This study was an assessor-blinded quasi-experimental trial with a historical control group. Consecutive participants admitted to the stroke center of the First Hospital of Jilin University were recruited. The hospital is located in the central part of Jilin Province, northeast China, with a population of 27 million and the highest age-standardized prevalence and incidence rate of stroke [5]. It also has the first province-wide stroke emergency map in China, named the Jilin Province Stroke Emergency Maps [18]. The transportation of patients between cities ensured the diversity of participants in our center. Patients admitted from March to June 2020 were assigned to the historical control group, and patients admitted from July to October 2020 were assigned to the intervention group.

This study was conducted in compliance with the principles of the Declaration of Helsinki. The study protocol was reviewed and approved by the ethical committee of the First Hospital of Jilin University (20K055-001) and was registered in the Chinese Clinical Trial Registry (ChiCTR2000040804). All participants signed an informed consent form.

Patients diagnosed with a first-ever ischemic stroke [19], confirmed by head computed tomography scanning or magnetic resonance imaging; aged≥18 years; able to understand instructions; and able to use mobile phones proficiently were included. Patients diagnosed with primary mental disorders or other serious organ dysfunction, such as liver and kidney dysfunction, were excluded.

Patients in the control group were provided with a health management manual with explanations and discharge advice during hospitalization; they conducted out-of-hospital follow-ups through an online WeChat public group and received monthly telephonic coaching sessions of approximately 15 minutes each, provided mainly by a qualified stroke health manager, as previously published [20]. In response to national policies, we canceled the health education courses held in our center during hospitalization to avoid social gatherings according to hospital rules.

Patients in the intervention group received 6 additional web-based educational sessions based on the HBM via Tencent Meeting, which is an audio and video conferencing application. The sessions were held every 2 weeks between 9 AM and 10 AM, with each session lasting approximately 40 minutes. The intervention was conducted in small groups, with about 8 to 10 people in each group. A qualified stroke health manager, skilled in health education, facilitated all the sessions. The first session was focused on perceived susceptibility. Approximately 20 minutes were allocated for free discussion to encourage participants to identify and express their own risk factors. The second session focused on perceived severity on the basis of the model. The third session focused on perceived benefits. Participants could ask questions and discuss the possible benefits of their behaviors. The fourth and fifth sessions mainly supplemented cues to actions. Participants were encouraged to put forward questions that they did not fully understand and received prompt explanations. The last session aimed to address patients’ perceived barriers to secondary prevention behaviors. Participants worked with others to correct incorrect beliefs and provide strategies to overcome these barriers (Table 1).

Participants received a notification on their mobile phones 1 hour before each session. We also implemented online reminders before each session and ensured that all participants attended the meetings on time. Patients who completed all sessions and participated in 3 or more discussions during the sessions were awarded an electronic version of the “Learning Star” honorary certificate to encourage them to continue to adhere to healthy behaviors.

The outcome assessor was a full-time physician who was professionally engaged in the follow-up of cerebrovascular disease and was not involved in any of the program implementations. Data on patients’ sociodemographic characteristics, complications, and stroke severity as well as BP, HbA_1c, and LDL-C results were retrieved from electronic medical records. The above physiological and biochemical indicators were collected twice: within 48 hours of admission and during the face-to-face outpatient clinic follow-up 3 months after stroke onset.

The primary outcomes were control of BP, HbA_1c, and LDL-C levels according to the recommendations for treatment targets in the Chinese Stroke Association guidelines [21]. The reference treatment targets for patients with ischemic stroke were as follows: BP<140/90 mmHg, LDL-C<2.6 mmol/L, and HbA_1c≤7%. BP of the nonhemiplegic side was measured after the patient rested for 5 minutes in a sitting position using a digital sphygmomanometer (Omron HBP-9020). The mean value of 3 consecutive measurements with 1-minute intervals was taken as the final result. LDL-C concentration and blood levels of HbA_1c were measured in fresh venous blood samples.

The outcome of self-reported medication adherence was measured by the Chinese version of the 8-item Morisky Medicine Adherence Scale (MMAS) [22,23,24]. The total score ranges from 0 to 8, with a score ≥6 classified as adherent [25]. We also used the modified Rankin Scale (mRS) to evaluate clinical outcomes in patients with stroke. mRS is a 7-level scale, ranging from 0 (no symptoms) to 6 (death), for assessing the recovery state of neurological function. Patients with an mRS score of 0-1 were defined as having favorable outcomes.

An a priori sample size calculation was performed for this study based on the outcome “the difference in BP at 3 months after discharge in patients with hypertensive ischemic stroke receiving a comprehensive reminder system intervention [26].” Assuming a 20% drop-out rate, at least 170 patients in each group were included to achieve a power of 0.90 with a 2-sided significance level of 5%. Baseline characteristics were reported as means (SDs), median (IQRs), and proportions, as appropriate. Means or probability and 95% CIs were used to report the differences between groups and were examined using the chi-square test for categorical variables and student 2-tailed t test or Mann-Whitney U test for continuous variables, respectively. Binary logistic regression was used to adjust for confounding variables and compare the proportion of vascular risk factors that reached the target value between the intervention and control groups. Differences in the levels of vascular risk factors between the 2 groups were compared using multiple linear regression, and clinical baseline parameters were considered as covariates. Two-sided P values less than .05 were regarded as statistically significant. Data analysis was performed using IBM SPSS Statistics (version 22.0; IBM Corp).

Our study demonstrated that a planned web-based education program based on the HBM model positively impacted the control of vascular risk factors in patients who had ischemic stroke and improved medication adherence. The theory-guided practice may help enhance the effectiveness of health management. A comprehensive reminder system based on the HBM also improved medication adherence and reduced BP of patients who had a stroke [26]. However, in the SMS4Stroke study, guided by the HBM with behavior change theory, BP did not show a significant change between the 2 groups [27]. A possible reason is that the intervention duration was relatively short at 2 months. The HBM predicts the impact on human behavior from 6 aspects, and participants may change their behavior through health education related to these themes.

Hypertension control is a cornerstone of secondary stroke prevention. Health education based on the HBM increased the BP control compliance rate of participants by 2 times at 3 months, but the level was still suboptimal compared with other studies [26,28]. One of the reasons may be that the cold weather in the northeast and the residents’ preference for pickles led to a BP level of 149/87 mmHg at baseline, which was significantly higher than that of patients from Southern and Northern China [29,30]. However, our data are almost consistent with those of the China National Stroke Registry, a multicenter database covering the whole country [31]. Therefore, there is room for improvement in the next period.

Higher LDL-C levels have been associated with stroke recurrence over the past 10 years. Unlike hypertension and hyperglycemia, dyslipidemia may not cause significant physical discomfort for patients. In addition, a lack of knowledge about statins also makes patients likely to ignore adherence to long-term use of this medicine [32,33]. After the web-based education intervention, LDL-C levels decreased to 0.6 mmol/L (95% CI −0.7 to −0.4), which was double the 0.3 mmol/L reported in a 12-week exercise and education program for nondisabling stroke at 6 months follow-up [34]. HbA_1c can objectively reflect the average glycemic control over the last 3 months. With the cues to actions we provided, through lifestyle changes and adherence to hypoglycemic agents, the mean HbA_1c of patients with diabetes in the intervention group fell within the normal range, and 148/160 (92.5%) patients achieved control goals, which was better than the results of a 3-month poststroke education delivered via 5-minute movies to stroke survivor and caregiver dyads in Pakistan [35]. This is likely because only a quarter of the patients in this study had diabetes, and the patients had better glycemic control at baseline.

The 3-months medication adherence in these 2 groups was higher than that of previous studies [9,36]. However, it is still lower than the 89.0% medication adherence reported in a study of 600 patients conducted in 3 stroke centers across Korea [37]. Only patients who continued to take all prescribed medications were further evaluated for adherence, resulting in the exclusion of patients who reported discontinuing all or some medications, which may explain this difference. A study conducted in 2020 found that the imposed COVID-19 lockdown is unlikely to affect the medication behavior of patients with high preestablished adherence [38]. Therefore, we purposefully established the medication beliefs of patients based on the HBM model, which helps to ensure patients’ medication adherence under the normalization of the pandemic.

Although statistically significant improvements were found in both vascular risk factor control and medication adherence in the intervention group compared with the control group, the functional outcomes were not statistically different. This is probably because the neurological condition was not severe in our participants. Multicenters with large samples and multifaceted efforts are required to prove the efficacy of our web-based health education program.

Our study has several limitations. Due to the shortage of medical staff and researchers during the lockdown, this study used a historical control method, which provides less quality evidence compared to randomized controlled trials. Outcomes after intervention were evaluated for 3 months, and long-term effects need to be further assessed. The study was conducted in a single center, and the included patients may have regional characteristics, having a certain impact on the results. We only investigated overall medication adherence, and the persistence of a specific medication was unknown. The assessment methods of self-report may have led to recall bias; a more objective measure of adherence would be preferable to confirm our findings. However, this study design allowed us to realize the potential benefits of web-based health education under real-life conditions and complement evidence from randomized controlled trials.

This study found that a web-based health education program under the HBM model during the COVID-19 lockdown in autumn 2020 in northeast China, designed to teach patients who experienced a stroke about secondary prevention, may result in improved vascular risk factors and medication adherence. It provides insights into achieving secondary stroke prevention goals, recommended in real-world guidelines. It also serves as a reference for the use of telemedicine to conduct disease prevention under the “new normal” of COVID-19 and the shortage of medical staff.