Identification of Patients with Low Life Expectancy (Massachusetts)

Project Details - Ended

Project Categories

Summary:

Clinical decision support (CDS) has been used to bring guidelines to clinicians at the point of care. However, CDS typically employs a one-size-fits-all methodology that may consider only a patient's age and gender, rarely taking into consideration a patient's comorbidities, medications, patient preferences, and importantly, life expectancy. Knowing a patient's life expectancy is crucial in caring for individuals, since benefits of many medical interventions, such as cancer screening or treatment of chronic illness, are predicated on patients living long enough for the interventions to have impact. While there are many published mortality prediction tools, they are infrequently integrated into electronic medical records (EMRs); when they are, data on functional status needed for the tools are rarely electronically available since these functional assessments are typically documented as part of narrative provider notes.

With a goal of identifying patients at high risk of death, this project will use a translational multidisciplinary approach and integrate two technologies--natural language processing and the artificial intelligence technique dynamic logic--to improve the ability to analyze narrative medical documents. From this, a high-fidelity model of risk of death will be created and then validated.

The specific aims of the project are to:

  • Determine whether a combination of artificial intelligence technology dynamic logic and NLP improves accuracy of identification of patients with low life expectancy (likely to die over the next 12 months) compared to the currently used methods. 
  • Determine whether the low life expectancy model developed on the general patient population is equally accurate in patients with chronic conditions on the example of patients with a) diabetes, b) hypertension, and c) osteoporosis. 
  • Develop open source software that will integrate dynamic logic and NLP to allow users to easily assess patients' life expectancy based on a combination of structured and narrative EMR data. 

Concepts and text characteristics will be abstracted from narrative notes in an estimated 550,000 patient records to identify data elements that are associated with an increased risk for low life expectancy. Algorithms will be used to identify subgroups of patients that share a set of characteristics associated with increased risk for death. The research team will look at two sub-populations of equal size: a development population that will be used to develop and test the algorithm for identification of patients with low life expectancy; and a validation population in which the developed algorithm will be formally validated. A secondary analysis will be done to identify patients who died within defined timeframes. The team will additionally validate the model with populations of patients in which intensive treatment is unlikely to bring benefit to patients with a life expectancy of 12 months or less: diabetes mellitus, hypertension, and osteoporosis.

The resulting model will be packaged into open-source software called the Limited Life Expectancy Predication Software, and made available to others to apply to their own datasets. The software will allow for the identification of patients at high risk for death in the analyzed timeframes and will give an output for each patient record ranging from 0 to 1 representing the calculated probability that the patient will die within the selected timeframe.

The goal of this project is to enable organizations and individual researchers to focus special attention on patients with limited life expectancy when comparing benefits of different treatment strategies.

This project does not have any related annual summary.
This project does not have any related publication.
This project does not have any related resource.
This project does not have any related survey.
This project does not have any related project spotlight.
This project does not have any related survey.
This project does not have any related story.
This project does not have any related emerging lesson.