One hundred studies were included (3,826 practitioners or practices, caring for more than 92,895 patients). Of these, 88 were RCTs.
The mean quality score was 7 (range: 2 to 10). Fifty-two per cent of the RCTs described an appropriate method of treatment allocation and 28% reported allocation concealment.
Overall, the CDSS improved outcomes in 62 (64%) of the 97 studies assessing practitioner performance and 7 (13%) of the 52 studies assessing patient outcomes.
Systems for diagnosis (10 studies): 4 of the 10 studies assessing practitioner performance found benefit from the CDSS. Successful systems were diagnostic systems for cardiac ischaemia in emergency department (2 studies), mood disorder screening in a post-traumatic stress clinic (1 study), and the diagnosis of acute bowel obstruction (1 study). None of the 5 studies assessing patient outcomes found success.
Reminder systems for prevention (21 studies): 16 of the 21 studies assessing practitioner performance found benefit from the CDSS. The outcomes included rates of screening, counselling, vaccination, testing use of medication, or identifying at-risk behaviours. Successful systems were reported mainly for ambulatory care. The single study assessing patient outcomes found no improvement in the primary analysis.
Systems for disease management (40 studies): 23 of the 37 studies assessing practitioner performance found benefit from the CDSS. Success with CDSS was found for: 5 of the 7 studies of diabetes care; 5 of the 13 studies of cardiovascular prevention; 6 of the 9 studies of urinary incontinence, human immunodeficiency virus management, functional assessment and acute respiratory distress syndrome; all 4 studies monitoring the effects of prescribed treatment; and 3 of the 4 studies aimed at reducing unnecessary health care utilisation. Five of the 27 studies assessing patient outcomes found benefit from the CDSS. Improvements were found for blood-pressure control (1 study), urinary incontinence (1 study), barotrauma and organ dysfunction in ventilated patients (1 study), patient-reported exacerbations of asthma (1 study), and length of hospital stay (1 study).
Systems for drug dosing and drug prescribing (29 studies): 15 of the 24 single-drug dosing studies assessing practitioner performance found benefit from the CDSS; 2 of the 18 single-drug studies assessing patient outcomes also found benefit from the CDSS. Four of the 5 multi-drug prescribing studies assessing practitioner performance found benefit from the CDSS, but none improved patient outcomes. Factors associated with CDSS success in improving practitioner performance were automatic prompting to use the system compared with user activated (odds ratio adjusted for study quality 3.0, 95% confidence interval: 1.2, 7.1) and studies where the study authors developed the system compared with those where the authors were not developers (odds ratio 6.6, 95% confidence interval: 1.7, 26.7).