Drug information center: challenges of the research process to answer enquiries in hospital pharmaceutical practices

Analysis of queries to a Swedish drug information centre identifies scientific knowledge gaps

Drug Information Centres (DICs) are providing clinicians with evidence-based support for rational drug treatment. However, knowledge gaps in the literature may hinder DICs from offering optimal guidance. This study examined the extent and nature of these knowledge gaps and their impact on clinical pharmacological advice, using real-world query data from a Swedish regional DIC. Data from 2022 at the Sahlgrenska University Hospital were analysed, focusing on queries outside off-label prescriptions and pharmaceutical aspects. A qualitative text content analysis identified phrases indicating a lack of evidence. Responses were categorized by the presence of these signalling phrases and the presence or absence of knowledge gaps. Among 209 responses, the majority were patient-specific (79%), posed by physicians (95%), and often related to adverse effects (37%). The analysis revealed that 23% of the responses had scientific knowledge gaps, and in 18% of the responses, no clinical pharmacological advice could be provided. Knowledge gaps were particularly associated with queries on adverse effects, long-term medication safety, specific patient cases and comorbidities, drug comparisons, or patient populations with limited literature coverage. This analysis highlights the potential of DICs to identify unmet clinical needs in drug treatment and to promote research for evidence-based, patient-centred drug treatment.

Retrospective analysis of adverse drug reaction enquiries to a hospital drug information service: lessons to be learned to increase in-hospital drug safety

OBJECTIVES

Adverse drug reactions (ADRs) are a major drug safety concern and a frequent topic of enquiries to hospital drug information services. Our goal was to analyse these enquiries regarding background, complexity, nature of ADR, and involved drug classes to improve in-hospital drug safety.

METHODS

Retrospectively, ADR enquiries to a German university hospital pharmacy drug information 2018-2022 were analysed regarding enquirer (profession, medical specialty) and enquiry details (drugs, suspected ADR/enquiry prior to drug initiation, ADR system organ class, probable cause identified, and enquiry complexity).

KEY FINDINGS

Of 543 enquiries, 516 (95%) were asked by physicians, 493 (91%) patient-specific, 390 (71%) on suspected ADRs, and 153 (28%) prior to drug initiation. Enquiries originated frequently from internal medicine (74/13.6%), paediatrics (71/13.1%), neurology (70/12.9%), and haemato-oncology (62/11.4%). Most frequent ADRs were haematologic (94/17%) and hepatic (72/13%). The median number of drugs per enquiry was three (range 0-37), 209 (38%) enquiries referred to one specific drug, 165 (30%) concerned ≥11 drugs. A probable cause for suspected ADRs was identified in 75 (36%) enquiries concerning one drug and 155 (94%) with ≥11 drugs. Most frequent drugs were antineoplastic (54/25.8%), nervous-system-drugs (42/20.1%), and anti-infective (40/19.1%). Most enquiries (342/63%) were complex (multiple/specialist resources).

CONCLUSIONS

Enquiries were usually asked by physicians referring to suspected ADRs in specific clinical situations. A probable cause was identified in many cases pointing to a direct positive impact on patient care. Enquiries prior to drug initiation should be encouraged to increase drug safety. Information on main ADR effects and drug classes helps with targeted counselling.

Factors Determining Quality of Drug Information by Hospital Pharmacies-Results from Five-Year Annual Quality Assessment

Drug information (DI) provided by hospital pharmacies aims to promote rational and safe drug therapy. While quality assessment for this task is recommended, more knowledge on the factors determining the quality is needed. We aimed to evaluate the impacts of different factors on the quality of DI provided by hospital pharmacies to healthcare professionals. Retrospectively, answers on fictitious enquiries about annual DI tests for German hospital pharmacies over five years were evaluated for content-related and structural requirements. Multivariate analysis was performed for the impact of the enquiry complexity, DI organization (specialized DI center; pharmacist responsible per day; DI on top of other routine tasks), and quality measures (second look; experience of answering pharmacist in DI/on ward; use of documentation database). In 2017-2021, 45, 71, 79, 118, and 122 hospital pharmacies participated. The enquiry complexity had a statistically significant impact on the content-related quality, with poor results for a higher complexity (years 2018/2021, OR 0.25/0.04, p < 0.01). The DI centers achieved better results regarding content-related quality than for a pharmacist responsible per day (OR 0.76/p = 0.65) or DI on top of routine tasks (OR 0.35/p = 0.02). The DI centers scored better in structural quality. The second look showed an overall trend of a better content-related and structural quality. In conclusion, specialized DI centers and second looks are recommended as quality-improving measures. Training for answering complex enquiries should be intensified.

Safety and quality of AI chatbots for drug-related inquiries: A real-world comparison with licensed pharmacists

Introduction

Pharmacists play a pivotal role in ensuring patients are administered safe and effective medications; however, they encounter obstacles such as elevated workloads and a scarcity of qualified professionals. Despite the prospective utility of large language models (LLMs), such as Generative Pre-trained Transformers (GPTs), in addressing pharmaceutical inquiries, their applicability in real-world cases remains unexplored.

Objective

To evaluate GPT-based chatbots' accuracy in real-world drug-related inquiries, comparing their performance to licensed pharmacists.

Methods

In this cross-sectional study, authors analyzed real-world drug inquiries from a Drug Information Inquiry Database. Two independent pharmacists evaluated the performance of GPT-based chatbots (GPT-3, GPT-3.5, GPT-4) against human pharmacists using accuracy, detail, and risk of harm criteria. Descriptive statistics described inquiry characteristics. Absolute proportion comparative analyses assessed accuracy, detail, and risk of harm. Stratified analyses were performed for different inquiry types.

Results

Seventy inquiries were included. Most inquiries were received from physicians (41%) and pharmacists (44%). Inquiries type included dosage/administration (34.2%), drug interaction (12.8%) and pregnancy/lactation (15.7%). Majority of inquires included adults (83%) and female patients (54.3%). GPT-4 had 64.3% completely accurate responses, comparable to human pharmacists. GPT-4 and human pharmacists provided sufficiently detailed responses, with GPT-4 offering additional relevant details. Both GPT-4 and human pharmacists delivered 95% safe responses; however, GPT-4 provided proactive risk mitigation information in 70% of the instances, whereas similar information was included in 25.7% of human pharmacists' responses.

Conclusion

Our study showcased GPT-4's potential in addressing drug-related inquiries accurately and safely, comparable to human pharmacists. Current GPT-4-based chatbots could support healthcare professionals and foster global health improvements.

Identifying Training Needs and Active Information Opportunities in Primary Care Through the Analysis of Drug Information Requests

INTRODUCTION

Drug information (DI) services should work toward efficiency by identifying knowledge gaps and actively creating resources to address those needs. The aim was to identify training needs and active information opportunities in primary care by analyzing DI requests and to calculate labor cost associated with DI requests addressable with training or active information.

METHODS

DI requests received in 2016 and 2017 by ambulatory care pharmacists were independently classified by 2 authors into: training (i.e., delivery of content meant to be retained as knowledge and used when needed); active information (i.e., resources created preemptively and consulted when needed); or passive information (i.e., not addressable with training or active information). Inter-rater reliability was calculated using Cohen's Kappa. Median time spent by category and across practice settings/professional types was compared using bivariate analysis. Thematic analysis categorized specific training and active DI requests and labor costs were calculated.

RESULTS

Of 2,041 DI requests, 330 (16.2%) were classified as training, 454 (22.2%) active information, and 1257 (61.6%) passive information (kappa = 0.769). Median (IQR) time to resolve requests was 5 (2-10) mins for training, 5 (3-11) active information, and 10 (4-15) passive information. Pharmacists spent 132.1 hrs = $8,956.98 answering questions addressable with training or active information. Areas warranting training or active information included: controlled substances, immunizations, patient assistance programs, policy/regulations, medication preparation/administration, storage/stability, disposal, availability/ordering medications, and patient-related resources.

CONCLUSION

Several opportunities for training and active information were identified. Despite the single-institution nature, the method described can serve as an example for other institutions.