Browse Prior Art Database

A Process and its Application for the Construction of Residual Quantitative Medical Models to Assess the Adherence, Effectiveness, and Accuracy of Doctors' Treatment Behaviors

IP.com Disclosure Number: IPCOM000248367D
Publication Date: 2016-Nov-21
Document File: 3 page(s) / 88K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to assess the effectiveness of doctors' treatment behaviors and with what level of accuracy those behaviors conform to existing medical knowledge.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 45% of the total text.

Page 01 of 3

A Process and its Application for the Construction of Residual Quantitative Medical Models to Assess the Adherence, Effectiveness, and Accuracy of Doctors' Treatment Behaviors

All doctors treat patients differently. The most obvious example of this is that doctors tend to specialize in specific medical sub-fields and a doctor's knowledge in a specific field is extensive and up-to-date. Even for doctors in the same field, however, past experiences, education, and biases affect treatment applications.

Prior to the recent (and ongoing) adoption of electronic medical record (EMR) tracking systems, it was difficult to assemble large amounts of data on doctors' practices. Further, the computing power and techniques necessary for such analysis has only recently begun to accommodate the volume of the datasets. This data includes both structured (e.g., lab tests) and unstructured (e.g., doctors' notes) components.

Additionally, the analysis of large databases of structured and unstructured medical knowledge is now a reality.

Efforts are currently underway to analyze the medical literature for relationships between drugs and health problems. For example, one such algorithm, Distributional Semantics, compares the distributions of the semantic usage of medical terms to establish such relationships. These relationships can only help to explain the data being analyzed, however. Applying these same methods to EMRs provides insight into

how doctors are treating patients. This can be aggregated at not only the doctor level, but also at the level of an entire partnership, department, or hospital as well.

A method is needed to assess the effectiveness of doctors' treatment behaviors and

with what level of accuracy those behaviors conform to existing medical knowledge.

Additionally, these assessments should be used to educate doctors. This feedback loop will improve patient care.

The novel contribution is a method of doctor assessment that:


• Constructs relationships between medical terms given various aggregation levels of medical literature, including EMRs


• Computes residual relationships across aggregations (e.g., a hospital's EMRs vs. a medical corpus)


• Provides analysis of these residual relationships to suggest where treatments may be better or more accurately applied. Analysis could also suggest future

directions of research (e.g., where treatments are effective but not documented

in the literature)

• Provides feedback to doctors regarding ways in which to improve patient care

Additional residual relationships could be generated by comparing historical "point in

time" medical or EMR corpora to current corpora to assess how quickly doctor treatments are adapting to new medical knowledge. This lag is likely to be different for different doctors.

Additional analysis could be applied to assess the sensitivity of medical treatments to

the precision of their application.

Current quantitative medical corpora analysis focuses on answering q...