Diagnosis-linked antibiotic prescribing quality indicators: demonstrating feasibility using practice-based routine primary care data, reliability, validity and their potential in antimicrobial stewardship.

BACKGROUND: Quality indicators (QIs) can be used to obtain valuable insights into prescribing quality. Five quantitative and nine diagnosis-linked QIs, aiming to provide general practitioners (GP) with feedback on their antibiotic prescribing quantity and quality, were previously developed and evaluated in a controlled study. OBJECTIVE: To confirm, in a larger non-controlled study, the feasibility of using routinely collected and extracted electronic patient records to calculate the diagnosis-linked QI outcomes for antibiotic prescribing, and their reliability and validity. METHODS: Retrospective study involving 299 Dutch general practices using routine care data (2018-2020). QIs describe total antibiotic and subgroup prescribing, prescribing percentages and first-choice prescribing for several clinical diagnoses. Practice variation in QI outcomes, inter-QI outcome correlations and sensitivity of QI outcomes to pandemic-induced change were determined. RESULTS: QI outcomes were successfully obtained for 278/299 practices. With respect to reliability, outcomes for 2018 and 2019 were comparable, between-practice variation in outcomes was similar to the controlled pilot, and inter-QI outcome correlations were as expected, for example: high prescribing of second choice antibiotics with low first-choice prescribing for clinical diagnoses. Validity was confirmed by their sensitivity to pandemic-induced change: total antibiotic prescribing decreased from 282 prescriptions/1000 registered patients in 2018 to 216 in 2020, with a decrease in prescribing percentages for upper and lower respiratory infections, from 26% to 18.5%, and from 28% to 16%. CONCLUSIONS: This study confirmed the fit-for-purpose (feasibility, reliability and validity) of the antibiotic prescribing QIs (including clinical diagnosis-linked ones) using routinely registered primary health care data as a source. This feedback can therefore be used in antibiotic stewardship programmes to improve GPs' prescribing routines.

Poor performance of ChatGPT in clinical rule-guided dose interventions in hospitalized patients with renal dysfunction.

PURPOSE: Clinical decision support systems (CDSS) are used to identify drugs with potential need for dose modification in patients with renal impairment. ChatGPT holds the potential to be integrated in the electronic health record (EHR) system to give such dosing advices. In this study, we aim to evaluate the performance of ChatGPT in clinical rule-guided dose interventions in hospitalized patients with renal impairment. METHODS: This cross-sectional study was performed at Tergooi Medical Center, the Netherlands. CDSS alerts regarding renal dysfunction were collected from the electronic health record (EHR) during a 2-week period and were presented to ChatGPT and an expert panel. Alerts were presented with and without patient variables. To evaluate the performance, suggested medication interventions were compared. RESULTS: In total, 172 CDDS alerts were generated for 80 patients. Indecisive responses by ChatGPT to alerts were excluded. For alerts presented without patient variables, ChatGPT provided "correct and identical" responses to 19.9%, "correct and different" responses to 26.7%, and "incorrect responses to 53.4% of the alerts. For alerts including patient variables, ChatGPT provided "correct and identical" responses to 16.7%, "correct and different" responses to 16.0%, and "incorrect responses to 67.3% of the alerts. Accuracy was better for newer drugs such as direct oral anticoagulants. CONCLUSION: The performance of ChatGPT in clinical rule-guided dose interventions in hospitalized patients with renal dysfunction was poor. Based on these results, we conclude that ChatGPT, in its current state, is not appropriate for automatic integration into our EHR to handle CDSS alerts related to renal dysfunction.

Women on diuretics have a higher risk of hospital admission because of hyponatremia than men.

PURPOSE: Recent studies suggest that women are more susceptible to diuretic-induced hyponatremia resulting in hospital admission than men. The aim of this study was to confirm whether these sex differences in hyponatremia-related hospital admissions in diuretic users remain after adjusting for several confounding variables such as age, dose, and concurrent medication. METHODS: In a case-control design nested in diuretic users, cases of hyponatremia associated hospital admissions between 2005 and 2017 were identified from the PHARMO Data Network. Cases were 1:10 matched to diuretic users as controls. Odds ratios (OR) with 95%CIs were calculated for women versus men and adjusted for potential confounders (age, number of diuretics, other hyponatremia-inducing drugs, chronic disease score) using unconditional logistic regression analysis. A subgroup analysis was performed for specific diuretic groups (thiazides, loop diuretics and aldosterone antagonists). RESULTS: Women had a statistically significantly higher risk of a hospital admission associated with hyponatremia than men while using diuretics (OR 1.86, 95%CI 1.64-2.11). Adjusting for the potential confounders resulted in an increased risk for women compared to men (ORadj 2.65, 95% CI 2.31-3.04). This higher risk in women was also seen in the three subgroup analyses after adjustment. CONCLUSION: Our findings show a higher risk of hyponatremia-related hospital admission in women than men while using diuretics. Further research is needed to understand the underlying mechanism of this sex difference to be able to provide sex-specific recommendations.

Assessment of drug-related problems at the emergency department in older patients living with frailty: pharmacist-led medication reviews within a geriatric care team.

BACKGROUND: Older patients are vulnerable to experiencing drug related problems (DRPs), which may result in emergency department (ED) visits. However, it is not standard practice to conduct medications reviews during ED visit. The aim of this study was to assess the number of DRPs in older patients living with frailty at the ED, identified through pharmacist-led medication reviews within a geriatric care team, and to determine the acceptance rate of pharmacists' recommendations among hospital physicians and general practitioners or elderly care specialists. METHODS: A retrospective observational study was performed in patients ≥ 70 years living with frailty at the ED at Tergooi Medical Center. Pharmacist-led medication reviews were conducted to identify and classify DRPs as part of a larger geriatric assessment. The acceptance rate of given recommendations was determined during follow-up. RESULTS: A total of 356 ED visits were included. The mean (standard deviation, SD) age of patients was 83 (6.8) years. About 76% of patients had at least one DRP. In total, 548 DRPs were identified with a mean of 1.5 DRP (SD 1.3) per patient. The acceptance rate of medication recommendations in admitted patients was 55%, and 32% among general practitioners/elderly care specialists in discharged patients. CONCLUSIONS: Pharmacist-led medication reviews as part of a geriatric care team identified DRPs in 76% of older patients living with frailty at the ED. The acceptance rate was substantially higher in admitted patients compared to discharged patients.

Total bodyweight and sex both drive pharmacokinetic variability of fluconazole in obese adults.

BACKGROUND: Fluconazole is commonly used to treat or prevent fungal infections. It is typically used orally but in critical situations, IV administration is needed. Obesity may influence the pharmacokinetics and therapeutic efficacy of a drug. In this study, we aim to assess the impact of obesity on fluconazole pharmacokinetics given orally or IV to guide dose adjustments for the obese population. METHODS: We performed a prospective pharmacokinetic study with intensive sampling in obese subjects undergoing bariatric surgery (n = 17, BMI ≥ 35 kg/m2) and non-obese healthy controls (n = 8, 18.5 ≤ BMI < 30.0 kg/m2). Participants received a semi-simultaneous oral dose of 400 mg fluconazole capsules, followed after 2 h by 400 mg IV. Population pharmacokinetic modelling and simulation were performed using NONMEM 7.3. RESULTS: A total of 421 fluconazole concentrations in 25 participants (total bodyweight 61.0-174 kg) until 48 h after dosing were obtained. An estimated bioavailability of 87.5% was found for both obese and non-obese subjects, with a 95% distribution interval of 43.9%-98.4%. With increasing total bodyweight, both higher CL and Vd were found. Sex also significantly impacted Vd, being 27% larger in male compared with female participants. CONCLUSIONS: In our population of obese but otherwise healthy individuals, obesity clearly alters the pharmacokinetics of fluconazole, which puts severely obese adults, particularly if male, at risk of suboptimal exposure, for which adjusted doses are proposed.

The 2021 Dutch Working Party on Antibiotic Policy (SWAB) guidelines for empirical antibacterial therapy of sepsis in adults.

BACKGROUND: The Dutch Working Party on Antibiotic Policy (SWAB) in collaboration with relevant professional societies, has updated their evidence-based guidelines on empiric antibacterial therapy of sepsis in adults. METHODS: Our multidisciplinary guideline committee generated ten population, intervention, comparison, and outcome (PICO) questions relevant for adult patients with sepsis. For each question, a literature search was performed to obtain the best available evidence and assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system. The quality of evidence for clinically relevant outcomes was graded from high to very low. In structured consensus meetings, the committee formulated recommendations as strong or weak. When evidence could not be obtained, recommendations were provided based on expert opinion and experience (good practice statements). RESULTS: Fifty-five recommendations on the antibacterial therapy of sepsis were generated. Recommendations on empiric antibacterial therapy choices were differentiated for sepsis according to the source of infection, the potential causative pathogen and its resistance pattern. One important revision was the distinction between low, increased and high risk of infection with Enterobacterales resistant to third generation cephalosporins (3GRC-E) to guide the choice of empirical therapy. Other new topics included empirical antibacterial therapy in patients with a reported penicillin allergy and the role of pharmacokinetics and pharmacodynamics to guide dosing in sepsis. We also established recommendations on timing and duration of antibacterial treatment. CONCLUSIONS: Our multidisciplinary committee formulated evidence-based recommendations for the empiric antibacterial therapy of adults with sepsis in The Netherlands.

Women are started on a lower daily dose of metoprolol than men irrespective of dose recommendations: A potential source of confounding by contraindication in pharmacoepidemiology.

PURPOSE: Current guidelines have no sex-specific dosage advice for metoprolol. To evaluate whether women and men are prescribed the same dose a cohort analysis was performed in the population-based Rotterdam Study (RS). Results were replicated in the Integrated Primary Care Information (IPCI) database of automated general practice data. METHODS: The mean daily starting doses of metoprolol in both sexes were compared with independent-samples t-tests and a linear regression analysis was used to adjust in the RS for co-variables, notably, cardiovascular comorbidity, migraine, age, SBP, DBP, BMI, socioeconomic status, use of other antihypertensive drugs, smoking, and alcohol. In the IPCI-database, adjustment was for age only. RESULTS: The mean daily starting dose was statistically significantly lower in women than in men in both the RS and IPCI database, with a mean difference of 4.8 mg (95%CI -7.8, -1.8) and 4.6 mg (95%CI -5.3,-4.0), respectively. Statistical significance remained after adjustment in both databases. CONCLUSIONS: Women received lower starting doses of metoprolol than men in two independent data collections despite non-sex specific cardiovascular guideline recommendations. This example of real-life pharmacotherapy can lead to a form of confounding by contraindication in pharmacoepidemiology.

Consecutive antibiotic use in the outpatient setting: an extensive, longitudinal descriptive analysis of antibiotic dispensing data in the Netherlands.

INTRODUCTION: Taking consecutive antibiotic use into account is of importance to obtain insight in treatment within disease episodes, use of 2nd- and 3rd-choice antibiotics, therapy failure and/or side effects. Nevertheless, studies dealing with consecutive antibiotic use are scarce. We aimed at evaluating switch patterns in antibiotic use in the outpatient setting in the Netherlands. METHODS: Outpatient antibiotic dispensing data was processed to antibiotic treatment episodes consisting of single prescriptions or consecutive prescriptions (2006 to 2014). Consecutive prescriptions were categorised into prolongations and switches. Switches were further analysed to obtain antibiotic switch percentages and trends over time. Outcomes were compared with recommendations of Dutch guidelines. RESULTS: A total of 43,179,867 antibiotic prescriptions were included in the analysis, consisting of single prescriptions (95%), prolongations (2%) and switches (3%). The highest switch percentages were found for trimethoprim (7.6%) and nitrofurantoin (5.4%). For fosfomycin, ciprofloxacin, flucloxacillin and trimethoprim we found the highest yearly increase in switching. Amoxicillin/clavulanic acid was most often used as second antibiotic in a switch. A surprisingly high number of 2nd- and 3rd-choice antibiotics are prescribed as first antibiotic in a treatment. CONCLUSIONS: Although the actual reason for a switch is unknown, switch patterns can reveal problems concerning treatment failure and guideline adherence. In general, switch percentages of antibiotics in the Netherlands are low. The data contributes to the knowledge regarding antibiotic switch patterns in the outpatient setting.

Diabetes and Glucose Dysregulation and Transition to Delirium in ICU Patients.

OBJECTIVES: To investigate whether diabetes and glucose dysregulation (hyperglycemia and/or hypoglycemia) are associated with ICU delirium. DESIGN: Prospective cohort study. SETTING: Thirty-two-bed mixed intensive care in a tertiary care center. PATIENTS: Critically ill patients admitted to the ICU with transitions of mental status from awake and nondelirious to delirious or remaining awake and nondelirious on the next day. Patients admitted because of a neurologic illness were excluded. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The study population consisted of 2,745 patients with 1,720 transitions from awake and nondelirious to delirious and 11,421 nontransitions remaining awake and nondelirious. Generalized mixed effects models with logit link function were performed to study the association between diabetes mellitus, glucose dysregulation, and delirium, adjusting for potential confounders. Diabetes was not associated with delirium (odds ratio adjusted, 0.93; 95% CI, 0.73-1.18). In all patients, the occurrence of hyperglycemia (odds ratio adjusted, 1.35; 95% CI, 1.15-1.59) and the occurrence of both hyperglycemia and hypoglycemia on the same day (odds ratio adjusted, 1.65; 95% CI, 1.12-2.28) compared with normoglycemia were associated with transition to delirium. Hypoglycemia was not associated with transition to delirium (odds ratio adjusted, 1.86; 95% CI, 0.73-3.71). In patients without diabetes, the occurrence of hyperglycemia (odds ratio adjusted, 1.41; 95% CI, 1.16-1.68) and the occurrence of both hyperglycemia and hypoglycemia on the same day (odds ratio adjusted, 1.87; 95% CI, 1.07-2.89) were associated with transition to delirium. In patients with diabetes, glucose dysregulation was not associated with ICU delirium. CONCLUSIONS: Diabetes mellitus was not associated with the development of ICU delirium. For hypoglycemia, only a nonsignificant odds ratio for ICU delirium could be noted. Hyperglycemia and the occurrence of hyperglycemia and hypoglycemia on the same day were associated with ICU delirium but only in patients without diabetes. Our study supports the institution of measures to prevent glucose dysregulation in nondiabetic ICU patients and contributes to the understanding of the determinants of delirium.

Sepsis calculator implementation reduces empiric antibiotics for suspected early-onset sepsis.

Significant overtreatment with antibiotics for suspected early onset sepsis (EOS) constitutes a persisting clinical problem, generating unnecessary risks, harms, and costs for many newborns. We aimed to study feasibility and impact of a sepsis calculator to help guide antibiotic for suspected EOS in a European setting. In this single-center study, the sepsis calculator was implemented as an addition to and in accordance with existing protocols. One thousand eight hundred seventy-seven newborns ≥ 35 weeks of gestational age were prospectively evaluated; an analogous retrospective control group (n = 2076) was used for impact analysis. We found that empirical treatment with intravenous antibiotics for suspected EOS was reduced from 4.8 to 2.7% after sepsis calculator implementation (relative risk reduction 44% (95% confidence interval 21.4-59.5%)). No evidence for changes in time to treatment start, treatment duration, or proven sepsis rates was found. Adherence to sepsis calculator recommendation was 91%. CONCLUSION: Pragmatic and feasible implementation of the sepsis calculator yields a 44% reduction of empirical use of antibiotics for EOS, without signs of delay or prolongation of treatment. These findings warrant a multicenter, nation-wide, randomized study evaluating systematic use of the sepsis calculator prediction model and its effects in clinical practice outside of the USA. What is known: • Significant overtreatment with antibiotics for suspected early-onset sepsis results in unnecessary costs, risks, and harms. • Implementation of the sepsis calculator in the USA has resulted in a significant decrease in empirical antibiotic treatment, without apparent adverse events. What is new: • Implementation of the sepsis calculator in daily clinical decision-making in a Dutch teaching hospital is feasible in conjunction with existing protocols, with high adherence. • Antibiotic therapy for suspected early-onset sepsis was reduced by 44% following implementation of the calculator.

Predictive value of prior colonization and antibiotic use for third-generation cephalosporin-resistant enterobacteriaceae bacteremia in patients with sepsis.

BACKGROUND: To prevent inappropriate empiric antibiotic treatment in patients with bacteremia caused by third-generation cephalosporin (3GC)-resistant Enterobacteriaceae (3GC-R EB), Dutch guidelines recommend ß-lactam and aminoglycoside combination therapy or carbapenem monotherapy in patients with prior 3GC-R EB colonization and/or recent cephalosporin or fluoroquinolone usage. Positive predictive values (PPVs) of these determinants are unknown. METHODS: We retrospectively studied patients with a clinical infection in whom blood cultures were obtained and empiric therapy with broad-spectrum ß-lactams and/or aminoglycosides and/or fluoroquinolones was started. We determined the PPVs of prior colonization and antibiotic use for 3GC-R EB bacteremia, and the consequences of guideline adherence on appropriateness of empiric treatment. RESULTS: Of 9422 episodes, 773 (8.2%) were EB bacteremias and 64 (0.7%) were caused by 3GC-R EB. For bacteremia caused by 3GC-R EB, PPVs of prior colonization with 3GC-R EB (90-day window) and prior usage of cephalosporins or fluoroquinolones (30-day window) were 7.4% and 1.3%, respectively, and PPV was 1.8% for the presence of any of these predictors. Adherence to Dutch sepsis guideline recommendations was 27%. Of bacteremia episodes caused by 3GC-R and 3GC-sensitive EB, 56% and 94%, respectively, were initially treated with appropriate antibiotics. Full adherence to guideline recommendations would hardly augment proportions of appropriate therapy, but could considerably increase carbapenem use. CONCLUSIONS: In patients receiving empiric treatment for sepsis, prior colonization with 3GC-R EB and prior antibiotic use have low PPV for infections caused by 3GC-R EB. Strict guideline adherence would unnecessarily stimulate broad-spectrum antibiotic use.

Risk of Hospitalization for Hypoglycemia in Older Patients with Diabetes Using Antipsychotic Drugs.

OBJECTIVE: Antipsychotics may disrupt metabolic regulation in patients with diabetes mellitus. The risk of hypoglycemia in older users of antipsychotics with diabetes is largely unknown. Therefore, we investigated the association between the use of antipsychotic drugs and hypoglycemia requiring hospital admission in older patients with diabetes. METHODS: In a nested case-control study using community pharmacy records linked to hospital admission data in the Netherlands (1998-2008), a cohort of 68,314 patients at least 65 years with diabetes was studied. Cases were patients from the study cohort with a first hospital admission for hypoglycemia; up to five comparison subjects were selected for each case. Exposure to antipsychotic drugs was the primary determinant of interest. Logistic regression analysis was performed to estimate the strength of the association between antipsychotic drug use and hypoglycemia, taking into account potential confounders. RESULTS: Eight hundred fifteen patients were admitted to hospital for hypoglycemia. Current use of antipsychotic drugs was associated with an increased risk of hypoglycemia compared with non-use (adjusted OR: 2.26; 95% CI: 1.45-3.52; Wald χ(2) = 13.08, df = 1, p ≤0.001), especially in the first 30 days of treatment (adjusted OR: 7.65; 95% CI: 2.50-23.41; Wald χ(2) = 12.72, df = 1, p ≤0.001) and with higher doses (adjusted OR: 8.20; 95% CI: 3.09-21.75; Wald χ(2) = 17.90, df = 1, p ≤0.001). CONCLUSION: Use of antipsychotic drugs by older patients with diabetes mellitus was associated with an increased risk of hospitalization for hypoglycemia. Our findings suggest that glucose levels should be monitored closely after initiation of antipsychotic drugs.

Sulfamethoxazole-Trimethoprim-Induced Hyperkalemia in Hospitalized Patients Using Potassium-Sparing Drugs: An Observational Study.

The objective of this analysis is to investigate the risk of hyperkalemia in hospitalized patients using sulfamethoxazole-trimethoprim (Co-trimoxazole) and a potassium-sparing drug (potassium-sparing diuretic or renin-angiotensin system [RAS]-inhibitor). Researchers conducted a nested case control study within a cohort of hospitalized patients using a potassium-sparing diuretic and/or a RAS-inhibitor from the PHARMO Database Network. Researchers estimated the odds ratios (ORs) and 95% confidence intervals (CI) for the risk of hyperkalemia in patients receiving both Co-trimoxazole and a potassium-sparing drug compared with patients only receiving a potassium-sparing drug. Among a cohort of 25,849 patients, researchers identified 2054 cases of hyperkalemia during hospitalization in patients also using a potassium-sparing drug. Using Co-trimoxazole in addition to a potassium-sparing drug was associated with an increased risk of hyperkalemia in hospitalized patients (ORadj = 1.65, 95% CI 1.26-2.16) compared with using only a potassium-sparing drug. There was a trend of a more pronounced association between hyperkalemia and the co-use of Co-trimoxazole and potassium-sparing drugs in patients with an estimated GFR of 15-29 mL/min (ORadj = 3.15, 95% CI 1.29-7.70). The number needed to harm for hyperkalemia induced by adding Co-trimoxazole to patients receiving a potassium-sparing drug is 19.5. Using the combination of Co-trimoxazole with a potassium-sparing drug in hospitalized patients increases the risk of hyperkalemia compared with using only a potassium-sparing drug. Physicians and other prescribers should be aware of hyperkalemia and routinely monitor serum potassium levels in hospitalized patients using this combination of drugs.

Performance of ChatGPT on Factual Knowledge Questions Regarding Clinical Pharmacy.

ChatGPT is a language model that was trained on a large dataset including medical literature. Several studies have described the performance of ChatGPT on medical exams. In this study, we examine its performance in answering factual knowledge questions regarding clinical pharmacy. Questions were obtained from a Dutch application that features multiple-choice questions to maintain a basic knowledge level for clinical pharmacists. In total, 264 clinical pharmacy-related questions were presented to ChatGPT and responses were evaluated for accuracy, concordance, quality of the substantiation, and reproducibility. Accuracy was defined as the correctness of the answer, and results were compared to the overall score by pharmacists over 2022. Responses were marked concordant if no contradictions were present. The quality of the substantiation was graded by two independent pharmacists using a 4-point scale. Reproducibility was established by presenting questions multiple times and on various days. ChatGPT yielded accurate responses for 79% of the questions, surpassing pharmacists' accuracy of 66%. Concordance was 95%, and the quality of the substantiation was deemed good or excellent for 73% of the questions. Reproducibility was consistently high, both within day and between days (>92%), as well as across different users. ChatGPT demonstrated a higher accuracy and reproducibility to factual knowledge questions related to clinical pharmacy practice than pharmacists. Consequently, we posit that ChatGPT could serve as a valuable resource to pharmacists. We hope the technology will further improve, which may lead to enhanced future performance.

Pooled Population Pharmacokinetic Analysis and Dose Recommendations for Ciprofloxacin in Intensive Care Unit Patients with Obesity.

Recent studies have explored the influence of obesity and critical illness on ciprofloxacin pharmacokinetics. However, variation across the subpopulation of individuals with obesity admitted to the intensive care unit (ICU) with varying renal function remains unexamined. This study aims to characterize ciprofloxacin pharmacokinetics in ICU patients with obesity and provide dose recommendations for this special population. Individual patient data of 34 ICU patients with obesity (BMI >30 kg/m2) from four studies evaluating ciprofloxacin pharmacokinetics in ICU patients were pooled and combined with data from a study involving 10 individuals with obesity undergoing bariatric surgery. All samples were collected after intravenous administration. Non-linear mixed effects modeling and simulation were used to develop a population pharmacokinetic model and describe ciprofloxacin exposure in plasma. Model-based dose evaluations were performed using a pharmacokinetic/pharmacodynamic target of AUC/MIC >125. The data from patients with BMI ranging from 30.2 to 58.1 were best described by a two-compartment model with first-order elimination and a proportional error model. The inclusion of Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) as a covariate on clearance reduced inter-individual variability from 57.3% to 38.5% (P < .001). Neither body weight nor ICU admission significantly influenced clearance or volume of distribution. Renal function is a viable predictor for ciprofloxacin clearance in ICU patients with obesity, while critical illness and body weight do not significantly alter clearance. As such, body weight and critical illness do not need to be accounted for when dosing ciprofloxacin in ICU patients with obesity. Individuals with CKD-EPI >60 mL/min/1.73 m2 may require higher dosages for the treatment of pathogens with minimal inhibitory concentration ≥0.25 mg/L.

Assessing accuracy of ChatGPT in response to questions from day to day pharmaceutical care in hospitals.

Background: The advent of Large Language Models (LLMs) such as ChatGPT introduces opportunities within the medical field. Nonetheless, use of LLM poses a risk when healthcare practitioners and patients present clinical questions to these programs without a comprehensive understanding of its suitability for clinical contexts. Objective: The objective of this study was to assess ChatGPT's ability to generate appropriate responses to clinical questions that hospital pharmacists could encounter during routine patient care. Methods: Thirty questions from 10 different domains within clinical pharmacy were collected during routine care. Questions were presented to ChatGPT in a standardized format, including patients' age, sex, drug name, dose, and indication. Subsequently, relevant information regarding specific cases were provided, and the prompt was concluded with the query "what would a hospital pharmacist do?". The impact on accuracy was assessed for each domain by modifying personification to "what would you do?", presenting the question in Dutch, and regenerating the primary question. All responses were independently evaluated by two senior hospital pharmacists, focusing on the availability of an advice, accuracy and concordance. Results: In 77% of questions, ChatGPT provided an advice in response to the question. For these responses, accuracy and concordance were determined. Accuracy was correct and complete for 26% of responses, correct but incomplete for 22% of responses, partially correct and partially incorrect for 30% of responses and completely incorrect for 22% of responses. The reproducibility was poor, with merely 10% of responses remaining consistent upon regeneration of the primary question. Conclusions: While concordance of responses was excellent, the accuracy and reproducibility were poor. With the described method, ChatGPT should not be used to address questions encountered by hospital pharmacists during their shifts. However, it is important to acknowledge the limitations of our methodology, including potential biases, which may have influenced the findings.

Major changes in chemotherapy regimens administered to breast cancer patients during 2000-2008 in the Netherlands.

There is little information available on the patterns of chemotherapy regimens administered in daily practice to patients with early stage and metastatic or recurrent breast cancer. To determine the trends in type of chemotherapy regimens used in breast cancer patients, newly diagnosed breast cancer patients in the period 2000-2008 who received chemotherapy were identified from the Eindhoven Cancer Registry (ECR) and linked to the PHARMO RLS, including data on, e.g., in- and outpatient drug use. Chemotherapy regimens were classified based on the received combinations and sequences. Trends in the distribution of adjuvant chemotherapy regimens (for early-stage breast cancer) and palliative chemotherapy regimens (for metastatic or recurrent breast cancer) were determined and stratified by Her2/neu status when possible. In this study, 422 patients diagnosed with early-stage breast cancer received adjuvant chemotherapy. The use of CMF (cyclophosphamide, methotrexate, and 5-fluorouracil) decreased from 90% in 2000 to almost none since 2005. Administration of regimens that included anthracyclines increased from 4% in 2000 to 96% in 2005, but decreased to 68% in 2008. The use of trastuzumab- and taxane-containing regimens (with or without anthracyclines) increased from 2005 onwards to 24% and 34%, respectively, in 2008. Among the 82 breast cancer patients who received palliative chemotherapy at diagnosis or after breast cancer recurrence, the use of CMF and anthracyclines (without taxanes) decreased, while the use of taxanes (with or without anthracyclines) increased (26% in 2008). Trastuzumab was used as palliative chemotherapy from 2003 onwards, with 22% of the metastatic breast cancer patients receiving trastuzumab-containing regimens in 2008, and bevacizumab was administered since 2007 with 19% of the patients receiving bevacizumab-containing regimens in 2008. In conclusion, major changes have taken place in the chemotherapeutic treatment of patients with early and recurrent breast cancer. These changes reflect the key findings from large clinical trials, as incorporated in the Dutch guidelines.

Ciprofloxacin Pharmacokinetics After Oral and Intravenous Administration in (Morbidly) Obese and Non-obese Individuals: A Prospective Clinical Study.

BACKGROUND AND OBJECTIVE: Ciprofloxacin is a fluoroquinolone used for empirical and targeted therapy of a wide range of infections. Despite the increase in obesity prevalence, only very limited guidance is available on whether the ciprofloxacin dose needs to be adjusted when administered orally or intravenously in (morbidly) obese individuals. Our aim was to evaluate the influence of (morbid) obesity on ciprofloxacin pharmacokinetics after both oral and intravenous administration, to ultimately guide dosing in this population. METHODS: (Morbidly) obese individuals undergoing bariatric surgery received ciprofloxacin either orally (500 mg; n = 10) or intravenously (400 mg; n = 10), while non-obese participants received semi-simultaneous oral dosing of 500 mg followed by intravenous dosing of 400 mg 3 h later (n = 8). All participants underwent rich sampling (11-17 samples) for 12 h after administration. Non-linear mixed-effects modelling and simulations were performed to evaluate ciprofloxacin exposure in plasma. Prior data from the literature were subsequently included in the model to explore exposure in soft tissue in obese and non-obese patients. RESULTS: Overall, 28 participants with body weights ranging from 57 to 212 kg were recruited. No significant influence of body weight on bioavailability, clearance or volume of distribution was identified (all p > 0.01). Soft tissue concentrations were predicted to be lower in obese individuals despite similar plasma concentrations compared with non-obese individuals. CONCLUSION: Based on plasma pharmacokinetics, we found no evidence of the influence of obesity on ciprofloxacin pharmacokinetic parameters; therefore, ciprofloxacin dosages do not need to be increased routinely in obese individuals. In the treatment of infections in tissue where impaired ciprofloxacin penetration is anticipated, higher dosages may be required. TRIAL REGISTRATION: Registered in the Dutch Trial Registry (NTR6058).

Narrow-spectrum antibiotics for community-acquired pneumonia in Dutch adults (CAP-PACT): a cross-sectional, stepped-wedge, cluster-randomised, non-inferiority, antimicrobial stewardship intervention trial.

BACKGROUND: Adults hospitalised to a non-intensive care unit (ICU) ward with moderately severe community-acquired pneumonia are frequently treated with broad-spectrum antibiotics, despite Dutch guidelines recommending narrow-spectrum antibiotics. Therefore, we investigated whether an antibiotic stewardship intervention would reduce the use of broad-spectrum antibiotics in patients with moderately severe community-acquired pneumonia without compromising their safety. METHODS: In this cross-sectional, stepped-wedge, cluster-randomised, non-inferiority trial (CAP-PACT) done in 12 hospitals in the Netherlands, we enrolled immunocompetent adults (≥18 years) who were admitted to a non-ICU ward and had a working diagnosis of moderately severe community-acquired pneumonia. All participating hospitals started in a control period and every 3 months a block of two hospitals transitioned from the control to the intervention period, with all hospitals eventually ending in the intervention period. The unit of randomisation was the hospital (cluster), and electronic randomisation (by an independent data manager) decided the sequence (the time of intervention) by which hospitals would cross over from the control period to the intervention period. Blinding was not possible. The antimicrobial stewardship intervention was a bundle targeting health-care providers and comprised education, engaging opinion leaders, and prospective audit and feedback of antibiotic use. The co-primary outcomes were broad-spectrum days of therapy per patient, tested by superiority, and 90-day all-cause mortality, tested by non-inferiority with a non-inferiority margin of 3%, and were analysed in the intention-to-treat population, comprising all patients who were enrolled in the control and intervention periods. This trial was prospectively registered at ClinicalTrials.gov, NCT02604628. FINDINGS: Between Nov 1, 2015, and Nov 1, 2017, 5683 patients were assessed for eligibility, of whom 4084 (2235 in the control period and 1849 in the intervention period) were included in the intention-to-treat analysis. The adjusted mean broad-spectrum days of therapy per patient were reduced from 6·5 days in the control period to 4·8 days in the intervention period, yielding an absolute reduction of -1·7 days (95% CI -2·4 to -1·1) and a relative reduction of 26·6% (95% CI 18·0-35·3). Crude 90-day mortality was 10·9% (242 of 2228 died) in the control period and 10·8% (199 of 1841) in the intervention period, yielding an adjusted absolute risk difference of 0·4% (90% CI -2·7 to 2·4), indicating non-inferiority. INTERPRETATION: In patients hospitalised with moderately severe community-acquired pneumonia, a multifaceted antibiotic stewardship intervention might safely reduce broad-spectrum antibiotic use. FUNDING: None.

Discontinuation of Antihypertensive Medications on the Outcome of Hospitalized Patients With Severe Acute Respiratory Syndrome-Coronavirus 2.

[Figure: see text].