gms | German Medical Science

Background: Elderly patients are at higher risk for PIM prescriptions compared to younger persons due to necessary medications for multiple chronic conditions and a wider range of drugs which are potentially inappropriate based on a physiologic decline in all steps of pharmacokinetics and -dynamics. Analyses exploring the association between PIM prescriptions and hospitalization rate revealed contradictory results.

Materials and Methods: PRISCUS criteria were used to define PIM prescriptions. We identified a cohort of 392,337 family physician-patients (60.7% female) in Baden-Württemberg who were aged ≥65 years (mean age 73.8 years, median age 73 years, the last quarter of the patients were 78 years of age or older) and had initiated treatment with PIMs between January 2009 and December 2010. A Cox proportional hazards model was used to compare rates of all-cause hospital admissions within180 days after treatment initiation. Treatment initiation was defined by the dispense date of the first PIM (study drug) or the first non-PIM (comparator drug) within the observational period of two years. We identified incident PIM users and incident hospitalizations in our claims data with a look-back period of half a year (the second half of 2008) free from any PIM dispenses or hospital stays. Non-PIM users were defined as patients who take medication out of the same drug groups as the PIMs according to ATC code (3-7 digits depending on the medication in question). We adjusted for potential confounders (age, sex, ethnicity, need of care, participant in disease management programs, number of ICD10-codings) measured in the half year before treatment initiation.

Results: A total of 92,243 (23.5%) incident PIM users and a total of 47,470 (12.1%) hospitalizations within 180 days after treatment initiation (dispense date of PIM or non-PIM) were observed among our 392,337 cohort patients. Most PIM users (58.4%) received a PIM prescription in one quarter of the year 2009 or 2010. Another 19.3% got PIM prescriptions in two quarters, and the remaining 22.3% in three or more quarters. Even after adjusting for confounders the hospitalization risk following the treatment initiation (PIM or non-PIM) is elevated in PIM users (hazard ratio (HR) = 1.397, 95% CI 1.368-1.427, p < 0.001) compared to non-PIM users.

Conclusion: The frequency of incident PIM prescriptions for older family-physician patients is not unexpected to this extent taking into account already published data for the prevalence of PIM use in older patients. Our results showed an increased risk of hospitalization for PIM users compared with non-PIM users.