Strategies to Improve Test Ordering in Primary Care

To the Editor: Dr Verstappen and colleagues found that a practice-based, multifaceted strategy using guidelines, feedback, and social interaction resulted in modest improvements in the quality of test ordering by Dutch primary care physicians. In the intervention, physicians received 3 personal feedback reports on 3 different clinical problems, discussed these in small group meetings, related them to evidence-based guidelines, and made plans for change. Such multifaceted interventions are time consuming, and therefore difficult to implement for the large spectrum of clinical problems that primary care physicians face. Furthermore, many studies have found that the effects of information and feedback are limited, and the effects may disappear soon after the end of intervention. In their literature review, for instance, Wensing et al concluded that some, but not all, multifaceted interventions are effective, but that information linked to performance generally was not effective. Large-scale interventions, related to sustained change in all possible test-ordering situations, are likely to be ineffective. Thus, I wonder if this is the way to proceed in improving test-ordering behavior.

Verstappen et al also concluded that multifaceted interventions are superior to single interventions. However, they made no comparison to any single intervention. Some of these interventions may in fact be quite effective. In a study of Dutch primary care physicians, my colleagues and I found that computer-guided decision support during the actual process of ordering the tests led to much larger effects than those reported by Verstappen et al. Such computer-based interventions are not time consuming, affect behavior at the moment the test is ordered, can be extended to all tests ordered, and can be sustained indefinitely.

I suggest that it is time to move forward to more powerful single and simple interventions rather than to more time-consuming and complex combinations of interventions with limited effectiveness.