What factors predict potentially inappropriate primary care prescribing in older people? Analysis of UK primary care patient record database

Background: Potentially inappropriate prescribing (PIP) in older people has been identified as a substantial problem, but few large population-based studies have investigated the underlying factors that predict it. Objective: To: (i) examine trends in PIP in UK older primary care patients; and (ii) assess factors associated with PIP. Methods: An analysis of routine, anonymized, computerized patient records of 201 UK general practices providing data to the DTN-LINK database between 1996 and 2005, which included approximately 230 000 registered patients per year aged >= 65 years. The main outcome measures were the number of different drugs prescribed per patient annually and the percentage of patients prescribed a PIP drug (modified 2003 Beers criteria). These were assessed for all drugs, and then for selected sub-classes (analgesics, antidepressants and sedatives/anxiolytics). Results: Whilst the number of drugs prescribed per patient increased, the percentage of subjects receiving a PIP drug declined from 32.2% in 1996 to 28.3% in 2005, largely due to a fall in co-proxamol (dextropropoxyphene/paracetamol [acetaminophen]) prescribing. In 2005, female gender, being older, more socioeconomically deprived or in a care home were strongly associated with PIP. However, the number of drugs prescribed was strongly associated with these variables and the strongest predictor of PIP; adjusting for number of drugs dramatically reduced the strength of all other associations except female gender with PIP. Factors predicting PIP in drug sub-groups were similarly reduced when adjusted for polypharmacy. However, some age trends remained: in the oldest group (aged >= 85 years), PIP of analgesics was less likely (odds ratio [OR] = 0.70, 95% CI 0.66, 0.75) while PIP of antidepressants was more likely (OR = 1.39, 95% CI 1.28, 1.51). Conclusion: PIP amongst older people in the UK, although declining, remains at a high level. The association of PIP with age, deprivation and care homes is largely explained by the higher overall prescribing rates in these groups. The overall rise in prescribing emphasizes that polypharmacy does not necessarily increase PIP and attempts to reduce PIP by focusing on polypharmacy have not been successful. Reductions in PIP have previously been achieved by introducing national guidelines (e.g. co-proxamol), but might also be achieved by alerting practitioners at the point of prescribing.