Identifying Functional Status Impairment in People Living With Dementia Through Natural Language Processing of Clinical Documents: Cross-Sectional Study.

BACKGROUND: Assessment of activities of daily living (ADLs) and instrumental ADLs (iADLs) is key to determining the severity of dementia and care needs among older adults. However, such information is often only documented in free-text clinical notes within the electronic health record and can be challenging to find. OBJECTIVE: This study aims to develop and validate machine learning models to determine the status of ADL and iADL impairments based on clinical notes. METHODS: This cross-sectional study leveraged electronic health record clinical notes from Mass General Brigham's Research Patient Data Repository linked with Medicare fee-for-service claims data from 2007 to 2017 to identify individuals aged 65 years or older with at least 1 diagnosis of dementia. Notes for encounters both 180 days before and after the first date of dementia diagnosis were randomly sampled. Models were trained and validated using note sentences filtered by expert-curated keywords (filtered cohort) and further evaluated using unfiltered sentences (unfiltered cohort). The model's performance was compared using area under the receiver operating characteristic curve and area under the precision-recall curve (AUPRC). RESULTS: The study included 10,000 key-term-filtered sentences representing 441 people (n=283, 64.2% women; mean age 82.7, SD 7.9 years) and 1000 unfiltered sentences representing 80 people (n=56, 70% women; mean age 82.8, SD 7.5 years). Area under the receiver operating characteristic curve was high for the best-performing ADL and iADL models on both cohorts (>0.97). For ADL impairment identification, the random forest model achieved the best AUPRC (0.89, 95% CI 0.86-0.91) on the filtered cohort; the support vector machine model achieved the highest AUPRC (0.82, 95% CI 0.75-0.89) for the unfiltered cohort. For iADL impairment, the Bio+Clinical bidirectional encoder representations from transformers (BERT) model had the highest AUPRC (filtered: 0.76, 95% CI 0.68-0.82; unfiltered: 0.58, 95% CI 0.001-1.0). Compared with a keyword-search approach on the unfiltered cohort, machine learning reduced false-positive rates from 4.5% to 0.2% for ADL and 1.8% to 0.1% for iADL. CONCLUSIONS: In this study, we demonstrated the ability of machine learning models to accurately identify ADL and iADL impairment based on free-text clinical notes, which could be useful in determining the severity of dementia.

How Do Risk Evaluation and Mitigation Strategies Impact Clinical Practice? A National Survey of Physicians.

The US Food and Drug Administration can require risk evaluation and mitigation strategy (REMS) programs for prescription drugs to ensure the benefits of use outweigh the risks. We conducted a national survey of physicians' experiences prescribing eight REMS-covered drugs: (1) ambrisentan; (2) bosentan; (3) clozapine; (4) isotretinoin; (5-7) the multiple myeloma (MM) drugs lenalidomide, pomalidomide, thalidomide; and (8) sodium oxybate. Between May 2022 and January 2023, we surveyed 5,331 physician prescribers of these drugs, and 1,295 (24%) returned surveys (range: 149 for bosentan to 226 for MM drugs). Although 765 (68%) respondents thought the certification process provided useful drug information, 757 (67%) wanted materials to include benefit data and 944 (84%) non-REMS-related risk data. A majority (704, 63%) thought the safe use requirements facilitated discussion with patients, but a similar number (637, 57%) attributed delayed medication access to these requirements. In multivariable modeling, MM drug and isotretinoin respondents were less likely than sodium oxybate respondents to agree that the certification process provided useful drug information (MM drug: odds ratio (OR) = 0.37, 95% confidence interval (CI) = 0.25-0.55; isotretinoin: OR = 0.39, 95% CI = 0.25-0.61), and isotretinoin, clozapine, and bosetan respondents were more likely than sodium oxybate respondents to agree that the safe use requirements often delayed medication access (isotretinoin: OR = 5.83, 95% CI = 3.70-9.19; clozapine: OR = 1.65, 95% CI = 1.08-2.54; bosentan: OR = 1.78, 95% CI = 1.12-2.85). Most physicians believe REMS programs convey useful drug safety information and facilitate discussion with patients but also seek information on benefits and non-REMS-related risks and better integration of REMS processes into clinical workflows.

Sodium-Glucose Cotransporter 2 Inhibitors and Nephrolithiasis Risk in Patients With Type 2 Diabetes.

Importance: Type 2 diabetes (T2D) is associated with an increased risk of kidney stones. Sodium-glucose cotransporter 2 inhibitors (SGLT2is) might lower the risk of nephrolithiasis by altering urine composition. However, no studies have investigated the association between SGLT2i use and nephrolithiasis risk in patients receiving routine care in the US. Objective: To investigate the association between SGLT2i use and nephrolithiasis risk in clinical practice. Design, Setting, and Participants: This new-user, active comparator cohort study used data from commercially insured adults (aged ≥18 years) with T2D who initiated treatment with SGLT2is, glucagon-like peptide 1 receptor agonists (GLP-1RAs), or dipeptidyl peptidase 4 inhibitors (DPP4is) between April 1, 2013, and December 31, 2020. The data were analyzed from July 2021 through June 2023. Exposure: New initiation of an SGLT2i, GLP-1RA, or DPP4i. Main Outcomes and Measures: The primary outcome was nephrolithiasis diagnosed by International Classification of Diseases codes in the inpatient or outpatient setting. New SGLT2i users were 1:1 propensity score matched to new users of a GLP-1RA or DPP4i in pairwise comparisons. Incidence rates, rate differences (RDs), and estimated hazard ratios (HRs) with 95% CIs were calculated. Results: After 1:1 propensity score matching, a total of 716 406 adults with T2D (358 203 pairs) initiating an SGLT2i or a GLP-1RA (mean [SD] age, 61.4 [9.7] years for both groups; 51.4% vs 51.2% female; 48.6% vs 48.5% male) and 662 056 adults (331 028 pairs) initiating an SGLT2i or a DPP4i (mean [SD] age, 61.8 [9.3] vs 61.7 [10.1] years; 47.4% vs 47.3% female; 52.6% vs 52.7% male) were included. Over a median follow-up of 192 (IQR, 88-409) days, the risk of nephrolithiasis was lower in patients initiating an SGLT2i than among those initiating a GLP-1RA (14.9 vs 21.3 events per 1000 person-years; HR, 0.69 [95% CI, 0.67-0.72]; RD, -6.4 [95% CI, -7.1 to -5.7]) or a DPP4i (14.6 vs 19.9 events per 1000 person-years; HR, 0.74 [95% CI, 0.71-0.77]; RD, -5.3 [95% CI, -6.0 to -4.6]). The association between SGLT2i use and nephrolithiasis risk was similar by sex, race and ethnicity, history of chronic kidney disease, and obesity. The magnitude of the risk reduction with SGLT2i use was larger among adults aged younger than 70 years vs aged 70 years or older (HR, 0.85 [95% CI, 0.79-0.91]; RD, -3.46 [95% CI, -4.87 to -2.05] per 1000 person-years; P for interaction <.001). Conclusions and Relevance: These findings suggest that in adults with T2D, SGLT2i use may lower the risk of nephrolithiasis compared with GLP-1RAs or DPP4is and could help to inform decision-making when prescribing glucose-lowering agents for patients who may be at risk for developing nephrolithiasis.

Common non-cardiovascular multimorbidity groupings and clinical outcomes in older adults with major cardiovascular disease.

BACKGROUND: Among older adults, non-cardiovascular multimorbidity often coexists with cardiovascular disease (CVD) but their clinical significance is uncertain. We identified common non-cardiovascular comorbidity patterns and their association with clinical outcomes in Medicare fee-for-service beneficiaries with acute myocardial infarction (AMI), congestive heart failure (CHF), or atrial fibrillation (AF). METHODS: Using 2015-2016 Medicare data, we took 1% random sample to create 3 cohorts of beneficiaries diagnosed with AMI (n = 24,808), CHF (n = 57,285), and AF (n = 36,277) prior to 1/1/2016. Within each cohort, we applied latent class analysis to classify beneficiaries based on 9 non-cardiovascular comorbidities (anemia, cancer, chronic kidney disease, chronic lung disease, dementia, depression, diabetes, hypothyroidism, and musculoskeletal disease). Mortality, cardiovascular and non-cardiovascular hospitalizations, and home time lost over a 1-year follow-up period were compared across non-cardiovascular multimorbidity classes. RESULTS: Similar non-cardiovascular multimorbidity classes emerged from the 3 CVD cohorts: (1) minimal, (2) depression-lung, (3) chronic kidney disease (CKD)-diabetes, and (4) multi-system class. Across CVD cohorts, multi-system class had the highest risk of mortality (hazard ratio [HR], 2.7-3.9), cardiovascular hospitalization (HR, 1.6-3.3), non-cardiovascular hospitalization (HR, 3.1-7.2), and home time lost (rate ratio, 2.7-5.4). Among those with AMI, the CKD-diabetes class was more strongly associated with all the adverse outcomes than the depression-lung class. In CHF and AF, differences in risk between the depression-lung and CKD-diabetes classes varied per outcome; and the depression-lung and multi-system classes had double the rates of non-cardiovascular hospitalizations than cardiovascular hospitalizations. CONCLUSION: Four non-cardiovascular multimorbidity patterns were found among Medicare beneficiaries with CHF, AMI, or AF. Compared to the minimal class, the multi-system, CKD-diabetes, and depression-lung classes were associated with worse outcomes. Identification of these classes offers insight into specific segments of the population that may benefit from more than the usual cardiovascular care.

Emulation of Randomized Clinical Trials With Nonrandomized Database Analyses: Results of 32 Clinical Trials.

Importance: Nonrandomized studies using insurance claims databases can be analyzed to produce real-world evidence on the effectiveness of medical products. Given the lack of baseline randomization and measurement issues, concerns exist about whether such studies produce unbiased treatment effect estimates. Objective: To emulate the design of 30 completed and 2 ongoing randomized clinical trials (RCTs) of medications with database studies using observational analogues of the RCT design parameters (population, intervention, comparator, outcome, time [PICOT]) and to quantify agreement in RCT-database study pairs. Design, Setting, and Participants: New-user cohort studies with propensity score matching using 3 US claims databases (Optum Clinformatics, MarketScan, and Medicare). Inclusion-exclusion criteria for each database study were prespecified to emulate the corresponding RCT. RCTs were explicitly selected based on feasibility, including power, key confounders, and end points more likely to be emulated with real-world data. All 32 protocols were registered on ClinicalTrials.gov before conducting analyses. Emulations were conducted from 2017 through 2022. Exposures: Therapies for multiple clinical conditions were included. Main Outcomes and Measures: Database study emulations focused on the primary outcome of the corresponding RCT. Findings of database studies were compared with RCTs using predefined metrics, including Pearson correlation coefficients and binary metrics based on statistical significance agreement, estimate agreement, and standardized difference. Results: In these highly selected RCTs, the overall observed agreement between the RCT and the database emulation results was a Pearson correlation of 0.82 (95% CI, 0.64-0.91), with 75% meeting statistical significance, 66% estimate agreement, and 75% standardized difference agreement. In a post hoc analysis limited to 16 RCTs with closer emulation of trial design and measurements, concordance was higher (Pearson r, 0.93; 95% CI, 0.79-0.97; 94% meeting statistical significance, 88% estimate agreement, 88% standardized difference agreement). Weaker concordance occurred among 16 RCTs for which close emulation of certain design elements that define the research question (PICOT) with data from insurance claims was not possible (Pearson r, 0.53; 95% CI, 0.00-0.83; 56% meeting statistical significance, 50% estimate agreement, 69% standardized difference agreement). Conclusions and Relevance: Real-world evidence studies can reach similar conclusions as RCTs when design and measurements can be closely emulated, but this may be difficult to achieve. Concordance in results varied depending on the agreement metric. Emulation differences, chance, and residual confounding can contribute to divergence in results and are difficult to disentangle.

Trends in the Use of Oral Anticoagulants for Adults With Venous Thromboembolism in the US, 2010-2020.

Importance: The introduction of direct oral anticoagulants (DOACs) has transformed the treatment of venous thromboembolism (VTE). Large health care databases offer valuable insight into how oral anticoagulants (OACs) are used in clinical practice and may aid in understanding reasons for changes in therapy. Objectives: To evaluate prescribing patterns of OACs for patients with VTE and identify clinical events that precede treatment changes. Design, Setting, and Participants: This retrospective cohort study used data from a public (Medicare fee-for-service) and a commercial (IBM MarketScan) health insurance database on 298 609 patients initiating OACs within 90 days of index VTE hospitalization from January 1, 2009, to December 31, 2020. Statistical analysis was conducted from April to August 2022. Exposures: Warfarin and the DOACs rivaroxaban, apixaban, dabigatran, and edoxaban. Main Outcomes and Measures: Characteristics of patients initiating different OACs, along with trends over time of patients initiating OACs, were compared. Time receiving continuous anticoagulant therapy, patterns of anticoagulant discontinuation (treatment gap of ≥30 days), and treatment switches were assessed. Clinical events in the 30 days preceding treatment modifications were identified. Results: A total of 203 378 individuals with Medicare (mean [SD] age, 76.9 [7.6] years; 122 554 women [60.3%]) and 95 231 with commercial insurance (mean [SD] age, 57.6 [15.8] years; 47 139 women [49.5%]) were included (N = 298 609). Warfarin was the most frequent OAC prescribed (163 044 [54.6%]), followed by rivaroxaban (66 882 [22.3%]) and apixaban (65 997 [22.1%]). The proportion of patients initiating DOACs increased from 0% in 2010 to 86.8% (22 420 of 25 817) in 2019 for patients with Medicare and 92.1% (4012 of 4357) in 2020 for commercially insured patients. Patients with chronic kidney disease were more likely to initiate warfarin (35 561 [11.9%]) or apixaban (16 294 [5.5%]) than rivaroxaban (10 136 [3.4%]), and those with a history of bleeding were more likely to initiate apixaban (5424 [1.8%]) than rivaroxaban (3007 [1.0%]). Overall, patients received persistent OAC treatment for approximately 6 months (Medicare: median, 175 days [IQR, 76-327 days]; commercial insurance: median, 168 days [IQR, 83-279 days]). A total of 33 011 patients (11.1%) switched anticoagulant therapy within a year. Switching to another anticoagulant was preceded most frequently by codes for a VTE diagnostic procedure (27.2% of all switchers [8983 of 33 011]). Conclusions and Relevance: This cohort study using data from 2 US health insurance databases suggests that most patients with VTE continued oral anticoagulant treatment for approximately 6 months. Clinical reasons for modifying anticoagulant therapy were identified in one-third of patients. Identifying reasons for treatment modification is crucial for generating valid evidence on drug safety and effectiveness.

Association of Antipsychotic Drug Exposure in Pregnancy With Risk of Neurodevelopmental Disorders: A National Birth Cohort Study.

Importance: Although antipsychotic drugs cross the placenta and animal data suggest potential neurotoxic effects, information regarding human neurodevelopmental teratogenicity is limited. Objective: To evaluate whether children prenatally exposed to antipsychotic medication are at an increased risk of neurodevelopmental disorders (NDD). Design, Settings, and Participants: This birth cohort study used data from the Medicaid Analytic eXtract (MAX, 2000-2014) and the IBM Health MarketScan Research Database (MarketScan, 2003-2015) for a nationwide sample of publicly (MAX) and privately (MarketScan) insured mother-child dyads with up to 14 years of follow-up. The MAX cohort consisted of 2 034 883 children who were not prenatally exposed and 9551 who were prenatally exposed to antipsychotic medications; the MarketScan consisted of 1 306 408 and 1221 children, respectively. Hazard ratios were estimated through Cox proportional hazards regression, using propensity score overlap weights for confounding control. Estimates from both cohorts were combined through meta-analysis. Exposures: At least 1 dispensing of a medication during the second half of pregnancy (period of synaptogenesis), assessed for any antipsychotic drug, at the class level (atypical and typical), and for the most commonly used drugs (aripiprazole, olanzapine, quetiapine, risperidone, and haloperidol). Main Outcomes and Measures: Autism spectrum disorder, attention-deficit/hyperactivity disorder, learning disability, speech or language disorder, developmental coordination disorder, intellectual disability, and behavioral disorder, identified using validated algorithms, and the composite outcome of any NDD. Data were analyzed from April 2020 to January 2022. Results: The MAX cohort consisted of 2 034 883 unexposed pregnancies and 9551 pregnancies with 1 or more antipsychotic drug dispensings among women with a mean (SD) age of 26.8 (6.1) years, 204 (2.1%) of whom identified as Asian/Pacific Islander, 2720 (28.5%) as Black, 500 (5.2%) as Hispanic/Latino, and 5356 (56.1%) as White. The MarketScan cohort consisted of 1 306 408 unexposed and 1221 exposed pregnancies among women with a mean (SD) age of 33.1 (5.0) years; race and ethnicity data were not available. Although the unadjusted results were consistent with an approximate 2-fold increased risk for most exposure-outcome contrasts, risks were no longer meaningfully increased after adjustment (eg, pooled unadjusted vs adjusted hazard ratios [95% CI] for any NDD after any antipsychotic exposure: 1.91 [1.79-2.03] vs 1.08 [1.01-1.17]), with the possible exception of aripiprazole (1.36 [1.14-1.63]). Results were consistent across sensitivity analyses. Conclusions and Relevance: The findings of this birth cohort study suggest that the increased risk of NDD seen in children born to women who took antipsychotic drugs late in pregnancy seems to be explained by maternal characteristics and is not causally related with prenatal antipsychotic exposure. This finding highlights the importance of closely monitoring the neurodevelopment of the offspring of women with mental illness to ensure early intervention and support. The potential signal for aripiprazole requires replication in other data before causality can be assumed.

Field-based body temperatures reveal behavioral thermoregulation strategies of the Atlantic marsh fiddler crab Minuca pugnax.

Behavioral thermoregulation is an important defense against the negative impacts of climate change for ectotherms. In this study we examined the use of burrows by a common intertidal crab, Minuca pugnax, to control body temperature. To understand how body temperatures respond to changes in the surface temperature and explore how efficiently crabs exploit the cooling potential of burrows to thermoregulate, we measured body, surface, and burrow temperatures during low tide on Sapelo Island, GA in March, May, August, and September of 2019. We found that an increase in 1°C in the surface temperature led to a 0.70-0.71°C increase in body temperature for females and an increase in 0.75-0.77°C in body temperature for males. Body temperatures of small females were 0.3°C warmer than large females for the same surface temperature. Female crabs used burrows more efficiently for thermoregulation compared to the males. Specifically, an increase of 1°C in the cooling capacity (the difference between the burrow temperature and the surface temperature) led to an increase of 0.42-0.50°C for females and 0.34-0.35°C for males in the thermoregulation capacity (the difference between body temperature and surface temperature). The body temperature that crabs began to use burrows to thermoregulate was estimated to be around 24°C, which is far below the critical body temperatures that could lead to death. Many crabs experience body temperatures of 24°C early in the reproductive season, several months before the hottest days of the year. Because the use of burrows involves fitness trade-offs, these results suggest that warming temperatures could begin to impact crabs far earlier in the year than expected.