Getting Pabrinex Right: A Multi-cycle Quality Improvement Study Reducing Waste in Thiamine Prescribing at a Scottish District General Hospital

Abstract

Background Patients with alcohol use disorders (AUD) are at increased risk of Wernicke's encephalopathy (WE) due to thiamine deficiency. At our centre all admitted patients with AUD are assessed for signs, symptoms and risk factors for WE. Such patients are prescribed oral or intravenous (IV) thiamine replacement in accordance with local guidelines; choice of dose and route of administration is determined based on the clinical features present at presentation. Failure to appropriately treat patients with WE may lead to potentially avoidable progression to Korsakoff syndrome. However, unnecessary use of IV formulations, high doses and extended durations of treatment can contribute to unnecessary administrations of IV thiamine, generating financial, environmental and workload costs. In the UK, IV thiamine is usually given in the form of compound vitamin B and C solution (Pabrinex). Aim This quality improvement (QI) project aims to quantify and reduce unnecessary prescription of IV Pabrinex in patients with AUD admitted to a district general hospital using a multi-cycle plan-do-study-act (PDSA) methodology. Materials and methods This retrospective study involved 106 patients admitted over a four-month period. Two education-based prescriber-targeted interventions were performed. Data collection was performed at baseline and after each intervention. Recommended thiamine dose was calculated by reviewing patients' admission documentation for presenting signs, symptoms and demographics. These data were then cross-referenced with local guidelines on the management of patients admitted with AUD to determine the recommended thiamine dosing regimen. The total number of IV Pabrinex doses recommended by guidelines was compared to the actual number prescribed for each patient to determine potentially unnecessary "excess" doses. Primary outcome measures recorded were the number of excess doses prescribed and proportion of patients receiving the correct initial dose. Results Baseline data showed 60% patients had received excess doses of Pabrinex, with a median of 13 excess doses per patient. This represented an excess cost of £35 per patient for medication alone. Following first and second interventions, median excess doses per patient in subsequent cycles fell to seven and then to 4.5, a significant reduction from baseline (p=0.0087). Proportion of patients receiving correct initial prescription showed a non-significant rise from 30% to 51.3% (p=0.12). Conclusions A targeted four-month multi-cycle QI project was carried out to quantify and reduce thiamine overprescription among AUD patients in a Scottish district general hospital. Interventions successfully reduced excess Pabrinex doses given per patient. This study highlights the potential for overprescription in the context of thiamine replacement and findings suggest that low-cost prescriber education strategies can be used to improve practice in this setting. Limitations included the single-centre design and short follow-up period, which limit generalisability. Clinical outcome measures were not recorded here but would be important to consider in future work. Caution should be exercised when aiming to reduce thiamine overprescription to mitigate the risk of underdosing patients who warrant treatment. Following completion of this project, a revised guideline has been implemented for this health board with simplified dosing recommendations, which may yield further improvements in practice.

Keywords: cost reduction; district general hospital; overprescription; pabrinex; prescriber; prescription; quality improvement; thiamine; waste.

Figure 1. Infographic-style poster developed during the first intervention cycle of this project.

Image credits: T Noteman. Information presented is based on regional guideline [5].

Figure 2. Median excess doses (pairs of vials) of Pabrinex prescribed per patient in the baseline, post-intervention 1 and post-intervention 2 data collection periods

Our baseline data showed a median of 13 excess doses prescribed per patient. Following one intervention this fell to seven and after the second intervention fell further to 3.5. ** p-value <0.01. Error bars show standard error of data for each collection period.

Figure 3. Weekly run chart showing median excess Pabrinex doses (pairs of vials) prescribed per patient in the baseline, post-intervention 1 and post-intervention 2 data collection periods

Following interventions, there was evidence of a reduction in excess Pabrinex prescribing. At baseline, a weekly median excess Pabrinex doses prescribed per patient of >15 was observed in three weeks out of five. In the post-intervention periods, weekly median did not surpass 11 at any point.

Figure 4. Proportion of patient prescribed he correct initial dose of thiamine replacement according to local guidelines in the baseline, post-intervention 1 and post-intervention 2 data collection periods

Proportion of patient prescribed correct initial dose of thiamine rose from 12/40 (30%) at baseline, to 9/27 (33%) after first intervention and 20/39 (51.3%) after intervention 2. There was no statistically significant difference between the three groups (X²=4.2074, p=0.12)