How does potentially inappropriate prescribing measured by STOPP and START relate to healthcare utilisation in older people? A cohort study
Medicines management can be challenging in older patients due to physiological changes in ageing and the prevalence of multimorbidity and polypharmacy. This has led to a focus on potentially inappropriate prescribing (PIP) in this age group, including both potentially inappropriate medicines and potential prescribing omissions. Although a number of explicit measures of PIP have been developed, evidence of a link between process measures of care and patient outcomes is important. This study aims to determine the association between PIP in older patients and healthcare utilisation.
This prospective cohort study included community-dwellers aged ≥65 years from The Irish Longitudinal Study on Ageing (TILDA) with linked administrative pharmacy claims data who were followed up after two years. PIP was assessed in the 12 months preceding outcome measurement using the Screening Tool for Older Persons Prescriptions (STOPP) and Screening Tool to Alert doctors to Right Treatment (START). The outcomes investigated were number of emergency department (ED) visits and GP visits in the previous 12 months reported by participants at their follow-up interview. Multivariate negative binomial regression was used to determine the association of PIP with these outcomes, adjusting for demographics, number of medicines and chronic conditions, and baseline outcome measurement. Given the longitudinal design, the sensitivity of these results to time-dependent confounding was assessed using marginal structural models. Statistical significance at p<0.05 was assumed.
Participants followed up (n=1,753) were mainly female (54.5%), had a mean age of 76.5 years (standard deviation (SD) 6), a median of 6 regular dispensed medicines and reported a median of 3 doctor-diagnosed chronic conditions. PIP was detected in 57% by STOPP and 41.8% by START, and 21.7% of all participants reported an ED visit while 96.1% visited their GP (median 4 visits, IQR 2.5-6). Those with any STOPP criterion had higher rates of ED visits (adjusted incident rate ratio (aIRR) 1.30, 95% confidence interval (CI) 1.02-1.66) and GP visits (aIRR 1.15, 95%CI 1.06-1.24). Patients with ≥2 START criteria had significantly more ED visits (aIRR 1.45, 95%CI 1.03-2.04) and GP visits (aIRR 1.13, 95%CI 1.01-1.27) than people with no START criteria. Accounting for time-varying confounding did not significantly alter these results.
Both STOPP and START were independently associated with increased healthcare utilisation, supporting the suggestion that PIP has a marginally unfavourable risk-benefit ratio. Evidence of a causal link is important if process measures of medication safety, like PIP, are being targeted to improve patient outcomes. However, effect sizes were modest and the magnitude of this should be considered by healthcare professionals when prioritising issues to devote time to with patients. Easily incorporating review of PIP criteria into routine care, for example using computer decision support systems, may provide an efficient way of reducing adverse patient outcomes.