Effect of an Educational Intervention for Primary Stroke Risk Reduction in Ghana & Nigeria: Pilot Randomized Controlled Trial

Abstract

Background:

Using tailored mobile health interventions to improve global vascular risk awareness and control is yet to be investigated for primary stroke prevention in Africa.

Methods:

This 2-arm pilot randomized controlled trial involved 100 stroke-free adults with at least 2 vascular risk factors for stroke. Eligible participants were assigned randomly to a control arm offering one time counselling (n=50) or a 2-month educational intervention arm (n=50) comprising a stroke video and riskometer app aimed at improving stroke risk factor awareness and health-seeking behavioural modification to control total vascular risk. Reduction in total stroke risk score was the primary outcome while feasibility and process measures were secondary outcomes.

Results:

All enrolled participants completed the 2-month follow-up (retention rate = 100%). The mean (standard deviation) age of participants was 59.5 (α12.5) years, 38% were males. The mean change in stroke risk score at 2 months was −11.9% (α14.2) in the intervention arm vs −1.2% (α9.1) in the control arm, p=0.0001. Stroke risk awareness improved by 16.1% (α24.7) in the intervention arm vs 8.9% (α24.7) in the control arm, p = 0.08. The intervention arm had 11.1 mmHg reduction in systolic blood pressure compared with 4.8 mmHg reduction in the control arm.

Conclusion:

The intervention demonstrated a positive signal of effect over a 2- month period. A definitive clinical trial with a longer duration of follow-up is warranted on the premise of these promising findings from this pilot RCT.

Graphical Abstract:

INTRODUCTION

Stroke is a major public health concern in Africa with high burden and poor outcomes.¹ With inadequate care facilities, primary prevention assumes foremost importance in the region.² Primary prevention is however thwarted by a myriad of factors, especially lack of awareness of stroke risk factors.^1,2

It has been suggested that effective implementation of population-wide strategies, including motivational digital health technologies can prevent up to 50–90% of stroke and cardiovascular disease events world-wide.² The objective of this pilot randomized controlled trial (RCT) was to test the feasibility and signal of effect of a mobile health (mHealth) intervention comprising the first-ever Afrocentric stroke risk estimation³ app and a stroke video in improving stroke risk factor awareness and global risk reduction among stroke-free adults compared with control group.

METHODS

Study design:

This was a 2-arm pilot, parallel, RCT involving 100 stroke-free adults with at least 2 vascular risk factors for stroke. Eligible participants were assigned randomly either to a control arm offered one time counselling (n=50) or an educational intervention arm for 2 months comprising stroke video and Afrocentric riskometer app aimed at improving stroke risk factor awareness and health seeking behavioural modification to control vascular risk (n=50). Institutional approval for the study was obtained from Ethics committees in Ghana and from the two Nigerian study sites, Ibadan and Zaria. Data for the study is available upon request from the principal investigator, MOO.

Details of the Methods are provided in the Supplemental Materials (including Tables S1, S2, S3 and S4).

Eligibility criteria:

Inclusion criteria:

(i) male or female; (ii) age ≥ 18 years; (iii) with at least 2 stroke risk factors based on the list of 11 topmost modifiable risk factors identified in the SIREN study (including hypertension, dyslipidemia, regular meat consumption, elevated waist-to-hip ratio, diabetes mellitus, low consumption of green leafy vegetables, stress, added table salt, cardiac disease, physical inactivity and cigarette smoking ⁴ (iv) ownership or access to smartphone in consenting stroke-free adults.

Randomization and masking

Randomization was at individual level, per site. Participants were not masked to group assignment due to the nature of the intervention. However, outcome measurements were performed by blinded assessors.

Outcome measures:

1. Signal of Efficacy measures: a) Change in total stroke risk score at end of month 2 calculated as difference between baseline score and month 2 score (primary outcome); b) Change in stroke risk factor awareness score; c) Change in systolic blood pressure, diastolic blood pressure, and body mass index over 2 months.

   The stroke risk factor awareness score was calculated out of a 15-item questionnaire with 11 items that are associated with stroke occurrence and 4 items that are not. The total number of correct responses out of 15 is multiplied by 100. (Table S1)

2. Feasibility measures: Proportion approached, proportion refused, proportion ineligible (reasons for ineligibility), proportion enrolled, proportion completing follow-up.

Power considerations and Statistical Analysis:

For this pilot trial, though our sample size focused on precision of estimates for a future study design, post hoc power estimation showed that the pilot trial had more than 90% power to detect the observed difference for change in total risk score. We compared means for continuous variables using the Student’s t-test while categorical variables were compared using the Chi-squared tests. A p<0.05 was considered a significant difference between the 2 groups.

Process measures:

included thematic analysis of recorded, transcribed key informant indepth interviews of intervention participants and providers.(Supplementary file)

RESULTS

Feasibility outcomes:

Of 985 consecutive stroke-free controls from the three participating sites screened for ≥2 risk factors, 565(57.4%) were eligible. Of those eligible, 20.0%(113) were randomly selected for the pilot and approached. They all consented. However, 13(11.5%) of those selected were not available at the time of study. One hundred stroke-free participants immediately available were enrolled, and randomized. All the 100 participants completed the 2 months follow-up giving a retention rate of 100%. See Figure 1 (CONSORT diagram).

Figure 1: CONSORT Diagram.

Baseline comparison of eligible study participants:

The mean ± SD age of study participants was 59.5 ± 12.5 years, 38% were males and 80% were Nigerians. Educational status was unevenly distributed by study arm( Table 1). Awareness of risk factors of stroke and baseline stroke riskometer scores were comparable between the 2 arms at baseline (Table S1).

Table 1.

Comparison of baseline characteristics of Intervention vs Control Arm Participants

	Intervention arm (n=50)	Control arm (n=50)	P-value
Age, mean(SD)	60.6 (9.6)	58.4(14.8)	0.381
Male, n(%)	21(42.0)	17 (34.0)	0.410
Nationality			0.999
Ghanaian	10 (20.0)	10(20.0)
Nigerian	40(80.0)	40(80.0)
Educational attainment			0.041
None	6 (12.0)	11(22.0)
Primary	16(32.0)	17(34.0)
Secondary	20 (40.0)	8(16.0)
Tertiary or higher	8(16.0)	14(28.0)
Monthly income level USD			0.795
0–100	32(68.1)	32(65.1)
101–250	12(25.5)	12(24.5)
>= 251	3(6.4)	5(10.2)
Type of domicile			0.999
Urban	40 (80.0)	40 (80.0)
Peri-urban / rural	10(20.0)	10(20.0)
Employment status			0.123
Employed	36 (72.0)	38(76.0)
Unemployed	5(10.0)	9 (18.0)
Retired	9(18.0)	3 (6.0)
Systolic BP, mm Hg, mean (SD)	156.6 (24.8)	151.4 (27.0)	0.317
Diastolic BP, mean(SD)	91.0 (15.7)	92.2 (15.9)	0.925
BMI, mean(SD)	24.7(5.5)	25.7(5.7)	0.385
Waist-to-hip ratio, mean (SD)	0.89(0.10)	0.93(0.11)	0.052
Fasting blood glucose mg/dl, mean (SD)	94.8(45.3)	87.9(24.5)	0.387
On antihypertensive, n(%)	26 (50.0)	18(36.0)	0.183
On Statin, n(%)	0(0.0)	0 (0.0)	-
On Antiglycemic, n(%)	1(2.0)	2(4.0)	0.588
On Aspirin, n(%)	0(0.0)	1(2.0)	-

Signal of efficacy measures:

All 100 participants completed the 2-month visit with analysis carried out on 50 participants per study arm. The mean ± SD change in stroke risk score at 2 months was −11.9% ± 14.2 in the intervention arm versus −1.2 ± 9.1 in the control arm, (p=0.0001). Change in stroke risk awareness was 16.1% ± 24.7 in the intervention arm versus 8.9% ± 24.7 in the control arm, (p = 0.08) (Tables S1 and S3). Changes in blood pressure, and BMI are shown in Table 2. A comparison of stroke risk factor profile at baseline and at month 2 by intervention arm is shown in Tables S2 and S3 with improvement in proportions of intervention participants with normal BP (without medication use) and normal waist-hip ratio at follow-up.

Table 2:

Study Outcomes:

	Intervention arm (n=50)	Control arm (n=50)	P-valuenp	Adjusted treatment effect (CI)#	Adjusted p-value
Change in stroke risk score (%), mean(sd)	−11.9(14.2)	−1.2(9.1)	0.0001	−10.4 (−15.5, −5.3)	0.0002
Change in stroke risk factor awareness (%), mean(sd)	16.1(24.7)	8.9(24.7)	0.083	8.0 (−2.2, 18.2)	0.121
Systolic blood pressure change from baseline, mean(sd)	−11.1(21.7)	−4.8(13.0)	0.051	−6.3 (−13.9, 1.4)	0.106
Diastolic blood pressure change from baseline, mean(sd)	−2.8(14.3)	−2.3(8.2)	0.727	−0.2 (−5.1, 4.8)	0.935
BMI change, mean(sd)	−0.48(1.17)	−0.62(3.20)	0.179	−0.02 (−1.0, 0.9)	0.971

^np: p-value estimated from non-parametric test (Wilcoxon rank-sum test) because of the skewed distribution of changes between baseline and follow-up measure of outcomes

^#treatment effect (difference between intervention and control arm), adjusted for education using an analysis of covariance (ANCOVA) model

Process measures:

All 16 clinical providers who provided feedback through key indepth informant interviews found the app acceptable, easy to use and expressed willingness to use the app to monitor their patients overall progress and risk factor control. Additionally, all 25 alternating intervention arm participants who provided feedback on their intervention experience via individual key indepth informant interviews reported the process was simple, easy to use, and convenient. They expressed willingness to use the app in future to help them reduce their risk of stroke. In addition, all control participants (100.0%) reported no exposure to the stroke video or app during the entire study duration.

No adverse events were reported.

DISCUSSION

In this pilot RCT involving individuals with at least two modifiable risk factors for stroke occurrence, a motivational intervention administered via mHealth was feasible in Ghana and Nigeria with 100% retention over 2 months. The combination of a stroke educational video and stroke risk estimation demonstrated a signal of efficacy with reduction in mean global stroke score of about ≈12% in the intervention arm vs ≈1% in the control arm from baseline. There was also a 16.1% versus 8.9% improvement in awareness of stroke risk factors in the intervention arm compared with usual care, though not statistically significant. By specific risk factors, the intervention arm had ≈11 mmHg reduction in systolic BP compared with ≈5 mmHg reduction in the control arm.

A proof-of-concept trial of a stroke riskometer app addressing stroke awareness and modifying stroke risk behaviors among a total of 50 Australian participants assigned to either the app or to usual care over a 6-month follow-up was also feasible with retention rate of 87% and non-signficant improvement total stroke risk( p=0.67).⁵

Limitations and future directions:

There was significant difference in educational attainment between the intervention and control arms. However, this was adjusted for in our analysis of the outcome measures (Table 2). Given the short duration of our study and the relatively small sample size of the pilot trial, further studies with a larger sample size and longer duration of follow-up of 12 months are warranted to assess sustainability of the signal of vascular risk reduction observed. Participant knowledge of their assignment to different interventions might have affected their behavior during the trial.

Conclusion:

A full scale clinical trial is clearly justified based on the promising findings from the present pilot trial including signal of improvement in overall stroke risk, blood pressure and waist-hip ratio.

Non standard abbreviations

BP

blood pressure

CONSORT

Consolidated Standards for Reporting Trials

Mhealth

mobile health

RCT

Randomized clinical trial