Use of computer decision support interventions to improve medication prescribing in older adults: a systematic review
Introduction
Medication related errors are high in adults. Adverse drug events occur in 20-35% of outpatients and 6-15 percent of hospitalized elderly and nursing home residents. Those 65 years and older are at an increased risk because of the number of medications they take and because of increased sensitivity to medications with age. In this study, they looked at how clinical computer decision support (CDS) within computerized provider order entry (CPOE) systems improved physicians medication prescribing.
Methods
The researchers looked for articles published between January 1980 and July 2007 using PubMed and EMBASE databases. For an article to be eligible, it needed to include a computer decision support system and ≥1 process or clinical outcome measurement related to medication prescribing in older adults. Studies were excluded if the CDS intervention only affected the patient’s knowledge of the medication (i.e. educational intervention). The following data were extracted: setting (inpatient or outpatient), study design, sample size, country, intervention, main outcome results and medication names.
They identified 285 potentially eligible articles based on keyword search. Of those 285 articles, only 10 articles met the eligibility criteria. All studies focused on adults ranging in age from ~60-90 years old and were either randomized controlled trials, cohort, pre/post intervention, or interrupted times series in design. Eight studies provided CDS directly to the physician at the point of care and 2 studies provided CDS indirectly to the physician (i.e. having a pharmacist receive feedback from the computer and then the pharmacist conveyed that to the physician).
Results
Effect on Process Outcomes
Eight of the 10 studies showed positive outcomes in regards to medication-related CDS interventions. The number needed to treat (NNT) to improve prescribing ranged from 5 to 1960 with a median value of 33. Six of those studies were direct CDS intervention and the remaining 2 were indirect CDS (i.e. pharmacist intermediary).
One study found that of 3718 hospitalized patients, 29% of patients were prescribed the recommended daily dose (versus 19% pre-CDS intervention) and there was also a decrease in the number of prescriptions for non-recommended drugs (7.6% vs 10.8% of total orders). In another study (N=311), real-time electronic reminders to providers resulted in 51.5% of post-fracture patients received either an osteoporosis medication or bone mineral density test compared to 5.9% of patients in usual care.
Effect on Clinical Outcomes
Only 2 studies reported any clinical outcomes. One study focused on improving psychotropic drug prescribing. They found that patients had a lower in-hospital fall rate, but CDS intervention had no effect on hospital length or stay of days of altered mental status. The other study looked to improve cardiovascular risk assessment and anti-hypertensive treatment. CDS intervention did not reduce cardiovascular risk or lower blood pressure.
Discussion/Conclusion
CDS intervention in older adults produced positive outcomes such as lower rates of inappropriate drug prescribing and a higher adherence to better drug choices and dosages. Adverse drug effects and inappropriate or inaccurate prescribing results in substantial health care costs, longer hospital lengths of stay and increased risk of death. Some of the studies recommended avoiding certain drugs, using a lower dose in older adults, avoiding drug interactions and choosing other (sometimes nonpharmacologic) therapies.
Noncomputer interventions such as clinical pharmacy and geriatric medicine service interventions have been shown to improve prescribing and decrease potential ADEs.
Yourman L, Concato J, Agostini JV. Use of computer decision support interventions to improve medication prescribing in older adults: a systematic review. Am J Geriatr Pharmacother. 2008 Jun;6(2):119-29.
