Arrow
use cases

Predict Drug Safety

Predict drug adverse effects with Artificial Intelligence, increasing patient health and satisfaction.

Problem

Every year, Adverse drug reactions (ADRs) – unintended, harmful events attributed to intended medicine use – result in more than 750,000 inpatient injuries or deaths, affect nearly two million hospital stays, and directly result in over one million ED visits and 125,000 hospitalizations. ADRs are estimated to cause at least 10% of all admissions in older adults.

Size of the Problem

  • 750,000 inpatient injuries or deaths are attributed to ADRs annually (1).
  • 2x increased incidence of ADRs in older adults (3).
  • 3x older adults are three times as likely to die of ADRs (3).
  • 20-60% is the prevalence of PIM usage among older adults (7).

Why it matters

Every year, adverse drug reactions (ADRs) – unintended, harmful events attributed to intended medicine use – result in more than 750,000 inpatient injuries or deaths, affect nearly two million hospital stays, and directly result in over one million ED visits and 125,000 hospitalizations (1,2). Older adults are more vulnerable to ADRs due to aging-related kidney and liver changes that create increased sensitivity and exposure to pharmaceuticals. They experience ADRs twice as frequently as their younger counterparts, and are four times as likely to be hospitalized (3). ADRs are estimated to cause at least 10% of all admissions in older adults, and between 10–39% of hospitalized older adults will experience an ADR (4,5). They are also more likely to die from ADRs; a recent study found that fatal outcomes were reported approximately three times more often for older adults (6).

One reason for increased risk of ADRs is the use of Potentially Inappropriate Medications (PIMs), particularly for older adults who may be taking multiple medications. Polypharmacy, commonly defined as regular use of five or more medications, and the prevalence of PIMs are strongly associated with increased risk of ADRs in older adults. Alarmingly, the prevalence of PIMs ranges from 20–60% of all older adults depending on healthcare setting and criteria used to define inappropriate prescribing (7). PIM use is associated with a 10–30% increased risk of hospitalization, and older adults with polypharmacy are roughly 80% more likely to be hospitalized within a year relative to equivalent patients without polypharmacy (7,8).

PIMs and polypharmacy can result in considerable cognitive impairment consistent with dementia and may lead to misdiagnosis and further prescriptions, potentially adding to an already-elevated ADR risk. Despite this, opportunities for medication reconciliation and deprescribing are frequently missed. A recent study found that 66% of hospitalized older patients had at least one PIM prescribed at discharge, 49% continued a previously prescribed PIM, 31% were prescribed a new PIM during hospitalization, and ultimately 36% visited the ED, were rehospitalized, or died within 30 days of discharge (9).

Solution

  1. Predictive Analytics for Identifying Risk Factors: AI algorithms utilize complex data analysis techniques to study historical health data, identifying patterns that might not be obvious to human analysts. By incorporating patient-specific data such as age, existing medical conditions, and current medication regimes, these models can predict potential ADRs with greater accuracy. This capability enables healthcare providers to tailor treatment plans to individual patient needs, potentially reducing the likelihood of adverse effects. Additionally, the insights gained from predictive analytics can guide pharmaceutical companies in enhancing drug safety profiles during the development phases (7).
  2. Real-Time Drug Interaction Alerts: AI-driven systems analyze the combination of medications prescribed to a patient to predict potential interactions that could lead to ADRs. By integrating with electronic health records (EHRs), these systems provide immediate feedback to healthcare providers at the point of prescribing. This proactive approach not only reduces the incidence of ADRs but also educates providers on safer prescribing practices, ultimately fostering a safer healthcare environment. Moreover, these alerts are dynamically updated as new data becomes available, ensuring that healthcare providers are informed about the latest drug interaction risks (10).
  3. Development of a Predictive Model for ADR Risk Level: Our predictive model, trained on a synthetic but realistically modeled database, utilizes key variables such as age, gender, liver and kidney function, and the number of medications to forecast the ADR risk level categorized as low, medium, or high. This model aids healthcare providers in proactively identifying patients at risk of ADRs, allowing for personalized adjustments to treatment plans that mitigate these risks and improve overall patient outcomes.
Discover more and interact with our AI!

Datasources

  • Rx Claims: Data extracted from health insurance pharmacy claims with details about each medication and its type, fill dates, days supply, pharmacy location, and prescribing clinician.
  • Medical Claims: Data extracted from health insurance medical claims with details about dates and place of service, diagnosis codes, key procedures, use of medical equipment, and provider specialties.
  • Health Risk Assessments: Self-reported data from health questionnaires that assess a person’s individual medical history, health risks, lifestyle, health behaviors, and quality of life.

Citations

  1. Sarah P,, Slight, et al. “The national cost of adverse drug events resulting from inappropriate medication-related alert overrides in the United States.” Journal of the American Medical Informatics Association, Volume 25, Issue 9, Sep. 2018, pp. 1183-1188.
  2. “Adverse Drug Events.” Department of Health and Human Services: Healthcare Quality, 2 Feb. 2020. Health.gov. Accessed 23 Jun. 2021.
  3. Beijer, H J M, and C J de Blaey. “Hospitalisations caused by adverse drug reactions (ADR): a meta-analysis of observational studies” Pharmacy World 8: Science, vol. 24, no. 2, 24 Apr. 2002, pp. 46-54. doi:10.1023/a:1015570104121.
  4. Jennings, Emma, et al. “Detection and Prevention of Adverse Drug Reactions in Multi-Morbid Older Patients” Age and Ageing, vol. 48, no.1, 12 Sep. 2018, pp. 10-13, academic.oup.com/ageing/article/48/1/10/5123812, 10.1093/ageing/afy157. Accessed 23 Jun 2021.
  5. Parameswaran, Nair N, et al. “Hospitalization in older patients due to adverse drug reactions -the need for a prediction tool.” Clin Interv Aging. 2016,11:497-505. 2 May 2016. doi:10.2147/C1A.599097.
  6. Dubrall, Diana et al. “Adverse drug reactions in older adults: a retrospective comparative analysis of spontaneous reports to the German Federal Institute for Drugs and Medical Devices.” BMC pharmacology € toxicology vol. 21, no. 25, 23 Mar. 2020, doi:10.1186/540360-020-0392-9.
  7. Weir, Daniala L., et al. “Both New and Chronic Potentially Inappropriate Medications Continued at Hospital Discharge Are Associated With Increased Risk of Adverse Events.” Journal of the American Geriatrics Society, vol. 68, no. 6, 31 Mar. 2020, pp. 1184-1192, pubmed.ncbi.nim.nih.gov/32232988/, 10.1111/j9516413. Accessed 23 Jun 2021.
  8. Finkelstein, Joseph et al. “Pharmacogenetic polymorphism as an independent risk factor for frequent hospitalizations in older adults with polypharmacy: a pilot study” Pharmacogenomics and personalized medicine vol. 9, 14 Oct. 2016, pp. 107-116. doi:10.2147/PGPM.S117014.
  9. Fick, Donna M. “Less Really Is More in Inappropriate Medication Use in Older Adults: How Can We Improve Prescribing and Deprescribing in Older Adults?" Journal of the American Geriatrics Society, vol. 68, no. 6, 4 May 2020, pp. 1175-1176, onlinelibrarywiley.com/doi/full/10.1111/jgs:16485, 10.1111/j95.16485. Accessed 23 Jun 2021.
  10. Gray, Shelly L., et al. “Meta-Analysis of Interventions to Reduce Adverse Drug Reactions in Older Adults." Journal of the American Geriatrics Society, vol. 66, no. 2, 19 Dec. 2017, pp. 282-288, pubmed.ncbi.nlm.nih.gov/29265170/, 10.1111/jg515195. Accessed 23 Jun 2021.

Book a Free Consultation

Trusted by the world's top healthcare institutions

Button Text