MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Errors and Data Mining in EMR

E Ford1*, S Terezakis2*, S Weintraub3*, (1) University of Washington, Seattle, WA, (2) Johns Hopkins Univ, Baltimore, MD, (3) Southcoast Hospitals Group, Fall River, MA




Presentations

(Thursday, 8/2/2018) 10:00 AM - 12:00 PM

Room: Karl Dean Ballroom A1

The electronic medical record (EMR) plays a key role in error in healthcare in two main ways. First, the EMR contains a wealth of information which may be used to identify errors and quality gaps. There are well-established methods for accomplishing this and new and emerging approaches as well. Second, the EMR itself can itself contribute to quality breakdowns and safety events. This is especially true in radiation oncology where there is a reliance on computer systems and a complex human-computer interface. This scientific-track symposium will examine these safety-related aspects of the EMR and will present new data on error patterns and new methods for data mining. Data mining of the EMR is an error-reduction technique that compliments other approaches such as incident learning or risk assessment, which, while valuable, require additional effort and are not incentivized by regulation or reimbursement. The challenge with EMR data mining is to extract information that is meaningful and timely. One method for this is the trigger indicator methodology in which the EMR is examined for specific “trigger� events which indicate that adverse event may have occurred. This has a long historical precedent in healthcare, and methods are now being extended to look beyond single isolated parameters into more complex and interrelated data in the EMR. Machine-learning has proven useful for this. However, none of these approaches are significant unless the health of patients can be improved. In this context, we will present one project in radiation oncology in which the EMR is being used to monitor health metrics of clinical importance. This allows quality gaps to be detected and addressed in real time. Finally the EMR itself may contribute to error. The human-computer interface in healthcare is not optimized for safety or quality and so it can itself drive error. We will review this in light of recent data from the national incident learning system in the US, RO-ILS®: Radiation Oncology Incident Learning System, where a pattern of errors has emerged which involve human factors engineering and the human computer interface. This data will help inform improvements in the EMR and also strategies that each clinic can employ to better utilize the EMR, thereby reducing error.

Learning Objectives:
1. Understand existing and emerging methods for analyzing electronic medical record (EMR) data to quantify gaps in the safety and quality of care.
2. Understand how the EMR can be used to continuously monitor the status of patients for clinically important endpoints.
3. Appreciate the role of the EMR and human-computer interaction in contributing to error.

Handouts

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email