In an article published this month in Health Affairs on the reasons why scientific evidence often has little effect on clinical practice, technology played a role--though not necessarily in a good way.
The authors, from the Rand Corp., examined six studies on comparative effectiveness research and its effects on practice. Among the problems cited were limited use of clinical decision support tools that could provide better alignment.
"Integrating clinical decision support tools seamlessly into clinical practice remains difficult because poorly designed tools can force clinicians to spend too much time engaging with computers rather than patients," the authors wrote. "Clinicians may also lack guidance on how to select an appropriate patient decision aid and may lack training in how and when to use the aids. Many providers that have experimented with decision aids have abandoned them because the providers had difficulty tailoring the aids to the work flow of their clinic."
The authors also noted a "pro-technology bias"--the uncritical tendency to believe that newer forms of technology are better, prompting a switch before clinical trials have been completed. They highlighted one trial in particular, which studied bare metal stents; the widespread use of drug-eluting stents by the time results were released, however, raised questions about the relevance of the results.
Among possible solutions, the authors propose developing objectives and standards first to better align research with outcomes.
Clinical decision support tools topped hospital IT leaders's wish lists, according to a recent survey by Black Book Rankings. Just 16 percent of hospital leader respondents said their facilities have the CDS tools they need for accountable care.
Meanwhile, researchers found that decision support in electronic health records produce the best return on investment for providers, according to a study published recently in the American Journal of Managed Care.
Still, other research has shown that systems need systems need flexibility, and to be tailored to the specific workflow for which they're being used, or they're likely to be ignored.
To learn more:
- read the abstract