Studies were eligible if they were clinical trials, randomised controlled trials (RCTs) or comparative studies. Eligible studies had to be conducted in an ambulatory setting, have sufficient information about the HIT intervention for assessment and examine laboratory test monitoring. Laboratory testing had to be for assessment of effectiveness, toxicity or side-effects. Only studies without cointerventions were included. Studies that targeted anticoagulants were excluded.
Most of the included studies were conducted in USA; one study was in Israel. Three quarters of the studies were conducted in large integrated healthcare delivery systems. All studies included patients nested within providers.
Interventions included electronic alerts to prescribing physicians and electronic alerts to pharmacists who could then order laboratory tests and contact the patient. All studies were carried out within healthcare systems with electronic records (half used their own systems and half used modifications of proprietary systems). Most interventions were built-in systems with computerised physician order entry. These were either interruptive (required the provider to respond to the alert) or non-interruptive (no action required). Most studies targeted a broad range of medications; one study targeted a single medication.
Outcomes included completion of laboratory test monitoring (75% of studies) and evaluation of physician test ordering (25% of studies). Baseline rates of appropriate laboratory monitoring varied greatly (14% to 95%) depending on the study drug.
Each study was independently assessed by two reviewers for inclusion. Disagreements were resolved by consensus.