Medication Safety Events After Acute Myocardial Infarction Among Veterans Treated at VA Versus Non-VA Hospitals.

INTRODUCTION: Fragmentation of health care across systems can contribute to mistakes in prescribing and filling medications among patients treated for myocardial infarction (MI). We sought to compare omissions, duplications, and delays in outpatient medications used for secondary prevention among veterans treated for MI at Veterans Affairs (VA) versus non-VA hospitals. METHODS: We utilized national VA and Centers for Medicare and Medicaid Services data (2012-2018) to identify veterans 65 years or older hospitalized for MI and measured the use of outpatient medications for secondary prevention in the 30 days after MI among those treated at VA versus non-VA hospitals. RESULTS: A total of 118,456 veterans experiencing MI were included; of which 102,209 were hospitalized at non-VA hospitals. An omission in any medication class occurred more frequently among veterans treated at non-VA versus VA hospitals (82.8% vs 67.8%, P < 0.001). In multivariable modeling, the odds of omissions in any medication class were higher among those treated at non-VA versus VA hospitals (odds ratio: 3.04; 95% CI: 2.88-3.20). Duplications occurred more frequently in veterans treated at non-VA versus VA hospitals: 1.9% versus 1.6% had 1 or more for non-VA versus VA hospitals ( P < 0.001). Veterans treated at non-VA hospitals were more likely to have delays of 3 days or more in prescription fills after hospital discharge (88.4% vs 70.6% across all classes, P < 0.001). CONCLUSIONS: Omissions, duplications, and delays in outpatient prescribing of secondary prevention medications were more common among 118,456 veterans treated at non-VA versus VA hospitals for MI. Interventions aimed at improving care transitions and optimizing medication use among veterans treated at non-VA hospitals should be implemented.

Geographic distribution of United States clinical trial sites utilized for guideline changing studies of cholesterol management.

BACKGROUND: This study aimed to evaluate the geographic distribution of United States (US) clinical trial sites utilizedfor guideline changing studies of cholesterol management. METHODS: Randomized trials evaluating pharmacologic interventions for cholesterol treatment and reporting location data (ie, zip code of trial sites) were identified. Location data was abstracted from ClinicalTrials.gov. RESULTS: Half of US counties were over 30 miles from a study site and, social determinants of health were more favorable in counties with versus without clinical trial sites. CONCLUSIONS: Stakeholders such as regulatory bodies andtrial sponsors should incentivize and support infrastructure that would enable a larger number of US counties to be utilized for clinical trial sites. TRIAL REGISTRATION: Not applicable.

Effectiveness and Safety of Twice- Versus Thrice-Daily Subcutaneous Heparin for Venous Thromboembolism Prophylaxis at a Large Academic Medical Center.

BACKGROUND: Dosing variation of subcutaneous unfractionated heparin (UFH) exist for venous thromboembolism prophylaxis (VTEP). OBJECTIVE: The purpose of this study was to compare the safety and effectiveness of thrice-daily (TID) versus twice-daily (BID) administration of UFH during a heparin shortage for VTEP. METHODS: A single-center retrospective analysis was conducted in patients with orders for BID subcutaneous UFH during a heparin shortage from September 1, 2019, to February 4, 2020. These patients were matched to patients with TID subcutaneous UFH orders from January 1, 2019, to May 31, 2019. The primary outcome was the incidence of deep-vein thrombosis or pulmonary embolism confirmed by imaging during hospitalization. The secondary outcome was the incidence of major or clinically relevant nonmajor bleeding events as defined by International Society on Thrombosis and Haemostasis (ISTH) definitions. RESULTS: A total of 277 patients with orders for BID UFH and meeting inclusion criteria were evaluated and matched to patients who received TID UFH. After the exclusion criteria were implemented, 510 patients remained in the TID group. The primary outcome occurred in 4% of patients in the BID group and 3% in the TID group (P = 0.645). Major bleeding or clinically relevant nonmajor bleeding events occurred in 10% of patients in the BID group and 8% in the TID group (P = 0.310). CONCLUSION AND RELEVANCE: There was no difference in effectiveness or safety of TID versus BID subcutaneous UFH for VTEP. During a heparin shortage, transitioning patients to BID UFH for VTEP to conserve supply may be considered.

Risk Factors for Antibiotic Resistant Urinary Pathogens in Patients Discharged From the Emergency Department.

Purpose: Urinary tract infections (UTIs) are one of the most common indications for antimicrobial use in the emergency department (ED). Appropriate empiric selection is crucial to ensure optimal care while limiting broad-spectrum antibiotic use. The primary objective of this study was to evaluate the relationship between patient-specific risk factors and drug resistant urinary pathogens in patients discharged from the ED and followed by Emergency Medicine Pharmacists (EMPs). Methods: This was a single-center, retrospective chart review of adult (≥18 years old) patients with positive urine cultures discharged from the ED. The association between risk factors and pathogen resistance to ≥1 classes of antibiotics was evaluated using multivariate logistic regression. Risk factors included the following: hospitalization within the previous 30 days, intravenous antibiotic use within 90 days, diabetes, clinical atherosclerotic cardiovascular disease, psychiatric disorder, dementia, current antibiotic use for any indication, previous lifetime history of UTIs, indwelling or intermittent catheterization, hemodialysis, previous lifetime history of a urologic procedure, urinary tract abnormality, immunosuppressive disease or medications, current residence in a nursing or rehabilitation facility, and history of a multidrug resistant organism (MDRO). Results: A total of 1018 patients were included. There was an increase in the odds of antibiotic resistance in patients with cystitis and ≥2 risk factors (Odds Ratio [OR] = 1.70, 95% CI = 1.24-2.32). In those with pyelonephritis, there was a non-significant increase in the odds of resistance for those with ≥2 risk factors (OR = 1.83, 95% CI = 0.98-3.42). Patients with pyelonephritis discharged on inappropriate antibiotics were more likely to return to the ED within 30 days (P = .03). Conclusions: For patients with cystitis discharged from the ED, those with ≥2 patient-specific risk factors had significantly increased odds of antibiotic resistance. Patients with pyelonephritis, but not cystitis, who were discharged on inappropriate antibiotics were more likely to return to the ED within 30 days. In conjunction with an EMP culture follow-up program, the identification of risk factors for antimicrobial resistance can be used to design more patient-specific empiric antibiotic selections.

Intravenous insulin for the management of non-emergent hyperglycemia in the emergency department.

PURPOSE: There is currently no consensus regarding the necessity of emergency department (ED) glucose reduction to manage hyperglycemia in patients presenting without a hyperglycemic emergency. Known consequences of intravenous (IV) insulin administration include hypoglycemia, hypokalemia, and increased ED length of stay. The primary objective of this study was to assess the impact of IV regular insulin on glucose reduction and ED length of stay in patients presenting to the ED with non-emergent hyperglycemia. Secondary objectives included the characterization of potential adverse events. METHODS: This was a retrospective, observational study of patients ≥18 years who received IV regular insulin and were discharged from the ED at a large academic Trauma Center. Univariate and multivariable regression analyses were utilized to determine if an association existed between IV insulin administration and blood glucose as well as ED length of stay. RESULTS: A total of 405 patients were included in the analysis. An insulin dose >5 units was associated with a greater reduction in blood glucose (difference = 37.4 mg/dL; p < .001) but no difference in ED length of stay relative to ≤5 units. Furthermore, 7.9% of patients developed hypokalemia and 0.4% developed hypoglycemia. CONCLUSION: The use of >5 units of IV regular insulin for the management of isolated hyperglycemia in the ED was associated with a modest reduction in blood glucose and no difference in ED length of stay compared with those that received ≤5 units. However, use of IV insulin for this purpose resulted in a 7.9% occurrence of hypokalemia.

Medication adherence to injectable glucagon-like peptide-1 (GLP-1) receptor agonists dosed once weekly vs once daily in patients with type 2 diabetes: A meta-analysis.

BACKGROUND: Suboptimal medication adherence has been associated with increased resource utilisation and mortality among patients with type 2 diabetes (T2D). Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are becoming increasingly important in the treatment of T2D. However, medications in this class differ considerably in their dosing frequency, which may impact adherence. We sought to perform a meta-analysis to compare adherence to injectable GLP-1RAs dosed once weekly vs once daily in patients with T2D. METHODS: Medline and Scopus were searched from 1/2005 to 7/2020 using keywords and MeSH terms pertaining to adherence and GLP-1RAs. Studies of adults with T2D were included if they compared adherence (as measured by proportion of days covered [PDC]) to injectable GLP-1RAs dosed once weekly vs once daily. A meta-analysis of non-overlapping studies was performed to evaluate the primary outcome of non-adherence, defined as the proportion of patients with a PDC < 80. RESULTS: A total of 7 studies evaluating 75 159 patients (range: 2886-30 097) with T2D were included. The follow-up periods of included studies ranged from 6 to 12 months. Injectable GLP-1RAs dosed once weekly were either dulaglutide, albiglutide or exenatide extended release; while liraglutide was the injectable once daily agent evaluated in all included studies. Upon meta-analysis, once weekly GLP-1RA dosing was associated with an 11% lower risk of non-adherence compared to once daily dosing (risk ratio = 0.89; 95% confidence interval = 0.83-0.95; I2  = 89%). CONCLUSION: Once weekly dosing of injectable GLP-1RAs was associated with better adherence vs once daily dosing among patients with T2D. These findings coupled with the known detrimental consequences of non-adherence suggest that dosing frequency is an important factor to consider when selecting a GLP-1RA.

Development of a Pharmacy Residency Research Certificate Program.

Background: Similar to teaching certificate programs, which have been incorporated in many pharmacy residencies to improve teaching skills, resident research certificate programs could provide residents with more exposure to biostatistics and research design. However, few research certificate programs have been described. Objective: To assess the impact of a research certificate program on residents' attitude toward, confidence in, and knowledge of research. Methods: A resident research certificate program was created for the 2018-2019 residency class at our institution, which included 33 pharmacy residents (15 postgraduate year-1 [PGY1] and 18 PGY2 residents). The program consisted of 7 sessions aimed at providing residents with research training. To earn a certificate, participants were required to attend 6 or more sessions and achieve a score greater than or equal to 70% on a postassessment. An optional questionnaire assessing attitude, confidence, and knowledge was also administered at baseline and following the certificate program. Results: Of the 33 residents participating in the research certificate program, 21 (9 PGY1 and 12 PGY2) residents completed both the baseline and follow-up questionnaire and were included in the analysis. All items assessing attitude and 3 of 6 items assessing confidence improved significantly following the certificate program when compared to baseline (P < .05 for all). Median knowledge scores were no different on the baseline versus follow-up questionnaire (P = .54). Conclusions: After completing a research certificate program, resident knowledge scores did not differ from baseline, but attitudes toward research and confidence appeared to improve. Research certificate programs could be an avenue for providing pharmacy residents with increased exposure to and comfort with biostatistics and research design.

Costs and factors associated with heart failure following kidney transplantation - a single-center retrospective cohort study.

The number of adults with heart failure (HF) will increase by ~50% between 2012 and 2030. Among kidney transplant recipients, HF accounts for 16% of all post-transplant admissions. We describe the burden of HF and predictors of healthcare utilization following kidney transplantation. We retrospectively identified adults who underwent kidney transplantation at our institution (01/2007-12/2017). Data were acquired from electronic health records, with healthcare utilization obtained from a statewide database. The HF incidence rate and prevalence were estimated for each year, total charges for HF and non-HF patients were compared, and logistic regression was employed for a 3-year predictive model of healthcare utilization associated with HF. Among 1731 kidney transplant recipients, the post-transplant HF incidence rate ranged from 1.91 (year 3) to 6.80 (year 10) per 100 person-years, while the prevalence increased from 31.7% (year 1) to 48.1% (year 10). Median charges were $75 837 (HF) compared to $42 940 (non-HF) per person-year (P < 0.001). Pretransplant HF [odds ratio (OR) = 3.12] and an eGFR < 45 (OR = 4.73) were the strongest predictors of HF encounters (P < 0.05 for both). We observed a high and increasing prevalence of HF, which was associated with twice the costs. Kidney transplant recipients would benefit from interventions aimed at mitigating HF risk factors.

Joint effect of race/ethnicity or location of residence and sex on low density lipoprotein-cholesterol among veterans with type 2 diabetes: a 10-year retrospective cohort study.

BACKGROUND: Cardiovascular (CV) disease is the leading cause of death among United States women. Rural residence and ethnic-minority status are individually associated with increased CV mortality. Managing dyslipidemia is important in the prevention of CV mortality. However, the impact of race/ethnicity and location of residence on sex differences in dyslipidemia management is not well understood. Therefore, we sought to understand the joint effects of race/ethnicity and location of residence on lipid management differences between veteran men and women with type 2 diabetes (T2D). METHODS: Veterans Health Administration and Centers for Medicare and Medicaid Services data were used to perform a longitudinal cohort study of veterans with T2D (2007-2016). Mixed effects logistic regression with a random intercept was used to model the association between sex and low-density lipoprotein (LDL) > 100 mg/dL and its interaction with race/ethnicity and location of residence after adjusting for all measured covariates. RESULTS: When female sex and rural location of residence were both present, they were associated with an antagonistic harmful effect on LDL. Similar antagonistic harmful effects on LDL were observed when the joint effect of female sex and several minority race/ethnicity groups were evaluated. After adjusting for measured covariates, the odds of LDL > 100 mg/dL were higher for urban women (OR = 2.66, 95%CI 2.48-2.85) and rural women (OR = 3.26, 95%CI 2.94-3.62), compared to urban men. The odds of LDL > 100 mg/dL was the highest among non-Hispanic Black (NHB) women (OR = 5.38, 95%CI 4.45-6.51) followed by non-Hispanic White (NHW) women (OR = 2.59, 95%CI 2.44-2.77), and Hispanic women (OR = 2.56, 95%CI 1.79-3.66). CONCLUSION: Antagonistic harmful effects on LDL were observed when both female sex and rural location of residence were present. These antagonistic effects on LDL were also present when evaluating the joint effect of female sex and several minority race/ethnicity groups. Disparities were most pronounced in NHB and rural women, who had 5.4 and 3.3 times the odds of elevated LDL versus NHW and urban men after adjusting for important covariates. These striking effect sizes in a population at high cardiovascular risk (i.e., older with T2D) suggest interventions aimed at improving lipid management are needed for individuals falling into one or more groups known to face health disparities.

Medication adherence to rivaroxaban and dabigatran in patients with non-valvular atrial fibrillation: a meta-analysis.

Several real-world studies have evaluated adherence to direct acting oral anticoagulants in patients with non-valvular atrial fibrillation (NVAF); however, these studies have not been systematically summarized. We performed a meta-analysis to compare adherence to rivaroxaban versus dabigatran therapy in United States (US) patients with NVAF in a real-world setting. Medline and Scopus were searched from January 2010 to August 2018 using keywords and MeSH terms related to adherence and oral anticoagulants. We included real-world studies of US adults with NVAF comparing adherence to dabigatran and rivaroxaban. Studies evaluating adherence by a measure other than proportion of days covered (PDC) were excluded. The proportion of patients with a PDC ≥ 80 (a commonly utilized definition of adherence) served as the primary outcome of interest. We conducted meta-analysis of non-overlapping studies using the Hartung-Knapp random-effects model to estimate risk ratios (RRs) with corresponding 95% confidence intervals (CIs). We included 5 studies evaluating 80,230 patients (range 2667-22,571). Median follow-up across studies was 6 months (range 3-12 months). The proportion of patients with a PDC ≥ 80 ranged from 59.5 to 83.5% for rivaroxaban users and 57.3 to 78.3% for dabigatran users. Upon meta-analysis, rivaroxaban use was associated with increased adherence compared with dabigatran use (RR 1.08; 95% CI 1.03-1.12; I2 = 88%). In conclusion, rivaroxaban was associated with increased adherence when compared to dabigatran in ~ 80,000 patients in a real-world setting. Possible explanations for this include dosing frequency or patient tolerance.

Weight-based versus non-weight-based diltiazem dosing in the setting of atrial fibrillation with rapid ventricular response.

PURPOSE: There is conflicting evidence to support the superiority of weight-based (WB) dosing of intravenous (IV) diltiazem over non-weight-based (NWB) dosing strategies in the management of atrial fibrillation (AFib) with rapid ventricular response (RVR). METHODS: A retrospective review evaluated patients presenting to the emergency department (ED) in AFib with RVR and receiving IV diltiazem from 2015 to 2018. Those receiving a NWB dose were compared with those receiving a WB dose based on actual body weight (ABW). Secondary analyses evaluated safety profiles of the regimens and compared response in groups defined by ABW or ideal body weight (IBW). RESULTS: A total of 371 patients were included in the analysis. No significant difference was observed in achieving a therapeutic response (66.5% vs. 73.1%, p = 0.18) or adverse events between the groups. Patients receiving a WB dose were significantly more likely to have a HR < 100 bpm than those receiving a NWB dose (40.9% vs. 53.5%, p = 0.01). When groups were defined by IBW, WB dosing was associated with a significantly higher incidence of achieving a therapeutic response (62.7% vs. 74.3%, p = 0.02). CONCLUSION: In patients presenting with AF with RVR, there was no significant difference in achieving a therapeutic response between the two strategies. A WB dosing approach did result in a greater proportion of patients with a HR < 100 bpm. The utilization of IBW for WB dosing may result in an increased achievement of a therapeutic response.

Hospital admissions and emergency department visits among kidney transplant recipients.

Reducing acute care utilization is a means of improving long-term patient outcomes. We sought to assess high inpatient (IP) admission and standalone emergency department (ED) utilization within a 9-month period post-kidney transplantation and to identify mutable factors to reduce utilization. In this ten-year retrospective study, 1599 adult kidney transplant recipients were identified. A previous transplant, graft loss, or death within 3 months post-transplantation excluded 319 patients. Comprehensive resource utilization data were obtained from a statewide database. Those with ≥2 IP admissions or standalone ED visits 4-12 months post-transplantation were classified as high utilizers. Multivariable logistic regression models were used for examining associations of predictors with high IP or ED utilization. Of 1280 kidney recipients, 209 and 183 were categorized as IP and ED high utilizers, respectively. Factors significantly associated with high IP utilization included valvular disease, body mass index ≥35, and IP or ED use <3 months post-transplantation; while factors associated with high ED utilization included IP or ED use <3 months post-transplantation, younger age, female, smoker, congestive heart failure, depression, and IP or ED use 1 year pre-transplantation. Inpatient and standalone ED utilization within a 9-month period after kidney transplantation is high and associated with sociodemographic factors, mutable comorbidities, and healthcare utilization.

Antibiotic Use and Respiratory Pathogens in Adults With Sickle Cell Disease and Acute Chest Syndrome.

Background: Acute chest syndrome (ACS) is an acute complication of sickle cell disease (SCD). Historically, the most common pathogens were Chlamydophila pneumoniae, Mycoplasma pneumoniae, and respiratory syncytial virus. Pediatric patients receiving guideline-adherent therapy experienced fewer ACS-related and all-cause 30-day readmissions compared with those receiving nonadherent therapy. This has not been evaluated in adults. Objectives: The primary objectives were to characterize antibiotic use and pathogens. The secondary objective was to assess the occurrence of readmissions associated with guideline-adherent and clinically appropriate treatment compared with regimens that did not meet those criteria. Methods: A retrospective cohort analysis was conducted for adults with SCD hospitalized between August 1, 2014, and July 31, 2017, with pneumonia (PNA) or ACS. The study was approved by the institutional review board. Results: A total of 139 patients with 255 hospitalizations were reviewed. Among 41 respiratory cultures, 3 organisms were isolated: Cryptococcus neoformans, Pseudomonas aeruginosa, and budding yeast. Respiratory panels were collected on 121 admissions, with 17 positive for 1 virus; all were negative for Chlamydophila pneumoniae and M pneumoniae. There were significantly more ACS-/PNA-related 7-day readmissions from patients on guideline-adherent regimens compared with nonadherent regimens (3.7% vs 0%; P = 0.04). Conclusion and Relevance: These findings challenge existing knowledge regarding the most common pathogens in adults with SCD with ACS or PNA. Routine inclusion of a macrolide may not be necessary. Future studies focused on pathogen characterization with standardized assessment are necessary to determine appropriate empirical therapy in this population.

External validation of three risk stratification rules in patients presenting with pulmonary embolism and cancer.

Numerous risk stratification rules exist to predict post-pulmonary embolism (PE) mortality; however, few were designed for use in cancer patients. In the EPIPHANY registry, adapted versions of common rules (the Hestia criteria, Pulmonary Embolism Severity Index [PESI], and simplified PESI [sPESI]) displayed high sensitivity for prognosticating mortality in PE patients with cancer. These adapted rules have yet to be externally validated. Therefore, we sought to evaluate the performance of an adapted Hestia criteria, PESI, and sPESI for predicting 30-day post-PE mortality in patients with cancer. We identified consecutive, adults presenting with objectively confirmed PE and cancer to our institution (November 2010 to January 2014). The proportion of patients categorized as low or high risk by these three risk stratification rules was calculated, and each rule's accuracy for predicting 30-day all-cause mortality was determined. Of the 124 patients with PE and active cancer identified, 25 (20%) experienced mortality at 30 days. The adapted Hestia criteria categorized 23 (19%) patients as low risk, while exhibiting a sensitivity of 88% (95% confidence interval [CI] = 68-97%), a negative predictive value NPV of 87% (95% CI = 65-97%), and a specificity of 20% (95% CI = 13-30%). A total of 38 (31%) and 30 (24%) patients were low risk by the adapted PESI and sPESI, with both displaying sensitivities of 92% and NPVs > 93%. Specificities were 36% (95% CI = 27-47%) and 28% (95% CI = 20-38%) for PESI and sPESI. In our external validation, the adapted Hestia, PESI, and sPESI demonstrated high sensitivity but low specificity for 30-day PE mortality in patients with cancer. Larger, prospective trials are needed to optimize strategies for risk stratification in this population.

Comorbid cancer and use of thrombolysis in acute pulmonary embolism.

Guidelines suggest the use of thrombolytic therapy for acute pulmonary embolism (PE) patients with hypotension who are not at high-risk of bleeding. Data describing the use of thrombolysis in patients with cancer are scarce. The aim of this study was to evaluate the relationship between cancer and the use of thrombolysis for acute PE. The 2013 and 2014 US National Inpatient Sample was used to identify admissions for acute PE. Identified admissions were stratified based on the presence or absence of cancer. Multivariable logistic regression was performed to determine the association between comorbid cancer and the odds of receiving thrombolysis after adjustment for patient- and hospital-level covariates. In those receiving thrombolysis, the association between cancer and in-hospital mortality was determined using logistic regression after adjusting for age ≥ 65 years and sex. We identified 72,546 admissions for acute PE; of which, 14.7% (n = 10,673) had comorbid cancer. A total of 3.4% (n = 2439) of patients received thrombolysis. Upon multivariable adjustment, cancer was associated with decreased odds of receiving thrombolysis (odds ratio = 0.55; 95% confidence interval = 0.48-0.64). When the population was restricted to PE admissions receiving thrombolysis, mortality occurred in 315 (12.9%) admissions; with no difference in in-hospital mortality observed between those with versus without cancer (p = 0.11). In this study of admissions for acute PE, comorbid cancer was associated with decreased odds of receiving thrombolysis. As PE is a common complication among patients with cancer, the risk-benefit profile of thrombolysis in this patient population should be determined.

Saddle pulmonary embolism and in-hospital mortality in patients with cancer.

PURPOSE: Saddle pulmonary embolism (PE) has been associated with an increased risk of 1 year mortality when compared to non-saddle PE among patients with cancer. We sought to evaluate the association between saddle PE and in-hospital outcomes among patients with comorbid cancer. METHODS: The 2013 and 2014 United States National Inpatient Sample was used to identify adult patients hospitalized for acute PE. Only patients with an International Classification Diseases, 9th Revision, Clinical Modification (ICD-9-CM) code indicating comorbid cancer were included. Identified admissions were stratified into the following 2 cohorts: saddle (defined as ICD-9-CM code = 415.13) and non-saddle PE. Multivariable logistic regression was performed to determine the association between saddle PE and the odds of in-hospital mortality after adjustment for age ≥ 80 years and sex. RESULTS: A total of 10,660 admissions for acute PE in patients with comorbid cancer were identified. Of which, 4.5% (n = 475) had a saddle PE. Median age was 67 years (interquartile range = 58-76) and 48.9% were male. In-hospital mortality occurred in 6.1% of patients. Upon multivariable adjustment, the odds of in-hospital mortality were higher in saddle versus non-saddle PE (odds ratio = 1.51; 95% confidence interval 1.08-2.10). CONCLUSION: In this retrospective study of admissions for acute PE in patients with comorbid cancer, saddle PE was associated with a higher odds of in-hospital mortality.

Automated detection of altered mental status in emergency department clinical notes: a deep learning approach.

BACKGROUND: Machine learning has been used extensively in clinical text classification tasks. Deep learning approaches using word embeddings have been recently gaining momentum in biomedical applications. In an effort to automate the identification of altered mental status (AMS) in emergency department provider notes for the purpose of decision support, we compare the performance of classic bag-of-words-based machine learning classifiers and novel deep learning approaches. METHODS: We used a case-control study design to extract an adequate number of clinical notes with AMS and non-AMS based on ICD codes. The notes were parsed to extract the history of present illness, which was used as the clinical text for the classifiers. The notes were manually labeled by clinicians. As a baseline for comparison, we tested several traditional bag-of-words based classifiers. We then tested several deep learning models using a convolutional neural network architecture with three different types of word embeddings, a pre-trained word2vec model and two models without pre-training but with different word embedding dimensions. RESULTS: We evaluated the models on 1130 labeled notes from the emergency department. The deep learning models had the best overall performance with an area under the ROC curve of 98.5% and an accuracy of 94.5%. Pre-training word embeddings on the unlabeled corpus reduced training iterations and had performance that was statistically no different than the other deep learning models. CONCLUSION: This supervised deep learning approach performs exceedingly well for the detection of AMS symptoms in clinical text in our environment. Further work is needed for the generalizability of these findings, including evaluation of these models in other types of clinical notes and other environments. The results seem promising for the ultimate use of these types of classifiers in combination with other information derived from the electronic health records as input for clinical decision support.

Should an LDL-Cholesterol Target-Based Approach Be Readopted?

OBJECTIVE: To review clinical trials driving the evolution of hyperlipidemic guidelines, discuss whether low-density lipoprotein (LDL) targets and adjunctive therapy on top of statins should be used, and summarize the pharmacist's role in helping achieve LDL goals. DATA SOURCES: MEDLINE search (1/1980-5/2017) using terms including LDL, lipid, and statin, with forward and backward citation tracking. STUDY SELECTION AND DATA EXTRACTION: English-language studies and guidelines assessing LDL-lowering therapy were included. DATA SYNTHESIS: In 2013, the American Heart Association and American College of Cardiology (AHA/ACC) hyperlipidemic guideline stepped back from LDL goals opting for statin monotherapy with an intensity of dosing predicated on baseline risk. This was driven by abundant clinical trial evidence for the statins, with adjunctive therapy on top of statins failing to show substantial benefit. However, recent evidence suggests that returning to LDL goals is warranted and adjuvant ezetimibe or PCSK9 inhibitor therapy may further reduce cardiovascular events. This is reflected in some society guidelines but not from the AHA/ACC. Pharmacists are well positioned to help achieve LDL goals, as they have positively affected LDL goal attainment across a multitude of settings. CONCLUSIONS: Statins are the mainstay of therapy but patients should have LDL targets, and if patients on maximally tolerated statin doses are not at goal, adjunctive therapy with ezetimibe and PCSK9 inhibitors may improve outcomes. In the emerging LDL goal paradigm, the pharmacist has a critical role to play.

Association of Inhaled Corticosteroids and Long-Acting Muscarinic Antagonists With Asthma Control in Patients With Uncontrolled, Persistent Asthma: A Systematic Review and Meta-analysis.

Importance: Long-acting muscarinic antagonists (LAMAs) are a potential adjunct therapy to inhaled corticosteroids in the management of persistent asthma. Objective: To conduct a systematic review and meta-analysis of the effects associated with LAMA vs placebo or vs other controllers as an add-on therapy to inhaled corticosteroids and the use of a LAMA as add-on therapy to inhaled corticosteroids and long-acting ß-agonists (LABAs; hereafter referred to as triple therapy) vs inhaled corticosteroids and LABA in patients with uncontrolled, persistent asthma. Data Sources: MEDLINE, EMBASE, Cochrane databases, and clinical trial registries (earliest date through November 28, 2017). Study Selection: Two reviewers selected randomized clinical trials or observational studies evaluating a LAMA vs placebo or vs another controller as an add-on therapy to inhaled corticosteroids or triple therapy vs inhaled corticosteroids and LABA in patients with uncontrolled, persistent asthma reporting on an outcome of interest. Data Extraction and Synthesis: Meta-analyses using a random-effects model was conducted to calculate risk ratios (RRs), risk differences (RDs), and mean differences (MDs) with corresponding 95% CIs. Citation screening, data abstraction, risk assessment, and strength-of-evidence grading were completed by 2 independent reviewers. Main Outcomes and Measures: Asthma exacerbations. Results: Of 1326 records identified, 15 randomized clinical trials (N = 7122 patients) were included. Most trials assessed adding LAMA vs placebo or LAMA vs LABA to inhaled corticosteroids. Adding LAMA vs placebo to inhaled corticosteroids was associated with a significantly reduced risk of exacerbation requiring systemic corticosteroids (RR, 0.67 [95% CI, 0.48 to 0.92]; RD, -0.02 [95% CI, -0.04 to 0.00]). Compared with adding LABA, adding LAMA to inhaled corticosteroids was not associated with significant improvements in exacerbation risk (RR, 0.87 [95% CI, 0.53 to 1.42]; RD, 0.00 [95% CI, -0.02 to 0.02]), or any other outcomes of interest. Triple therapy was not significantly associated with improved exacerbation risk vs inhaled corticosteroids and LABA (RR, 0.84 [95% CI, 0.57 to 1.22]; RD, -0.01 [95% CI, -0.08 to 0.07]). Conclusions and Relevance: In this systematic review and meta-analysis, the use of LAMA compared with placebo as add-on therapy to inhaled corticosteroids was associated with a lower risk of asthma exacerbations; however, the association of LAMA with benefit may not be greater than that with LABA. Triple therapy was not associated with a lower risk of exacerbations.