Usability of Graphical User Interfaces With Semiautomatic Delabeling Feature to Improve Drug Allergy Documentation.

BACKGROUND: The quality of allergy documentation in electronic health records is frequently poor. OBJECTIVE: To compare the usability of 3 graphical user interfaces (GUIs) for drug allergy documentation. METHODS: Physicians tested 3 GUIs by means of 5 fictional drug allergy scenarios: the current GUI (GUI 0), using mainly free-text, and 2 new coded versions (GUI 1 and GUI 2) asking information on allergen category, specific allergen, symptom(s), symptom onset, timing of initial reaction, and diagnosis status with a semiautomatic delabeling feature. Satisfaction was measured by the System Usability Scale questionnaire, efficiency by time to complete the tasks, and effectiveness by a task completion score. Posttest interviews provided more in-depth qualitative feedback. RESULTS: Thirty physicians from 7 different medical specialties and with varying degrees of experience participated. The mean System Usability Scale scores for GUI 1 (77.25, adjective rating "Good") and GUI 2 (78.42, adjective rating "Good") were significantly higher than for GUI 0 (56.58, adjective rating "OK") (Z, 6.27, Padj < .001 and Z, 6.62, Padj < .001, respectively). There was no significant difference in task time between GUIs. Task completion scores of GUI 1 and GUI 2 were higher than for GUI 0 (Z, 9.59, Padj < .001 and Z, 11.87, Padj < .001, respectively). Quantitative and qualitative findings were combined to propose a GUI 3 with high usability. CONCLUSIONS: The usability and quality of allergy documentation was higher for the newly developed coded GUIs with a semiautomatic delabeling feature without being more time-consuming.

Added value of patient- and drug-related factors to stratify drug-drug interaction alerts for risk of QT prolongation: Development and validation of a risk prediction model.

AIMS: Many clinical decision support systems trigger warning alerts for drug-drug interactions potentially leading to QT prolongation and torsades de pointes (QT-DDIs). Unfortunately, there is overalerting and underalerting because stratification is only based on a fixed QT-DDI severity level. We aimed to improve QT-DDI alerting by developing and validating a risk prediction model considering patient- and drug-related factors. METHODS: We fitted 31 predictor candidates to a stepwise linear regression for 1000 bootstrap samples and selected the predictors present in 95% of the 1000 models. A final linear regression model with those variables was fitted on the original development sample (350 QT-DDIs). This model was validated on an external dataset (143 QT-DDIs). Both true QTc and predicted QTc were stratified into three risk levels (low, moderate and high). Stratification of QT-DDIs could be appropriate (predicted risk = true risk), acceptable (one risk level difference) or inappropriate (two risk levels difference). RESULTS: The final model included 11 predictors with the three most important being use of antiarrhythmics, age and baseline QTc. Comparing current practice to the prediction model, appropriate stratification increased significantly from 37% to 54% appropriate QT-DDIs (increase of 17.5% on average [95% CI +5.4% to +29.6%], padj = 0.006) and inappropriate stratification decreased significantly from 13% to 1% inappropriate QT-DDIs (decrease of 11.2% on average [95% CI -17.7% to -4.7%], padj ≤ 0.001). CONCLUSION: The prediction model including patient- and drug-related factors outperformed QT alerting based on QT-DDI severity alone and therefore is a promising strategy to improve DDI alerting.

Machine Learning Techniques Outperform Conventional Statistical Methods in the Prediction of High Risk QTc Prolongation Related to a Drug-Drug Interaction.

In clinical practice, many drug therapies are associated with prolongation of the QT interval. In literature, estimation of the risk of prescribing drug-induced QT prolongation is mainly executed by means of logistic regression; only one paper reported the use of machine learning techniques. In this paper, we compare the performance of both techniques on the same dataset. High risk for QT prolongation was defined as having a corrected QT interval (QTc) ≥ 450 ms or ≥ 470 ms for respectively male and female patients. Both conventional statistical methods (CSM) and machine learning techniques (MLT) were used. All algorithms were validated internally and with a hold-out dataset of respectively 512 and 102 drug-drug interactions with possible drug-induced QTc prolongation. MLT outperformed the best CSM in both internal and hold-out validation. Random forest and Adaboost classification performed best in the hold-out set with an equal harmonic mean of sensitivity and specificity (HMSS) of 81.2% and an equal accuracy of 82.4% in a hold-out dataset. Sensitivity and specificity were both high (respectively 75.6% and 87.7%). The most important features were baseline QTc value, C-reactive protein level, heart rate at baseline, age, calcium level, renal function, serum potassium level and the atrial fibrillation status. All CSM performed similarly with HMSS varying between 60.3% and 66.3%. The overall performance of logistic regression was 62.0%. MLT (bagging and boosting) outperform CSM in predicting drug-induced QTc prolongation. Additionally, 19.2% was gained in terms of performance by random forest and Adaboost classification compared to logistic regression (the most used technique in literature in estimating the risk for QTc prolongation). Future research should focus on testing the classification on fully external data, further exploring potential of other (new) machine and deep learning models and on generating data pipelines to automatically feed the data to the classifier used.

Interdisciplinary knowledge gaps on intravenous fluid management in adult patients: Survey among physicians and nurses of a university hospital.

RATIONALE: Intravenous (IV) fluids are frequently involved in iatrogenic complications in hospitalized patients. Knowledge of IV fluids seems inadequate and is not covered sufficiently in standard medical education. METHODS: Two surveys were developed, based on the 2016 British National Institute for Health and Care Excellence guideline 'IV fluid therapy in adults in hospital', to provide insight on the learning needs and expectations of physicians and nurses. Each survey focused on profession-specific practice and consisted of three parts: demographics, knowledge questions and evaluation of current habits. Physicians and nurses practicing in a Belgian university hospital were invited to complete the survey electronically, respectively, in January and May 2018. RESULTS: A total of 103 physicians (19%) and 259 nurses (24%) participated. Although every indication for fluid therapy may require a specific fluid and electrolyte mixture, and hence, knowledge of their exact composition, most physicians and nurses did not know the composition of commonly prescribed solutions for IV infusion. Senior physicians did not score better than juniors did on questions concerning the daily needs of a nil-by-mouth patient. The availability of an IV fluid on the ward guides physicians to prescribe IV fluids (17%). Nurses (56%) feel they share responsibility in fluid management as they frequently intervene in urgent situations. More than half of participants (70% of physicians, 79% of nurses) indicated a need for additional information. CONCLUSIONS: A clear need for more structured information on IV fluids was identified. Both physicians and nurses struggle with fluid therapy. Continuing education on IV fluid management, emphasizing multidisciplinary collaboration, and monitoring evidence-based practice is essential to support the clinical decision process in daily practice.

Evaluation of an optimized context-aware clinical decision support system for drug-drug interaction screening.

OBJECTIVE: Evaluation of the effect of six optimization strategies in a clinical decision support system (CDSS) for drug-drug interaction (DDI) screening on alert burden and alert acceptance and description of clinical pharmacist intervention acceptance. METHODS: Optimizations in the new CDSS were the customization of the knowledge base (with addition of 67 extra DDIs and changes in severity classification), a new alert design, required override reasons for the most serious alerts, the creation of DDI-specific screening intervals, patient-specific alerting, and a real-time follow-up system of all alerts by clinical pharmacists with interventions by telephone was introduced. The alert acceptance was evaluated both at the prescription level (i.e. prescription acceptance, was the DDI prescribed?) and at the administration level (i.e. administration acceptance, did the DDI actually take place?). Finally, the new follow-up system was evaluated by assessing the acceptance of clinical pharmacist's interventions. RESULTS: In the pre-intervention period, 1087 alerts (92.0 % level 1 alerts) were triggered, accounting for 19 different DDIs. In the post-intervention period, 2630 alerts (38.4 % level 1 alerts) were triggered, representing 86 different DDIs. The relative risk forprescription acceptance in the post-intervention period compared to the pre-intervention period was 4.02 (95 % confidence interval (CI) 3.17-5.10; 25.5 % versus 6.3 %). The relative risk for administration acceptance was 1.16 (95 % CI 1.08-1.25; 54.4 % versus 46.7 %). Finally, 86.9 % of the clinical pharmacist interventions were accepted. CONCLUSION: Six concurrently implemented CDSS optimization strategies resulted in a high alert acceptance and clinical pharmacist intervention acceptance. Administration acceptance was remarkably higher than prescription acceptance.

GABA is a mediator of brain AT1 and AT2 receptor-mediated blood pressure responses.

The nucleus tractus solitarius (NTS), paraventricular nucleus (PVN), and rostral ventrolateral medulla (RVLM) are the most targeted regions of central blood pressure control studies. Glutamate and gamma-aminobutyric acid (GABA) interact within these brain regions to modulate blood pressure. The brain renin-angiotensin system also participates in central blood pressure control. Angiotensin II increases blood pressure through the stimulation of angiotensin II type 1 (AT1) receptors within the PVN and RVLM and attenuates baroreceptor sensitivity, resulting in elevated blood pressure within the NTS. Angiotensin II type 2 (AT2) receptors in cardiovascular control centers in the brain also appear to be involved in blood pressure control and counteract AT1 receptor-mediated effects. The current review is focused on the interaction of GABA with AT1 and AT2 receptors in the control of blood pressure within the RVLM, PVN and NTS. Within the NTS, GABA is released from local GABAergic interneurons that are stimulated by local AT1 receptors and mediates a hypertensive response. In contrast, the local increase in GABA levels observed after AT2 receptor stimulation within the RVLM, likely from GABAergic nerve endings originating in the caudal ventrolateral medulla, is important in the mediation of the hypotensive response. Preliminary results suggest that the hypertensive response to AT1 receptor stimulation within the RVLM is associated with a reduction in GABA release. The current experimental evidence therefore indicates that GABA is an important mediator of brainstem responses to AT1 and AT2 receptor stimulation and that increased GABA release may play a role in hypertensive and hypotensive responses, depending on the site of action.

Evaluation of context-specific alerts for potassium-increasing drug-drug interactions: A pre-post study.

OBJECTIVE: To investigate whether context-specific alerts for potassium-increasing drug-drug interactions (DDIs) in a clinical decision support system reduced the alert burden, increased alert acceptance, and had an effect on the occurrence of hyperkalemia. MATERIALS AND METHODS: In the pre-intervention period all alerts for potassium-increasing DDIs were level 1 alerts advising absolute contraindication, while in the post-intervention period the same drug combinations could trigger a level 1 (absolute contraindication), a level 2 (monitor potassium values), or a level 3 alert (informative, not shown to physicians) based on the patient's recent laboratory value of potassium. Alert acceptance was defined as non-prescription or non-administration of the interacting drug combination for level 1 alerts and as monitoring of the potassium levels for level 2 alerts. RESULTS: The alert burden decreased by 92.8%. The relative risk (RR) for alert acceptance based on prescription rates for level 1 alerts and monitoring rates for level 2 alerts was 15.048 (86.5% vs 5.7%; 95% CI 12.037-18.811; P < 0.001). With alert acceptance for level 1 alerts based on actual administration and for level 2 alerts on monitoring rates, the RR was 3.597 (87.6% vs 24.4%; 95% CI 3.192-4.053; P < 0.001). In the generalized linear mixed model the effect of the intervention on the occurrence of hyperkalemia was not significant (OR 1.091, 95% CI 0.172-6.919). CONCLUSION: The proposed strategy seems effective to get a grip on the delicate balance between over- and under alerting.

Investigation of the Role of AT2 Receptors in the Nucleus Tractus Solitarii of Normotensive Rats in Blood Pressure Control.

AIM: The nucleus tractus solitarii (NTS) densely expresses angiotensin II type 2 receptors (AT2R), which are mainly located on inhibitory gamma-aminobutyric acid (GABA) neurons. Central AT2R stimulation reduces blood pressure, and AT2R stimulation in the rostral ventrolateral medulla (RVLM), mediates a hypotensive response through a GABAergic mechanism. We aimed to test the hypothesis that an AT2R mediated inhibition of the GABA release within the NTS might be involved in this hypotensive response, by assessing possible alterations in blood pressure and heart rate, as well as in GABA levels in normotensive Wistar rats. METHODS: In vivo microdialysis was used for measurement of extracellular GABA levels and for perfusion of the selective AT2R agonist, Compound 21, within the NTS. Our set-up allowed to determine simultaneously the excitatory glutamate dialysate levels. The mean arterial pressure and heart rate responses were monitored with a pressure transducer. RESULTS: Local perfusion of Compound 21 into the NTS did not modify blood pressure and heart rate, nor glutamate and GABA levels compared to baseline concentrations. A putative effect was also not unmasked by concomitant angiotensin II type 1 receptor blockade with candesartan. Positive control experiments confirmed that the experimental set up had enough sensitivity to detect a reduction in GABA dialysate levels and blood pressure. CONCLUSION: The results did not provide evidence for a role of the AT2R within the NTS in the control of blood pressure, nor for an interaction with local GABAergic signaling in normotensive rats.

AT1 Receptor Mediated Hypertensive Response to Ang II in the Nucleus Tractus Solitarii of Normotensive Rats Involves NO Dependent Local GABA Release.

AIM: It is well-established that angiotensin II exerts a dampening effect on the baroreflex within the nucleus tractus solitarii (NTS), the principal brainstem site for termination of baroreceptor afferents and which is densely populated with gamma-aminobutyric acid (GABA)ergic neurons and nerve terminals. The present study was designed to investigate whether local release of GABA is involved in the effects mediated by local angiotensin II within the NTS. METHODS: In vivo microdialysis was used for measurement of extracellular glutamate and GABA levels and for infusion of angiotensin II within the NTS of conscious normotensive Wistar rats. The mean arterial pressure (MAP) and heart rate response to local infusion of angiotensin II were subsequently monitored with a pressure transducer under anesthesia. The angiotensin II type 1 receptor (AT1R) antagonist, candesartan, was used to assess whether responses were AT1R dependent and the nitric oxide (NO) synthase inhibitor, N(ω)-nitro-L-arginine methyl ester (L-NAME), was used to assess the involvement of NO in the evoked responses by infusion of angiotensin II. The MAP and heart rate responses were monitored with a pressure transducer. RESULTS: Local infusion into the NTS of angiotensin II induced a significant to ninefold significantly increase in extracellular GABA levels; as well as MAP was increased by 15 mmHg. These responses were both abolished by co-infusion of either, the angiotensin II type 1 receptor antagonist, candesartan, or the NO synthase inhibitor, L-NAME, demonstrating that the effect is not only AT1R dependent but also NO dependent. The pressor response to angiotensin II was reversed by co-infusion with the GABAA receptor antagonist, bicuculline. Local blockade of NO synthase decreased both, GABA and glutamate concentrations. CONCLUSION: Our results suggest that the AT1R mediated hypertensive response to angiotensin II within the NTS in normotensive rats is GABA and NO dependent. Nitric oxide produced within the NTS tonically potentiates local GABA and glutamate release.

Clinical Decision Support Systems for Drug Allergy Checking: Systematic Review.

BACKGROUND: Worldwide, the burden of allergies-in particular, drug allergies-is growing. In the process of prescribing, dispensing, or administering a drug, a medication error may occur and can have adverse consequences; for example, a drug may be given to a patient with a documented allergy to that particular drug. Computerized physician order entry (CPOE) systems with built-in clinical decision support systems (CDSS) have the potential to prevent such medication errors and adverse events. OBJECTIVE: The aim of this review is to provide a comprehensive overview regarding all aspects of CDSS for drug allergy, including documenting, coding, rule bases, alerts and alert fatigue, and outcome evaluation. METHODS: The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed as much as possible and searches were conducted in 5 databases using CPOE, CDSS, alerts, and allergic or allergy as keywords. Bias could not be evaluated according to PRISMA guidelines due to the heterogeneity of study types included in the review. RESULTS: Of the 3160 articles considered, 60 met the inclusion criteria. A further 9 articles were added based on expert opinion, resulting in a total of 69 articles. An interrater agreement of 90.9% with a reliability Κ=.787 (95% CI 0.686-0.888) was reached. Large heterogeneity across study objectives, study designs, study populations, and reported results was found. Several key findings were identified. Evidence of the usefulness of clinical decision support for drug allergies has been documented. Nevertheless, there are some important problems associated with their use. Accurate and structured documenting of information on drug allergies in electronic health records (EHRs) is difficult, as it is often not clear to healthcare providers how and where to document drug allergies. Besides the underreporting of drug allergies, outdated or inaccurate drug allergy information in EHRs poses an important problem. Research on the use of coding terminologies for documenting drug allergies is sparse. There is no generally accepted standard terminology for structured documentation of allergy information. The final key finding is the consistently reported low specificity of drug allergy alerts. Current systems have high alert override rates of up to 90%, leading to alert fatigue. Important challenges remain for increasing the specificity of drug allergy alerts. We found only one study specifically reporting outcomes related to CDSS for drug allergies. It showed that adverse drug events resulting from overridden drug allergy alerts do not occur frequently. CONCLUSIONS: Accurate and comprehensive recording of drug allergies is required for good use of CDSS for drug allergy screening. We found considerable variation in the way drug allergy are recorded in EHRs. It remains difficult to reduce drug allergy alert overload while maintaining patient safety as the highest priority. Future research should focus on improving alert specificity, thereby reducing override rates and alert fatigue. Also, the effect on patient outcomes and cost-effectiveness should be evaluated.

Hypotensive Response to Angiotensin II Type 2 Receptor Stimulation in the Rostral Ventrolateral Medulla Requires Functional GABA-A Receptors.

Objectives: Angiotensin II, glutamate and gamma-aminobutyric acid (GABA) interact within the rostral ventrolateral medulla (RVLM) and the paraventricular nucleus (PVN) modulating the central regulation of blood pressure and sympathetic tone. Our aim was to assess the effects of local angiotensin II type 2 receptor stimulation within the RVLM and the PVN on neurotransmitter concentrations and mean arterial pressure (MAP). Methods:In vivo microdialysis was used for measurement of extracellular glutamate and GABA levels and for local infusion of the angiotensin II type 2 receptor agonist Compound 21 in the RVLM and the PVN of conscious normotensive Wistar rats. The MAP response to local Compound 21 was monitored with a pressure transducer under anaesthesia. Angiotensin II type 2 receptor selectivity was assessed using the angiotensin II type 2 receptor antagonist PD123319; the GABA-A receptor antagonist bicuculline was used to assess the involvement of GABA-A receptors. Results: Infusion of Compound 21 (0.05 µg/µl/h) in the RVLM significantly increased GABA levels and lowered blood pressure. These effects were abolished by co-infusion with PD123319. No changes in neurotransmitter levels or effects on blood pressure were seen with PD123319 infusion alone. Co-infusion of bicuculline abolished the Compound 21 evoked decrease in MAP. Infusion of Compound 21 within the PVN did not change extracellular neurotransmitter levels nor MAP. Conclusion: Selective stimulation of angiotensin II type 2 receptor within the RVLM by local Compound 21 infusion reduces blood pressure and increases local GABA levels in normotensive rats. This hypotensive response requires functional GABA-A receptors, suggesting that GABAergic neurons are involved in the sympatho-inhibitory action underlying this hypotensive response.

A comparative study of medication use after stroke in four countries.

OBJECTIVES: The use of medication plays an important role in secondary stroke prevention and treatment of post-stroke comorbidities. The Collaborative Evaluation of Rehabilitation in Stroke across Europe (CERISE) was set up to investigate the inpatient stroke rehabilitation process in four centres, each in a different European country: Belgium, Germany, United Kingdom and Switzerland. PATIENTS AND METHODS: Patients' medication use 5 years post-stroke was compared between countries. Focus was put on cerebrovascular secondary prevention, including (a) adequate antithrombotic treatment, (b) treatment of cardiovascular comorbidities and diabetes, and (c) the use of lipid-lowering drugs; as well as on the treatment of stroke-related disorders such as depression, anxiety and pain. RESULTS: Medication data were available for 247 patients. Data about depression and anxiety were available for 233. CONCLUSION: There were no significant differences between the four centres in antithrombotic treatment and in the treatment of cardiovascular comorbidities and diabetes. However, significantly more patients from the UK were treated with lipid-lowering drugs compared to Belgian patients. Significant differences were also observed between the centres in the prevalence and treatment of depression. More Belgian patients suffered from depression compared to German patients and significantly more Belgian patients took antidepressants than patients in Germany. This was in contrast to the prevalence and treatment of anxiety and pain, for which no significant differences between the centres were seen. Related to pain treatment, it was observed that almost 40% of all patients suffering from pain, used no specific medication.

Pilot evaluation of an optimized context-specific drug-drug interaction alerting system: A controlled pre-post study.

OBJECTIVES: Clinical decision support (CDS) systems are frequently used to reduce unwanted drug-drug interactions (DDIs) but often result in alert fatigue. The main objective of this study was to investigate whether a newly developed context-specific DDI alerting system would improve alert acceptance. METHODS: A controlled pre-post intervention study was conducted in 4 departments in a university hospital. After a 7-month pre-intervention period, the new system was activated in the intervention departments, while the old system remained activated in the control departments. Post-intervention data was collected for a 7-month period. RESULTS: A significant increase of the overall acceptance rate was observed between the pre- and post-intervention period (2.2% versus 52.4%; p<0.001) for the intervention departments and between the intervention and control departments (2.5% versus 52.4%; p<0.001) in the post-intervention period. There were no significant differences in acceptance rates between the pre- and post-intervention period in the control departments and also not between the control and intervention departments in the pre-intervention period. CONCLUSIONS: The improvement was probably related to several optimization strategies including the customization of the severity classification, the creation of individual screening intervals, the inclusion of context factors for risk assessment, the new alert design and the creation of a follow-up system. The marked increase in alert acceptance looks promising and should be further evaluated after hospital wide implementation. System aspects that require further optimization were identified and will be developed. Further research is warranted to develop context-aware algorithms for complex class-class interactions.

Hypotensive and sympathoinhibitory responses to selective central AT2 receptor stimulation in spontaneously hypertensive rats.

The type 2 angiotensin receptor (AT2R) has been suggested to counterbalance the type 1 angiotensin receptor (AT1R) in the central regulation of blood pressure and sympathetic tone. In the present study we investigated the blood pressure responses to stimulation of central AT2Rs by the selective agonist Compound 21 in conscious spontaneously hypertensive rats (SHRs) and normotensive Wistar Kyoto rats (WKY rats). We also assessed the impact on noradrenaline [norepinephrine (NE)] plasma levels, autonomic function, spontaneous baroreflex sensitivity, and the possible involvement of the nitric oxide (NO) pathway and the AT1Rs. Chronic intracerebroventricular Compound 21 infusion lowered blood pressure and NE plasma levels in both rat strains. The night-time hypotensive effect was greater in SHRs compared with WKY rats. Compound 21 improved spontaneous baroreflex sensitivity more in SHRs than in WKY rats. These effects were abolished by co-administration of the AT2R antagonist PD123319 or the NO synthase inhibitor Nω-nitro-L-arginine methyl ester hydrochloride (L-NAME). Central AT1R blockade did not enhance the hypotensive response to Compound 21. Chronic selective stimulation of central AT2Rs lowers blood pressure through sympathoinhibition, and improves spontaneous baroreflex sensitivity more in SHRs than in WKY rats. These responses appear to require a functioning central NO pathway, but are not modified by central AT1R blockade. Collectively, the data demonstrate specific beneficial effects of stimulation of central AT2Rs in hypertension associated with increased sympathetic tone, and suggest that central AT2Rs may represent a potential new therapeutic target for the treatment of neurogenic hypertension.

Determinants of medication use in a multi-ethnic population of pregnant women: a cross-sectional study.

OBJECTIVES: To investigate determinants of medication use among multi-ethnic pregnant women. METHODS: A total of 641 pregnant women participated in this cross-sectional study in a Brussels university hospital. A questionnaire was used to obtain data on socio-demographic characteristics and medication use. Chi-squared tests and binary logistic regression analyses were performed using SPSS 19. RESULTS: Medication use during pregnancy (37%) was positively associated with age older than 35 years, Western origin, being born in Belgium, high education and employment status. Highly educated Western women had a greater exposure to drugs with an unknown safety profile than Arab/Turkish and 'Other (non-Western) origins' women. In the latter two groups, low education and nulliparity were the most important determinants of lower drug use. Nulliparous Arab/Turkish women used significantly less medications (17%) during pregnancy than parous women with the same ethnic background (34%; p = 0.024). CONCLUSIONS: Medication use during pregnancy is considerable but differs according to ethnicity. Age, parity, educational level, occupational status and duration of stay in Belgium are important determinants that should be taken into account for risk assessment and preventive measures targeting pregnant women.

Performance of a clinical decision support system and of clinical pharmacists in preventing drug-drug interactions on a geriatric ward.

BACKGROUND: Drug-drug interactions (DDIs) can lead to adverse drug events and compromise patient safety. Two common approaches to reduce these interactions in hospital practice are the use of clinical decision support systems and interventions by clinical pharmacists. OBJECTIVE: To compare the performance of both approaches with the main objective of learning from one approach to improve the other. SETTING: Acute geriatric ward in a university hospital. METHODS: Prospective single-centre, cohort study of patients admitted to the geriatric ward. An independent pharmacist compared the clinical decision support alerts with the DDIs identified by clinical pharmacists and evaluated their interventions. Contextual factors used by the clinical pharmacists for evaluation of the clinical relevance were analysed. Adverse drug events related to DDIs were investigated and the causality was evaluated by a clinical pharmacologist based on validated criteria. MAIN OUTCOME MEASURE: Number of alerts, interventions and the acceptance rates. RESULTS: Fifty patients followed by the clinical pharmacists, were included. The clinical pharmacists identified 240 DDIs (median of 3.5 per patient) and advised a therapy change for 16 of which 13 (81.2 %) were accepted and three (18.8 %) were not. The decision support system generated only six alerts of which none were accepted by the physicians. Thirty-seven adverse drug events were identified for 29 patients that could be related to 55 DDIs. For two interactions the causality was evaluated as certain, for 31 as likely, for ten as possible and for 12 as unlikely. Mainly intermediate level interactions were related to adverse drug events. Contextual factors taken into account by the clinical pharmacists for evaluation of the interactions were blood pressure, international normalised ratio, heart rate, potassium level and glycemia. Additionally, the clinical pharmacists looked at individual administration intervals and drug sequence to determine the clinical relevance of the interactions. CONCLUSION: Clinical pharmacists performed better than the decision support system mainly because the system screened only for high level DDIs and because of the low specificity of the alerts. This specificity can be increased by including contextual factors into the logic and by defining appropriate screening intervals that take into account the sequence in which the drugs are given.