Using machine learning to develop a clinical prediction model for SSRI-associated bleeding: a feasibility study.

INTRODUCTION: Adverse drug events (ADEs) are associated with poor outcomes and increased costs but may be prevented with prediction tools. With the National Institute of Health All of Us (AoU) database, we employed machine learning (ML) to predict selective serotonin reuptake inhibitor (SSRI)-associated bleeding. METHODS: The AoU program, beginning in 05/2018, continues to recruit ≥ 18 years old individuals across the United States. Participants completed surveys and consented to contribute electronic health record (EHR) for research. Using the EHR, we determined participants who were exposed to SSRIs (citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline, vortioxetine). Features (n = 88) were selected with clinicians' input and comprised sociodemographic, lifestyle, comorbidities, and medication use information. We identified bleeding events with validated EHR algorithms and applied logistic regression, decision tree, random forest, and extreme gradient boost to predict bleeding during SSRI exposure. We assessed model performance with area under the receiver operating characteristic curve statistic (AUC) and defined clinically significant features as resulting in > 0.01 decline in AUC after removal from the model, in three of four ML models. RESULTS: There were 10,362 participants exposed to SSRIs, with 9.6% experiencing a bleeding event during SSRI exposure. For each SSRI, performance across all four ML models was relatively consistent. AUCs from the best models ranged 0.632-0.698. Clinically significant features included health literacy for escitalopram, and bleeding history and socioeconomic status for all SSRIs. CONCLUSIONS: We demonstrated feasibility of predicting ADEs using ML. Incorporating genomic features and drug interactions with deep learning models may improve ADE prediction.

Sociodemographic characteristics differ across routine adult vaccine cohorts: An All of Us descriptive study.

BACKGROUND: The National Institutes of Health All of Us (AoU) Research Program is currently building a database of 1million+ adult subjects. With it, we describe the characteristics of those with documented vaccinations. OBJECTIVES: To describe the sociodemographic, health status, and lifestyle factors associated with vaccinations. METHODS: This is a retrospective study involving data from the AoU program (R2020Q4R2, N = 315,297). Five vaccine cohorts [influenza, hepatitis B (HBV), pneumococcal <65 years old, pneumococcal ≥65 years old, and human papillomavirus (HPV)] were generated based on vaccination history. The influenza cohort comprised participants with documented influenza vaccinations in electronic health records (EHRs) from September 2017 to May 2018. Other vaccine cohorts comprised participants with ≥1 lifetime record(s) of vaccination documented in the EHR by December 2018. The vaccine cohorts were compared to the overall AoU cohort. Descriptive statistics were generated using EHR- and survey-based sociodemographic, health, and lifestyle information. The SAMBA (0.9.0) R package was utilized to adjust for EHR selection and outcome misclassification biases to infer sources of disparity for pneumococcal vaccinations in older adults. RESULTS: Cohort counts were as follows: influenza (n = 15,346), HBV (n = 6323), pneumococcal <65 (n = 15,217), pneumococcal ≥65 (n = 15,100), and HPV (n = 2125). All vaccine cohorts had higher proportions of White and non-Hispanic/Latino participants compared to the overall AoU cohort. The largest differences were found in pneumococcal age ≥65, with 80.2% White participants compared to 52.9% in the overall study population. Multivariable analysis revealed that race/ethnic disparities in pneumococcal vaccination among older adults were explained by biological sex, income, health insurance, and education-related variables. CONCLUSION: Racial, ethnic, education, and income characteristics differ across the vaccine cohorts among AoU participants. These findings inform future utilization of large health databases in vaccine epidemiology research and emphasize the need for more targeted interventions that address differences in vaccine uptake.

Foodborne botulism treated with heptavalent botulism antitoxin.

OBJECTIVE: To report a case of foodborne botulism and subsequent use of the investigational heptavalent botulism antitoxin (H-BAT). CASE SUMMARY: A 60-year-old man was hospitalized with blurred vision, diplopia, and dysarthria. On hospital day 2, the patient was transferred to the intensive care unit for progressive fatigable weakness with ptosis, dysphagia, dysarthria, and nausea. Secondary to worsening respiratory distress, the patient was intubated and placed on a ventilator. The patient could open his eyes only with assistance but still had normal strength in all extremities. H-BAT was administered 48 hours after presentation for possible botulism. The patient then revealed that he consumed home-canned corn several days prior to admission. On hospital day 8, botulinum neurotoxin was confirmed in the patient's serum and the home-canned corn. The patient slowly regained muscle strength and was discharged to a long-term acute care facility on hospital day 22. DISCUSSION: Foodborne botulism is caused by a neurotoxin from Clostridium botulinum and usually occurs after the consumption of improperly prepared home-canned food. Botulism is characterized by symmetrical descending paralysis that may progress to respiratory arrest. The standard confirmatory test for botulism is a mouse bioassay to prove the presence of botulinum neurotoxin. Outside of supportive care, the treatment options for botulism are limited. Individuals with botulism often require intensive care unit monitoring and potentially ventilatory support. H-BAT, the only treatment available for botulism in patients older than 1 year, is a purified and despeciated equine-derived immunoglobulin active against all known botulinum neurotoxins. H-BAT's despeciation significantly reduces the risk of hypersensitivity reactions, anaphylaxis, and serum sickness. CONCLUSIONS: In a confirmed case of foodborne botulism treated with H-BAT, the patient tolerated H-BAT and did not develop any hypersensitivity reactions or serum sickness.

Health Disparities in Pharmacy Practice Within the Community: Let's Brainstorm for Solutions.

Health disparity is defined as a type of health difference that is closely linked with social, economic and/or environmental disadvantage. Over the past two decades, major efforts have been undertaken to mitigate health disparities and promote health equity in the United States. Within pharmacy practice, health disparities have also been identified to play a role in influencing pharmacists' practice across various clinical settings. However, well-characterized solutions to address such disparities, particularly within pharmacy practice, are lacking in the literature. Recognizing that a significant amount of work will be necessary to reduce or eliminate health disparities, the University of California, Irvine (UCI) School of Pharmacy and Pharmaceutical Sciences held a webinar in June 2021 to explore pertinent issues related to this topic. During the session, participants were given the opportunity to propose and discuss innovative solutions to overcome health disparities in pharmacy practice. The goal of this perspective article is to distill the essence of the presentations and discussions from this interactive session, and to synthesize ideas for practical solutions that can be translated to practice to address this public health problem.

Evaluation of standard versus reduced dose apixaban for the treatment of venous thromboembolism in patients with severe renal disease (ESRD-VTE).

BACKGROUND: There are no clear dosing recommendations when using apixaban for venous thromboembolism (VTE) treatment in patients with severe or end-stage renal disease; clinical trials excluded patients with a creatinine clearance (CrCl) <25 mL/min or on dialysis. This study compares bleeding rates in patients with severe or end-stage renal disease taking standard versus reduced dose apixaban for VTE treatment. MATERIALS AND METHODS: This was a multicenter, retrospective cohort study using electronic medical records between January 1, 2013, and August 31, 2021. This study included patients 18 years or older who had severe or end-stage renal disease when prescribed apixaban for VTE treatment. Severe or end-stage renal disease was defined as at least one of the following: CrCl <25 mL/min, SCr >2.5 mg/dL, CKD stage 4 or 5, or on dialysis. The primary endpoint was rate of clinically relevant bleeding within six months of starting apixaban. Secondary endpoints were VTE recurrence within six months of starting apixaban, time to clinically relevant bleed, and time to VTE recurrence. RESULTS: A total of 203 patients were included in the final analysis (n = 125 on 5 mg; n = 78 on 2.5 mg). Clinically relevant bleeding rate was significantly higher in the standard dose group (14.4 % vs 3.8 %, p = 0.02). Rates of VTE recurrence appear similar (6.4 % vs 7.7 %, p = 0.21). CONCLUSIONS: A reduced dose of apixaban may be considered when treating VTE in patients with severe or end-stage renal disease.

Improved Patient-Reported Medication Adherence, Patient Satisfaction, and Glycemic Control in a Collaborative Care Pharmacist-Led Diabetes "Tune-Up" Clinic.

Diabetes complications remain a leading cause of death, which may be due to poor glycemic control resulting from medication nonadherence. The relationship between adherence status and HbA1c (glycemic control) has not been well-studied for clinical pharmacist interventions. This study evaluated medication adherence, patient satisfaction, and HbA1c, in a collaborative pharmacist-endocrinologist diabetes clinic over 6 months. Of 127 referred, 83 patients met the inclusion criteria. Mean medication adherence scores, considered "good" at baseline, 1.4 ± 1.2, improved by 0.05 points (p = 0.018), and there was a 26% increase in patients with good adherence. A significant improvement of 0.40 percentage points (95% CI: -0.47, -0.34) was observed in mean HbA1c across the three time points (p < 0.001). Mean total satisfaction scores were high and increased, with mean 91.3 ± 12.2 at baseline, 94.7 ± 9.6 at 3 months, and 95.7 ± 10.8 at 6 months (p = 0.009). A multimodal personalized treatment approach from a pharmacist provider significantly and positively impacted glycemic control regardless of self-reported medication adherence, and patient satisfaction remained high despite changing to a clinical pharmacist provider and increased care intensity.

Estimated Cost-Effectiveness, Cost Benefit, and Risk Reduction Associated with an Endocrinologist-Pharmacist Diabetes Intense Medical Management "Tune-Up" Clinic.

BACKGROUND: In 2012 U.S. diabetes costs were estimated to be $245 billion, with $176 billion related to direct diabetes treatment and associated complications. Although a few studies have reported positive glycemic and economic benefits for diabetes patients treated under primary care physician (PCP)-pharmacist collaborative practice models, no studies have evaluated the cost-effectiveness of an endocrinologist-pharmacist collaborative practice model treating complex diabetes patients versus usual PCP care for similar patients. OBJECTIVE: To estimate the cost-effectiveness and cost benefit of a collaborative endocrinologist-pharmacist Diabetes Intense Medical Management (DIMM) "Tune-Up" clinic for complex diabetes patients versus usual PCP care from 3 perspectives (clinic, health system, payer) and time frames. METHODS: Data from a retrospective cohort study of adult patients with type 2 diabetes mellitus (T2DM) and glycosylated hemoglobin A1c (A1c) ≥ 8% who were referred to the DIMM clinic at the Veterans Affairs San Diego Health System were used for cost analyses against a comparator group of PCP patients meeting the same criteria. The DIMM clinic took more time with patients, compared with usual PCP visits. It provided personalized care in three 60-minute visits over 6 months, combining medication therapy management with patient-specific diabetes education, to achieve A1c treatment goals before discharge back to the PCP. Data for DIMM versus PCP patients were used to evaluate cost-effectiveness and cost benefit. Analyses included incremental cost-effectiveness ratios (ICERs) at 6 months, 3-year estimated total medical costs avoided and return on investment (ROI), absolute risk reduction of complications, resultant medical costs, and quality-adjusted life-years (QALYs) over 10 years. RESULTS: Base case ICER results indicated that from the clinic perspective, the DIMM clinic costs $21 per additional percentage point of A1c improvement and $115-$164 per additional patient at target A1c goal level compared with the PCP group. From the health system perspective, medical cost avoidance due to improved A1c was $8,793 per DIMM patient versus $3,506 per PCP patient (P = 0.009), resulting in an ROI of $9.01 per dollar spent. From the payer perspective, DIMM patients had estimated lower total medical costs, a greater number of QALYs gained, and appreciable risk reductions for diabetes-related complications over 2-, 5- and 10-year time frames, indicating that the DIMM clinic was dominant. Sensitivity analyses indicated results were robust, and overall conclusions did not change appreciably when key parameters (including DIMM clinic effectiveness and cost) were varied within plausible ranges. CONCLUSIONS: The DIMM clinic endocrinologist-pharmacist collaborative practice model, in which the pharmacist spent more time providing personalized care, improved glycemic control at a minimal cost per additional A1c benefit gained and produced greater cost avoidance, appreciable ROI, reduction in long-term complication risk, and lower cost for a greater gain in QALYs. Overall, the DIMM clinic represents an advanced pharmacy practice model with proven clinical and economic benefits from multiple perspectives for patients with T2DM and high medication and comorbidity complexity. DISCLOSURES: No outside funding supported this study. The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Preliminary versions of the study data were presented in abstract form at the American Pharmacists Association Annual Meeting & Exposition; March 27, 2015; San Diego, California, and the Academy of Managed Care Pharmacy Annual Meeting; April 21, 2016; San Francisco, California. Study concept and design were contributed by Hirsch, Bounthavong, and Edelman, along with Morello and Morreale. Arjmand, Ourth, Ha, Cadiz, and Zimmerman collected the data. Data interpretation was performed by Ha, Morreale, and Morello, along with Cadiz, Ourth, and Hirsch. The manuscript was written primarily by Hirsch and Zimmerman, along with Arjamand, Ourth, and Morello, and was revised by Hirsch and Cadiz, along with Bounthavong, Ha, Morreale, and Morello.