Audit of Interventions to improve discharge prescribing
Medications are frequently omitted from discharge prescriptions and discharge summaries in Saint James’s Hospital. Given the substantial body of evidence linking these omissions to adverse drug events and subsequent readmission, as well as inappropriate propagation of inaccurate prescribing in the community, these errors represent a source of preventable iatrogenic risk to patients.
Impact of an educational intervention
The educational intervention did not significantly impact discharge prescribing error rates. While there was a trend towards increased communication around medication changes observed in the Post-Intervention cohort, this was not statistically significant.
This lack of impact can be rationalised by the type of educational intervention used. Our educational intervention comprised of a Didactic Teaching Session and Information Bulletins. Both interventions rely on passive dissemination of information, and there is very little evidence to support this as an effective means of improving patient outcomes[24, 29].
Impact of EPR
The introduction of EPR significantly reduced discharge prescribing error, with a significant reduction in the primary outcome of proportion of PAM which are prescribed or discontinued with explanation on discharge. This may be explained by EPR design. Prior to introduction of EPR, less than half (41.94%) of interns always referred to admission MR when completing discharge prescriptions, and inpatient medication lists (including current prescriptions) were likely the only or primary source used for discharge prescribing. Following introduction of EPR, the discharge prescribing process integrated patient medication lists with admission MR. This ensures that the discharge prescriber is aware of medications which have been intentionally held, either during admission or on discharge, and this information was not readily available on inpatient medication lists.
The above findings are in line with other studies which examine impact of EPR systems on discharge prescribing error[24, 25]. Direct comparison between studies is often limited by inter-study variability, although a trend towards improved prescribing has been validated in a number of papers[25, 30].
Understanding results in the context of EPR design and prescribing habits
In order to rationalise findings, results were contextualised with similar studies, survey data, and discussion with the SJH Informatics Department[24, 25]. Despite improvement in patient safety, a similar study identified a new form of sociotechnical error that was unique to EPR prescribing[25]. Another study noted issues with EPR rigidity as a cause of error [31]. In this study, despite improvement in the primary composite outcome, issues were identified. Deliberate discontinuation of a medication on the discharge prescription and discharge summary was significantly reduced following introduction of EPR. Two explanations for this negative finding were identified.
A significant separation of process for discharge prescribing and discharge discontinuation was introduced with EPR. Prior to EPR, paper discharge prescriptions contained an integrated box in which prescribers were prompted to list and rationalise discontinued medications. Issues accessing documentation, and difficulty interpreting documentation were frequently cited by intern prescribers as sources of error in the survey conducted. On SJH EPR, there are now separate tabs for “Discharge Medication Reconciliation” and “Medication Changes and Comments”, and the latter tab is frequently not used when discharging patients. This separation of processes increases number of steps required to list and rationalise discontinued medications.
Access to documentation is important in identifying and rationalising medications which have been discontinued, and the EPR introduction may have reduced this accessibility. When completing the Medications Reconciliation tab, it is not possible to access clinical notes. When completing the Medications Changes and Comments tab, it is not possible to concurrently access any information from the patient’s chart. When asked to identify causes of prescribing an intentionally discontinued medication on discharge, interns frequently cited issues such as “discontinuation not clearly documented”, “poor admission notes” and “reason for discontinuation not clearly documented”. As demonstrated in Table 4, prescribers frequently rely on multiple sources of information when generating discharge prescriptions. Expecting prescribers to navigate in and out of these tabs may detract from the prescribing task at hand and increase the time taken to complete it. Of note, time pressure is frequently listed as a cause of prescribing error in survey data. As some pre-configured Electronic Patient Record Systems may not lend themselves to reconfiguration, local awareness of limitations of EPR, education of prescribers about these limitations and trials of work-around solutions have potential to improve patient care. It is essential to ensure such systems are audited appropriately.
Efforts to improve discharge prescribing should focus on making the relevant resources quickly available to the prescriber, in a manner which allows them to identify relevant information. Future quality improvement efforts should attempt to integrate clinical notes and input from senior colleagues (in addition to admission MR) in the discharge prescribing process. Providing teaching to non-consultant hospital doctors on this topic is unlikely to be adequate in developing a proactive service. The role of clinical pharmacists in medication safety is essential in all health systems, irrespective of e-health infrastructure. In the survey of intern doctors, they reported that medication omissions on admission were the most likely cause of omissions at discharge. This highlights the importance a pharmacist led MR on admission. Pharmacist-led medication reconciliation is effective at reducing discharge prescribing errors in health systems Electronic Patient Records are instituted [32]. This is a cost-effective method of reducing error, and clinicians should advocate for discharge MR implementation where possible[17].
Strengths and Limitations
Strengths of this study include assessment of two interventions in the same institution. The study is unique in examining interventions in terms of changes to hospital-wide clinical practice in a pragmatic setting, rather than surrogate markers like clinician performance in a standardised assessment. It assesses two interventions in the same institution, and includes an educational intervention which is passive in nature, reflecting the most widely available mechanism for addressing discharge prescribing issues.
Limitations of this study include variation between study groups. There was a relative reduction in number of PAM per patient in the Post-EPR study group, compared to the Pre-Intervention and Post-Education study groups. Given that number of PAM did not correlate significantly with primary or secondary outcomes, this did not likely alter our interpretation of the study results. Further, this study was conducted on a sample of 1,218 PAM (162 discharged patients) in one hospital system. Results may not be directly applicable to other hospital systems.