Key points from the evidence

1. We identified six systematic reviews (all with meta-analyses), one network meta-analysis and one quasi experimental study on the impact of HBPM on BP control.

2. Six systematic reviews reported statistically significant reductions in BP in HBPM interventions compared with routine care in people with diagnosed hypertension.

• Five systematic reviews reported that for the intervention group (which included HBPM with and without co-interventions such as telemonitoring) compared with routine care, the mean difference (MD) in the reduction in diastolic BP (DBP) ranged from ‑0.10 to ‑4mm Hg and for systolic BP (SBP) the MD ranged from ‑0.21 to ‑8.09 mm Hg. One of the systematic reviews (n=21,053) assessed HBPM interventions that were similar to the Connect Me BP programme in Scotland. The systematic review found DBP and SBP were significantly reduced in the HBPM group, compared with routine care.
• One systematic review with meta-analysis of four studies (n=1,741) reported statistically significant reductions in SBP in the telemonitoring group compared with routine care at 6 and 12 months follow up.

3. A network meta-analysis of electronic health interventions (eHealth), defined in the study as mobile health technologies and telemedicine such as BP telemonitoring or HBPM, reported reductions in MD for DBP and SBP compared with routine care across the multiple interventions included (n=15,308). Interventions that included more than two types of eHealth interventions (phone calls, BP telemonitoring, websites or smartphone applications) were statistically significantly more effective than routine care in reducing DBP. All types of eHealth interventions assessed (covering all BP telemonitoring or HBPM interventions) were statistically significantly more effective in reducing SBP compared with routine care.

4. One quasi experimental implementation study (n=7,429) explored the feasibility and impact of implementing a hypertension telemonitoring system in primary care (Scale-Up BP, a precursor to Connect Me BP), in people with diagnosed hypertension in NHS Lothian, Scotland. In the study, the decreases noted in SBP were greater in people whose hypertension was initially uncontrolled in the study. In people whose SBP was ≥135 mm Hg, the mean reduction was 15.06 mm Hg (interquartile ratio [IQR] 6 to 23), but for people whose BP was <135 mm Hg, the mean reduction was -1.18 mm Hg (IQR ‑7 to 7).

5. The extent to which the reported results for BP control are clinically significant is unclear. There is no agreement within the literature on what a clinically meaningful difference is with opinions varying from 2 to 10 mmHg difference for SBP and 2 to 5 mm Hg difference for DBP.

6. Evidence for the impact of HBPM on prediction of morbidity and mortality risk comes from three systematic reviews, one meta-analysis, one prospective cohort study and one unpublished cohort study.

7. A systematic review (n=4,830) reported that HBPM and ambulatory BPM (ABPM) had similar hazard ratios (HR) for predicting CVD events for SBP and DBP. Both HBPM and ABPM were better predictors of CVD events than measurements of BP taken in a healthcare setting.

8. A systematic review (n=21,053) that included similar interventions to the Connect Me BP programme in Scotland and a prospective 5 year follow up (n=45) reported no statistically significant differences in morbidity or mortality risk between the HBPM intervention (with or without co-interventions) and routine care groups.

9. A systematic review (n=10,443) reported that for SBP, higher BP as measured by HBPM was not a predictor of CVD events and mortality, yet for DBP, HBPM was a predictor of CVD events and mortality in one of two cohorts included in the study (mixed population). The study authors concluded that there was no evidence to suggest that either HBPM or ABPM was better for predicting CVD and mortality risk.

10 There are limited data on the direct effects of HBPM interventions on CVD events. An unpublished cohort study reported that users of the Connect Me BP service were statistically significantly less likely to experience acute coronary syndrome, stroke or uncontrolled heart failure, compared with patients with hypertension in the routine care group (BP telemonitoring n=5,249; routine care n=163,848).

11. Medication adherence is recognised as a factor underlying inadequate control of BP.22 Evidence for the impact of HBPM on medication adherence comes from four systematic reviews and one network meta-analysis. Evidence for the impact of HBPM on medication initiation comes from two systematic reviews.

12. One systematic review (n=1,495) reported statistically significant improvements in medication adherence through use of mobile health applications (defined as mobile devices used to monitor BP, as well as other wireless technologies), but this was not specific to HBPM.

13. One network-meta-analysis (n=15,308) of eHealth interventions reported that the use of smartphone applications alongside BPM (three studies) statistically significantly improved medication adherence (risk ratio [RR]=0.55, 95% CI 0.33 to 0.93, I2=82%). No other intervention was statistically significant (more than two types of intervention, phone calls, BP telemonitoring, emails, short messaging service (SMS) or website).

14 Three systematic reviews reported mixed findings:

• One systematic review (n=1,550) included one study (BP telemonitoring, with pharmacist case management and telephone support) that showed improved adherence to medication at 6 months that was statistically significant (+10.7% in the BP telemonitoring group compared with -5.9% in the usual care group, p<0.05), but not at 12 or 18 months follow up. No statistically significant improvements in adherence were reported in the four other included studies that reported adherence outcomes.
• One systematic review (n=8,933) included 24 studies, of which 12 studies reported on medication adherence outcomes. Seven of the 12 studies reported statistically significant improvements in medication adherence in the intervention groups (that included HBPM with and without co-interventions). Five included studies did not report statistically significant effects on medication adherence.
• One systematic review (n=not reported) of self-management interventions for BP in people with hypertension (including HBPM with and without co-interventions) reported that in nine out of 14 interventions, practical support with adherence had no or a small statistical effect (n for the nine studies and any associated subgroup analyses not reported).

15. For the impact of HBPM (and associated co-interventions) on medications initiation, two systematic reviews (with meta-analyses) reported increases in number of medications prescribed.

• One systematic review (n=1,550) reported an increase in the number of BP medications prescribed in a pharmacist-led HBPM intervention group, compared with routine care, in three of the six included studies. The three remaining studies did not report a statistically significant difference in the number of medications prescribed between study groups.
• In one systematic review (n=21,053) that included interventions similar to the Connect Me BP programme in Scotland, eleven (n=3,987) out of 65 included studies assessed the impact of HBPM on medication initiation. All studies reported an increase in the number of anti-hypertensive medicines prescribed. The number of medications prescribed (unclear whether per person) increased by 0.17 in the HBPM compared with routine care group. As the number of medications increased, so did BP (SBP: intercept of regression slope= -1.72 mm Hg, p=0.0085; DBP: intercept of regression slope=-1.40 mm Hg, p=0.012).

16. Evidence for the impact of HBPM on the length of time spent in healthcare appointments comes from one quasi experimental study of the implementation of hypertension HBPM telemonitoring at eight primary care practices in NHS Lothian (n=7,429). The HBPM telemonitoring intervention used (Scale-Up BP) was a precursor programme to Connect Me BP in Scotland. Compared with the previous year (pre-implementation of HBPM telemonitoring), time spent in all clinical appointments was reduced statistically significantly in the HBPM telemonitoring group compared with non-telemonitoring (adjusted MD 16.1 minutes, 95% CI 0.1 to 32.1 minutes, p=0.048). A reduction in time spent in face-to-face appointments in the HBPM telemonitoring group compared with non-telemonitoring was not statistically significant (adjusted MD 12.7 minutes, 95% CI -0.5 to 25.9 minutes, p=0.059).

17. Evidence for safety outcomes comes from two systematic reviews (with meta-analysis) and two randomised controlled trials (RCTs). One systematic review with meta-analysis and two RCTS included measurement of HBPM that was similar to the Connect Me BP programme in Scotland.

18. No statistically significant differences (or narrative, as described by study authors when significance was not assessed) in the number or risk of adverse events between the HBPM intervention (with or without co-interventions) and routine care were reported in the evidence reviewed.

19. Evidence for patient perspectives comes from one systematic review and six primary studies (four qualitative, one quasi experimental, one national evaluation from England). Six of the seven studies included measurement of HBPM that was similar to or the same as the Connect Me BP programme in Scotland.

20. Patient perspectives cover a broad range of views. No weighting of importance is given to the benefits or concerns summarised from the evidence.

• Patients with hypertension felt that HBPM supported greater autonomy and motivation to look after their own health and provided them with reassurance because measurement of BP at home was considered to be more accurate than BP measurement taken in a healthcare setting. Patients with hypertension reported feeling calmer and that measuring their BP at home reduced their anxiety associated with clinic measurements. Conversely, measuring BP at home may increase feelings of confusion and stress when BP readings are high, contribute to uncertainty around BP targets, and create a concern about inaccurate readings.
• Patients with hypertension felt that measurement of BP at home was easy, enabling quick responses to their BP readings and saving time for clinicians. Concerns raised included the burden of measurement and feelings of being overwhelmed, costs of potentially have to buy a monitor, and health literacy-related concerns such as a lack of technical skills in using the equipment and uncertainty over how to interpret results.
• Patients being monitored for hypertension reported feeling more in control of their condition and, as a result, were better able to discuss their results with their GP in an informed way. Clinicians also felt that HBPM supported the relationship between patient and clinician, linked to higher levels of trust. Feelings of confusion and stress were triggered in patients with diagnosed hypertension when clinicians did not provide feedback or were not supportive.

21. The economic literature was limited but indicated that compared with routine care in managing hypertension, the use of remote BPM monitoring integrated with professional support may improve clinical outcomes, enhance patient empowerment and provided reduced healthcare expenditure.

22. A de novo economic evaluation was developed to assess the costs and benefits of the Connect Me BP programme versus routine care. We modelled the proportion of people with controlled or uncontrolled BP in each arm and used this to inform healthcare utilisation and estimated CVD outcomes in each model arm. The analysis was informed by Scottish patient data from the Scale-Up BP evaluation, ensuring relevance to the Scottish healthcare context:

• the model demonstrated that it is a dominant strategy compared to routine care, meaning it is estimated to have lower costs and better health outcomes for patients compared with routine care
• a scenario analysis excluding cardiovascular events showed that the intervention remained dominant, supporting the robustness of the base case findings.