Using computers to improve prescription practices

Recommendations directly in the electronic patient record

During a hospital stay 30-50% of patients receive antibiotics. Nevertheless, doctors always find it very challenging to prescribe the right drug at the right dosage for the right duration, not least because new pathogens that develop resistance to antibiotics are constantly emerging, often in locally limited outbreaks. However, the use of antibiotics that fail to achieve the desired effect promotes the development and spread of antibiotic resistance and exposes patients to other unwanted side effects. 

To counteract this, a team led by Benedikt Huttner from the University of Geneva is relying on computerised decision support systems that include the latest treatment recommendations. In the COMPASS (COMPuterised Antibiotic Stewardship Study) project, they have integrated one such tool directly in the in-house patient records at three hospitals in Geneva, Lugano and Bellinzona. This helps doctors select both the right antibiotics and the appropriate treatment period directly during consultations. They must also state clear reasons whenever their prescribing practice deviates from the local guidelines. Every three months they receive feedback on their prescribing practice.

No effect on the quantity, but indications of better quality of prescriptions

In their study, the researchers arranged for the tool to be used by doctors in various hospital departments for a year, and then compared their prescribing practices with a group of doctors who were working in the same period in other departments of the same hospitals, but who did not use the tool. The departments were allocated randomly. During the study period, there were over 11,000 referrals to the departments in the test group and over 9,500 referrals to departments in the comparator group. Overall, around four in ten patients received antibiotics.

In their analysis, the researchers took into account factors that can influence antibiotic prescription, such as age, type of department and concomitant illnesses. However, there was no overall effect revealed on the amount of antibiotics prescribed: doctors who used the tool did not significantly reduce their use of the drugs. However, further analyses suggested that the tool did have an effect on the quality of prescribing, e.g. patients in departments using the tool were more likely to be switched to oral antibiotics. Nevertheless, the tested intervention failed to achieve the important goal of reducing antibiotic use. As possible reasons for this, the researchers refer to the inadequate acceptance of the tool by the doctors and the insufficient user-friendliness of the tool. It may also be possible that antibiotic prescribing in the tested hospitals is relatively good and any improvements are therefore more difficult to achieve in this specific case.

Better user-friendliness should increase acceptance

The project has shown, however, that stewardship programmes promoting the more targeted use of antibiotics in hospitals using digital tools can basically be implemented with reasonable effort. Accordingly, the researchers also expect that such tools will be better accepted, and thus more effective, in the long term. In order to explore how their user-friendliness and flexibility can be improved, Benedikt Huttner's team is working in an international project designed to investigate a smartphone app with the same goals developed by a Canadian company. This is currently being tested in hospitals in Switzerland, the Netherlands and Sweden. Although the study has not yet been concluded, it has already shown that the demand for this user-friendly tool is very high, particularly among younger doctors.

November 2021

COMPASS study (COMPuterized Antibiotic Stewardship Study)