Arrow
use cases

Predict Opioid Abuse

Prevent patients from consuming an overprescribed amount of opioids and predict which patients are likely to suffer from opioid abuse

Problem

Improve the care you offer by predicting which patients with opioid prescriptions are likely to suffer from opioid abuse. The opioid epidemic is a public health crisis endangering countless lives. While many factors contribute to this crisis, main drivers of the epidemic include overprescription and misuse of prescriptions. A study by Johns Hopkins which examined about 350,000 prescriptions given by 20,000 physicians revealed that on average all doctors overprescribed across seven common surgeries. As a result of this mass over prescription, patients fall at risk of continuing to abuse their prescriptions beyond the necessary amounts.

Size of the Problem

  • 1.7 million people suffer from substance abuse disorders related to opioid medication.
  • Every day, 128 of these people die from an opioid-related overdose.
  • Opioid overdoses grew by 30% across 45 states in the US.

Why it matters

The opioid epidemic has taken a devastating toll on the United States, resulting in an increasingly concerning number of fatalities. According to CBHSQ estimates, 1.7 million Americans suffer from substance abuse disorders related to prescribed opioids--and tragically 128 people perish each day due to overdose-related issues. The situation grows more dire by the year; data obtained from 45 states reveals that overdoses grew 30% between 2016 and 2017 alone. Overprescription and misuse are largely seen as primary drivers behind this crisis which Johns Hopkins research supports, finding that 350K prescriptions given by 20K doctors were generally overprescribed for 7 common surgeries on average.

Solution

Although the fight against the opioid epidemic is multifaceted, one effective way to stop patients from falling into substance abuse is to prevent them from consuming an overprescribed amount of opioids as early as possible. AI helps physicians improve the care they offer by predicting which patients with opioid prescriptions are likely to suffer from opioid abuse. These predictions can help inform physicians of the considerations that should be made prior to prescribing a specific quantity of opioid pain medication to a patient. A common challenge here is that not all providers may have historical data on which patients have suffered from opioid abuse. Targeted marketing techniques used by pharmaceutical companies have also been shown to influence physicians’ prescribing choices. In addition to identifying which patients are at risk of substance abuse, AI can also predict the likelihood that a physician will overprescribe opioid medications. This transparency enables intervention at the provider facility level in cases where a physician has a high propensity to overprescribe.

Datasources

  • Electronic Health Records (EHR): They contain detailed information about patients' medical histories, including opioid prescriptions and substance abuse-related diagnoses.
  • Prescription Drug Monitoring Program (PDMP): It records prescriptions for controlled substances, including opioids, allowing for tracking prescription patterns and detecting potential cases of abuse.
  • Substance Abuse Treatment Databases: These databases contain information about patients who have received treatment for substance abuse disorders, including opioids, providing data on abuse and recovery histories.
  • Healthcare Claims Databases: They contain information about medical insurance claims, which may include records of opioid prescriptions and substance abuse-related treatments.
  • Health Surveys and Epidemiological Studies: These sources provide population-level data on opioid consumption and the prevalence of substance abuse disorders, informing trends and patterns of abuse at the community level.

Citations

  1. Matero M, Giorgi S, Curtis B, Ungar LH, Schwartz HA. Opioid death projections with AI-based forecasts using social media language. NPJ Digit Med. 2023 Mar 8;6(1):35. doi: 10.1038/s41746-023-00776-0. Erratum in: NPJ Digit Med. 2023 Mar 17;6(1):45. PMID: 36882633; PMCID: PMC9992514.

  1. Gadhia S, Richards GC, Marriott T, et alArtificial intelligence and opioid use: a narrative reviewBMJ Innovations 2023;9:78-96.

  1. Pitfalls and solutions for using AI to predict opioid use disorder. (n.d.). https://www.fau.edu/newsdesk/articles/machine-learning-opioid-use-disorder.php

  1. Rutherford, G. (n.d.). Machine learning predicts risk of opioid use disorder for individual patients. https://www.ualberta.ca/folio/2022/12/machine-learning-predicts-risk-of-opioid-use-disorder.html

  1. McCue, M. (n.d.). Intelligent help for the Opioid Crisis lifecycle: The need for an Artificial Intelligence & Informatics-Based strategy to address the opioid crisis. New Jersey State Policy Lab. https://policylab.rutgers.edu/intelligent-help-for-the-opioid-crisis-lifecycle-the-need-for-an-artificial-intelligence-informatics-based-strategy-to-address-the-opioid-crisis/

Book a Free Consultation

Trusted by the world's top healthcare institutions