Epidemiology of Suboptimal Prescribing in Older, Community Dwellers A Two-Wave, Population-Based Survey in Dicomano, Italy

Background: Because of co-morbidity, older persons are often exposed to use of an excessive number of drugs, which per se implies also use of inappropriate drugs or of potentially interacting drugs ('suboptimal prescribing'). Time trends of suboptimal prescribing in older, community dwellers have been poorly investigated, particularly in Italy. Objective: To evaluate the time-course modification of suboptimal prescribing in older, community dwellers. Methods: We conducted a study on an Italian cohort of older (aged >= 65 years), community dwellers for whom data were collected in a two-wave (1995 and 1999), population-based survey. Suboptimal prescribing was defined as occurrence of polypharmacy (>= 5 medications), prescription of inappropriate medications (according to 1991 Beers' criteria) and prescription of potentially interacting drugs (as identified by the Micromedex (R) Drug-Reax (R) system). All outcome variables were modelled as continuous and dichotomous. Results: In 568 participants (59.9% women, mean standard error age 72.7 +/- 0.2 years), polypharmacy and potentially interacting drugs were more prevalent in 1999 than in 1995, while prevalence of inappropriate drugs was lower in 1999. The proportion of participants receiving polypharmacy was nearly 3-fold greater in 1999 than in 1995 (21.6% vs 8.8%; p<0.001). After adjustment for disability, coronary artery disease/stroke, heart failure and other co-morbidities, polypharmacy was twice as prevalent in 1999 as in 1995, with a mean increase of 0.5 drugs per participant. In contrast, in models adjusted for the same variables and also for polypharmacy, inappropriate prescribing was reduced (60% lower prevalence, 0.06 mean reduction in prescriptions per participant) in 1999 compared with 1995. In multivariable models, no change was observed in the number of potentially interacting prescriptions. Conclusions: Between 1995 and 1999, the number of prescriptions received by a cohort of older, community dwellers increased substantially, even after controlling for co-morbidity. On the other hand, when controlling for polypharmacy, other indicators of suboptimal prescribing remained unaffected or even decreased, suggesting that even increasing levels of polypharmacy do not necessarily imply other forms of suboptimal prescribing among older persons.