Intended for healthcare professionals

  1. Research
  2. Learning...
  3. Learning implementation of a guideline based decision support system to improve hypertension treatment in primary care in China: pragmatic cluster randomised controlled trial
CCBYNC Open access
Research

Learning implementation of a guideline based decision support system to improve hypertension treatment in primary care in China: pragmatic cluster randomised controlled trial

BMJ 2024; 386 doi: https://doi.org/10.1136/bmj-2023-079143 (Published 23 July 2024) Cite this as: BMJ 2024;386:e079143

Linked Editorial

Improving hypertension management in primary care
  1. Jiali Song, cardiologist1,
  2. Xiuling Wang, cardiologist1,
  3. Bin Wang, postdoctoral fellow1,
  4. Yilan Ge, cardiologist1,
  5. Lei Bi, cardiologist1,
  6. Fuyu Jing, cardiologist1,
  7. Huijun Jin, cardiologist1,
  8. Teng Li, cardiologist1,
  9. Bo Gu, senior research fellow2,
  10. Lili Wang, cardiologist1,
  11. Jun Hao, statistician3,
  12. Yanyan Zhao, intermediate statistician3,
  13. Jiamin Liu, associate professor1,
  14. Haibo Zhang, associate professor1,
  15. Xi Li, professor1 2,
  16. Jing Li, professor1,
  17. Wenjun Ma, professor4,
  18. Jiguang Wang, professor5,
  19. Sharon-Lise T Normand, professor6 7,
  20. Jeph Herrin, associate professor8,
  21. Jane Armitage, professor9 10,
  22. Harlan M Krumholz, professor8 11 12,
  23. Xin Zheng, professor1 2 13 14
  24. on behalf of the LIGHT Collaborative Group
  1. 1National Clinical Research Centre for Cardiovascular Diseases, State Key Laboratory of Cardiovascular Disease, Chinese Academy of Medical Sciences and Peking Union Medical College, Fuwai Hospital, National Centre for Cardiovascular Diseases, Beijing, China
  2. 2National Clinical Research Centre for Cardiovascular Diseases, Shenzhen, Fuwai Shenzhen Hospital, Chinese Academy of Medical Sciences, Shenzhen, China
  3. 3Medical Research and Biometrics Centre, State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Centre for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
  4. 4Hypertension Centre, State Key Laboratory of Cardiovascular Disease, Chinese Academy of Medical Sciences and Peking Union Medical College, Fuwai Hospital, Beijing, China
  5. 5The Shanghai Institute of Hypertension, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China
  6. 6Department of Health Care Policy, Harvard Medical School, Boston, MA, USA
  7. 7Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, MA, USA
  8. 8Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT, USA
  9. 9Clinical Trial Service Unit and Epidemiological Studies Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK
  10. 10MRC Population Health Research Unit, Nuffield Department of Population Health, University of Oxford, Oxford, UK
  11. 11Centre for Outcomes Research and Evaluation, Yale-New Haven Hospital, New Haven, CT, USA
  12. 12Department of Health Policy and Management, Yale School of Public Health, New Haven, CT, USA
  13. 13Coronary Artery Disease Ward 2, Fuwai Shenzhen Hospital, Chinese Academy of Medical Sciences, Shenzhen, China
  14. 14Clinical Trial Centre, Fuwai Shenzhen Hospital, Chinese Academy of Medical Sciences, Shenzhen, China
  1. Correspondence to: X Zheng zhengxin{at}fuwai.com
  • Accepted 10 June 2024

Abstract

Objective To evaluate the effectiveness of a clinical decision support system (CDSS) in improving the use of guideline accordant antihypertensive treatment in primary care settings in China.

Design Pragmatic, open label, cluster randomised trial.

Setting 94 primary care practices in four urban regions of China between August 2019 and July 2022: Luoyang (central China), Jining (east China), and Shenzhen (south China, including two regions).

Participants 94 practices were randomised (46 to CDSS, 48 to usual care). 12 137 participants with hypertension who used up to two classes of antihypertensives and had a systolic blood pressure <180 mm Hg and diastolic blood pressure <110 mm Hg were included.

Interventions Primary care practices were randomised to use an electronic health record based CDSS, which recommended a specific guideline accordant regimen for initiation, titration, or switching of antihypertensive (the intervention), or to use the same electronic health record without CDSS and provide treatment as usual (control).

Main outcome measures The primary outcome was the proportion of hypertension related visits during which an appropriate (guideline accordant) treatment was provided. Secondary outcomes were the average reduction in systolic blood pressure and proportion of participants with controlled blood pressure (<140/90 mm Hg) at the last scheduled follow-up. Safety outcomes were patient reported antihypertensive treatment related events, including syncope, injurious fall, symptomatic hypotension or systolic blood pressure <90 mm Hg, and bradycardia.

Results 5755 participants with 23 113 visits in the intervention group and 6382 participants with 27 868 visits in the control group were included. Mean age was 61 (standard deviation 13) years and 42.5% were women. During a median 11.6 months of follow-up, the proportion of visits at which appropriate treatment was given was higher in the intervention group than in the control group (77.8% (17 975/23 113) v 62.2% (17 328/27 868); absolute difference 15.2 percentage points (95% confidence interval (CI) 10.7 to 19.8); P<0.001; odds ratio 2.17 (95% CI 1.75 to 2.69); P<0.001). Compared with participants in the control group, those in the intervention group had a 1.6 mm Hg (95% CI −2.7 to −0.5) greater reduction in systolic blood pressure (−1.5 mm Hg v 0.3 mm Hg; P=0.006) and a 4.4 percentage point (95% CI −0.7 to 9.5) improvement in blood pressure control rate (69.0% (3415/4952) v 64.6% (3778/5845); P=0.07). Patient reported antihypertensive treatment related adverse effects were rare in both groups.

Conclusions Use of a CDSS in primary care in China improved the provision of guideline accordant antihypertensive treatment and led to a modest reduction in blood pressure. The CDSS offers a promising approach to delivering better care for hypertension, both safely and efficiently.

Trial registration ClinicalTrials.gov NCT03636334.

Introduction

Hypertension is the leading modifiable risk factor for cardiovascular disease and death, affecting an estimated 1.3 billion people worldwide in 2019.123 Based on the data from a nationwide survey in 2012-15, around 245 million Chinese adults had hypertension, and the treatment and control rates were low at 41% and 15%, respectively.4 The use of evidence based antihypertensive treatments, a key strategy for lowering blood pressure and reducing cardiovascular risk,5678 is suboptimal in China. Inappropriate prescribing behaviour and clinical inertia are common.91011 Among patients treated for hypertension, 68% were taking only one drug,4 and about 8% of drugs used were not recommended by guidelines.11

Improving the performance of primary care doctors—the mainstay of the hypertension care workforce in China—is a key first step towards improving blood pressure control and patient outcomes.9 However, China is facing multiple challenges to achieve this goal, given the constrained and unevenly distributed resources.91012 In 2021, primary care doctors provided half of the outpatient care (4.3 billion visits) in China.13 Despite such heavy workloads, they often have inadequate training. In 2021, 56% of doctors in community health centres had an education level below medical college.14 Continuing education and financial incentives for these doctors are also insufficient.10 An affordable and scalable strategy is needed to improve blood pressure management.

A clinical decision support system (CDSS), characterised by integrating patient data and guideline recommendations at the point of care, has the potential to improve the performance of primary care doctors.915 Despite its promise of delivering scalable and sustainable care, the CDSS has been rarely adopted in China.10 Previous studies in resource constrained countries such as India have provided evidence of the effectiveness of CDSS for the management of hypertension in primary care settings.1617 These studies tested CDSS’s effect on blood pressure change and showed inconsistent results.1617 Few studies have been primarily designed to systematically and quantitatively assess the effect of CDSS on guideline accordant treatment prescribing.18

We developed a CDSS that could generate a tailored antihypertensive regimen and tested whether the use of a CDSS could improve the provision of guideline accordant antihypertensive treatment compared with usual care in Chinese primary care settings. We also assessed the impact of the CDSS on blood pressure change and blood pressure control.

Methods

Trial design

The Learning Implementation of Guideline-based decision support system for Hypertension Treatment (LIGHT) trial was a pragmatic, cluster randomised trial conducted in 94 primary care practices in four urban regions of China: Luoyang (central China), Jining (east China), and Shenzhen (south China, including two regions). To ensure the wide feasibility of using electronic health record data and the CDSS, the trial was designed to recruit primary care practices in urban regions. Details of the trial rationale, design, and methods have been described previously.19 In each region, after screening during a three month baseline period, eligible primary care practices were randomised to receive either CDSS guided treatment or usual care. Eligible patients were enrolled during the first three months after site randomisation. They were asked to attend the clinic in the primary care practice at least every three months, as recommended by the National Basic Public Health Services Programme (see supplementary appendix section S2)20 and Chinese Guidelines for Hypertension Prevention and Management in Primary Care.21 At each visit, a 1-3 month supply of antihypertensive drugs could be dispensed.22 The duration of follow-up was nine months. Owing to the covid-19 pandemic in 2020, we extended the scheduled follow-up period in each region for about one month to four months according to the policies of local Centers for Disease Control and Prevention.

Recruitment and participants

Primary care practices were eligible for the trial if they stored at least one agent from each of four classes of antihypertensive drugs for hypertension treatment (angiotensin converting enzyme inhibitors or angiotensin receptor blockers, beta blockers, calcium channel blockers, diuretics); had computers and internet access to use the bespoke electronic health record for this trial; and had at least 100 local patients registered for hypertension management in the National Basic Public Health Services Programme.20

Patients were eligible for the trial if they had an established diagnosis of hypertension and were registered for hypertension management in the National Basic Public Health Services Programme20; had a systolic blood pressure <180 mm Hg and diastolic blood pressure <110 mm Hg; and were taking 0-2 classes of antihypertensive drugs. The diagnosis of hypertension was based on a systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg, or both, from three measurements on different days, or the patients were taking antihypertensive drugs according to the Chinese Guidelines for Hypertension Prevention and Management in Primary Care.21 Patients were excluded if they had a self-reported history of coronary artery disease, chronic kidney disease, or heart failure, or were intolerant to two or more classes of antihypertensive drugs (see supplementary appendix section S3).

Interventions

The CDSS was developed using the hypertension treatment guideline for primary care in China,21 which is generally consistent with international guidelines.2324 Based on the measured blood pressure, current use of antihypertensive drugs, specific clinical indications, contraindications, and patient reported drug adverse effects or intolerance, the CDSS would generate recommendations to escalate treatment until patients achieved blood pressure control; recommend guideline accordant regimen, including the class and dose of antihypertensive drugs; and optimise treatment for patients with specific indications (eg, diabetes), intolerance, or contraindication. To ensure feasibility and patient safety, the CDSS would not recommend a regimen for patients with complications such as coronary artery disease, chronic kidney disease, or heart failure.

Supplementary appendix section S4 depicts use of the CDSS. Before prescribing took place during each visit, the doctors had to click an icon to obtain CDSS recommendations before receiving a regimen. We encouraged doctors to discuss the CDSS recommended regimen with patients. The final prescription was up to the shared decision between doctors and patients. If doctors did not follow the recommendations, a pop-up alert reminded them to adjust the prescription. If doctors refused to follow the alert, the relevant reasons were collected. Additional alerts were triggered if prescriptions involved contraindicated drugs, under-dosage, or over-dosage. Doctors in the control group used the same electronic health record to collect data but did not receive recommendations or alerts. Prescribing decisions were based on their knowledge and experience.

Randomisation and blinding

Primary care practices were randomised to intervention group or control group (1:1 ratio) using a central computerised randomisation program. Randomisation was performed in four regions sequentially (see supplementary appendix section S5) and stratified by the proportion of hypertension related visits with guideline accordant treatment during the baseline period, and the characteristics of the primary care practice (including the hospital to which the practice was affiliated, the type of primary care practice, or district; see supplementary appendix section S6 for details of the stratifying factors for each region). Given the nature of the intervention, it was not possible to blind the practice allocation to the doctors. Participants were not informed of their allocation, as consent for the study was waived. To minimise the bias for outcome analyses, the independent statistician was fully blinded to practice allocation.

Effectiveness outcomes

The primary outcome was the proportion of hypertension related visits at which appropriate treatment was provided. Appropriate treatment was defined as a prescription in line with prespecified specifications (see supplementary appendix section S7). This specification was based on the current countrywide guideline21 and adapted to the treatment scenarios for this trial. Briefly, these included initiating and titrating antihypertensive treatments for patients with inadequate blood pressure control, switching to guideline accordant antihypertensive drugs, and refraining from prescribing contraindicated drugs. The appropriateness of the prescription was automatically assessed using a computerised algorithm. Secondary outcomes included the proportion of hypertension related visits with acceptable treatment, defined as either appropriate or non-appropriate but with acceptable reasons for failing to titrate antihypertensive treatment (acceptable reasons were self-reported and self-measured home blood pressure within the acceptable range, or possible antihypertensive treatment related events (syncope, injurious fall, hypotension, or bradycardia)); outcome measures with average change in systolic blood pressure from baseline to the last scheduled follow-up; and the proportion of participants with controlled blood pressure (<140/90 mm Hg) at the last scheduled follow-up. An exploratory outcome was the proportion of participants with vascular events (a composite of cardiac death, non-fatal stroke, and non-fatal myocardial infarction).

Safety outcome

The safety outcome was patient reported antihypertensive treatment related events of syncope, injurious fall, symptomatic hypotension or systolic blood pressure <90 mm Hg, or bradycardia (see supplementary appendix section S8).

Data collection

We provided a bespoke electronic health record for each primary care practice. Baseline characteristics of practices and participants were collected through questionnaires in this electronic health record. For sites not equipped to use electronic health records, we installed the bespoke electronic health record in their computers. For sites equipped to use local electronic health records, we embedded our electronic health record into the existing one as a module to improve workflow. At each visit, blood pressure was measured twice with an automated sphygmomanometer (Omron HBP-1300) after five minutes of rest with an interval of 1-2 minutes, and the mean of two measurements was recorded. In the electronic health record, doctors recorded the measured blood pressure, medical history, antihypertensive drug use, self-measured home blood pressure (if provided), patient reported treatment related adverse events (see supplementary appendix section S8), drug intolerance, drug adherence (see supplementary appendix section S9), and vascular events.

Sample size

We conducted a pragmatic randomised trial to assess the effectiveness of a CDSS in different settings. Owing to the lack of similar previous studies for reference, we estimated statistical power based on the number of potentially eligible sites in the Luoyang region initially. We assumed that at least 10 sites would be needed in the intervention group and 10 in the control group and that the baseline appropriate rate of treatment would be 55%. With a moderate intra-site correlation of 0.05 and a within patient correlation of 0.1, under the maximum type I error of two sided α=0.05 and statistical power of 90%, we determined that three hypertension related visits per patient would be needed for 50 patients at each site to detect an 18% absolute difference in the proportion of appropriate treatments (ie, 55% appropriate treatments in the control group, 73% in the intervention group). Subsequent enrolment and randomisation were to be carried out in four regions involving 94 participating sites. Under the same assumptions as above and maximum 25% loss to follow-up of patient at each visit, we determined that about 12 000 patients would provide at least 90% power to detect a difference of 4% in the appropriate treatment rate—an average effect of CDSS use reported in a previous study.25

Statistical analysis

All analyses were based on the intention-to-treat principle. After cluster randomisation of primary care practices, all the data of enrolled patients from each visit during the study period were included for analyses, regardless of whether the primary care practice or patient completed the study. We summarised the characteristics of the practices, doctors, and participants by study group. Standardised differences between the two groups were calculated by generalised mixed effects regression models, with sites as the random effect. A standardised difference >0.25 indicated imbalance.2627 The analysis unit for the appropriate or acceptable treatment rates was visit, and for the other outcomes was participant. We used generalised linear mixed effect regression models with a logit and identity link function for the binary and continuous response variables, respectively. In these models, we included both practice and participant as random intercepts for visit level analysis and included practice as random intercepts for participant level analysis. The estimates of the intervention effect were obtained by averaging over the random effects. The region, baseline rate of appropriate antihypertensive treatment (median or higher or less than median), and calendar time were included as fixed effects. No data were missing for covariates. The analysis of the primary outcome and secondary outcome of acceptable appropriate treatment rate was based on available data of visits. The analyses of average change in systolic blood pressure and the proportion of participants with controlled blood pressure were based on the last scheduled visit, and without imputation. As a sensitivity analysis, we conducted multiple imputation for the missing blood pressure data of the last scheduled follow-up. We also conducted an analysis of blood pressure control defined as a systolic blood pressure <130 mm Hg and a diastolic blood pressure <80 mm Hg. Subgroup analyses were performed by implementation region; baseline tertiles of site appropriate treatment rates; education level of doctors; and age, sex, education level, use of antihypertensive drugs, and baseline systolic blood pressure of participants. In addition, we conducted additional analyses for outcomes among the patients with baseline blood pressure ≥140/90 mm Hg.

We considered a P value <0.05 (two sided test) to be statistically significant for the primary outcome. We also used a significance level of 0.05 for other outcomes, but these findings should be interpreted with caution as the analyses were not statistically powered. Additional details on statistical analyses are provided in the statistical analysis plan and in supplementary appendix sections S10 and S11 and tables S2 and S3. SAS version 9.4 (SAS Institute) was used for all statistical analyses.

Patient and public involvement

No patients were involved in the design of the study or review of our manuscript. We were unable to involve patients and members of the public in this study owing to lack of funding and expertise in conducting patient and public involvement focus groups. Although no patients or members of the public were directly involved in this study, the clinical investigators’ clinical practice with patients informed the design and rationale of this study.

Results

Characteristics of practices and participants

A total of 94 primary care practices were randomised (46 to CDSS and 48 to usual care), of which two in the CDSS group withdrew after enrolment. During the baseline period, median appropriate treatment rates of primary care practices in the CDSS group and usual care group were 63.0% and 60.0%, respectively. Between August 2019 and 2021, 12 137 participants were enrolled in the trial: 5755 in the CDSS group (median 123 participants per cluster) and 6382 participants in the usual care group (median 135 participants per cluster). The median duration of follow-up was 11.6 months, during which a total of 23 113 visits (median 4.0 visits per participant) in the CDSS group and 27 868 visits (median 4.0 visits per participant) in the usual care group were included. Overall, 86.0% (4952/5755) of participants in the CDSS group and 91.6% (5845/6382) in the usual care group completed follow-up (fig 1, also see supplementary appendix table S4).

Fig 1
Fig 1

Enrolment, randomisation, and follow-up. CDSS=clinical decision support system

Baseline characteristics of primary care practices, doctors, and participants were well balanced between the two groups (table 1). The mean age of the doctors was 46 (standard deviation (SD) 12) years, 59.1% were women, 91.0% had an educational attainment of medical college level or higher, and all were licensed. The mean age of the participants was 61 (SD 13) years and 42.5% were women. The mean systolic blood pressure of the participants was 134.1 (SD 14.8) mm Hg, and 92.3% were using at least one class of antihypertensive drug.

Table 1

Baseline characteristics of primary care practices, doctors, and participants. Values are number (percentage) unless stated otherwise

View this table:

Primary outcome

The proportion of hypertension related visits with appropriate treatment was significantly higher in the CDSS group (77.8% v 62.2%; absolute difference 15.2 percentage points (95% confidence interval (CI) 10.7 to 19.8); P<0.001; odds ratio 2.17 (95% CI 1.75 to 2.69); P<0.001) compared with the usual care group (table 2, fig 2, also see supplementary appendix section S11). Subgroup analyses showed that the CDSS improved appropriate treatment across regions and subgroups (fig 3).

Table 2

Primary, secondary, and exploratory outcomes. Values are number with outcome/total number in group (percentage) unless stated otherwise

View this table:
Fig 2
Fig 2

Distribution of proportion of hypertension related visits with appropriate (guideline based) treatment

Fig 3
Fig 3

Subgroup analysis of proportion of hypertension related visits with appropriate (guideline based) treatment. CDSS=clinical decision support system; CI=confidence interval

Secondary and exploratory outcomes

The proportion of hypertension related visits with acceptable treatment was higher in the CDSS group (84.7% v 70.4%, absolute difference 12.6 percentage points (95% CI 8.1 to 17.2); P<0.001) than in the usual care group. Participants in the CDSS group had a 1.6 mm Hg (95% CI −2.7 to −0.5) greater reduction in systolic blood pressure (−1.5 (SD 16.2) mm Hg v 0.3 (SD 16.1) mm Hg; P=0.006) and a non-significant improvement in blood pressure control of 4.4 percentage points (95% CI −0.7 to 9.5) (69.0% v 64.6%; P=0.07) compared with those in the usual care group. When analysis was limited to those with baseline blood pressure ≥140/90 mm Hg, participants in the CDSS group had a 1.9 mm Hg (95% CI −3.6 to −0.3; P=0.02) greater reduction in systolic blood pressure and an improvement in blood pressure control of 5.0 percentage points (95% CI −1.0 to 11.1; P=0.08) compared with those in the usual care group (see supplementary appendix table S5). After multiple imputation for missing blood pressure data from the last scheduled follow-up, the use of CDSS significantly improved the rate of blood pressure control of 4.1 percentage points (95% CI 3.5 to 4.8; P<0.001) (see supplementary appendix table S6). In addition, when using a definition for blood pressure control as systolic blood pressure <130 mm Hg and diastolic blood pressure <80 mm Hg, the effectiveness of the CDSS on blood pressure control was consistent with the main analysis (see supplementary appendix table S7). The proportion of patients with good adherence to prescribed drugs in the CDSS group was lower than in the usual care group (3343/5706 (58.6%) v 4642/6291 (73.8%), standardised difference 0.33). CDSS use resulted in better blood pressure control among patients with good drug adherence, but not among those with poor drug adherence (see supplementary appendix table S8). The proportion of participants reporting vascular events between the two groups was similar (1.0% v 0.7%; absolute difference 0.1 percentage points (95% CI −0.3 to 0.5; P=0.43) (table 2).

Safety outcomes

The rates of patient reported antihypertensive treatment related adverse effects were low and similar between the two groups (see supplementary appendix table S9).

Discussion

In this large, cluster randomised trial conducted in 94 primary care practices in China, the use of a CDSS led to a 15 percentage point absolute increase in the proportion of hypertension related visits with guideline accordant prescribed treatment, and a modest reduction in systolic blood pressure of 1.6 mm Hg. Patient reported treatment related adverse events were rare. The CDSS appeared effective across regions and sites with varying baseline guideline accordant treatment rates, and in settings with different educational attainment of the primary care doctors.

Comparison with other studies

In this trial we quantified the effectiveness of CDSS for enhancing primary hypertension care using comprehensive guideline based metrics. The observed effect of the CDSS corresponded to an estimated 26% relative increase28 in guideline accordant treatment. Previous studies have assessed the effectiveness of CDSSs on a range of process outcomes for hypertension care, including drug prescribing, clinical tests completed, and number of clinic visits, and reported absolute 2 to 20 percentage point improvements in the corresponding process outcome,29303132 but few provided evidence on the extent of guideline based prescribing. A trial conducted in 14 US hospitals, assessed a CDSS designed to improve appropriate drug prescribing.18 The CDSS could recommend a pharmacologically appropriate drug class based on several patient characteristics, including age, race, and disease (ie, diabetes, coronary artery disease, and heart failure) but did not provide dosage to intensity treatment. The CDSS in the current study was not designed to guide treatment for coronary artery disease or heart failure for two reasons—firstly, because patients with these conditions are more likely to follow regimens provided by doctors from hospitals rather than primary care practices, and, secondly, because of concern about the complexity of an individualised and specific algorithm and patient safety.

This trial provided insights into the effectiveness of CDSS as a single factor intervention for improving the control of hypertension in primary care. Several previous CDSS studies focused on change in blood pressure, and the results were mixed,161732 partly because blood pressure could be affected by multiple factors. In the current study, the marginal effect of using a CDSS on blood pressure control should be interpreted in light of several points. Unlike other studies that enrolled patients with uncontrolled blood pressure,1617 this trial included two thirds of patients with blood pressure under control at baseline. Therefore, the overall effect of a CDSS on blood pressure could have been diluted. This point can be supported by the finding of the subgroup analysis, which showed a larger CDSS related reduction in systolic blood pressure among patients with uncontrolled blood pressure than those with controlled blood pressure at baseline. Moreover, the potential benefit of the CDSS on blood pressure control in this study may be offset by poor adherence of patients to prescribed drugs, as we found that CDSS use led to significant blood pressure control among patients with good adherence, but not in the overall population. As this trial was primarily aimed at enhancing doctors’ prescribing behaviour through CDSS use, it did not involve multicomponent interventions. In future research, integrating a patient engagement approach with CDSS use for enhancing shared decision making might improve patients’ adherence to treatment and their long term outcomes.3334 Furthermore, the benefit of some appropriate treatments, such as switching to angiotensin converting enzyme inhibitors or angiotensin receptor blockers for patients with diabetes and blood pressure under control, was more reflected by the improvement of long term outcomes rather than reduction in blood pressure. Notably, a 4.4 percentage point improvement in blood pressure control rate, as observed in this study, could translate into a considerable reduction of cardiovascular mortality and morbidity in larger populations.35

Strengths of this study

The overall design of this study was distinguished by several unique features, which were considered to improve implementation, acceptance, and generalisability of CDSS use in primary care settings. Firstly, the CDSS was integrated into the clinical workflow by being embedded into the electronic health record that collected the relevant information for usual hypertension care. Without interruption to routine clinical care, simple data collection minimised the burden of CDSS use. As primary care doctors need to click an icon only once, the CDSS could provide immediate, specific, and tailored recommendations at the point of care without unnecessary alerts. These features rendered the CDSS more acceptable to primary care doctors,15 even if they had little experience of using a CDSS.

Secondly, we included primary care practices, involving a large sample of hypertensive patients, across three urban cities with a broad range of baseline appropriate treatment rates and doctors with different levels of education. The large sample size enabled us to assess the effectiveness of CDSS across regions and various subgroups. We observed the effects of CDSS were different across regions, partly owing to the regional variations in appropriate treatment and blood pressure control rates at baseline. Nevertheless, substantial improvement in appropriate treatment was found in the overall population after accounting for heterogeneity in different sites from four regions, suggesting that CDSS had been well implemented and accepted in diverse clinical settings. Subgroup analyses indicated that CDSS could be generalised to other similar primary care settings in the future, particularly for underserved regions with poorer blood pressure control.

Thirdly, the CDSS was designed to serve as a tool to provide a decision aid for improving the process of shared decision making for hypertension care. The doctors were not mandated to follow recommendations generated from CDSS—the final prescription was based on a shared decision between doctors and patients. The reasons for CDSS recommendations not being followed were collected. The findings may provide valuable information for improving adherence to CDSS recommendations in the future.

Limitations of this study

Some limitations need to be considered. Firstly, we used a cluster randomised design to avoid contamination across groups. Doctors in the control group were, however, aware of the purpose of the trial and thus this might have had a positive impact on provision of usual care. If that was the case, the effect of CDSS might be underestimated. Secondly, this trial was conducted in primary care practices in urban China where most doctors have attained an education level of medical college or higher; the effectiveness of the CDSS in rural areas remains to be confirmed. It may be speculated that the CDSS would be equally or more effective in rural areas, since we observed similar improvements in practices with lower appropriate treatment rates at baseline or among doctors with lower education levels. Thirdly, the use of the CDSS was restricted to participants without coronary artery disease, chronic kidney disease, or heart failure, which limited the generalisability of the results. Fourthly, the trial was unable to evaluate the sustainability of the CDSS intervention and was underpowered to assess its effectiveness on cardiovascular events. Fifthly, we did not assess safety outcomes such as electrolyte disorders or renal dysfunction. Finally, we did not collect doctors’ satisfaction with the intervention, which could help its implementation outside of the trial. This will be investigated in future studies.

Policy implications

Our study has implications for strengthening the primary care systems in China and other resource restricted regions or countries with heavy burdens of cardiovascular disease.363738 Unlike other multicomponent intervention strategies394041 that inherently require more resources to implement and are more complex to scale-up, CDSS could potentially serve as a low cost, efficient, scalable, and sustainable means to improve access and equity to high quality care of hypertension. Primary care practices included in this study might differ in some ways from those in other regions of China and other resource restricted regions or countries, such as the availability of health information systems, antihypertensive drugs, and qualification of primary care providers. Despite these limitations, in rural areas where computers are not equipped to use a CDSS, the CDSS could support decision making through mobile phones or tablets, where highly compatible electronic health records for hypertension management can be installed. Another difference is that in this study the antihypertensive drugs were dispensed every 1-3 months, but in many primary care settings from other resource restricted regions or countries, a shorter prescription period may be common since a reliable supply chain of antihypertensive drugs may be difficult owing to problems with storage capacity.42 CDSS could be more useful when more frequent visits and prescriptions are needed. Of note, CDSS in this study was used by primary care doctors with higher education levels. As non-doctor health workers play an important role in task sharing or shifting strategies to mitigate the under-qualification and deficiency problems of primary care doctors,4143 future investigations could assess the effectiveness of CDSS among these staff with appropriate training and supervision.

Conclusions

The use of CDSS for hypertension management statistically significantly improved guideline accordant primary care for antihypertensive treatment and led to a modest reduction in blood pressure. This strategy offers a promising approach to delivering high quality care for hypertension efficiently and safely, particularly for resource constrained regions with a heavy burden of cardiovascular diseases like China.

What is already known on this topic

  • Hypertension care in China is suboptimal, with a large geographical variation

  • Improving the performance of primary care doctors by ensuring their prescribing behaviour follows current guidelines is a key step towards improving blood pressure control and patient outcomes

  • Although clinical decision support systems have the potential to improve hypertension care in a low cost and efficient way, evidence for their effectiveness in improving the use of guideline accordant antihypertensive treatment in primary care is limited

What this study adds

  • The use of a clinical decision support system in primary care in China statistically significantly improved guideline accordant antihypertensive treatment and led to a modest reduction in blood pressure

Web extra Extra material supplied by authors

Ethics statements

Ethical approval

This trial was approved by the central ethics committee at Fuwai Hospital (ID No 2017-939) and the collaborating centres. Individual level informed consent for use of the clinical decision support system was waived because of the minimal risk to the participants.

Data availability statement

Data described in the manuscript will not be made publicly available. Data collected for the study will be made available publicly upon reasonable request. For further detailed data access policy and procedure, contact zhengxin@fuwai.com.

Acknowledgments

We thank the participants, doctors, and administrative staff for their contributions, and the members of the trial steering committee.

Footnotes

  • Contributors: JS, XW, and BW are joint first authors and contributed equally to this article. JS, XW, and BW wrote the first draft of the manuscript. HMK, XZ, and JA are joint senior authors and designed the study. YG, LB, FJ, HJ, TL, BG, LW, JH, YZ, JL, HZ, XL, JL, WM, JW, STN, JH, JA, HMK, and XZ critically revised the manuscript. YZ and JH conducted analyses. XZ supervised and obtained funding for the study. JS, XW, BW, YZ, JH, and XZ had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. XZ is the guarantor. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

  • Funding: The trial was supported by the Chinese Academy of Medical Sciences innovation fund for medical science (2016-I2M-1-006) and the Sanming Project of Medicine in Shenzhen (SZSM201811093). The funders had no role in the study design, data collection, data analysis, data interpretation, writing of the report, or decision to submit the article for publication. All authors confirm that they worked independently from the funders.

  • Competing interests: All authors have completed the ICMJE uniform disclosure form at https://www.icmje.org/disclosure-of-interest/ and declare: support from the Chinese Academy of Medical Sciences and the Sanming Project of Medicine in Shenzhen; in the past three years, HMK received options for Element Science and Identifeye and payments from F-Prime for advisory roles, and he is a cofounder of, and holds equity in, Hugo Health, Refactor Health, and Ensight-AI, and he is associated with research contracts through Yale University from Janssen, Kenvue, and Pfizer; no other relationships or activities that could appear to have influenced the submitted work.

  • Transparency: The corresponding author (XZ) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

  • Dissemination to participants and related patient and public communities: The findings of this study will be presented to researchers, doctors, and students through a conference. We will also increase the impact of the study by sharing the findings with policy makers, research communities, and the general public through social media. We will also present the findings in a blog to explain the results.

  • Provenance and peer review: Not commissioned; externally peer reviewed.

http://creativecommons.org/licenses/by-nc/4.0/

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

References

  1. ↵
    1. Vaduganathan M,
    2. Mensah GA,
    3. Turco JV,
    4. Fuster V,
    5. Roth GA
    . The Global Burden of Cardiovascular Diseases and Risk: A Compass for Future Health. J Am Coll Cardiol2022;80:2361-71. doi:10.1016/j.jacc.2022.11.005 pmid:36368511
    OpenUrlCrossRefPubMed
  2. ↵
    1. Mills KT,
    2. Stefanescu A,
    3. He J
    . The global epidemiology of hypertension. Nat Rev Nephrol2020;16:223-37. doi:10.1038/s41581-019-0244-2 pmid:32024986
    OpenUrlCrossRefPubMed
  3. ↵
    1. Zhou B,
    2. Carrillo-Larco RM,
    3. Danaei G,
    4. et al.,
    5. NCD Risk Factor Collaboration (NCD-RisC)
    . Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019: a pooled analysis of 1201 population-representative studies with 104 million participants. Lancet2021;398:957-80. doi:10.1016/S0140-6736(21)01330-1 pmid:34450083
    OpenUrlCrossRefPubMed
  4. ↵
    1. Wang Z,
    2. Chen Z,
    3. Zhang L,
    4. et al.,
    5. China Hypertension Survey Investigators
    . Status of Hypertension in China: Results From the China Hypertension Survey, 2012-2015. Circulation2018;137:2344-56. doi:10.1161/CIRCULATIONAHA.117.032380 pmid:29449338
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Xu W,
    2. Goldberg SI,
    3. Shubina M,
    4. Turchin A
    . Optimal systolic blood pressure target, time to intensification, and time to follow-up in treatment of hypertension: population based retrospective cohort study. BMJ2015;350:h158. doi:10.1136/bmj.h158 pmid:25655523
    OpenUrlAbstract/FREE Full Text
  6. ↵
    1. Maddox TM,
    2. Ross C,
    3. Tavel HM,
    4. et al
    . Blood pressure trajectories and associations with treatment intensification, medication adherence, and outcomes among newly diagnosed coronary artery disease patients. Circ Cardiovasc Qual Outcomes2010;3:347-57. doi:10.1161/CIRCOUTCOMES.110.957308 pmid:20488918
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Fontil V,
    2. Pacca L,
    3. Bellows BK,
    4. et al
    . Association of Differences in Treatment Intensification, Missed Visits, and Scheduled Follow-up Interval With Racial or Ethnic Disparities in Blood Pressure Control. JAMA Cardiol2022;7:204-12. doi:10.1001/jamacardio.2021.4996 pmid:34878499
    OpenUrlCrossRefPubMed
  8. ↵
    1. Kim BJ,
    2. Cho YJ,
    3. Hong KS,
    4. et al
    . Treatment Intensification for Elevated Blood Pressure and Risk of Recurrent Stroke. J Am Heart Assoc2021;10:e019457. doi:10.1161/JAHA.120.019457 pmid:33787300
    OpenUrlCrossRefPubMed
  9. ↵
    1. Li X,
    2. Krumholz HM,
    3. Yip W,
    4. et al
    . Quality of primary health care in China: challenges and recommendations. Lancet2020;395:1802-12. doi:10.1016/S0140-6736(20)30122-7 pmid:32505251
    OpenUrlCrossRefPubMed
  10. ↵
    1. Li X,
    2. Lu J,
    3. Hu S,
    4. et al
    . The primary health-care system in China. Lancet2017;390:2584-94. doi:10.1016/S0140-6736(17)33109-4 pmid:29231837
    OpenUrlCrossRefPubMed
  11. ↵
    1. Su M,
    2. Zhang Q,
    3. Bai X,
    4. et al
    . Availability, cost, and prescription patterns of antihypertensive medications in primary health care in China: a nationwide cross-sectional survey. Lancet2017;390:2559-68. doi:10.1016/S0140-6736(17)32476-5 pmid:29102087
    OpenUrlCrossRefPubMed
  12. ↵
    1. Zhou T,
    2. Wang Y,
    3. Zhang H,
    4. et al.,
    5. China PEACE Collaborative Group
    . Primary care institutional characteristics associated with hypertension awareness, treatment, and control in the China PEACE-Million Persons Project and primary health-care survey: a cross-sectional study. Lancet Glob Health2023;11:e83-94. doi:10.1016/S2214-109X(22)00428-4 pmid:36521957
    OpenUrlCrossRefPubMed
  13. ↵
    National Health Commission of the People’s Republic of China. The statistical bulletin on the development of China’s medical care industry in 2022. 2022. http://www.nhc.gov.cn/guihuaxxs/s3586s/202310/5d9a6423f2b74587ac9ca41ab0a75f66.shtml (accessed March 17 2024).
  14. ↵
    National Health Commission of the People’s Republic of China. 2022 China Health Statistics Yearbook. 2022. http://www.nhc.gov.cn/mohwsbwstjxxzx/tjtjnj/202305/6ef68aac6bd14c1eb9375e01a0faa1fb.shtml.
  15. ↵
    1. Lu Y,
    2. Melnick ER,
    3. Krumholz HM
    . Clinical decision support in cardiovascular medicine. BMJ2022;377:e059818. doi:10.1136/bmj-2020-059818 pmid:35613721
    OpenUrlAbstract/FREE Full Text
  16. ↵
    1. Prabhakaran D,
    2. Jha D,
    3. Prieto-Merino D,
    4. et al.,
    5. Members of the Research Steering Committee,Investigators,Members of the Data Safety and Monitoring Board
    . Effectiveness of an mHealth-Based Electronic Decision Support System for Integrated Management of Chronic Conditions in Primary Care: The mWellcare Cluster-Randomized Controlled Trial. Circulation2019;139:380-91. doi:10.1161/CIRCULATIONAHA.118.038192 pmid:30586732
    OpenUrlCrossRefPubMed
  17. ↵
    1. Anchala R,
    2. Kaptoge S,
    3. Pant H,
    4. Di Angelantonio E,
    5. Franco OH,
    6. Prabhakaran D
    . Evaluation of effectiveness and cost-effectiveness of a clinical decision support system in managing hypertension in resource constrained primary health care settings: results from a cluster randomized trial. J Am Heart Assoc2015;4:e001213. doi:10.1161/JAHA.114.001213 pmid:25559011
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Hicks LS,
    2. Sequist TD,
    3. Ayanian JZ,
    4. et al
    . Impact of computerized decision support on blood pressure management and control: a randomized controlled trial. J Gen Intern Med2008;23:429-41. doi:10.1007/s11606-007-0403-1 pmid:18373141
    OpenUrlCrossRefPubMedWeb of Science
  19. ↵
    1. Song J,
    2. Wang X,
    3. Wang B,
    4. et al.,
    5. LIGHT Collaborative Group
    . Effectiveness of a clinical decision support system for hypertension management in primary care: study protocol for a pragmatic cluster-randomized controlled trial. Trials2022;23:412. doi:10.1186/s13063-022-06374-x pmid:35578345
    OpenUrlCrossRefPubMed
  20. ↵
    National Health and Family Planning Commission. National Basic Public Health Service Standards (Third Edition) 2017. http://www.nhc.gov.cn/jws/s3578/201703/d20c37e23e1f4c7db7b8e25f34473e1b.shtml (accessed March 11 2024).
  21. ↵
    1. Office for Management of Hypertension in Primary Health Care NEPHSP
    . Chinese Guidelines for Hypertension Prevention and Management in Primary Care. Chinese Circulation Journal2017;11:1041-48.
    OpenUrl
  22. ↵
    National Health Commission of the People’s Republic of China. Notice on Issuing the Long Term Prescription Management Standards (Trial) 2021. http://www.nhc.gov.cn/yzygj/s7659/202108/e8fbc0fa4bb6450b8ed0a8e9055a5b5e.shtml (accessed March 16 2024).
  23. ↵
    1. Whelton PK,
    2. Carey RM,
    3. Aronow WS,
    4. et al
    . ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults: Executive Summary: A Report of the American College of Cardiology/American Heart Association Task F. Circulation2017;2018:138.
    OpenUrl
  24. ↵
    1. James PA,
    2. Oparil S,
    3. Carter BL,
    4. et al
    . 2014 evidence-based guideline for the management of high blood pressure in adults: report from the panel members appointed to the Eighth Joint National Committee (JNC 8). JAMA2014;311:507-20. doi:10.1001/jama.2013.284427 pmid:24352797
    OpenUrlCrossRefPubMed
  25. ↵
    1. Shojania KG,
    2. Jennings A,
    3. Mayhew A,
    4. Ramsay C,
    5. Eccles M,
    6. Grimshaw J
    . Effect of point-of-care computer reminders on physician behaviour: a systematic review. CMAJ2010;182:E216-25. doi:10.1503/cmaj.090578 pmid:20212028
    OpenUrlAbstract/FREE Full Text
  26. ↵
    1. Linden A,
    2. Samuels SJ
    . Using balance statistics to determine the optimal number of controls in matching studies. J Eval Clin Pract2013;19:968-75. doi:10.1111/jep.12072 pmid:23910956
    OpenUrlCrossRefPubMed
  27. ↵
    1. Rubin DB
    . Using propensity scores to help design observational studies: application to the tobacco litigation. Health Serv Outcomes Res Methodol2001;2:169-88doi:10.1023/A:1020363010465.
    OpenUrlCrossRef
  28. ↵
    1. Zhang J,
    2. Yu KF
    . What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA1998;280:1690-1. doi:10.1001/jama.280.19.1690 pmid:9832001
    OpenUrlCrossRefPubMedWeb of Science
  29. ↵
    1. Borbolla D,
    2. Giunta D,
    3. Figar S,
    4. Soriano M,
    5. Dawidowski A,
    6. de Quiros FG
    . Effectiveness of a chronic disease surveillance systems for blood pressure monitoring. Stud Health Technol Inform2007;129:223-7.pmid:17911711
    OpenUrlPubMed
  30. ↵
    1. Barnett GO,
    2. Winickoff RN,
    3. Morgan MM,
    4. Zielstorff RD
    . A computer-based monitoring system for follow-up of elevated blood pressure. Med Care1983;21:400-9. doi:10.1097/00005650-198304000-00003 pmid:6341724
    OpenUrlCrossRefPubMedWeb of Science
  31. ↵
    1. Rogers JL,
    2. Haring OM,
    3. Wortman PM,
    4. Watson RA,
    5. Goetz JP
    . Medical information systems: assessing impact in the areas of hypertension, obesity and renal disease. Med Care1982;20:63-74. doi:10.1097/00005650-198201000-00005 pmid:7078276
    OpenUrlCrossRefPubMedWeb of Science
  32. ↵
    1. Montgomery AA,
    2. Fahey T,
    3. Peters TJ,
    4. MacIntosh C,
    5. Sharp DJ
    . Evaluation of computer based clinical decision support system and risk chart for management of hypertension in primary care: randomised controlled trial. BMJ2000;320:686-90. doi:10.1136/bmj.320.7236.686 pmid:10710578
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Sarkar U,
    2. Samal L
    . How effective are clinical decision support systems?BMJ2020;370:m3499. doi:10.1136/bmj.m3499 pmid:32943393
    OpenUrlFREE Full Text
  34. ↵
    1. Chen Z,
    2. Liang N,
    3. Zhang H,
    4. et al
    . Harnessing the power of clinical decision support systems: challenges and opportunities. Open Heart2023;10:e002432. doi:10.1136/openhrt-2023-002432 pmid:38016787
    OpenUrlAbstract/FREE Full Text
  35. ↵
    1. Zhou B,
    2. Perel P,
    3. Mensah GA,
    4. Ezzati M
    . Global epidemiology, health burden and effective interventions for elevated blood pressure and hypertension. Nat Rev Cardiol2021;18:785-802. doi:10.1038/s41569-021-00559-8 pmid:34050340
    OpenUrlCrossRefPubMed
  36. ↵
    1. Bitton A,
    2. Fifield J,
    3. Ratcliffe H,
    4. et al
    . Primary healthcare system performance in low-income and middle-income countries: a scoping review of the evidence from 2010 to 2017. BMJ Glob Health2019;4(Suppl 8):e001551. doi:10.1136/bmjgh-2019-001551 pmid:31478028
    OpenUrlAbstract/FREE Full Text
  37. ↵
    1. Arshad V,
    2. Samad Z,
    3. Das J,
    4. et al
    . Prescribing Patterns of Antihypertensive Medications in Low- and Middle-Income Countries: A Systematic Review. Asia Pac J Public Health2021;33:14-22. doi:10.1177/1010539520965280 pmid:33084371
    OpenUrlCrossRefPubMed
  38. ↵
    1. Schutte AE,
    2. Srinivasapura Venkateshmurthy N,
    3. Mohan S,
    4. Prabhakaran D
    . Hypertension in Low- and Middle-Income Countries. Circ Res2021;128:808-26. doi:10.1161/CIRCRESAHA.120.318729 pmid:33793340
    OpenUrlCrossRefPubMed
  39. ↵
    1. Jafar TH,
    2. Gandhi M,
    3. de Silva HA,
    4. et al.,
    5. COBRA-BPS Study Group
    . A Community-Based Intervention for Managing Hypertension in Rural South Asia. N Engl J Med2020;382:717-26. doi:10.1056/NEJMoa1911965 pmid:32074419
    OpenUrlCrossRefPubMed
  40. ↵
    1. Schwalm J-D,
    2. McCready T,
    3. Lopez-Jaramillo P,
    4. et al
    . A community-based comprehensive intervention to reduce cardiovascular risk in hypertension (HOPE 4): a cluster-randomised controlled trial. Lancet2019;394:1231-42. doi:10.1016/S0140-6736(19)31949-X pmid:31488369
    OpenUrlCrossRefPubMed
  41. ↵
    1. Sun Y,
    2. Mu J,
    3. Wang DW,
    4. et al.,
    5. CRHCP Study Group
    . A village doctor-led multifaceted intervention for blood pressure control in rural China: an open, cluster randomised trial. Lancet2022;399:1964-75. doi:10.1016/S0140-6736(22)00325-7 pmid:35500594
    OpenUrlCrossRefPubMed
  42. ↵
    1. Marklund M,
    2. Cherukupalli R,
    3. Pathak P,
    4. et al
    . Hypertension treatment capacity in India by increased workforce, greater task-sharing, and extended prescription period: a modelling study. Lancet Reg Health Southeast Asia2022;10:100124. doi:10.1016/j.lansea.2022.100124 pmid:37383361
    OpenUrlCrossRefPubMed
  43. ↵
    1. Anand TN,
    2. Joseph LM,
    3. Geetha AV,
    4. Prabhakaran D,
    5. Jeemon P
    . Task sharing with non-physician health-care workers for management of blood pressure in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Glob Health2019;7:e761-71. doi:10.1016/S2214-109X(19)30077-4 pmid:31097278
    OpenUrlCrossRefPubMed
Back to top