(133) Prescription Sequence Symmetry Analysis to Identify Prescription Cascades for Dementia Drug Treatments in the United Kingdom

Background: Dementia drug treatments have adverse drug events (ADEs) such as nausea, diarrhea, insomnia and hallucinations meaning these patients are vulnerable to prescribing cascades. Prescription sequence symmetry analysis (PSSA) is a signal detection method for ADEs, but its capacity to detect them in the context of dementia treatments has not been tested.

Objectives: To investigate potential prescription cascades for dementia drug treatments.

Methods: We performed PSSA using data from the United Kingdom (CPRD GOLD) between 2005 and 2022. We included patients with prescriptions for newly initiated index drugs (i.e. medications potentially associated with ADEs) and marker drugs (i.e. medications that can be used to treat ADEs). Index drugs were acetylcholinesterase inhibitors (AChE) (donepezil, rivastigmine and galantamine) and marker drugs were antiemetics, antiepileptics, antipropulsives and psycholeptics. Drugs were identified based on ATC 4th level codes. Patients with < 1 year of prior history at the earliest marker/index prescription or those who received index and marker drugs on the same date were excluded. To assess the validity of the PSSA results, we included a positive (amiodarone to thyroxin) and a negative (amiodarone to allopurinol) benchmarks, and a positive control (AChE inhibitors to memantine). The look-back time to exclude prevalent users and the maximum time allowed between the start of the index and the marker medications was set at 1 year. Adjusted sequence ratios (aSR) taking into account prescribing trends were calculated, with an aSR>1 indicating a prescription cascade. PSSA analysis was performed using the CohortSymmetry R package (GitHub - oxford-pharmacoepi/CohortSymmetry).

Results: Antiemetrics showed an aSR of 2.30 (95% CI 1.48-3.64) for A04AA and 1.50 (1.32-1.73) for A04AD. Antipropulsives (A07DA) showed an aSR of 1.71 (1.58-1.84). For psycholeptics, N05AA, N05AB and N05AC (combined), N05AD, and N05AL, showed an aSR of 1.15 (1.07-1.247), 1.49 (1.30-1.72), and 1.22 (1.04-1.44), respectively, whereas N05AF, N05AN, and N05AH showed an aSR < 1. The only antiepileptics with an aSR>1 were N03AG (1.55 [1.25-1.92]). Positive and negative benchmarks showed an aSR of 8.60 (7.86 to 9.43) and 0.97 (0.85-1.12), respectively. The positive control showed an aSR of 9.19 (8.13-10.4).

Conclusions: Our results show antiemetics, antipropulsives, and certain psycholeptics and antiepileptics are potential prescription cascades in those being prescribed dementia drug treatments. This study supports the application of PSSA to detect prescribing cascades for dementia drug treatments.