Background Patients with high blood pressure (hypertension) in the community frequently fail to meet treatment goals: a condition labelled as ‘uncontrolled’ hypertension. The optimal way to organise and deliver care to hypertensive patients has not been clearly identified.
Aim To determine the effectiveness of interventions to improve control of blood pressure in patients with hypertension.
Design of study Systematic review of randomised controlled trials.
Setting Primary and ambulatory care.
Method Interventions were categorised as following: self-monitoring; educational interventions directed to the patient; educational interventions directed to the health professional; health professional- (nurse or pharmacist) led care; organisational interventions that aimed to improve the delivery of care; and appointment reminder systems. Outcomes assessed were mean systolic and diastolic blood pressure, control of blood pressure and proportion of patients followed up at clinic.
Results Seventy-two RCTs met the inclusion criteria. The trials showed a wide variety of methodological quality. Self-monitoring was associated with net reductions in systolic blood pressure (weighted mean difference [WMD] −2.5mmHg, 95%CI = −3.7 to −1.3 mmHg) and diastolic blood pressure (WMD −1.8mmHg, 95%CI = −2.4 to −1.2 mmHg). An organised system of regular review allied to vigorous antihypertensive drug therapy was shown to reduce blood pressure and all-cause mortality in a single large randomised controlled trial.
Conclusion Antihypertensive drug therapy should be implemented by means of a vigorous stepped care approach when patients do not reach target blood pressure levels. Self-monitoring is a useful adjunct to care while reminder systems and nurse/pharmacist -led care require further evaluation.
Hypertension is largely managed in primary care and is an important public health problem in terms of associated stroke and cardiovascular events. It is mostly of unknown aetiology, easy to diagnose, and readily preventable by blood pressure reduction. Extensive epidemiological data have strengthened the well-recognised relationship between blood pressure and risk of cardiovascular disease, and have confirmed the importance of systolic blood pressure as a determinant of risk.1 However, blood pressure goals are achieved in only 25–40% of the patients who take antihypertensive drug treatment,2,3 which is something that has remained unchanged for the last 40 years.4
Use of self-monitoring of blood pressure by patients and professionals has gained popularity and is now recommended in particular patients in certain national and international guidelines; a recent meta-analysis of randomised trials on the subject did suggested a benefit in terms of mean blood pressure and blood pressure control.5 This systematic review aims to update and build upon previous reviews,6,7 by summarising the evidence from randomised controlled trials (RCTs) that evaluate non-pharmacological interventions to improve the management of hypertension in primary care.
How this fits in
The majority of patients with hypertension who take blood pressure medication fail to reach treatment goals. There is a strong evidence base for the benefits of antihypertensive drug therapy, but there is a lack of clarity on how to organise and deliver care for patients with hypertension in the community. This systematic review of 72 randomised controlled trials shows that self-monitoring leads to a significant decline in systolic and diastolic blood pressure and, as such, may be a useful adjunct to care and is likely to lead to a reduction in cardiovascular events. An organised system of regular review allied to vigorous antihypertensive drug therapy was shown to reduce blood pressure and allcause mortality. Nurse- or pharmacist-led care and appointment-reminder systems may be a promising way of improving blood pressure control, but require further evaluation.
Searching and study selection
Studies of patients aged >18 years with essential hypertension in an ambulatory setting were included. The interventions comprised all those that aimed to improve blood pressure control by non-pharmacological means and were classified as:
educational interventions directed to the patient;
educational interventions directed to the health professional;
nurse- or pharmacist-led care;
organisational interventions that aimed to improve the delivery of care; or
The outcomes assessed were mean systolic and diastolic blood pressure, control of blood pressure, and the proportion of patients followed up at clinic.
Original RCTs were identified by an all-language search in February 2008 of all articles (any year) in the Cochrane Controlled Trials Register and Medline (search strategy shown in Appendix 1); articles dated from January 1980 were searched on Embase. Included studies had to be RCTs with a contemporaneous control group, where patient care in the intervention group(s) was compared with either no intervention or usual care.
Data extraction and analysis
Two of the authors assessed lists of citations and abstracts independently. Differences were resolved by discussion and final adjudication was performed by an additional two authors. Reprints of all potentially relevant citations were obtained and data were independently extracted in duplicate using a structured data-collection form. Study quality was assessed by collection of data on inclusion and exclusion criteria; randomisation procedure; allocation concealment; blinding of participants, providers of care, and outcome assessors; and losses to follow-up.8
The effects on blood pressure outcomes of the six pre-defined intervention categories outlined above were examined. For the outcomes of mean systolic and diastolic blood pressure, pressure differences from baseline to final follow-up in the intervention and control groups were compared and pooled using the weighted mean difference approach.7 For the outcomes of blood pressure control and clinic attendance at follow-up, statistical and clinical significance was evaluated by means of estimating odds ratios (ORs) with 95% confidence intervals (CIs). Individual study definitions of control of blood pressure and attendance at clinic were used. For both continuous and categorical outcomes, the meta-analyses for heterogeneity were checked by visual inspection and by Cochran's C test. Pooled ORs and their 95% CIs were calculated with The Cochrane Collaboration RevMan software (version 5.02).
Trial flow, study characteristics, and quality assessment
The flow of studies through the stages of the systematic review is shown in Figure 1. A total of 72 trials were included in this systematic review (Characteristics of included randomized controlled trials are described in Appendix 2). The reported methodological quality of included studies was generally poor to moderate. The randomisation process was described in 30 (42%) of the 72 trials included, while only 14 (19%) had adequate allocation concealment. In 15 studies (21%), the outcome assessors were blind to the treatment allocation and losses to follow-up of 20% or more occurred in 18 (25%) of studies.
The impact of interventions is summarised in Table 1 (full data available from authors). There was substantial heterogeneity for several interventions and outcomes. In these situations, pooled data are not reported but the range of results from individual RCTs are presented.
With regard to self-monitoring (n = 18 RCTs), pooled data from 12 RCTs that reported on differences in mean systolic blood pressure8–19 showed that self-monitoring was associated with a significant reduction of −2.5 mmHg (95% CI = −3.7 to −1.3 mmHg). Pooled data from 14 RCTs on difference of mean diastolic blood pressure,9–22 showed that self-monitoring was associated with a reduction of −1.8 mmHg (95% CI = −2.4 to −1.2 mmHg). In the six RCTs that reported on control of blood pressure,10,12,18,22–24 there was no significant improvement in blood pressure control seen (OR 1.0, 95% CI = 0.8 to 1.2).
Educational interventions directed to the patient involved 20 RCTs. Eleven RCTs reported mean difference systolic blood pressure,25–35 13 RCTs reported mean difference diastolic blood pressure,25–38 and seven reported blood pressure control.22,23,30,32,33,39,40 For mean difference in systolic blood pressure and diastolic blood pressure outcomes, pooling of results from individual RCTs produced substantial heterogeneity, so pooled mean differences are not valid. The reported mean difference in systolic blood pressure ranged from −16 mmHg to 1 mmHg, and from −9 mmHg to 7 mmHg for mean difference in diastolic blood pressure. In terms of blood pressure control, there was a trend towards improved blood pressure control and this was significant (OR 0.83; 95% CI = 0.75 to 0.91). Educational interventions directed towards the physician (n = 10 RCTs)30,41–49 were not associated with a significant decrease in mean systolic blood pressure or diastolic blood pressure; control of blood pressure produced heterogeneous results (OR ranged from 0.8 to 1.1).
Nurse- or pharmacist-led care
For nurse- or pharmacist-led care (n = 12 RCTs),50–61 the majority of RCTs were associated with improved blood pressure control. However, for all three outcomes, the pooling of results from individual RCTs produced substantial heterogeneity, so pooled mean differences may not be valid. The range of mean difference was from −13 mmHg to 0 mmHg in mean systolic blood pressure (n = 10 RCTs) and from −8 mmHg to 0 mmHg for diastolic blood pressure (n = 12 RCTs); control of blood pressure (n = six RCTs) produced ORs ranging from 0.1 to 0.9.
Organisational interventions that aimed to improve the delivery of care were described in nine RCTs.40,41,60–69 For all three outcomes, pooling of results from individual RCTs produced heterogeneous results, so pooled mean differences may not be valid. Of note, the largest RCT, the Hypertension Detection and Follow–Up Program (HDFP),64 produced substantial reductions in systolic blood pressure and diastolic blood pressure across the three groups (weighted mean difference −8.2/−4.2 mmHg, −11.7/−6.5 mmHg, −10.6/−7.6 mmHg for the three strata of entry blood pressure).
For appointment-reminder systems (n = 8 RCTs),70–77 the pooled results – although favouring the intervention for follow-up of patients (OR of being lost to follow-up 0.4, 95% CI = 0.3 to 0.5) – are heterogeneous because of the single outlying RCT, and the pooled results should be treated with caution. Pooled data from two small RCTs – one a three-armed study of telephone reminder, mailed reminder, and usual care,76 and the other a parallel study of SMS reminder versus usual care77 – gave heterogeneous results in terms of systolic and diastolic blood pressure, but did show a significant improvement in blood pressure control, OR 0.5 (95% CI = 0.4 to 0.7).
In this systematic review, self-monitoring was associated with a significant decline in systolic blood pressure (−2.5 mmHg) and diastolic blood pressure (−1.8 mmHg). Although this blood pressure reduction does not appear substantial in clinical terms, it would, nonetheless, appear to be a useful adjunct to care and is likely to lead to a reduction in mortality and cardiovascular events. This appears to be confirmed in the HDFP study65,66 where an organised system of regular review allied to vigorous antihypertensive drug therapy was shown to reduce blood pressure as well as all-cause mortality.
At 5-year follow-up, the reductions in blood pressure (∼10 mmHg for systolic blood pressure and 5 mmHg for diastolic blood pressure) seen in this study were associated with a significant reduction in all-cause mortality (6.4% versus 7.8%, absolute risk reduction = 1.4%, numbers needed to treat = 71). Nurse- or pharmacist-led care and appointment-reminder systems may be a promising way of improving blood pressure control, but require further evaluation.
A previous meta-analysis of self-monitoring produced similar findings to the current study of modest, but potentially important, benefit in systolic and diastolic blood pressure.5 This is important in light of the fact that self-monitoring is now practised by up to two-thirds of the population that has hypertension in the US and Europe.78
There are also other elements identified from this review that appear to be associated with improved blood pressure control and are consistent with findings from observational studies and previous systematic reviews. A more recent observational study showed that antihypertensive drug therapy was initiated or changed in only 38% of episodes of care, despite documented uncontrolled hypertension for at least 6 months.79 Lack of practice organisation is associated with a failure to achieve treatment surrogate goals in hypertension, diabetes, and secondary prevention of coronary heart disease.80
This review had several limitations. Several RCTs included patients with hypertension who were treated and untreated and had differential rates of antihypertensive drug prescribing.8,18,47,76 Many RCTs contained multifaceted interventions that did not fit into a single intervention category.40,51,67 Consequently, it has been difficult to attribute how far single elements that make up complex interventions exert their independent effect on blood pressure control. Finally, several of the RCTs did not make any recommendations about the need for adjustment of target blood pressure readings when self-monitoring was the intervention being assessed, nor did they appear to anticipate lower blood pressure readings in the self-monitoring group.16,21,22,24 This may have attenuated the impact of self-monitoring on blood pressure control because of failure to intensify treatment. Self-monitoring of blood pressure by patients and blood pressure management by allied healthcare professionals both require further development and evaluation in larger RCTs and prospective studies, including cardiovascular outcomes.
This systematic review does, however, confirm that the most effective way to manage hypertension in the community is through a structured approach combining systematic identification and follow-up, which will include patient self-monitoring allied with appropriate treatment with antihypertensive medications.
We are very grateful to Margaret Burke (Cochrane Heart Group) for help with searching and to Shah Ibrahim who was an author on the original review. Our thanks also to Alison Blenkinsopp, Barry Carter, Sandy Logan, Frank Sullivan, Hayden Bosworth, Brian Haynes, David Jewell, Jim Krieger, Richard McManus, Steven Ornstein, Mike Phelan, Mary Rogers, Lin Song, Kelly Zarnke and Peter Whincup concerning clarification about individual RCTs and providing additional data. Thanks to Craig Ramsay for advice concerning factorial trials. We are grateful to Curt Furberg for facilitating contact with the investigators of US-based studies. Our particular thanks to Charlie Ford for information regarding the Hypertension Detection and Follow-Up Program (HDFP) study. Lastly, we are grateful to Debbie Farrell for administrative support.
This study was supported by a Cochrane Fellowship awarded to the lead author by the Health Research Board of Ireland through competitive peer review.
The authors have stated that there are none.
Discuss this article
Contribute and read comments about this article on the Discussion Forum: http://www.rcgp.org.uk/bjgp-discuss
- Received March 16, 2010.
- Revision received April 14, 2010.
- Accepted May 4, 2010.
- © British Journal of General Practice, 2010.