Integrating Contextual Factors into Clinical Decision Support to Reduce Contextual Error and Improve Outcomes in Ambulatory Care

Effective care planning must account for patient context: the circumstance and behaviors of the patient. Contextual errors occur when a care plan is inappropriate because of inattention to patient context. For example, increasing the dosage of a patient’s medication to manage deterioration of a chronic condition is a contextual error when the disease symptoms are due to something in a patient’s circumstances, such as a change in health insurance coverage, loss of social support, or competing responsibilities. Contextual errors can lead to diminished healthcare outcomes and an increase in healthcare costs related to overuse and misuse of medical services. When a clinician more easily recognizes clues—“contextual red flags”—that represent challenges a patient is experiencing that impact their health and healthcare, it allows them to develop a contextualized care plan that addresses these contextual factors. These red flags come both from the patient directly and from the medical record.

This research examined whether the use of clinical decision support (CDS) can be used to alert clinicians to both red flags and factors to avoid contextual errors.

The specific aim of the research was as follows:

• Design and implement contextualized CDS and assess through a randomized controlled intervention trial.

A CDS intervention was developed represented by a contextual care box containing information from a patient questionnaire designed to capture contextual factors, as well as relevant information about the patient from the electronic health record (EHR). Passive and active interruptive alerts were used to direct clinicians toward a contextualized care plan and away from contextual error. The research team conducted two studies: a randomized controlled trial with real patients who agreed to carry hidden recorders during their visits, and a second with unannounced standardized patients (USPs) using four case scenarios. USPs are actors trained in a particular scenario that allow for isolation of a variable for study: in this case, the CDS. For the former study, the researchers looked at outcome measures with a proportion of red flags that resolved 6 months after the initial encounter, the numbers of flags probed, and the proportion of contextual factors addressed. For the USP study, the researchers looked at costs attributed to overuse, misuse, and underuse of medical services.

In the randomized controlled trial, contextual red flags were not more likely to resolve with the intervention, but there was an increase in asking patients questions about their circumstances, “contextual probing,” and the contextualization of care. In the USP study, the intervention did not increase the rate of probing for contextualized factors, although probing across study arms was associated with a greater likelihood of plans that had been contextualized. The CDS did increase contextualization of plans, over and above the effect of probing. The intervention encounters had a significantly higher proportion of contextualized items per case than controls.

The researchers concluded that further research is needed to determine if contextualized CDS leads to improved health outcomes, but that this study supports the use of CDS drawing on data from the patient prior to a visit and data from the EHR to improve the contextualization of care plans.