Communication outcomes of critical imaging results in a computerized notification system
Identification of Problem: As a clinician, one is always concerned about missing radiology reports that recommend additional imaging studies. If there is a pulmonary nodule which subsequently develops into a cancer this can lead to great harm for the patient. This study attempts to look at this problem by analyzing how many imaging “alerts” which are sent to ordering clinicians are acknowledged by them using a computerized notification system. They also try to determine reasons why some alert are not acknowledged.
Source of Study/Study Design: The study was conducted at the outpatient clinics in the Michael DeBakey Veterans Affairs Medical Center in Houston, Texas. The VA uses a CPOE system that notifies physicians of non-life-threatening abnormalities on radiology tests ordered via a View Alert notification that is sent to the ordering provider. At the DeBakey Center, the radiology department has set up codes to reflect “significant unexpected findings” in their readings of the images ordered. These codes are sent immediately to the ordering physician as well as occasionally to an alternate provider. Clicking on an alert provides notification to the EHR that the alert was seen. If the alert did not seem to have been read, a system was set up to ensure that someone did get notification of this alert.
Results: The study looked at 20,680 radiographic studies done over approximately a 3 month period. Of these studies, 1,017 alerts were sent and the authors looked at the 402 unacknowledged alerts. Of these 402 alerts, 293 were acknowledged although not electronically. 40 (3.9% of the alerts sent) of the remaining alerts were believed to be alerts where the provider was truly unaware of them 4 weeks after the test was performed. All of these alerts involved findings which could lead to a delayed or missed diagnosis and 73% of these might have been some type of malignancy.
Conclusion: Use of a computerized notification system to alert physicians about abnormal radiographic studies can still lead to cases where the primary provider is seemingly unaware of abnormal test results. This rate appears to be improved when compared to non-computerized notification systems although the number of studies looking at this are limited. Additional studies are needed to see if these results can be carried over to other EHR systems as well as why some alerts are still missed,
Comments: This study is a little disconcerting when you realize that even with electronically sent alerts being generated, about 4% of the alerts did not seem to be acknowledged or the provider was not aware of reception. Although not looked at, one has to wonder whether the studies which were acknowledged as being received were acted upon. In our emergency department we frequently receive reports about abnormalities which need to be correlated with previous studies (nodule seen on CXR, is it new or old) or in which follow up studies are recommended to be done in the ensuing months to rule out some other potentially serious problem. Are these recommended studies performed? I believe that this is as much a potential problem, if not more, for the ordering physician and would be an interesting follow-up to this study. KOC