Abstract
Background: The negative predictive value and accuracy of d-dimer assays for venous thromboembolism (VTE) worsens with patient age, meaning additional testing is not avoided in the majority of older patients. We evaluated the effectiveness of a computerized alert to reduce the number of tests ordered for these patients.
Methods: We hypothesized that an alert at the time of d-dimer ordering for patients ≥65-years-old would decrease the rate of d-dimer testing for these patients. We randomized two groups of clinics. We analyzed the baseline rate of d-dimer orders/1000 patient visits, the rate in the intervention group, and the rate in the control group. After 18 months, the alert was activated for control clinic providers.
Results: Rate of d-dimers orders/1000 clinic visits in intervention clinics before and after the alert was significantly reduced. We saw a reduced rate in the control clinics also, but it was smaller (possibly due to secular trend). After the alert was turned on in the original control clinics the rate was significantly smaller. The reduced rate persisted in the intervention clinic throughout the study. No significant rate changes were seen in patients <65-years-old.
| d-dimer orders/1000 patient visits | ||
|---|---|---|
| ≥65-years-old | Baseline rate (no alerts) | Rate difference (95% CI) |
| Intervention clinics | 4.91 | −3.36 (−4.05, −2.66) |
| Rate difference (95% CI) | −1.30 (−2.18, −0.42) | |
| Control clinics | 3.61 | −1.23 (−1.91, −0.55) |
| <65-years-old | No alerts | |
| Intervention clinics | 4.07 | 0.30 (−0.18, 0.77) |
| Rate difference (95% CI) | −0.27 (−0.81, 0.27) | |
| Control clinics | 3.8 | 1.53 (−0.31, 0.61) |
Conclusions: A targeted electronic alert can change the ordering behavior of providers, decrease inappropriate test ordering, and may foster adherence to evidence-based guidelines for management of patients with suspected VTE.
- Received September 11, 2008.
