Knowledge-based, computerized, patient clinical decision support system for perioperative pain, nausea and constipation management: a clinical feasibility study.

Opioid administration is particularly challenging in the perioperative period. Computerized-based Clinical Decision Support Systems (CDSS) are a promising innovation that might improve perioperative pain control. We report the development and feasibility validation of a knowledge-based CDSS aiming at optimizing the management of perioperative pain, postoperative nausea and vomiting (PONV), and laxative medications. This novel CDSS uses patient adaptive testing through a smartphone display, literature-based rules, and individual medical prescriptions to produce direct medical advice for the patient user. Our objective was to test the feasibility of the clinical use of our CDSS in the perioperative setting. This was a prospective single arm, single center, cohort study conducted in Strasbourg University Hospital. The primary outcome was the agreement between the recommendation provided by the experimental device and the recommendation provided by study personnel who interpreted the same care algorithm (control). Thirty-seven patients were included in the study of which 30 (81%) used the experimental device. Agreement between these two care recommendations (computer driven vs. clinician driven) was observed in 51 out 54 uses of the device (94.2% [95% CI 85.9-98.4%]). The agreement level had a probability of 86.6% to exceed the 90% clinically relevant agreement threshold. The knowledge-based, patient CDSS we developed was feasible at providing recommendations for the treatment of pain, PONV and constipation in a perioperative clinical setting.Trial registration number & date The study protocol was registered in ClinicalTrial.gov before enrollment began (NCT05707247 on January 26th, 2023).

Deep learning application to automated classification of recommendations made by hospital pharmacists during medication prescription review.

PURPOSE: Recommendations to improve therapeutics are proposals made by pharmacists during the prescription review process to address suboptimal use of medicines. Recommendations are generated daily as text documents but are rarely reused beyond their primary use to alert prescribers and caregivers. If recommendation data were easier to summarize, they could be used retrospectively to improve safeguards for better prescribing. The objective of this work was to train a deep learning algorithm for automated recommendation classification to valorize the large amount of recommendation data. METHODS: The study was conducted in a French university hospital, at which recommendation data were collected throughout 2017. Data from the first 6 months of 2017 were labeled by 2 pharmacists who assigned recommendations to 1 of the 29 possible classes of the French Society of Clinical Pharmacy classification. A deep neural network classifier was trained to predict the class of recommendations. RESULTS: In total, 27,699 labeled recommendations from the first half of 2017 were used to train and evaluate a classifier. The prediction accuracy calculated on a validation dataset was 78.0%. We also predicted classes for unlabeled recommendations collected during the second half of 2017. Of the 4,460 predictions reviewed, 67 required correction. When these additional labeled data were concatenated with the original dataset and the neural network was retrained, accuracy reached 81.0%. CONCLUSION: To facilitate analysis of recommendations, we have implemented an automated classification system using deep learning that achieves respectable performance. This tool can help to retrospectively highlight the clinical significance of daily medication reviews performed by hospital clinical pharmacists.

Safety and efficacy of colchicine in crystal-induced arthritis flare in 54 patients with severe chronic kidney disease.

INTRODUCTION: Colchicine, commonly used in gout flare, is contraindicated in severe chronic kidney disease (CKD) (estimated glomerular filtration rate <30 mL/min). However, in this context, there are few alternatives, and colchicine use persists. We evaluated the tolerance of colchicine and its efficacy in patients with severe CKD. PATIENTS AND METHODS: All prescriptions of colchicine for managing crystal-induced arthritis flare (gout or calcium pyrophosphate deposition (CPPD) disease) in a hospitalised patient with severe CKD were screened from September 2020 to September 2021. After patient consent and treatment information, clinical and biological safety and efficacy data were prospectively collected from day 1 (D1) to D11. RESULTS: We included 54 patients (median age 75 years (IQR 67-83)) with 62 colchicine prescriptions (cases). Twelve (22%) patients were on dialysis. The main reason for hospitalisation was heart failure (31.5%), acute renal failure (22.2%), infection (18.5%) or an acute joint episode (9.3%). In total, 59.3% of patients had diabetes. The prescriptions concerned 58 cases of gout flares, 1 case of CPPD and 3 cases of both. Initial colchicine dosages were ≤0.5 mg/day in 47/62 (75.8%) cases; no dosage exceeded 1 mg/day (median duration of 6 days (IQR 3-11)). Colchicine was well tolerated in 47/61 (77%) cases. No serious adverse event was reported. Colchicine was considered completely effective by the medical team in 48/58 (83%) of cases. CONCLUSION: The use of colchicine, at reduced doses, was mostly effective to treat crystal-induced arthritis flare in 54 patients with severe CKD and was well tolerated, without any serious adverse events.

Using machine learning or deep learning models in a hospital setting to detect inappropriate prescriptions: a systematic review.

OBJECTIVES: The emergence of artificial intelligence (AI) is catching the interest of hospital pharmacists. A massive collection of health data is now available to train AI models and hold the promise of disrupting codes and practices. The objective of this systematic review was to examine the state of the art of machine learning or deep learning models that detect inappropriate hospital medication orders. METHODS: A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. MEDLINE and Embase databases were searched from inception to May 2023. Studies were included if they reported and described an AI model intended for use by clinical pharmacists in hospitals. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool (PROBAST). RESULTS: 13 articles were selected after review: 12 studies were judged to have high risk of bias; 11 studies were published between 2020 and 2023; 8 were conducted in North America and Asia; 6 analysed orders and detected inappropriate prescriptions according to patient profiles and medication orders; and 7 detected specific inappropriate prescriptions, such as detecting antibiotic resistance, dosage abnormality in prescriptions, high alert drugs errors from prescriptions or predicting the risk of adverse drug events. Various AI models were used, mainly supervised learning techniques. The training datasets used were very heterogeneous; the length of study varied from 2 weeks to 7 years and the number of prescription orders analysed went from 31 to 5 804 192. CONCLUSIONS: This systematic review points out that, to date, few original research studies report AI tools based on machine or deep learning in the field of hospital clinical pharmacy. However, these original articles, while preliminary, highlighted the potential value of integrating AI into clinical hospital pharmacy practice.

Do proton pump inhibitors alter the response to immune checkpoint inhibitors in cancer patients? A meta-analysis.

Introduction: Gut microbiota can significantly affect the effectiveness of immune checkpoint inhibitors (ICIs) in cancer patients. Recently, antibiotics were shown to decrease survival rate of patients treated by ICIs. Proton pump inhibitors (PPIs) can indeed modulate microbiota's diversity, therefore altering ICIs response. A meta-analysis was performed based on published data to verify this hypothesis. Methods: In this study, over 41 publications, exploring the impact of concomitant PPI treatment on outcomes of ICI-treated patients, were analyzed. Evaluated endpoints were overall survival (OS) and progression-free survival (PFS). Pooled hazard ratios (HRs) with a 95% confidence interval (CI) were reported in ICIs in PPI users versus non-PPI users. Subgroup analyses were performed to minimize the impact of study heterogeneity and to investigate the influence of PPI on the different groups of interest. There was no evidence of publication bias for OS and PFS analysis in subgroup analysis. Results: Forty-one studies were included in the meta-analysis, including a total of 20,042 patients. OS of patients receiving ICIs was negatively correlated in patients concomitantly treated with PPI (HR=1.37; 95%CI, 1.23-1.52). PFS of cancer patients receiving ICIs was also negatively correlated with PPI treatment (HR=1.28; 95%CI, 1.15-1.42). PPI and ICI use was associated with worst OS and PFS not only for non-small-cell lung cancer (NSCLC) or urothelial cancer patients but also for patients treated with anti PD-1 (OS) and anti PD-L1 (OS and PFS) immunotherapies when administered in non-first line and when PPI was received as baseline treatment or in 60 days before ICI initiation. PPI and ICI use also showed the worst OS and PFS for patients from Europe and Asia. Conclusion: This meta-analysis suggests that PPI treatment leads to significantly worse outcomes in advanced cancer patients treated by ICIs in terms of PFS and OS.

Prescribing practices of lopinavir/ritonavir, hydroxychloroquine and azithromycin during the COVID-19 epidemic crisis and pharmaceutical interventions in a French teaching hospital.

OBJECTIVE: The aims of this study were to describe prescribing practices of lopinavir/ritonavir, hydroxychloroquine and azithromycin during the COVID-19 epidemic crisis (primary endpoint), then to characterise pharmaceutical interventions (PIs) targeted to these medications and evaluate the impact of these PIs on prescribers' practices (secondary end-points). METHODS: This retrospective observational study was carried out at the University Hospital of Strasbourg (France) from March to April 2020. The analysed population excluded patients from intensive care units but included all other adult patients with COVID-19 who received at least one dose of lopinavir/ritonavir combination, hydroxychloroquine or azithromycin, while inpatients. Analyses were performed by using data extracted from electronic medical records. RESULT: During the study period, 278 patients were included. A rapid decrease in lopinavir/ritonavir prescriptions was observed. This was accompanied by an increase in hydroxychloroquine and azithromycin prescriptions until the end of March, followed by a decrease leading to the disappearance of these two medications in April. The pharmaceutical analysis of the prescriptions resulted in 59 PIs of which 21 were associated with lopinavir/ritonavir, 32 with hydroxychloroquine and 6 with azithromycin. Regarding the medication-related problems, the most frequent ones were incorrect treatment durations (n=32 (54.2%)), drug interactions with potential torsadogenic reactions (n=14 (23.7%)) and incorrect dosing (n=6 (10.2%)). From the 59 PIs, 48 (81.4%) were accepted and physicians adjusted the medication regimens in a timely manner. CONCLUSION: This study demonstrated the value-even more meaningful in a crisis situation-of a strong synergy between physicians and pharmacists for patient-safety focused practices.

Anticoagulant and antiplatelet combined therapy in patients 75 years and over with atrial fibrillation: a prospective observational study assessing adherence to clinical guidelines.

Objective: According to current guidelines on atrial fibrillation (AF), the addition of an antiplatelet therapy to an anticoagulant for a stable vascular disease does not decrease the ischaemic hazard but increases the risk of bleeding. The aim of the study was to assess compliance of practices with existing clinical guidelines concerning the use of anticoagulant-antiplatelet combined therapy in patients 75 years and over with AF. Methods: This prospective observational study was carried out at the University Hospital of Strasbourg (France) between August 2016 and January 2017 with data collection on 1 day of every month. To be included, the patient had to be 75 years and over with AF and treated with anticoagulant-antiplatelet therapy. The population included all the patients admitted at the hospital excluding those from the Gynaecology-Obstetrics and Paediatrics departments. With regard to clinical ongoing guidelines (French, European, American and Canadian), the patients were sorted into three groups. Group 1: combined therapy in compliance with recommendations; Group 2: combined therapy debatable as to benefit-risk ratio; and Group 3: combined therapy not compliant with recommendations. Result: Ninety-three out of 3307 patients 75 years and over received anticoagulant-antiplatelet combined therapy prior to their hospital admission. Thirty-two patients (34.4% - Group 1) had experienced an acute event and/or revascularisation within the past year. Twenty-four patients (25.8% - Group 2) had not experienced recent revascularisation and had stable coronary disease but were suffering from peripheral artery disease. Group 3 consisted of 37 patients (39.8%), none of which had experienced recent revascularisation or had unstable coronary disease. For all groups, the main dual therapy was acetylsalicylic acid + fluindione (59.1%). Conclusion: In our study, 37 antiplatelet (39.8%) treatments could have been stopped. These results should spur prescribers into regular reassessment of combination antithrombotic therapy since it contributes to polypharmacy and increases the risk of adverse events.

Iatrogenic hypoglycemia-related hospital admissions identified through databases: economic burden and causes.

Background Hypoglycemia is an acute and frequent complication of diabetes. Objectives To assess the number of hospital admissions due to iatrogenic hypoglycemia in Alsace (France) over a year, to estimate the associated economic burden and to identify causes. Method A retrospective analysis was performed using data extracted from hospital databases. Costs were calculated from French official tariffs. Setting 31 public and private hospitals. A review of the medical records of patients with iatrogenic hypoglycemia-related hospital admissions was performed at the University Hospital of Strasbourg. Main outcome measures Hypoglycemia-related hospital admissions: number, costs and causes. Results Out of 42,381 hospitalizations, 147 iatrogenic hypoglycemia-related hospital admissions (0.4%) were identified; 41 patients with type 1 diabetes mellitus and 106 with type 2. The total cost associated to the 147 events was € 407,441. The median cost per patient was € 1,224.6 [563.0-2,505.7 (interquartile range)] for type 1 diabetes mellitus and € 3,670.9 [2,505.7-3,670.9] for type 2. Forty-six patients over the 147 were coming from the University Hospital of Strasbourg. In this hospital, the most common origin of the hypoglycemia was missed meals (n = 7), the second was a mismatch between antidiabetic medicines and carbohydrate intake (n = 6), the third was an incorrect use of antidiabetic medicines (n = 5). Conclusions 147 hospitalizations due to iatrogenic hypoglycemia were identified with an estimated global cost of € 407,441. Optimizing therapy with low-risk hypoglycemic medicines, improving access to continuous glucose monitoring systems and offering adequate education, could help address the causes of hypoglycemia.

[Evaluation of educational interventions with dialysis patient]. / Évaluation d'interventions éducatives auprès du patient dialysé.

The treatment of end-stage renal disease requires a significant number of drug treatments. At patient level, daily management is somewhat difficult: Number of prescribed pills, medication side effects, treatment of asymptomatic diseases… The objective of the study was to investigate the effect of guidance tailored to each patient receiving hemodialysis, performed by the pharmacist (educational interventions). Adult haemodialysis patients with hyperphosphatemia despite phosphate binders were eligible for study entry. The study was controlled with a retrospective group. The primary end point was a change in serum phosphate levels. The secondary end points were therapy adherence, knowledge regarding phosphate management and patient satisfaction with the programme. Sixteen patients in each group participated in the study. The mean serum phosphate level at endpoint was decreased by 0.25 mmol/L in the intervention group (0.41 mmol/L for patients with expectancy for this reduction) and by 0.11 mmol/L in the control group. Five patients normalized their serum phosphate level in the intervention group against three patients in the control group. The mean score of adherence decreased from 1.75 to 1.50. The main factors affecting adherence were forgetfulness or carelessness in taking medications and number of daily doses. This study showed the feasibility of an improvement in serum phosphate level and adherence driven by therapeutic education, though effect was highly amplified by the motivation induced by pharmaceutical guidance. Patients emphasize the importance of the involvement of pharmacist in their care.

Gastrointestinal perforations in patients treated with erlotinib: A report of two cases with fatal outcome and literature review.

Erlotinib has been approved as second-line treatment in patients with non-small cell lung cancer (NSCLC) experiencing relapse after first-line platinum-based chemotherapy. Herein, we report two occurrences of erlotinib-associated gastrointestinal perforation (GIP) in NSCLC patients. Two patients aged 60 and 79 years received erlotinib as third- and second-line NSCLC treatment, respectively. GIP occurred following 3 weeks and 6 months of erlotinib treatment, leading to death a few days later in both patients, neither of whom had any intestinal metastasis. Risk factors related to erlotinib-induced GIP were concomitant oral corticosteroid therapy and ciprofloxacin administration, which may result in erlotinib overexposure. GIP is a severe adverse drug reaction of erlotinib, infrequently described in the literature, compared to other targeted therapies. The lethal risk of erlotinib-associated GIP should be taken into account when evaluating the benefit-risk balance of erlotinib in patients without epidermal growth factor receptor activating mutations.

A UV-Raman spectrometry method for quality control of anticancer preparations: Results after 18 months of implementation in hospital pharmacy.

In France, chemotherapy preparation units of hospital pharmacy compound cytotoxic infusion bags adapted to each patient. The narrow therapeutic index of these preparations led us to implement qualitative and quantitative control for patients' safety. To this aim, we calibrated an equipment combining UV-vis spectrometry and Raman spectroscopy (QC Prep+) and monitored 14 different molecule-solvent combinations over a 18 months period. This rapid and specific method allowed the qualitative and quantitative analysis of 1 mL sample tests in less than 2 min. On 5742 anticancer preparations, we obtained accepted results with more than 99.4% solvent identification, 99.6% drug identification and only 1.52% of preparations not matching quantitative specifications (±15% of theoretical concentration). This quantitative control enabled us to pinpoint some critical points of production for two of the most common preparations. We thus updated the procedures of reconstitution and preparation, increasing the quality of final product. UV-Raman spectrometry is thus an effective tool to control chemotherapy infusions and to improve good practices of preparation.

Development and multi-centre evaluation of a method for assessing the severity of potential harm of medication reconciliation errors at hospital admission in elderly.

BACKGROUND: Medication reconciliation is a powerful process to correct medication errors (ME) resulting from miscommunicated information at transitions of care. This study aims to develop and evaluate a scoring method for assessing the severity of potential harm of ME intercepted by medication reconciliation at hospital admission in elderly. METHODS: The development of the scoring method was based on a literature search and the creation of a list of high-risk drugs used in outpatient care. The evaluation of the method was carried out in 7 French hospitals and was based on two criteria: the inter-rater reliability and acceptability. The assessment of the inter-rater reliability was based on intra-class correlation coefficient (ICC) calculations. Each hospital prospectively enrolled the 10 first patients aged 65 or older presenting with at least one ME. Seven blocks of 10 patients were formed. After randomization, each block was rated by practitioners from 3 hospitals. The assessment of the acceptability was based on a satisfaction questionnaire. RESULTS: A clinical algorithm was developed. The inter-rater reliability of the method was validated by the overall agreement of the 7 hospitals ratings. The agreement was at least substantial (ICC>0.60) and in most of cases almost perfect (ICC>0.80). The acceptability of the method was judged as satisfactory. CONCLUSION: This multi-centre project has validated an instrument for assessing the severity of potential harm of ME intercepted by medication reconciliation. This will allow studies to be conducted with large cohorts of patients in order to develop epidemiological databases of ME of potential clinical significance.

[Mecanisms of pharmacokinetic interactions involving oral anticancer agents]. / Mécanismes des interactions pharmacocinétiques impliquant les agents anticancéreux oraux.

Oral anticancer agents and particularly kinase inhibitors are subject to pharmacokinetic drug interactions in relation to absorption and elimination phases. Interacting factors are food, fruit juices, cigarette smoke, acid-reducing agents and inducers/inhibitors. Some anticancer agents are inducers and/or inhibitors and can also perpetrate drug interactions. This review emphasizes the mechanisms of pharmacokinetic drug interactions involving oral anticancer agents.

Impact of clinical pharmacy services in a hematology/oncology inpatient setting.

BACKGROUND/AIM: Clinical pharmacists are contributing to safe medication use by providing comprehensive management to patients and medical staff. However, little is known regarding their impact in oncology. The aim of this study was to document and evaluate the role of clinical pharmacy services in a hematology/oncology department. PATIENTS AND METHODS: A prospective, descriptive, observational study was carried out from May 2012 to May 2013. Medication reviews concerning hospitalized adult cancer patients were performed twice a week. Medication problems, pharmaceutical interventions and acceptance rate by the oncologists were recorded by a clinical pharmacist. RESULTS: A total of 4,393 prescriptions (including chemotherapy and support) of 489 adult cancer patients (mean age=63 years) were analyzed. The pharmacist identified 552 drug-related problems (12.6% of the prescriptions) primarily related to anti-infective agents (59.5%). Medication problems included inappropriate medications (20.6%), untreated indications (14.8%), inappropriate administrations (14.1%), underdosing (11.7%), drug-drug interactions (14.3%), lack of monitoring (9.6%), overdosing (8.9%), administration omissions (3.5%) and side-effects (2.5%). Interventions (n=552) led to treatment discontinuation (26.2%), drug dosing adjustments (21.5%), drug additions (16.9%), alternate routes of administration (11.7%), replacement of a drug by another one (10.7%), therapeutic drug monitoring (10.3%) and optimizing administration (2.6%). Most (96%) of the interventions were accepted and implemented by the medical staff. CONCLUSION: The integration of clinical pharmacy services resulted in drug-specific interventions in 12.6 % of the prescriptions of hospitalized adult patients with cancer. Medication problems mostly concerned anti-infective agents. The intervention acceptance rate by oncologists was high. The outcome of care in the hematology/oncology inpatient setting remains to be measured.

Erlotinib: another candidate for the therapeutic drug monitoring of targeted therapy of cancer? A pharmacokinetic and pharmacodynamic systematic review of literature.

Erlotinib is currently marketed at fixed standard dosage against pancreatic cancer and non-small-cell lung carcinoma. However, erlotinib pharmacokinetics (PK) is characterized by significant variability that may affect efficacy and tolerability. The aim of this review is to assess evidence that would justify therapeutic drug monitoring (TDM) and provide key information for the interpretation of erlotinib plasma concentrations. Literature was systematically reviewed to evaluate the standard criteria defining the potential clinical usefulness of TDM. Assessment was focused on the existence of unpredictable and wide PK variability and of consistent PK-pharmacodynamic relationships. PK parameters actually show marked variability (apparent clearance estimated to 4.85 ± 4.71 L/h, elimination half-life to 21.86 ± 28.35 hours, and apparent volume of distribution to 208 ± 133 L). Many covariates influence these parameters (CYP3A4 inducers or inhibitors, food, age, liver impairment), but most sources of variability still have to be identified. Some studies have demonstrated a relationship between exposure to erlotinib and clinical outcomes or skin toxicity. Erlotinib activity and target concentrations furthermore depend on tumor characteristics (eg, mutations on epidermal growth factor receptor and on K-ras). These results are in favor of TDM in addition to treatment adjustment for tumor biomarkers, but prospective clinical trials validating its clinical benefits are lacking. This review provides all the relevant information available to assist clinical interpretation of erlotinib plasma measurements. PK percentile curves and consideration to covariates yield information on whether a concentration measured is expected, whereas half maximal inhibitory concentration values determined in vitro provide preliminary insights on target concentration values to reach. Eventually, dosage adaptation might be considered in patients with intolerable toxicity because of excessive plasma levels or conversely nonresponse imputable to insufficient exposure.

Fatal stimulation of acute myeloid leukemia blasts by pegfilgrastim.

UNLABELLED: We herein report the case of a male patient with acute myeloid leukemia with fatal outcome attributable to pharmacokinetics of pegfilgrastim. CASE REPORT: An unexplained blast proliferation in a patient with acute myeloid leukemia following cytotoxic induction chemotherapy was investigated in depth. Myeloblast hyperstimulation was likely related to pegfilgrastim, the long half-life of which extended the duration of side-effects, resulting in massive and rapidly fatal leukemia cell proliferation. CONCLUSION: Pegfilgrastim can cause unexpected deleterious effects in acute myeloid leukemia. We, thus, recommend administering drugs with a shorter half-life, such as filgrastim or lenograstim, to reduce infection incidence in patients receiving myelosuppressive chemotherapy associated with a clinically significant incidence of febrile neutropenia.

New anticancer agents: role of clinical pharmacy services.

Clinical pharmacy (or clinical pharmacy services) aims to contribute to safe medication use by providing comprehensive management to patients and medical staff, both in the community and the hospital. In oncology, these services include comprehensive medication reviews integrating chemotherapy, supportive care and ambulatory treatment for co-morbidities, medication information for the medical staff and patients, therapeutic drug monitoring (anticancer agents, anti-infective agents, immunosuppressive drugs in recipients of allogeneic stem cell transplantation), supportive care counseling (nutritional support, pain management, chemotherapy side-effects prophylaxis and treatment), elaboration of therapeutic guidelines, optimal use of economic resources. With regard to new anticancer agents, pharmacists both in the community and in hospitals are faced with a growing body of complex information as well as the development of ambulatory treatment (oral agents, subcutaneous administration). Clinical pharmacists with oncology training have the potential to optimize drug use both in the hospital and the community. With the understanding and recognition of drug interactions and side-effects, pharmacists can provide timely interventions and information to health providers, as well as counseling to patients.

Potential clinical impact of medication discrepancies at hospital admission.

BACKGROUND: Medication errors at the interfaces of care are highly prevalent. This study aims to identify unintentional medication discrepancies at hospital admission and to explore their potential clinical impact in elderly patients. METHOD: The study was conducted in an Internal Medicine Department. Patients ≥ 65 years admitted through the emergency department were eligible. Best possible medication histories, obtained from different sources by pharmacists, were compared to admission medication prescriptions to identify and correct unintentional discrepancies. A three-category scale was used to rate errors for their potential to cause harm: Level (L) 1 "no potential harm", L2 "monitoring or intervention potentially required to preclude harm", and L3 "potential harm". This scale was also designed to take into account patient's clinical characteristics and high-risk drugs. RESULTS: 256 patients were included. Mean age was 82.2 ± 7.2 years old. 85 patients (33.2%) had ≥ 1 unintentional discrepancies. Overall, there were 173 unintentional discrepancies. The 3 most common drug classes involved in errors were nervous system (22.0%), gastrointestinal (20.0%) and cardiovascular (18.0%) medications. The most common types of errors were "omission" (87.9%) and "incorrect dose" (8.1%). Among the unintentional discrepancies, 20.8% had the potential to require increased monitoring or intervention to preclude harm (L2) and 6.4% had the potential to cause clinical deterioration (L3). CONCLUSION: More than 25% of the identified errors presented a potential clinical impact. These results show that a combined intervention of pharmacists and physicians in a collaborative medication reconciliation process has a high potential to reduce clinical relevant errors at hospital admission.