The impact of polypharmacy on health outcomes in older patients: a retrospective cohort study using CPRD database.

Application Number
Lay Summary

This study tries to find out if the use of multiple medications, taken by patients aged 75 years and older, is independently related with adverse health outcomes, such as hospitalisations, unwanted reactions to medication, falls, and a decrease in survival, over a period of one and five years. Given that up to 40% of 65 years and older individuals are on 5 or more medications, this problem is a public health issue. We will analyse the data from primary care practices and determine the number of medications the patients are on. Then, we will study if those patients with more than four, seven and ten medications suffer adverse health outcomes more frequently than those without polypharmacy. In the analyses, we will account for the number and severity of diseases that patients suffer from, and also for the degree of dependence and memory problems the patients have, because these characteristics, and not just the number of medications, could be related with adverse health outcomes.

Technical Summary

The main objective is to analyse the relationship between polypharmacy and one and five-year mortality, hospitalisations, adverse drug reactions and falls requiring hospital admission, adjusted for relevant confounders in patients aged 75 years and older. Secondary objectives are to analyse potentially inappropriate prescribing, and to study factors related with polypharmacy. These objectives will be achieved through a cross- sectional and retrospective cohort study. We will retrieve the number of drugs prescribed at the index record date. Polypharmacy will be defined by different cut-off points of 5, 8 and 11 prescribed drugs. Potentially inappropriate prescribing will be assessed with the Screening Tool Of Older People's Prescriptions (STOPP) criteria, and under-prescribing with the Screening Tool To Alert To Right Treatment (START) criteria. Study outcomes include mortality, urgent hospitalisations, adverse drug reactions (ADR) and falls requiring hospital admission. Confounders will include age, sex, previous falls, previous hospitalisations, comorbidities, cognitive impairment and the e-frailty index. Relationship between polypharmacy and outcomes will be summarised with odd ratios (OR) and hazard ratios (HR) with their 95% confidence interval obtained from logistic and Cox regression analysis adjusted for confounders. To avoid over-adjusting, we will include in models only those comorbidities not included in e-frailty index.

Health Outcomes to be Measured

Hospitalisation; Adverse drug reactions; Mortality; Falls.


Carlos Rodriguez Pascual - Chief Investigator - University Of Lincoln
Carlos Rodriguez Pascual - Corresponding Applicant - University Of Lincoln
Irene Boateng - Collaborator - University Of Lincoln
Paul Grassby - Collaborator - University Of Lincoln
Zahid Asghar - Collaborator - University Of Lincoln


HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data