Skip navigation

Using computers to manage hypertension in primary care

 

Hypertension is a major contributor to cardiovascular morbidity and mortality but its control in the community remains a problem. Blood pressure and other risk factors need to be considered routinely and modified. A systematic approach is required for case finding, recall and review. These tasks are well nigh impossible if reliance has to be placed on manual record systems. However, the use of computer-based decision-support systems is growing in primary care and should make the task do-able. But, are IT systems helping to improve the management of hypertension? A systematic review [1] of has sought an answer to that question by evaluating the early lessons from the use of computers in the management of high blood pressure.

Reference

Alan A Montgomery, Tom Fahey. A systematic review of the use of computers in the management of hypertension. J Epidemiol Community Health 1998 52:520-525.

How was the question tackled?

Criteria to guide a search for relevant trials included the use of computers and clinical decision support systems (CDSS) in the administration and management of hypertension. Focus was on outcomes in three areas - patient administration, physician performance and blood pressure control. The aim was to identify randomised controlled trials that allocated patients, GPs or practices to computer based interventions. The review adopted the Cochrane search strategy and covered the Cochrane Library (1997: issue 3) as well as Medline and other relevant databases. The search also sought previous reviews in the same subject. The review team also contacted the study authors for additional unpublished information. The search included studies published in non-English language journals.

What did the review find?

Seven trials were identified which met the inclusion criteria, with most examining more than one outcome. Each study was assessed according to the quality criteria in the Cochrane Collaboration Handbook. All the trials that met the criteria were of similar methodological quality. Because of the diverse outcomes and relatively small number of trials examining particular outcome no quantitative summary measures for the outcomes were possible.

Three areas were examined. Patient uptake administration was evaluated in 5 trials - 4 of which reported significant improvements using a computer. Physician performance was evaluated in 3 trials - 2 of which reported significant improvements using a computer. Blood pressure control was evaluated in 6 trials 2 of, which reported significant improvements using a computer. However, positive findings from three of the trials need to be regarded cautiously because the cluster randomisation may have been responsible for the results. If this was the case the numbers reporting improvement in each of the three areas would be 2 of 5 (instead of 4 of 5), 1 of 3 (2 of 3) and 1 of 6 (2 of 6).

Effect on patient administration

Uptake: three of the trials showed that where a reminder was generated for the doctor by computer the percentage of patients whose blood pressure was recorded increased.

Follow-up: results from the three trials varied. While one showed that the use of the system generated more home visits it had no effect in another trial. Overall there is evidence that a higher proportion of patients in the computer group were being offered regular surveillance and monitoring.

Physician performance

Knowledge . One trial tested the effect of computer feedback when linked with, or provided separately from an educational programme. Doctors who were involved in both activities improved their scores more than those who were only given the computer based feedback.

Recording information . One trial showed that recording of 15 key data items was significantly better in computer based records.

Prescribing . One trial showed that computer generated recommendations did not indicate any advantage over normal care.

Blood pressure control.

The results were inconclusive. Two trials reported a greater proportion of patients in the computer group having controlled blood pressure. However, two other trials reported that computer generated feedback did not result in any further decrease in blood pressure compared with usual care.

What has the study told us?

The review confirmed that the use of computers could have a favourable influence on the administration of systems for hypertension management although it did not indicate any benefit from using computers in terms of clinician performance and blood pressure control. But the review was a snapshot in time as the development and use of computers systems within health care management continues to grow at a rapid rate.

The results of this review are limited however, because:

These results contradict two previous reviews that have shown more positive results. One looking at computer use in primary care reported that all 21 studies examining clinician performance demonstrated an improvement in outcomes. Another review of clinical decision support systems in a broad variety of interventions has shown that improvement can occur. But both of these reviews could be criticised for lack of methodological rigour.

Notwithstanding these results there is a growing use of computers and clinical decisions support systems across the NHS, particularly in the management of chronic conditions, such at the management of diabetes, of asthma, and of anticoagulation therapy. Similarly expert systems are being developed for the treatment of heart failure and depression. There are cautionary signals from this review, which should encourage more careful evaluation of new systems as they develop.