due to financial and time constraints, repeat blood investigations (i.e., total cholesterol and fasting blood sugar levels) were not performed (Fig. 1). The review at 12 weeks was timed to coincide with the stroke patients’ follow up schedule at KLS.

The trial was prospectively registered in the Australia-New Zealand Clinical Trial Registry (ACTRN12618002050235, 21/12/2018).

Primary objective was to evaluate the impact of Stroke Riskometer Application (SRA™) on lifestyle changes among stroke caregivers using Life Simple 7 scores. Secondary outcomes were to assess caregivers’ stroke risk and relative risk of stroke at 5 and 10 years in the intervention group and Conventional Care group, using SRA™ and to determine lifestyle changes of caregivers within 3 months after knowledge of personal stroke risk, using the LS7 scores. Basically, calculation of absolute and relative risk reduction of stroke were performed in both intervention and conventional group.

Intention to treat principles were used, with last observation carried forward employed for participants who did not turn up for exit assessment. Demographic and risk factor profiles for participants under SRA™ and conventional care were analysed using SPSS version 25 software [21]. All effect estimates were adjusted for prognostic risk factors at baseline, and 95% confidence intervals (CI) were provided. The level of statistical significance was set at p-value of < 0.05. Wilcoxon Sign rank test was used to compare pre and post intervention scores of all participants and Mann Whitney test to compare changes between both groups. Multilinear regression analysis was conducted to identify factors which influenced changes in stroke risk, with significance level set at p < 0.05.

Stroke risk: Changes between both groups were not significant (refer Table 2), although within groups, there was a 24.5% reduction in stroke risk at 5 years in the SRA™ group, compared to 16.7% reduction in the conventional care group (Refer table 3). However, these were statistically not significant when we compared between the two groups (Refer table 3). Similar findings were noted in the improvement of stroke risk in the next 10 years (SRA™ 7.1%, conventional care 6.7%).

Relative risk of stroke: Reduction was noted in both groups at 5- and 10-years (RR in 5 years- SRA™ 16.5%, Control 15.0%; RR in 10 years- SRA™ 9.3%, Control 18.4%).

Lifestyle changes: At exit assessment, the LS7 total scores showed improvement by 53.8% in the app group with median (IQR) from baseline of 6.5 (2.0) to 10.0 (3.0). As for participants in the control group, their LS7 score showed less improvement from 6 (2.5) baseline to 8 (2.0) at exit, which is a 33.3% improvement. Adjustment was made to analyse the LS7 scores to compensate for unrecorded exit fasting blood sugar and total cholesterol levels.

None of the participants repeated formal blood test at exit interview. Only four participants did capillary blood glucose test at outside facilities i.e. pharmacy/ home glucometer. Readings were informed verbally as “normal value”. Despite the adjustment of using only 5 of the 7 variables, (i.e., LS5 scores), the results remained consistent, with an improvement of 75% seen in the app group while conventional care group showed improvement of 50% in total scores.

Based on multilinear regression analysis, factors which were associated with changes in LS7 scores were body mass index (p = 0.001), smoker status (p < 0.001) and presence of diabetes at the point of entry into the study (p = 0.021). Refer Table 4.

One of the strengths of this study was the inclusion of an ethnically diverse urban group of stroke caregivers, who are at increased risk of stroke i.e. the immediate family members of stroke patients. The heterogeneity of the sample population may increase generalizability of results to cardiovascular disease prevention in other settings. However, there were limitations where the short duration of follow up for our study (three months) was inadequate to observe any sustainable improvement in the intervention group. Secondly there was a possible chance of contamination between intervention and control group who might be able to download the app upon exposure during baseline assessment. It is possible that they may have accessed other mobile health application that is similar to SRA™ during the study period (i.e.,7 out of 37 in control arm had downloaded the SRA™). Caregivers among the same family members may be more inspired or motivated to reduce their risk of stroke by adopting a healthier lifestyle and managing their risk factors together as a group.

The main processes that may have led to contamination in this study were health professionals delivering both active and comparator treatments and communication between clinicians and participants from the different trial arms. Contamination can be avoided by carrying out a clustered study design, where clusters were defined by geography, holding treatment sessions at different times in different locations, staggering the arrangement of data collection schedules so that participants do not meet in the waiting room [26]. Communication between participants was thought to be most likely in environments in which participants came into close contact. Examples of this included interaction between participants who were family members and patients in the waiting room. The third limitation was recall bias in self-reporting of lifestyle changes. However, the use of standardized questionnaires and uniformly trained research assistants reduced the likelihood of such bias. The fourth limitation could be attributed to the contact the participants had with the study team, who provided some technical support or guidance to use the app. However, the participants did not require much support hence, it is unlikely that the contact may have contributed to the improvement outcome or response desirability bias.

O Meldevdev and colleagues concluded that the SRA™ consistently captures stroke risk across various countries including Asian. Thus, users can depend on its capability to measure their stroke risks, but as it is a novel stroke prevention approach, the SRA™ should continue to evolve and more research is required to further augment precision and validity of this advantageous tool [26].

Future work should also focus on a qualitative study to explore how the participants felt the mobile application helped in initiating or maintaining the lifestyle changes. A cohort-based study measuring sustainability of life style changes and its impact on stroke risk reduction among Malaysians would support further evidence for its usability and acceptability.