6A-05 - Application of Signal Detection Theory to Identify the Association between Potentially Inappropriate Medications and Unplanned Hospitalization: A Case-case-time-control Study

Background: Potentially inappropriate medication (PIM) increases the risk of adverse drug reactions, which can further contribute to unplanned hospitalizations. Since many of these hospitalizations are avoidable, identifying PIMs that require monitoring is essential.

Objectives: This study aims to investigate the association between PIM and the risk of unplanned hospitalization in the older population and to identify high-risk criteria that require special attention.

Methods: A case-case-time-control study was conducted using the Taiwan National Health Insurance Claims Database from 2016 to 2022. Patients aged 65 years or above with a record of unplanned hospitalization were included as cases. Unplanned hospitalization was identified for patients with records of emergency visits and hospitalization on the same day, and the event date was defined as the index date (day 0). Each case was assigned a hazard period (day -1 to -30) and a reference period (day -61 to -90), and 399 criteria in PIM-Taiwan were applied to identify PIM. The association between PIM and unplanned hospitalization was assessed by conditional logistic regression and presented as odds ratios (ORs) with 95% confidence intervals (95%CIs). A false discovery rate was adopted to adjust p values for each criterion to avoid type-I errors due to multiple testing.

Results: Overall, 557,927 cases were included with a mean age of 77.6 years on the index date, and 13.8% and 12.3% of them were diagnosed with neurologic disorders (e.g., dementia or cognitive impairment) or respiratory diseases (e.g., asthma). The average number of prescriptions was 10.1 in the hazard period and 7.7 in the reference period. Compared to the reference period, the hazard period was more likely to receive PIM in both case-crossover (OR: 2.02, 95%CI: 2.00–2.04) and control-crossover analysis (OR: 1.05, 95%CI: 1.04–1.06), resulting in a case-case-time-control OR of 1.93 (95%CI 1.90–1.95). Among the 399 criteria, 50 criteria showed a significant association, and the prescribing of diclofenac (OR 1.58, p < 0.001), ibuprofen (OR 1.79, p < 0.001), and ketorolac (OR 3.55, p < 0.001) were the top 3 criteria associated with an increased risk of unplanned hospitalization.

Conclusions: PIM was associated with an increased risk of unplanned hospitalization, with NSAIDs prescribing indicating the strongest association. Further studies adjusting unmeasurable confounders are needed to validate the association and to facilitate clinical evidence implication.