Is hospitalisation associated with the unintentional discontinuation of appropriate long-term medication in the GP record?

Problem

Medication errors at transitions of care are known to be extremely common. Previous studies have reported prescription errors, most commonly omissions of appropriate medications, occur at hospital discharge. The objective of this study was to determine the association between the discontinuation of evidence based long term medication in general practice prescriptions and hospitalisation by comparing the discontinuation rate to a non-hospitalised group.

Approach

A retrospective cohort study was undertaken of patients ≥65 years in 44 Irish general practices using the Irish Primary Care Research Network. Data were gathered on prescribing from the GP electronic record and hospitalisation exposure via Healthlink discharge message notification. Four cohorts were defined by patients continuously prescribed antiplatelets/anticoagulants, statins, thyroid medication and respiratory inhalers over a one-year period. The outcome of discontinuation was defined as the absence of the specified chronic medication in the six-months post discharge for the hospitalised group, and the absence of the specified chronic medication over a random six-month period post enrolment for the non-hospitalised participants. Multilevel logistic regression was used for all analyses, with adjustments made for relevant confounders such as age, number of medications prescribed, multi-morbidity, and potentially inappropriate prescribing (PIP).

Findings

A total of 19,914 patients were enrolled (3,905 (19.6%) of whom had experienced hospitalisation). The odds of discontinuation were not significant for three of the predefined groups (antiplatelets/anticoagulants (OR 0.91, 95%CI (0.77, 1.08), p=0.289); statins (1.04, 95%CI (0.87,1.24), p<0.662; thyroid medications (OR0.69, 95%CI (0.44, 1.10)) comparing those who had been hospitalised to those who had not been hospitalised. The odds of discontinuation in the respiratory inhalers group were less likely in the hospitalised group (0.60, 95%CI (0.45, 0.79), p<0.001). These effects remained after adjusting for confounding variables. Increasing age (p<0.001) was also associated with the odds of discontinued medication.

Consequences

Unintentional omission of chronic medications commonly occurs on hospital discharge prescriptions. These have the potential to impact a patient's health following discharge if not subsequently identified and corrected. This study examined the likelihood of discontinuity of chronic medications in the GP record post hospitalisation. The study’s findings questions the heretofore assumed significance of hospitalisation, and potential discharge prescription omissions, in the disruption of patients’ long term medication regimes.