(P-128) Comparisons of Effect Estimates of 12 Prescribing Cascades Using Prescription Sequence Symmetry Analysis

Background: Prescribing cascades have long been considered as important clinical issue in older adults because they may increase use of potentially inappropriate medications and subsequent adverse outcomes. However, population-level effect estimation of prescribing cascades are relatively scarce. Particularly, most of existing studies were limited to single pair of prescribing cascade and were unable to do comparisons of effect estimates across distinct prescribing cascades.

Objectives: Using Taiwan’s National Health Insurance Research Database (NHIRD), this study aims to compare the effect estimates of 12 prescribing cascades from prescription sequence symmetry analysis (PSSA).

Methods: We analyzed 14 PSSA sets of index drugs and marker drugs, including two confirmatory analyses to check the validity of PSSA and 12 exploratory analyses, which have been reported as suspected prescribing cascade in the previous studies. In the confirmatory analysis, we selected amiodarone induced hypothyroidism, a well-documented adverse event, as positive control and amiodarone induced allopurinol prescription as a negative control. In each cohort of PSSA set, we identified patients who both newly initiated index drug and marker drug in the outpatient setting between 2014 to 2016 from the NHIRD. To identify the new user of index drugs and marker drugs, individuals who had ever received index drugs or marker drugs in an outpatient or inpatient setting in the previous year or on the same date were excluded. The observation period for sequences of incident marker drug use was restricted to 1 year before and after incident index drug use to reduce potential impact from within-subject confounding in a long study period. The results of PSSA were performed as the crude sequence ratio (SR) and adjusted SR (aSR) with 95% confidence intervals (CI).

Results: In the confirmatory analyses, amiodarone was significantly associated with hypothyroidism (aSR: 3.77, 95% CI: 3.43-4.14). In the contrast, amiodarone was not associated with acute gout (aSR: 0.83, 95% CI: 0.76-0.90). In the exploratory analyses, among 12 potential prescription cascade related drug pairs, 9 of them reached statistical significance and the aSR were ranged from 1.02 to 1.46. Dihydropyridine calcium channel blockers induced edema showed highest aSR (aSR: 1.46, 95% CI: 1.45-1.48) and statins induced muscle pain showed lowest aSR (aSR: 1.02, 95% CI: 1.02-1.03) within these results.

Conclusions: The effect estimates of the 12 potential prescribing cascade varied and some of them did not reach statistical significance. More studies are warranted to investigate whether these prescribing cascade link to clinical outcomes.