ClinicalTrials.gov
ClinicalTrials.gov Menu

AI for Alcohol Misuse in the EHR

The safety and scientific validity of this study is the responsibility of the study sponsor and investigators. Listing a study does not mean it has been evaluated by the U.S. Federal Government. Know the risks and potential benefits of clinical studies and talk to your health care provider before participating. Read our disclaimer for details.
ClinicalTrials.gov Identifier: NCT03833804
Recruitment Status : Not yet recruiting
First Posted : February 7, 2019
Last Update Posted : February 7, 2019
Sponsor:
Information provided by (Responsible Party):
Majid Afshar, Loyola University

Brief Summary:
The investigators propose to develop an open-source, publicly available AI software package that health systems could download and apply to their electronic health record data marts to screen for alcohol misuse in their patients. The investigators hypothesize that the natural language processing algorithm can provide a standardized and interoperable approach for an automated daily screen on all hospitalized patients and provide better implementation fidelity for screening, brief intervention, and referral to treatment.

Condition or disease Intervention/treatment
Alcohol Drinking Alcohol; Harmful Use Alcohol Abuse Trauma Other: Processing of clinical notes in the EHR data collected during routine care

Detailed Description:

Alcohol misuse accounts for 1 in 10 deaths in the United States and rates of misuse have risen 9% between 2002 and 2012 with 18% of United States population reporting a pattern of regular binges. As many as 33% of patients with trauma encounters have alcohol misuse.3 This burden can be reduced, however - screening, brief intervention, and referral to treatment (SBIRT) programs reduce trauma recidivism by nearly 50%.

The National Academy of Sciences has recommended the Alcohol Use Disorders Identification Test (AUDIT) for screening in the electronic health record (EHR). However, the investigators performed a survey across multiple health systems in Chicago and found few are using the AUDIT and substantial heterogeneity exists in the format of data capture for alcohol. At Loyola University's Trauma Center which uses the AUDIT in its SBIRT program, full-time dedicated screeners completed the screen in 56% of patients arriving in the emergency department. The most common reason for missing patients was inability to screen during off-duty hours.

The clinical narrative in the EHR captured during routine care is a rich source of information that is not currently used in a formal manner to screen for alcohol misuse. For example, the social history in the providers' notes typically incorporate patients' answers to questions about alcohol use, but there are no automated methods currently available to make use of these data. Natural Language Processing (NLP) and machine learning are subfields of artificial intelligence (AI) that provide powerful tools to analyze unstructured data, such as free text, in the EHR. NLP is a set of computational methods for deriving meaning from human-generated texts that feeds into machine learning algorithms to learn and predict. The role of NLP for case identification in alcohol misuse is in its infancy; no peer-reviewed publications have yet examined NLP for this purpose.

The investigators hypothesize that the NLP algorithm can provide a standardized and interoperable approach for an automated daily screen on all hospitalized patients and provide better implementation fidelity for SBIRT.

The aim is to prospectively test the effectiveness of the NLP algorithm as a pre-screen tool for alcohol misuse in a trauma center. The investigators will use a pre-post comparison design at a Level I Trauma Center. After 12 months of using the NLP tool for daily automated pre-screens, the investigators will assess implementation fidelity of alcohol screening at the index hospital and examine health outcomes


Study Type : Observational
Estimated Enrollment : 1504 participants
Observational Model: Cohort
Time Perspective: Prospective
Official Title: Artificial Intelligence for the Identification of Alcohol Misuse From the Electronic Health Record
Estimated Study Start Date : January 2021
Estimated Primary Completion Date : March 2023
Estimated Study Completion Date : April 2023

Group/Cohort Intervention/treatment
Conventional pre-screen
Two nurses using routinely collected blood alcohol laboratory data on all trauma encounters to a Level I trauma center to identify individuals at-risk for alcohol misuse to receive standard-of-care screening, brief intervention, or referral to treatment (SBIRT) intervention.
NLP (natural language processing) pre-screen
Automated processing of clinical notes collected during routine care in first 24 hours of trauma admission to identify individuals at-risk for alcohol misuse to receive standard-of-care screening, brief intervention, or referral to treatment (SBIRT) intervention.
Other: Processing of clinical notes in the EHR data collected during routine care
Clinical notes collected in the first day of trauma admission during usual care as input to natural language processing and machine learning algorithm.




Primary Outcome Measures :
  1. Proportion of patients that received SBIRT (screening, brief intervention, or referral to treatment) [ Time Frame: 12 months ]
    The primary outcome is the proportion of patients who received SBIRT after pre-screening positive for being at-risk for alcohol misuse.



Information from the National Library of Medicine

Choosing to participate in a study is an important personal decision. Talk with your doctor and family members or friends about deciding to join a study. To learn more about this study, you or your doctor may contact the study research staff using the contacts provided below. For general information, Learn About Clinical Studies.


Ages Eligible for Study:   18 Years to 89 Years   (Adult, Older Adult)
Sexes Eligible for Study:   All
Accepts Healthy Volunteers:   No
Sampling Method:   Non-Probability Sample
Study Population
Adult emergency department encounters for trauma care at a Level I trauma center
Criteria

Inclusion Criteria:

  • Ages 18 years old to 89 years old
  • Primary trauma admission
  • Length of stay greater than 24 hours

Exclusion Criteria:

  • Cannot participate in the usual care SBIR intervention
  • Death or obtunded during first 24 hours of admission
  • Discharged against medical advice
  • Transferred from another acute care hospital
  • Transferred to another acute care hospital

Information from the National Library of Medicine

To learn more about this study, you or your doctor may contact the study research staff using the contact information provided by the sponsor.

Please refer to this study by its ClinicalTrials.gov identifier (NCT number): NCT03833804


Contacts
Contact: Majid Afshar, MD 7083279017 majid.afshar@lumc.edu

Sponsors and Collaborators
Loyola University

Additional Information:
Publications of Results:
Responsible Party: Majid Afshar, Physician-Scientist Assistant Professor, Loyola University
ClinicalTrials.gov Identifier: NCT03833804     History of Changes
Other Study ID Numbers: LU211165
First Posted: February 7, 2019    Key Record Dates
Last Update Posted: February 7, 2019
Last Verified: February 2019
Individual Participant Data (IPD) Sharing Statement:
Plan to Share IPD: Yes
Plan Description: The patient data are protected health information and unavailable to public but the algorithm will be shared. The investigators will serialize our best models developed using either pickle (a Python native mechanism for object serialization) or joblib (https://pythonhosted.org/joblib/) and write software that will be capable of reloading them and making predictions. The software will be distributed via github.com or similar web-based software hosting service.
Supporting Materials: Analytic Code
Time Frame: 12 months after completion of study and available for at least five years on github.com
URL: http://github.com

Studies a U.S. FDA-regulated Drug Product: No
Studies a U.S. FDA-regulated Device Product: No

Keywords provided by Majid Afshar, Loyola University:
natural language processing
machine learning
artificial intelligence
clinical decision support
alcohol misuse
unhealthy alcohol use
excessive alcohol use
trauma patient

Additional relevant MeSH terms:
Alcoholism
Alcohol Drinking
Alcohol-Related Disorders
Substance-Related Disorders
Chemically-Induced Disorders
Mental Disorders
Drinking Behavior
Ethanol
Anti-Infective Agents, Local
Anti-Infective Agents
Central Nervous System Depressants
Physiological Effects of Drugs