Patient Perceptions of Audio-Only Versus Video Telehealth Visits: A Qualitative Study Among Patients in an Academic Medical Center Setting.

Introduction: Telehealth utilization surged during the COVID-19 pandemic, offering expanded health care access. Audio-only visits emerged as a crucial tool for patients facing technology or connectivity barriers to still use telehealth. This qualitative study aims to better understand patient perceptions of audio-only versus video telehealth visits during the COVID-19 pandemic, and how patients perceive the role of each in their overall health care. Methods: Semi-structured interviews were conducted with 14 adult patients seeking care at an academic medical center located in the Southeast region of the United States. Patients had experienced both an audio-only and video telehealth visit within the past 6 months. Topics covered in the interview included comfort, preference, quality, and communication during each type of visit. Interviews were transcribed verbatim, coded, and analyzed using a general inductive approach. Results: Participants valued having both modalities available largely due to convenience and saw these visits as supplemental or supporting their in-person care. Preferences for visit types were varied among participants and were context-specific, influenced by visit purpose and provider rapport. Patients viewed audio-only visits favorably for informational follow-ups and highlighted their convenience, particularly for multitasking and caregiving duties. In contrast, video visits were seen as more effective for communication due to visual cues and better suited for demonstrating health conditions. Audio-only visits were also seen as less technology-dependent and served as a vital back-up to failed video encounters. Discussion: Despite varied preferences, patients perceived both modalities as complementary to in-person care. Concerns around the quality of care were mitigated by patients' and providers' judicious use of visit types based on clinical appropriateness and existing rapport. The results emphasize the necessity and flexibility of audio-only visits in ensuring equitable access to telehealth, especially for those with technology limitations or demanding responsibilities. To maintain the access and convenience afforded by telehealth and ensure these benefits are offered equitably, policy makers and health care organizations must continue to provide flexible telehealth options, including audio-only visits.

County Mammogram Uptake can be Predicted from Social Determinants of Health, but Patterns Do Not Hold for Individual Patients.

Purpose: The objective of this study is to describe patterns in barriers to breast cancer screening uptake with the end goal of improving screening adherence and decreasing the burden of mortality due to breast cancer. This study looks at social determinants of health and their association to screening and mortality. It also investigates the extent that models trained on county data are generalizable to individuals. Methods: County level screening uptake and age adjusted mortality due to breast cancer are combined with the Centers for Disease Controls Social Vulnerability Index (SVI) to train a model predicting screening uptake rates. Patterns learned are then applied to de-identified electronic medical records from individual patients to make predictions on mammogram screening follow through. Results: Accurate predictions can be made about a county's breast cancer screening uptake with the SVI. However, the association between increased screening, and decreased age adjusted mortality, doesn't hold in areas with a high proportion of minority residents. It is also shown that patterns learned from county SVI data have little discriminative power at the patient level. Conclusion: This study demonstrates that social determinants in the SVI can explain much of the variance in county breast cancer screening rates. However, these same patterns fail to discriminate which patients will have timely follow through of a mammogram screening test. This study also concludes that the core association between increased screening and decreased age adjusted mortality does not hold in high proportion minority areas. Objective: The objective of this study is to describe patterns in social determinants of health and their association with female breast cancer screening uptake, age adjusted breast cancer mortality rate and the extent that models trained on county data are generalizable to individuals.

Differences in rehabilitation evaluation access for rural and socially disadvantaged stroke survivors.

BACKGROUND: Most stroke survivors have ongoing deficits and report unmet needs. Despite evidence that rehabilitation improves stroke survivors' function, access to occupational and physical therapy is limited. Describing access to care for disadvantaged communities for different levels of stroke severity will provide proportions used to create Markov economic models to demonstrate the value of rehabilitation. OBJECTIVES: The objective of this study was to explore differences in the frequency of rehabilitation evaluations via outpatient therapy and home health for Medicare Part B ischemic stroke survivors in rural and socially disadvantaged locations. METHODS: We completed a retrospective, descriptive cohort analysis using the 2018 and 2019 5% Medicare Limited Data Sets (LDS) from the Centers for Medicare and Medicaid Services using STROBE guidelines for observational studies. We extracted rehabilitation Current Procedural Terminology (CPT) codes for those who received occupational or physical therapy to examine differences in therapy evaluations for rural and socially disadvantaged populations. RESULTS: Of the 9,076 stroke survivors in this cohort, 44.2% did not receive any home health or outpatient therapy. Of these, 64.7% had a moderate or severe stroke, indicating an unmet need for therapy. Only 2.0% of stroke survivors received outpatient occupational therapy within the first year Rural and socially disadvantaged communities accessed rehabilitation evaluations at lower rates than general stroke survivors. CONCLUSIONS: These findings describe the poor access to home health and outpatient rehabilitation for stroke survivors, particularly in traditionally underserved populations. These results will influence future economic evaluations of interventions aimed at improving access to care.

Projecting the Clinical and Economic Impacts of Changes to HIV Care Among Adolescents and Young Adults in the United States: Lessons From the COVID-19 Pandemic.

BACKGROUND: During the COVID-19 pandemic, many US youth with HIV (YHIV) used telehealth services; others experienced disruptions in clinic and antiretroviral therapy (ART) access. METHODS: Using the Cost-effectiveness of Preventing AIDS Complications (CEPAC)-Adolescent HIV microsimulation model, we evaluated 3 scenarios: 1) Clinic: in-person care; 2) Telehealth: virtual visits, without CD4 or viral load monitoring for 12 months, followed by return to usual care; and 3) Interruption: complete care interruption with no ART access or laboratory monitoring for 6 months (maximum clinic closure time), followed by return to usual care for 80%. We assigned higher 1-year retention (87% vs 80%) and lower cost/visit ($49 vs $56) for Telehealth vs Clinic. We modeled 2 YHIV cohorts with non-perinatal (YNPHIV) and perinatal (YPHIV) HIV, which differed by mean age (22 vs 16 years), sex at birth (85% vs 47% male), starting CD4 count (527/µL vs 635/µL), ART, mortality, and HIV-related costs. We projected life months (LMs) and costs/100 YHIV over 10 years. RESULTS: Over 10 years, LMs in Clinic and Telehealth would be similar (YNPHIV: 11 350 vs 11 360 LMs; YPHIV: 11 680 LMs for both strategies); costs would be $0.3M (YNPHIV) and $0.4M (YPHIV) more for Telehealth than Clinic. Interruption would be less effective (YNPHIV: 11 230 LMs; YPHIV: 11 620 LMs) and less costly (YNPHIV: $1.3M less; YPHIV: $0.2M less) than Clinic. Higher retention in Telehealth led to increased ART use and thus higher costs. CONCLUSIONS: Telehealth could be as effective as in-person care for some YHIV, at slightly increased cost. Short interruptions to ART and laboratory monitoring may have negative long-term clinical implications.

Examining Utilization of an Outpatient Telenutrition Service Across Primary Care Clinics in South Carolina.

Introduction: Limited research exists on outpatient telenutrition, and more evidence is needed on service utilization and program evaluation. This study explored service utilization trends and patterns of the Medical University of South Carolina (MUSC) Outpatient Telehealth Nutrition (OT Nutrition) service. Methods: De-identified patient service utilization data were obtained from MUSC's OT Nutrition administrative files (2012-2020). Service utilization (referrals, visits scheduled, consultations, no-shows, no-show rate) was measured at the clinic level and stratified by patient type (adult/pediatric) and clinic rurality (rural/urban). Data were analyzed using descriptive statistics and a K-means cluster analysis. Results: Service utilization (2012-2020) reflected 6,212 referrals, 3,993 visits scheduled, and 2,880 consultations across 56 clinics. Yearly utilization trends showed high variability with no statistically significant differences observed on univariate comparisons of patient type or clinic rurality. The introduction of the Direct-To-Consumer modality mitigated a 36.7% decrease in consultations during the COVID-19 pandemic in 2020. Results of a K-means cluster analysis (p < 0.001) indicated 7% (n = 4) of clinics were very high and high utilizers, 36% (n = 21) were moderate and low utilizers, and 53% (n = 31) were very low utilizers. Discussion: Telenutrition can be delivered effectively to patients without requiring travel outside patients' medical homes or residences. Although continued advocacy is necessary for South Carolina to expand telenutrition coverage, more research is needed to evaluate the OT Nutrition service. Cluster analysis is an effective tool for identifying statistically significant groupings of clinics based on service utilization and could be used with implementation science in future program evaluation.

Addressing COVID-19 vaccine hesitancy in rural community pharmacies: a protocol for a stepped wedge randomized clinical trial.

BACKGROUND: Uptake of COVID-19 vaccines remains problematically low in the USA, especially in rural areas. COVID-19 vaccine hesitancy is associated with lower uptake, which translates to higher susceptibility to SARS-CoV-2 variants in communities where vaccination coverage is low. Because community pharmacists are among the most accessible and trusted health professionals in rural areas, this randomized clinical trial will examine implementation strategies to support rural pharmacists in delivering an adapted evidence-based intervention to reduce COVID-19 vaccine hesitancy. METHODS: We will use an incomplete stepped wedge trial design in which we will randomize 30 rural pharmacies (unit of analysis) to determine the effectiveness and incremental cost-effectiveness of a standard implementation approach (consisting of online training that describes the vaccine hesitancy intervention, live webinar, and resource website) compared to adding on a virtual facilitation approach (provided by a trained facilitator in support of the delivery of the vaccine hesitancy counseling intervention by pharmacists). The intervention (ASORT) has been adapted from an evidence-based vaccine communication intervention for HPV vaccines through a partnership with rural pharmacies in a practice-based research network in seven southern US states. ASORT teaches pharmacists how to identify persons eligible for COVID-19 vaccination (including a booster), solicit and address vaccine concerns in a non-confrontational way, recommend the vaccine, and repeat the steps later if needed. The primary trial outcome is fidelity to the ASORT intervention, which will be determined through ratings of recordings of pharmacists delivering the intervention. The secondary outcome is the effectiveness of the intervention, determined by rates of patients who agree to be vaccinated after receiving the intervention. Other secondary outcomes include feasibility, acceptability, adoption, reach, and cost. Cost-effectiveness and budget impact analyses will be conducted to maximize the potential for future dissemination and sustainability. Mixed methods will provide triangulation, expansion, and explanation of quantitative findings. DISCUSSION: This trial contributes to a growing evidence base on vaccine hesitancy interventions and virtual-only facilitation of evidenced-based practices in community health settings. The trial will provide the first estimate of the relative value of different implementation strategies in pharmacy settings. TRIAL REGISTRATION: NCT05926544 (clinicaltrials.gov); 07/03/2023.

Predicting Mammogram Screening Follow Through with Electronic Health Record and Geographically Linked Data.

Cancer is the second leading cause of death in the United States, and breast cancer is the fourth leading cause of cancer-related death, with 42,275 women dying of breast cancer in the United States in 2020. Screening is a key strategy for reducing mortality from breast cancer and is recommended by various national guidelines. This study applies machine learning classification methods to the task of predicting which patients will fail to complete a mammogram screening after having one ordered, as well as understanding the underlying features that influence predictions. The results show that a small group of patients can be identified that are very unlikely to complete mammogram screening, enabling care managers to focus resources. SIGNIFICANCE: The motivation behind this study is to create an automated system that can identify a small group of individuals that are at elevated risk for not following through completing a mammogram screening. This will enable interventions to boost screening to be focused on patients least likely to complete screening.

Mammogram Uptake from Social Determinants of Health Can Be Lost in Translation to Individual Patients.

Purpose: The objective of this study is to describe patterns in barriers to breast cancer screening uptake with the end goal of improving screening adherence and decreasing the burden of mortality due to breast cancer. This study looks at social determinants of health and their association to screening and mortality. It also investigates the extent that models trained on county data are generalizable to individuals. Methods: County level screening uptake and age adjusted mortality due to breast cancer are combined with the Centers for Disease Controls Social Vulnerability Index (SVI) to train a model predicting screening uptake rates. Patterns learned are then applied to de-identified electronic medical records from individual patients to make predictions on mammogram screening follow through. Results: Accurate predictions can be made about a county's breast cancer screening uptake with the SVI. However, the association between increased screening, and decreased age adjusted mortality, doesn't hold in areas with a high proportion of minority residents. It is also shown that patterns learned from county SVI data have little discriminative power at the patient level. Conclusion: This study demonstrates that social determinants in the SVI can explain much of the variance in county breast cancer screening rates. However, these same patterns fail to discriminate which patients will have timely follow through of a mammogram screening test. This study also concludes that the core association between increased screening and decreased age adjusted mortality does not hold in high proportion minority areas.

Cardiovascular Risk Assessment Among Adolescents and Youths Living With HIV: Evaluation of Electronic Health Record Findings and Implications.

BACKGROUND: The HIV epidemic remains a major public health concern, particularly among youths living with HIV. While the availability of antiretroviral therapy has significantly improved the health outcomes of people living with HIV, there is growing evidence that youths living with HIV may be at increased risk of cardiovascular disease. However, the underlying mechanisms linking HIV and cardiovascular disease among youths living with HIV remain poorly understood. One potential explanation is that HIV-related biomarkers, including detectable viral load (VL) and low cluster of differentiation 4 (CD4) lymphocyte counts, may contribute to increased cardiovascular risk. Despite the potential importance of these biomarkers, the relationship between HIV-related biomarkers and cardiovascular risk among youths living with HIV has been understudied. OBJECTIVE: To address this gap, we examined whether detectable VL and low CD4 lymphocyte counts, both of which are indications of unsuppressed HIV, were associated with cardiovascular risk among youths living with HIV. METHODS: We analyzed electronic health record data from 7 adolescent HIV clinics in the United States (813 youths living with HIV). We used multivariable linear regression to examine the relationship between detectable VL and CD4 lymphocyte counts of ≤200 and cardiovascular risk scores, which were adapted from the gender-specific Framingham algorithm. RESULTS: In our study, nearly half of the participants (366/766, 47.8%) had detectable VL, indicating unsuppressed HIV, while 8.6% (51/593) of them had CD4 lymphocyte counts of ≤200, suggesting weakened immune function. We found that those with CD4 lymphocyte counts of ≤200 had significantly higher cardiovascular risk, as assessed by Cardiac Risk Score2, than those with CD4 lymphocyte counts of >200 (P=.002). After adjusting for demographic and clinical factors, we found that for every 1000-point increase in VL copies/mL, the probability of having cardiovascular risk (Cardiac Risk Score2) increased by 38%. When measuring the strength of this connection, we observed a minor effect of VL on increased cardiovascular risk (ß=.134, SE 0.014; P=.006). We obtained similar results with Cardiac Risk Score1, but the effect of CD4 lymphocyte counts of ≤200 was no longer significant. Overall, our findings suggest that detectable VL is associated with increased cardiovascular risk among youths living with HIV, and that CD4 lymphocyte counts may play a role in this relationship as well. CONCLUSIONS: Our study highlights a significant association between unsuppressed HIV, indicated by detectable VL, and increased cardiovascular risk in youths living with HIV. These findings emphasize the importance of implementing interventions that address both VL suppression and cardiovascular risk reduction in this population. By tailoring interventions to meet the unique needs of youths, we can promote overall well-being throughout the HIV care continuum and across the life span. Ultimately, these efforts have the potential to improve the health outcomes and quality of life of youths living with HIV. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.2196/11185.

Annual Prevalence of Opioid Receipt by South Carolina Medicaid-Enrolled Children and Adolescents: 2000-2020.

Understanding patterns of opioid receipt by children and adolescents over time and understanding differences between age groups can help identify opportunities for future opioid stewardship. We conducted a retrospective cohort study, using South Carolina Medicaid data for children and adolescents 0-18 years old between 2000-2020, calculating the annual prevalence of opioid receipt for medical diagnoses in ambulatory settings. We examined differences in prevalence by calendar year, race/ethnicity, and by age group. The annual prevalence of opioid receipt for medical diagnoses changed significantly over the years studied, from 187.5 per 1000 in 2000 to 41.9 per 1000 in 2020 (Cochran-Armitage test for trend, p < 0.0001). In all calendar years, older ages were associated with greater prevalence of opioid receipt. Adjusted analyses (logistic regression) assessed calendar year differences in opioid receipt, controlling for age group, sex, and race/ethnicity. In the adjusted analyses, calendar year was inversely associated with opioid receipt (aOR 0.927, 95% CI 0.926-0.927). Males and older ages were more likely to receive opioids, while persons of Black race and Hispanic ethnicity had lower odds of receiving opioids. While opioid receipt declined among all age groups during 2000-2020, adolescents 12-18 had persistently higher annual prevalence of opioid receipt when compared to younger age groups.

Exploring the Therapeutic Potential of Elastase Inhibition in Age-Related Macular Degeneration in Mouse and Human.

Abnormal turnover of the extracellular matrix (ECM) protein elastin has been linked to AMD pathology. Elastin is a critical component of Bruch's membrane (BrM), an ECM layer that separates the retinal pigment epithelium (RPE) from the underlying choriocapillaris. Reduced integrity of BrM's elastin layer corresponds to areas of choroidal neovascularization (CNV) in wet AMD. Serum levels of elastin-derived peptides and anti-elastin antibodies are significantly elevated in AMD patients along with the prevalence of polymorphisms of genes regulating elastin turnover. Despite these results indicating significant associations between abnormal elastin turnover and AMD, very little is known about its exact role in AMD pathogenesis. Here we report on results that suggest that elastase enzymes could play a direct role in the pathogenesis of AMD. We found significantly increased elastase activity in the retinas and RPE cells of AMD mouse models, and AMD patient-iPSC-derived RPE cells. A1AT, a protease inhibitor that inactivates elastase, reduced CNV lesion sizes in mouse models. A1AT completely inhibited elastase-induced VEGFA expression and secretion, and restored RPE monolayer integrity in ARPE-19 monolayers. A1AT also mitigated RPE thickening, an early AMD phenotype, in HTRA1 overexpressing mice, HTRA1 being a serine protease with elastase activity. Finally, in an exploratory study, examining archival records from large patient data sets, we identified an association between A1AT use, age and AMD risk. Our results suggest that repurposing A1AT may have therapeutic potential in modifying the progression to AMD.

Determining the Marginal Cost Differences of a Telehealth Versus an In-person Occupational Therapy Evaluation Session for Stroke Survivors Using Time-driven Activity-based Costing.

OBJECTIVE: To estimate the marginal cost differences and care delivery process of a telerehabilitation vs outpatient session. DESIGN: This study used a time-driven activity-based costing approach including (1) observation of rehabilitation sessions and creation of manual time stamps, (2) structured and recorded interviews with 2 occupational therapists familiar with outpatient therapy and 2 therapists familiar with telerehabilitation, (3) collection of standard wages for providers, and (4) the creation of an iterative flowchart of both an outpatient and telerehabilitation session care delivery process. SETTING: Telerehabilitation and outpatient therapy evaluation. PARTICIPANTS: Three therapists familiar with care deliver for telerehabilitation or outpatient therapy (N=3). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Marginal cost difference between telerehabilitation and outpatient therapy evaluations. RESULTS: Overall, telerehabilitation ($225.41) was more costly than outpatient therapy ($168.29) per session for a cost difference of $57.12. Primary time drivers of this finding were initial phone calls (0 minutes for OP therapists vs 35 minutes for TR) and post documentation (5 minutes for OP vs 30 minutes for TR) demands for telerehabilitation. CONCLUSIONS: Telerehabilitation is an emerging platform with the potential to reduce costs, improve health care inequities, and facilitate better patient outcomes. Improvements in documentation practices, staffing, technology, and reimbursement structuring would allow for a more successful translation.

Implementation and Evaluation of a Wraparound Virtual Care Program for Children with Medical Complexity.

Objectives: Children and adolescents with medical complexity benefit from care coordination and specialized pediatric care, but many access barriers exist. We implemented a virtual wraparound model to support patients with medical complexity and their families and used an economic framework to measure outcomes. Methods: Children with medical complexity were identified and enrolled in a virtual complex care program with a dedicated multidisciplinary team, which provided care coordination, education, parental support, acute care triage, and virtual visits. A retrospective pre- and postanalysis of data obtained from the Hospital Industry Data Institute (HIDI) database measured inpatient, outpatient, and emergency department (ED) utilization and charges before implementation and during the 2-year program. Results: Eighty (n = 80) children were included in the economic evaluation, and 75 had sufficient data for analysis. Compared to the 12 months before enrollment, patients had a 35.3% reduction in hospitalizations (p = 0.0268), a 43.9% reduction in emergency visits (p = 0.0005), and a 16.9% reduction in overall charges (p = 0.1449). Parents expressed a high degree of satisfaction, with a 70% response rate and 90% satisfaction rate. Conclusions: We implemented a virtual care model to provide in-home support and care coordination for medically complex children and adolescents and used an economic framework to assess changes in utilization and cost. The program had high engagement rates and parent satisfaction, and a pre/postanalysis demonstrated statistically significant reduction in hospitalizations and ED visits for this high-cost population. Further economic evaluation is needed to determine sustainability of this model in a value-based payment system.

Contribution of Continuous Virtual Monitoring to Hospital Safety, Quality, and Value of Care for COVID-19 Patients.

Introduction: The rapid onset of the COVID-19 pandemic increased hospital admissions and shortages for personal protective equipment (PPE) used to slow the spread of infections. In addition, nurses treating COVID-19 patients have time-consuming guidelines to properly don and doff PPE to prevent the spread. Methods: To address these issues, the Medical University of South Carolina repurposed continuous virtual monitoring (CVM) systems to reduce the need for staff to enter patient rooms. The objective of this study was to identify the economic implications associated with using the CVM program for COVID-19 patients. We employed a time-driven activity-based costing approach to determine time and costs saved by implementing CVM. Results: Over the first 52 days of the pandemic, the use of the CVM system helped providers attend to patients needs virtually while averting 19,086 unnecessary in-person interactions. The estimated cost savings for the CVM program for COVID-19 patients in 2020 were $419,319, not including potential savings from avoided COVID-19 transmissions to health care workers. A total of 19,086 PPE changes were avoided, with savings of $186,661. After accounting for cost of the CVM system, the net savings provided an outstanding return on investment of 20.6 for the CVM program for COVID-19 patient care. Conclusion: The successful and cost saving repurposing of CVM systems could be expanded to other infectious disease applications, and be applied to high-risk groups, such as bone marrow and organ transplant patients.

An examination of predictive markers for successful sperm extraction procedures: a linear model and systematic review.

The authors performed a comprehensive review of current literature to create a model comparing commonly evaluated variables in male factor infertility, for example, follicle-stimulating hormone (FSH), testicular volume (TV), and testosterone (T), to better predict sperm retrieval rate (SRR). Twenty-nine studies were included, 9 with data on conventional testicular sperm extraction (cTESE) for a total of 1227 patients and 20 studies including data on microdissection testicular sperm extraction (mTESE) for a total of 4760 patients. A weighted-means value of SRR, FSH, T, and TV was created, and a weighted linear regression was then used to describe associations among SRR, type of procedure, FSH, T, and TV. In this study, weighted-means values demonstrated mTESE to be superior to cTESE with an SRR of 51.9% vs 40.1%. Multiple weighted linear regressions were created to describe associations among SRR, procedure type, FSH, T, and TV. The models showed that for every 1.19 mIU ml-1 increase in FSH, there would be a significant decrease in SRR by 1.0%. Seeking to create a more clinically relevant model, FSH values were then divided into normal, moderate elevation, and significant elevation categories (FSH <10 mIU ml-1, 10-19 mIU ml-1, and >20 mIU ml-1, respectively). For an index patient undergoing cTESE, the retrieval rates would be 57.1%, 44.3%, and 31.2% for values normal, moderately elevated, and significantly elevated, respectively. In conclusion, in a large meta-analysis, mTESE was shown to be more successful than cTESE for sperm retrievals. FSH has an inverse relationship to SRR in retrieval techniques and can alone be predictive of cTESE SRR.

Not all phenotypes are created equal: covariates of success in e-phenotype specification.

BACKGROUND: Electronic (e)-phenotype specification by noninformaticist investigators remains a challenge. Although validation of each patient returned by e-phenotype could ensure accuracy of cohort representation, this approach is not practical. Understanding the factors leading to successful e-phenotype specification may reveal generalizable strategies leading to better results. MATERIALS AND METHODS: Noninformaticist experts (n = 21) were recruited to produce expert-mediated e-phenotypes using i2b2 assisted by a honest data-broker and a project coordinator. Patient- and visit-sets were reidentified and a random sample of 20 charts matching each e-phenotype was returned to experts for chart-validation. Attributes of the queries and expert characteristics were captured and related to chart-validation rates using generalized linear regression models. RESULTS: E-phenotype validation rates varied according to experts' domains and query characteristics (mean = 61%, range 20-100%). Clinical domains that performed better included infectious, rheumatic, neonatal, and cancers, whereas other domains performed worse (psychiatric, GI, skin, and pulmonary). Match-rate was negatively impacted when specification of temporal constraints was required. In general, the increase in e-phenotype specificity contributed positively to match-rate. DISCUSSIONS AND CONCLUSIONS: Clinical experts and informaticists experience a variety of challenges when building e-phenotypes, including the inability to differentiate clinical events from patient characteristics or appropriately configure temporal constraints; a lack of access to available and quality data; and difficulty in specifying routes of medication administration. Biomedical query mediation by informaticists and honest data-brokers in designing e-phenotypes cannot be overstated. Although tools such as i2b2 may be widely available to noninformaticists, successful utilization depends not on users' confidence, but rather on creating highly specific e-phenotypes.

Assessment of the Cost Burden of Episodic Recurrent Vestibular Vertigo in the US.

Importance: Understanding of the economic burden of recurrent vestibular causes of vertigo and areas contributing to the cost is needed. Objective: To analyze and identify the factors contributing to the direct medical costs associated with Ménière disease (MD), vestibular migraine (VM), and benign paroxysmal positional vertigo (BPPV). Design, Setting, and Participants: This economic evaluation used MarketScan Commercial Database claims data from 2018 to identify the non-Medicare patient population with the diagnoses of MD, VM, or BPPV. Data were analyzed January 1 to December 31, 2018. Main Outcomes and Measures: The total direct medical costs associated with MD, VM, and BPPV. Results: A total of 53 210 patients (mean [SD] age, 47.8 [11.8] years; 67.6% female) were included in this study, with 34 738 normal comparisons. There were 5783 (10.9%) patients with MD, 3526 (6.6%) patients with VM, and 43 901 (82.5%) patients with BPPV in the data set. Mean age and sex were different across the different vestibular disorders. Across the different groups, patients with no comorbidities or with a Charlson Comorbidity Index score of zero ranged from 98.4% to 98.8%. Around 5% of patients were hospitalized with inpatient stay lasting between 4.6 and 5.2 days. After adjusting for age, sex, and comorbidities, there were large differences in mean adjusted annual payments/direct costs across the different groups (MD, $9579; VM, $11 371; and BPPV, $8247). This equated to a total incremental estimated cost of $60 billion compared with the normal population. The number of outpatient brain magnetic resonance imaging or computed tomography scans per patient ranged anywhere from 1 to 20, while the number of inpatient brain magnetic resonance imaging or computed tomography scans per patient ranged anywhere from 1 to 6. A heat map of the total cost expenditure indicated that the costs were concentrated around the Midwest, Lake Michigan, and the East Coast. Conclusions and Relevance: In this economic evaluation, the 3 most common causes of recurrent vertigo-MD, VM, and BPPV-had considerable medical costs associated with them. Extraneous imaging orders and vestibular testing are factors to consider for cost reduction. However, further research and widespread education is needed to optimize the diagnosis, treatment, and care of patients presenting with vestibular disorders or dizziness.

Dabigatran and Wet AMD, Results From Retinal Pigment Epithelial Cell Monolayers, the Mouse Model of Choroidal Neovascularization, and Patients From the Medicare Data Base.

Background: Age-related macular degeneration (AMD), the leading cause of irreversible blindness in elderly Caucasian populations, includes destruction of the blood-retina barrier (BRB) generated by the retinal pigment epithelium-Bruch's membrane complex (RPE/BrM), and complement activation. Thrombin is likely to get access to those structures upon BRB integrity loss. Here we investigate the potential role of thrombin in AMD by analyzing effects of the thrombin inhibitor dabigatran. Material and Methods: MarketScan data for patients aged ≥65 years on Medicare was used to identify association between AMD and dabigatran use. ARPE-19 cells grown as mature monolayers were analyzed for thrombin effects on barrier function (transepithelial resistance; TER) and downstream signaling (complement activation, expression of connective tissue growth factor (CTGF), and secretion of vascular endothelial growth factor (VEGF)). Laser-induced choroidal neovascularization (CNV) in mouse is used to test the identified downstream signaling. Results: Risk of new wet AMD diagnosis was reduced in dabigatran users. In RPE monolayers, thrombin reduced TER, generated unique complement C3 and C5 cleavage products, led to C3d/MAC deposition on cell surfaces, and increased CTGF expression via PAR1-receptor activation and VEGF secretion. CNV lesion repair was accelerated by dabigatran, and molecular readouts suggest that downstream effects of thrombin include CTGF and VEGF, but not the complement system. Conclusions: This study provides evidence of association between dabigatran use and reduced exudative AMD diagnosis. Based on the cell- and animal-based studies, we suggest that thrombin modulates wound healing and CTGF and VEGF expression, making dabigatran a potential novel treatment option in AMD.

Incidence of Poststroke Depression in Patients With Poststroke Dysphagia.

PURPOSE: Poststroke dysphagia and poststroke depression (PSD) can have devastating effects on stroke survivors, including increased burden of care, higher health care costs, poor quality of life, and greater mortality; however, there is a dearth of research examining depression in patients diagnosed with dysphagia after stroke. Thus, we aimed to study the incidence of PSD in patients with poststroke dysphagia to provide foundational knowledge about this patient population. METHOD: We conducted a retrospective, cross-sectional study of individuals with a primary diagnosis of acute ischemic stroke (AIS) and secondary diagnoses of dysphagia and/or depression using administrative claims data from the 2017 Medicare 5% Limited Data Set. RESULTS: The proportion of depression diagnosis in patients with poststroke dysphagia was significantly higher than the proportion of depression diagnosis in those without poststroke dysphagia during acute hospitalization: 12.01% versus 9.52%, respectively (p = .003). CONCLUSIONS: Our results demonstrated that persons with poststroke dysphagia were as, or slightly more, likely to have PSD compared to the general stroke population, and to our knowledge, they establish the first reported incidence of PSD in Medicare patients with dysphagia after AIS. Future research is warranted to further explore the effects of PSD on poststroke dysphagia.

Automatically identifying opioid use disorder in non-cancer patients on chronic opioid therapy.

Background: Using the International Classification of Diseases (ICD) codes alone to record opioid use disorder (OUD) may not completely document OUD in the electronic health record (EHR). We developed and evaluated natural language processing (NLP) approaches to identify OUD from the clinal note. We explored the concordance between ICD-coded and NLP-identified OUD.Methods: We studied EHRs from 13,654 (female: 8223; male: 5431) adult non-cancer patients who received chronic opioid therapy (COT) and had at least one clinical note between 2013 and 2018. Of eligible patients, we randomly selected 10,218 (75%) patients as the training set and the remaining 3436 patients (25%) as the test dataset for NLP approaches.Results: We generated 539 terms representing OUD mentions in clinical notes (e.g., "opioid use disorder," "opioid abuse," "opioid dependence," "opioid overdose") and 73 terms representing OUD medication treatments. By domain expert manual review for the test dataset, our NLP approach yielded high performance: 98.5% for precision, 100% for recall, and 99.2% for F-measure. The concordance of these NLP and ICD identified OUD was modest (Kappa = 0.63).Conclusions: Our NLP approach can accurately identify OUD patients from clinical notes. The combined use of ICD diagnostic code and NLP approach can improve OUD identification.