Home Blood Pressure Telemonitoring and Hypertension Management in Kenya: A Feasibility Study (HBPT-K)

Design

Single-arm non-randomized feasibility study.

Setting, participants, and recruitment

This six-month feasibility study was conducted at the Mwai Kibaki Kenyatta National Hospital Annex, which is a 350-bed capacity, level six national teaching referral hospital located in Othaya Sub County, Nyeri County. We aimed to include adult patients with a new diagnosis of hypertension or with known uncontrolled hypertension. An office BP threshold of > 140/90 was chosen as the most widely reported definition of (uncontrolled) hypertension in both national and international guidelines. See Table 1 for an overview of the in- and exclusion criteria.

All individuals were recruited sequentially at the study site by research assistants and/or treating physicians during regular outpatient clinic follow-up visits for hypertension.

HBPT intervention

After giving informed consent, patients were provided with an Omron M3 Upper Arm BP monitor blood pressure machine and access to the Luscii (14) Home Blood Pressure Telemonitoring application for at least six months. The hypertension program and Luscii (14) application used in this study have been available since 2021, and have been described and evaluated in multiple studies (15, 16). The program (measurement frequencies, home blood pressure monitoring technique, and threshold values) is developed in line with the latest European Society of Hypertension (ESH) guidelines (17). The Luscii (14) application consists of a smartphone or tablet application and an online healthcare provider dashboard (web page). BP data is transmitted directly to the healthcare provider dashboard via the smartphone- or tablet application. Once transmitted, the BP data is analyzed based on predefined algorithms that will subsequently generate alerts (see details below) which are only visible in the healthcare provider dashboard. In the user application, patients can review their BP data and have access to relevant educational modules and self-care documents, which, for example, emphasize the need for smoking cessation and antihypertensive drug adherence. These so-called self-care documents and educational lessons were provided to the patients both in English as well as in Swahili. See Multimedia Appendix 2 for an overview of all included educational lessons and self-care documents.

Patients measure their BP twice in the morning and twice in the evening during measurement weeks. The program consists of four acute measurement algorithms and one chronic measurement algorithm. The acute algorithms have higher frequencies of measurement weeks and stricter threshold values and are used during the initial titration phase of achieving BP control. Stable patients with BP values closer to the desired BP target values are included in the chronic algorithm and measure one week per month. See Multimedia Appendix 2 for a complete overview of the measurement algorithms.

The telemonitoring program generates three different types of alerts (Figure 1) following BP data transmission by the patients. First, there are simple alerts, which are triggered by a one-time measurement exceeding predefined thresholds—for example, a very high systolic BP. Second are complex alerts, which are triggered if a series of measurements are off target (e.g., multiple consecutive off-target BP measurements). In practice, these alerts are used for clinical decisions, such as drug treatment adjustments. Third are alarms that are triggered when patients either miss a single scheduled measurement (overdue alert) or when they have missed all measurements within a certain time frame (inactive alert).

Figure 1

Alerts can be processed automatically (e.g., a single overdue alert) by the telemonitoring platform based on automated workflows created by the responsible healthcare providers. These workflows can contain automated feedback messages that are delivered to the user automatically when the alert is processed by the monitoring platform (e.g., your BP has improved, your target has changed). Besides auto processing by the telemonitoring platform, alerts can also be processed manually (e.g., a complex alert for off-target BP) by a healthcare provider. This processing is usually done by a dedicated E-nurse (a nurse who is responsible for processing the alerts and contacting the patients if needed) and a similar set-up was used in this study. An alert is processed manually with an administrative action if no treatment adjustment is performed, and processed with a clinical action if the antihypertensive treatment is adjusted, for example.

Whether alerts need manual processing depends on choices made by healthcare providers in the telemonitoring algorithm. Prior to processing the alerts, healthcare providers can decide to contact patients to gather relevant clinical information (e.g., adherence to antihypertensive drugs of lifestyle interventions). They can also provide feedback to the patients—for example, following antihypertensive treatment adjustment. Patients were also contacted by the E-nurse following inactivity alerts.

Outcomes and data collection

The primary outcome for this study was the difference in the proportion of BP control (% with a BP < 140/90) between baseline and T = 6 months, as measured with a SPRINT standardized in-office BP measurement (18). Secondary endpoints included:

• – Differences in drug adherence as measured with the MARS-5 (19) scale between baseline and T = 6 months. Scores range from 5 to 25, where a higher MARS-5 score indicates higher self-reported adherence.

• – Patient satisfaction as measured with the validated Telehealth Usability Questionnaire (TUQ) (20) and mHealth App Usability Questionnaire (MAUQ) (21) at T = 6 months. Scores range from 1 to 7, where higher TUQ and MAUQ scores indicate higher satisfaction.

• – Healthcare provider satisfaction as measured with the validated mHealth App Usability Questionnaire (MAUQ) at T = 6 months. Scores range from 1 to 7, where a higher MAUQ score indicates higher satisfaction.

Additionally, at baseline, we collected demographic data (age, sex, medical history, current use of antihypertensive drugs, and smoking status). To estimate the feasibility in terms of using the HBPT program and its potential workload, we extracted telemonitoring data after the final study participant completed the T = 6-month visit. This data included the total telemonitoring participation (active days in the program), adherence to the measurement schedules, number of alerts, and data on the telemonitoring alert processing, including the time spent processing alerts. To calculate the time spent processing alerts, we assumed a workload of five minutes for a digital message, five minutes for a phone call, 10 minutes for a clinical action, and two minutes for an administrative action. We also extracted the number of physical consultations during the study period (excluding the study visits).

Sample size

We estimated the level of BP control to increase, based on available HBPT studies conducted outside the African continent, by 20% from 0% at baseline in our study population (22). For a McNemar test to detect a 20% difference between marginal proportions at T = 6 months with a power of 80% and a two-sided significance of 5%, at least 42 individuals would need to be recruited. As this is a feasibility study and relevant data on potential dropouts and potential effects on BP control in this specific setting and population is not available, we opted for a more pragmatic approach and decided to recruit 100 study participants.

Statistical analysis

All statistical analyses were performed following the per-protocol principle and data was only used from patients that completed the T = 6-month follow-up. Descriptive statistics were used for the demographic data. Data was reported as proportions (%), means (standard deviation), or medians (interquartile range). For the primary outcome, a McNemar test for proportions was used, and to compare the average BP (absolute systolic and diastolic values) between baseline and T = 6 months, a paired T-test was used. All statistical analyses were performed using Microsoft Excel (23) or R (24).

Study participants and baseline data

Between March 2024 and January 2025, a total of 100 patients gave informed consent to participate in the study. Out of these, 16 were lost to follow-up during the study period and therefore, a total of 84 participants were included in the final analysis. See Figure 2 for the study flowchart.

Figure 2

The mean age was 54 (SD = 14) and 73% were females. All patients had a known diagnosis of hypertension and 6% also reported having diabetes. Only two (2%) of the included patients reported being active smokers. See Table 2 for a complete overview of the baseline demographic data.

Blood pressure control, antihypertensive drug use, and consultations

At baseline, 0% of the included patients had a BP < 140/90 mmHg as measured using the SPRINT standardized in-office BP measurements (18). This increased to 72% after six months (P < 0.0001) (Table 3). Absolute BP values improved from 157 mmHg (SD 18) systolic to 128 mmHg (SD 12), and 97 mmHg (SD 8) to 82 mmHg (SD 12) after six months (p < 0.0001 for both systolic and diastolic BPs). The median number of antihypertensives used increased from 2 (IQR 2–3) to 3 (IQR 2–3) during the study period and the median number of physical consultations (excluding study visits) was 1 (IQR 0–2).

Drug adherence

MARS-5 questionnaires were completed by 99% and 88% of the study participants at baseline and six months, respectively. Median MARS-5 scores at baseline were 25 (IQR 25–25) and remained 25 (IQR 25–25) at T = 6 month follow-up.

Patient- and healthcare provider satisfaction

Both TUQ and MAUQ questionnaires were completed by 89% of the study participants. Overall patient satisfaction scores were high with a median MAUQ of 7 (IQR 6.97–7) and a median TUQ score of 6.95 (IQR 6.86–7). Four healthcare providers completed the MAUQ questionnaire for healthcare providers and satisfaction was high with a median MAUQ of 7 (IQR 6.05–7).

Telemonitoring data

See Table 4 for a complete overview of the telemonitoring data. Telemonitoring data were extracted after the final patient completed the T = 6 month follow-up. Therefore, the total average duration of the telemonitoring program was 276 days (9.2 months). Each individual patient transmitted an average of 114 BP measurements during the study period. These measurements triggered a total of 1677 alerts, of which the majority were simple alerts (930, 55%). The HBPT platform automatically processed 35% (579) of all alerts and 65% were processed manually (1098). The total time investment based on the average alert processing for each individual patient was 51.7 minutes over 9.2 months.

Exploring the effects of HBPT

We believe the positive effects of HBPT on hypertension control rates to be multi-factorial.

HBPT is a complex and multidimensional intervention and the combination of measurements, remote monitoring, and digital lifestyle interventions can all contribute individually to improved clinical outcomes (8, 9). Given the small size and design of our study, we were unable to specifically point towards a specific HBPT effect. As existing studies (8, 9) have shown, the identified positive effect on BP control is most likely the result of a combination of increased healthcare provider disease insight and awareness, reduced physician inertia (26), increased patient disease insight, increased drug adherence, and lifestyle modification (22). This was also partly underscored by an increased number of antihypertensive drugs, which could indicate that the HBPT program addressed physician inertia, which is known to be an important cause of ongoing uncontrolled hypertension (26). These findings are consistent with other studies that demonstrated HBPT to be associated with improved systolic and diastolic blood pressure, an increase in the number of antihypertensive medications prescribed, but with unchanged drug adherence and clinic visits as compared to usual care (11, 27).

Feasibility of HBPT in underserved areas

Implementing a digital healthcare system like HBPT in an underserved area appeared to be feasible based on our study findings and could thereby improve healthcare access, which is in line with similar results from HBPT studies in underserved areas outside of the African continent (10). This conclusion was further supported by a low drop-out rate and fewer patients experiencing connectivity and technology-related issues during the study period. The involved healthcare providers also did not experience any implementation or technical issues with data transmission or the telemonitoring process itself. Additionally, the overall HBPT set-up and telemonitoring infrastructure could be further improved: for example, with the use of a dedicated remote patient monitoring center (6). Further automation of the alerts—for example, by using artificial intelligence—would also allow to further scale-up the current system. Using multiple E-nurses in a remote patient monitoring center would also improve the vulnerability of the required human intervention in managing the alerts, as we only used a single E-nurse in this study. On the other hand, we did experience a relatively low adherence to the measurement schedule, which might negatively affect long-term sustainability. The introduction of specific measures like more flexible and shorter measurement schedules and the introduction of an element of gamification with rewards could help improve this.

Study limitations

The small sample size and design need to be considered when interpreting results of this study. Specifically, the lack of a control group affects the interpretation of the clinical effect of the HBPT intervention as compared to usual care without HBPT. Additionally, to estimate drug adherence, we used a self-reporting adherence tool (MARS-5). Self-reports tend to overestimate adherence compared to more objective assessment methods like measuring antihypertensive drug concentrations in blood or urine samples. These advanced measurement methods, however, were not available given the study setting.

Additionally, we did not incorporate any cost-effective or cost-utility analysis. Patients were provided with a BP machine and the HBPT app was provided free of cost. True long-term adherence and feasibility might depend on the availability of the BP machines, the funding of the HBPT application, and the continuous availability of mobile internet to transmit BP values. Also, we used a cut-off of 140/90 mmHg for all studies irrespective of cardiovascular risk, which might be inappropriate for certain patients. Further studies should therefore include a specific telemonitoring program with a 135/85 mmHg target where applicable. Finally, the adherence to the measurement schedule was low, which could have significantly impacted the effects of the HBPT intervention on multiple outcome measures.