Clinical reminder systems have emerged as one effective means of translating research findings into practice by improving compliance with established clinical guidelines. Computer implementation holds promise because of the possibility of tailoring the reminders to the particular patient, presenting them "on-screen" at the time of the patient encounter, and facilitating the ordering of indicated laboratory tests and/or procedures. The VA computer system has been designed with these considerations in mind. New software has been distributed to accommodate a broad range of reminders within the VA Computerized Patient Record System (CPRS). An increasing number of reminders are being developed for national distribution that address important issues in patient care. In addition, many individual VA stations have developed numerous clinical reminders of their own. However, little is known about the types of factors that may influence whether or not providers appropriately respond to such reminders.
The objective of this study was to investigate how various factors affect the success of computerized clinical reminders in the VA.
The research design involved database analyses as well as a mailed survey to primary care providers. The principal source of data was existing VA databases and returned surveys, and the principal type of analysis was outcomes assessment. Specific outcomes analyzed were the presence of responses to computerized clinical reminders in association with primary care clinic visits. Those studied included patients seen by selected primary care providers at four VA medical centers over a 9-month period.
he four sites had substantial differences in the number of reminders (n=13, 22, 30 and 63) principally directed to primary care providers. The overall response rate to the mailed survey was 67.8% (59 of 87) and did not significantly vary among sites.
On the survey, there was a relatively strong tendency for providers to indicate that there are too many tasks that primary care providers are expected to accomplish during the allotted time for a patient visit, that there are too many Clinical Reminders, and that responding to Clinical Reminders is too time-consuming. There was a less strong tendency for providers to indicate that the computerized Clinical Reminder system is helpful in caring for patients and that the Clinical Reminder system makes errors in displaying reminders for patients for whom the reminder does not apply. Providers seemed divided on the issue of the extent that Clinical Reminders were easy to use and the extent that Clinical Reminders were taken care of by ancillary staff. There were strong tendencies for providers to indicate that compliance with clinical guidelines related to Clinical Reminders was being evaluated by VA management or administration, that they received feedback on compliance with clinical guidelines related to Clinical Reminders, that their department or primary care team encourage the use of the Clinical Reminder system, and that their peers are using the Clinical Reminder system (“reinforcing factors”). From survey responses, a greater perception of such “reinforcing factors” as well as a higher number of weekly outpatients were both associated with more negative attitudes with regard to reminders and time constraints in clinical encounters (p<.05). (Both the “reinforcing” and reminder/time constraint measures were based on aggregate scores of relevant survey questions).
Regarding provider responses to “clinical management” reminders, there were no overarching factors that were consistently associated with resolution of all reminders either within or across sites. However, specific factors were found to be associated with reminder resolution (p<.05) for certain reminders at each site. For example, in a number of the reminder analyses, a more favorable attitude with regard to reminders and time constraints in clinical encounters was associated (p<.05) with a higher likelihood of resolution. Also, for a number of reminders, a higher number of cumulative patient diagnoses (suggesting greater clinical complexity) was associated with a reduced likelihood of reminder resolution (p<.05). The association of the aggregate measure of “reinforcing factors” on reminder resolution was significant (p<.05) for a number of the reminders, however, the directionality in such associations was variable.
This study has contributed to our knowledge of provider attitudes regarding reminders and the degree to which certain factors may be associated with resolution of specific reminders. Findings of this study will be important in helping to understand “best practices” in reminder implementation and in addressing ongoing efforts to improve the VA clinical reminder system and, ultimately, to improve patient outcomes.
External Links for this Project
None at this time.