A general framework for developing computable clinical phenotype algorithms.

OBJECTIVE: Present a general framework providing high-level guidance to developers of computable algorithms for identifying patients with specific clinical conditions (phenotypes) through a variety of approaches, including but not limited to machine learning and natural language processing methods to incorporate rich electronic health record data. MATERIALS/METHODS: Drawing on extensive prior phenotyping experiences and insights derived from three algorithm development projects conducted specifically for this purpose, our team with expertise in clinical medicine, statistics, informatics, pharmacoepidemiology, and healthcare data science methods conceptualized stages of development and corresponding sets of principles, strategies, and practical guidelines for improving the algorithm development process. RESULTS: We propose five stages of algorithm development and corresponding principles, strategies, and guidelines: 1) assessing fitness-for-purpose, 2) creating gold standard data, 3) feature engineering, 4) model development, and 5) model evaluation. DISCUSSION/CONCLUSION: This framework is intended to provide practical guidance and serve as a basis for future elaboration and extension.

Mortality with concurrent treatment with gabapentin and opioids among people with spine diagnoses in the U.S. Medicare population: a propensity-matched cohort study.

IMPORTANCE: Given the negative impact of opioid use on population health, prescriptions for alternative pain-relieving medications, including gabapentin, have increased. Concurrent gabapentin and opioid prescriptions are commonly reported in retrospective studies of opioid-related overdose deaths. OBJECTIVE: To determine whether people who filled gabapentin and opioid prescriptions concurrently ('gabapentin + opioids') had greater mortality than those who filled an active control medication (tricyclic antidepressants [TCAs] or duloxetine) and opioids concurrently ('TCAs/duloxetine + opioids'). We hypothesized that people treated with gabapentin + opioids would have higher mortality rates compared to people treated with TCAs/duloxetine + opioids. DESIGN: Propensity score-matched cohort study with an incident user, active control design. The median (maximum) follow-up was 45 (1093) days. SETTING: Population-based. PARTICIPANTS: Medicare beneficiaries with spine-related diagnoses 2017-2019. The primary analysis included those who concurrently (within 30 days) filled at least 1 incident gabapentin + at least 1 opioid or at least 1 incident TCA/duloxetine + at least 1 opioid. EXPOSURES: People treated with gabapentin + opioids (n=67,133) were matched on demographic and clinical factors in a 1:1 ratio to people treated with TCAs/duloxetine + opioids (n=67,133). MAIN OUTCOMES AND MEASURES: The primary outcome was mortality at any time. A secondary outcome was occurrence of a major medical complication at any time. RESULTS: Among 134,266 participants (median age 73.4 years; 66.7% female), 2360 died before the end of follow-up. No difference in mortality was observed between groups (adjusted hazard ratio (HR) and 95% confidence interval (CI) for gabapentin + opioids was 0.98 (0.90, 1.06); p=0.63). However, people treated with gabapentin + opioids were at slightly increased risk of a major medical complication (1.02 (1.00, 1.04); p=0.03) compared to those treated with TCAs/duloxetine + opioids. Results were similar in analyses (a) restricted to less than or = 30-day follow-up and (b) that required at least 2 fills of each prescription. CONCLUSIONS AND RELEVANCE: When treating pain in older adults taking opioids, the addition of gabapentin did not increase mortality risk relative to addition of TCAs or duloxetine. However, providers should be cognizant of a small increased risk of major medical complications among opioid users initiating gabapentin compared to those initiating TCAs or duloxetine.

Genetic and Congenital Anomalies in Infants With Hypoxic-Ischemic Encephalopathy.

BACKGROUND: Infants with hypoxic ischemic encephalopathy (HIE) may have underlying conditions predisposing them to hypoxic-ischemic injury during labor and delivery. It is unclear how genetic and congenital anomalies impact outcomes of HIE. METHODS: Infants with HIE enrolled in a phase III trial underwent genetic testing when clinically indicated. Infants with known genetic or congenital anomalies were excluded. The primary outcome, i.e., death or neurodevelopmental impairment (NDI), was determined at age two years by a standardized neurological examination, Bayley Scales of Infant Development, Third Edition (BSID-III), and the Gross Motor Function Classification Scales. Secondary outcomes included cerebral palsy and BSID-III motor, cognitive, and language scores at age two years. RESULTS: Of 500 infants with HIE, 24 (5%, 95% confidence interval 3% to 7%) were diagnosed with a genetic (n = 15) or congenital (n = 14) anomaly. Infants with and without genetic or congenital anomalies had similar rates of severe encephalopathy and findings on brain magnetic resonance imaging. However, infants with genetic or congenital anomalies were more likely to have death or NDI (75% vs 50%, P = 0.02). Among survivors, those with a genetic or congenital anomaly were more likely to be diagnosed with cerebral palsy (32% vs 13%, P = 0.02), and had lower BSID-III scores in all three domains than HIE survivors without such anomalies. CONCLUSIONS: Among infants with HIE, 5% were diagnosed with a genetic or congenital anomaly. Despite similar clinical markers of HIE severity, infants with HIE and a genetic or congenital anomaly had worse neurodevelopmental outcomes than infants with HIE alone.

Using Effect Scores to Characterize Heterogeneity of Treatment Effects.

This JAMA Guide to Statistics and Methods article explains effect score analyses, an approach for evaluating the heterogeneity of treatment effects, and examines its use in a study of oxygen-saturation targets in critically ill patients.

Esophageal Versus Rectal Temperature Monitoring During Whole-Body Therapeutic Hypothermia for Hypoxic-Ischemic Encephalopathy: Association with Short- and Long-Term Outcomes.

OBJECTIVE: To compare the short- and long-term outcomes of infants with hypoxic-ischemic encephalopathy (HIE) treated with whole-body therapeutic hypothermia (TH), monitored by esophageal vs rectal temperature. STUDY DESIGN: We conducted a secondary analysis of the multicenter High-Dose Erythropoietin for Asphyxia and Encephalopathy (HEAL) trial. All infants had moderate or severe HIE and were treated with whole-body TH. The primary outcome was death or neurodevelopmental impairment (NDI) at 22-36 months of age. Secondary outcomes included seizures, evidence of brain injury on magnetic resonance imaging, and complications of hypothermia. Logistic regression was used with adjustment for disease severity and site as clustering variable because cooling modality differed by site. RESULTS: Of the 500 infants who underwent TH, 294 (59%) and 206 (41%) had esophageal and rectal temperature monitoring, respectively. There were no differences in death or NDI, seizures, or evidence of injury on magnetic resonance imaging between the 2 groups. Infants treated with TH and rectal temperature monitoring had lower odds of overcooling (OR 0.52, 95% CI 0.34-0.80) and lower odds of hypotension (OR 0.57, 95% CI 0.39-0.84) compared with those with esophageal temperature monitoring. CONCLUSIONS: Although infants undergoing TH with esophageal monitoring were more likely to experience overcooling and hypotension, the rate of death or NDI was similar whether esophageal monitoring or rectal temperature monitoring was used. Further studies are needed to investigate whether esophageal temperature monitoring during TH is associated with an increased risk of overcooling and hypotension.

Contemporary definitions of infant growth failure and neurodevelopmental and behavioral outcomes in extremely premature infants at two years of age.

BACKGROUND: Associations of 2-year neurodevelopmental and behavioral outcomes with growth trajectories of preterm infants are unknown. METHODS: This secondary analysis of a preterm cohort examined in-hospital and discharge to 2-year changes in anthropometric z-scores. Two-year follow-up included Bayley Scales of Infant Development (BSID-III) and Child Behavior Checklist. RESULTS: Among 590 infants, adjusted in-hospital growth was not associated with any BSID-III subscale. Occipitofrontal circumference (OFC) growth failure (GF) in-hospital was associated with increased adjusted odds of attention problems (aOR 1.65 [1.03, 2.65]), aggressive behavior (aOR 2.34 [1.12, 4.89]), and attention-deficit-hyperactivity symptoms (aOR 1.86 [1.05, 3.30]). Infants with OFC GF at 2 years had lower adjusted BSID-III language scores (-4.0 [-8.0, -0.1]), increased odds of attention problems (aOR 2.29 [1.11, 4.74]), aggressive behavior (aOR 3.09 [1.00, 9.56]), and externalizing problems (aOR 3.01 [1.07, 8.45]) compared to normal OFC growth cohort. CONCLUSION: Infants with OFC GF are at risk for neurodevelopmental and behavioral impairment. CLINICAL TRIAL REGISTRATION: This study is a secondary analysis of pre-existing data from the PENUT Trial Registration: NCT01378273.

Description and initial validation of a novel measure of pain intensity: the Numeric Rating Scale of Underlying Pain without concurrent Analgesic use.

ABSTRACT: Although many individuals with chronic pain use analgesics, the methods used in many randomized controlled trials (RCTs) do not sufficiently account for confounding by differential post-randomization analgesic use. This may lead to underestimation of average treatment effects and diminished power. We introduce (1) a new measure-the Numeric Rating Scale of Underlying Pain without concurrent Analgesic use (NRS-UP(A))-which can shift the estimand of interest in an RCT to target effects of a treatment on pain intensity in the hypothetical situation where analgesic use was not occurring at the time of outcome assessment; and (2) a new pain construct-an individuals' perceived effect of analgesic use on pain intensity (EA). The NRS-UP(A) may be used as a secondary outcome in RCTs of point treatments or nonpharmacologic treatments. Among 662 adults with back pain in primary care, participants' mean value of the NRS-UP(A) among those using analgesics was 1.2 NRS points higher than their value on the conventional pain intensity NRS, reflecting a mean EA value of -1.2 NRS points and a perceived beneficial effect of analgesics. More negative values of EA (ie, greater perceived benefit) were associated with a greater number of analgesics used but not with pain intensity, analgesic type, or opioid dose. The NRS-UP(A) and EA were significantly associated with future analgesic use 6 months later, but the conventional pain NRS was not. Future research is needed to determine whether the NRS-UP(A), used as a secondary outcome may allow pain RCTs to target alternative estimands with clinical relevance.

Platelet Transfusion and Death or Neurodevelopmental Impairment in Children Born Extremely Preterm.

Importance: Infants born extremely preterm receive transfusions at higher platelet count thresholds than older children and adults due to concerns for intracranial hemorrhage. A recent randomized trial comparing 2 platelet transfusion thresholds showed the higher threshold was associated with increased risk of long-term adverse neurodevelopmental outcomes. Objective: To evaluate the association of platelet transfusion exposure with death and severe neurodevelopmental impairment (NDI) at 2 years' corrected age in a cohort of infants born extremely preterm. Design, Setting, and Participants: An observational cohort study and secondary analysis of the Preterm Erythropoietin Neuroprotection Trial, a randomized, placebo-controlled clinical trial of erythropoietin neuroprotection in neonates born extremely preterm, was conducted in 30 neonatal intensive care units in the US from December 1, 2013, to September 31, 2016. This analysis included 819 infants born extremely preterm at 24 to 27 completed weeks of gestation who had a documented outcome (death or neurodevelopmental assessment). Analysis was performed in April 2023. Exposures: Any platelet transfusion during neonatal intensive care unit hospitalization. Main Outcomes and Measures: The primary composite outcome was death or severe NDI evaluated at 2 years' corrected age using the Bayley Scales of Infant Development-Third Edition (BSID-III) and the Gross Motor Function Classification System and was defined as the presence of severe cerebral palsy or a BSID-III composite motor or cognitive score 2 SDs below the mean. Confounding by indication for platelet transfusion was addressed with covariate adjustment and propensity score methods. Results: Of the 819 infants included in the analysis (429 [52.4%] male; mean [SD] gestational age, 25.5 [1.1] weeks), 245 (30.0%) received at least 1 platelet transfusion during their initial hospitalization. The primary outcome occurred in 46.5% (114 of 245) of infants exposed to a platelet transfusion and 13.9% (80 of 574) of nonexposed infants with a corresponding odds ratio of 2.43 (95% CI, 1.24-4.76), adjusted for propensity score, gestational age at birth, and trial treatment group. The individual components of death and severe NDI were directionally consistent with the overall composite outcome. Conclusions and Relevance: The findings of this study suggest that platelet transfusion in infants born extremely preterm may be associated with an increased risk of death or severe NDI at 2 years' corrected age, although the possibility of residual confounding by indication cannot be excluded.

Effectiveness of Existing Insomnia Therapies for Patients Undergoing Hemodialysis : A Randomized Clinical Trial.

BACKGROUND: Chronic insomnia is common in patients undergoing in-center hemodialysis, yet there is limited evidence on effective treatments for this population. OBJECTIVE: To compare the effectiveness of cognitive behavioral therapy for insomnia (CBT-I), trazodone, and placebo for insomnia in patients undergoing long-term hemodialysis. DESIGN: Randomized, multicenter, double-blinded, placebo-controlled trial. (ClinicalTrials.gov: NCT03534284). SETTING: 26 dialysis units in Albuquerque, New Mexico, and Seattle, Washington. PARTICIPANTS: Patients with Insomnia Severity Index (ISI) score of 10 or greater, with sleep disturbances on 3 or more nights per week for 3 or more months. INTERVENTION: Participants were randomly assigned to 6 weeks of CBT-I, trazodone, or placebo. MEASUREMENTS: The primary outcome was the ISI score at 7 and 25 weeks from randomization. RESULTS: A total of 923 patients were prescreened, and of the 411 patients with chronic insomnia, 126 were randomly assigned to CBT-I (n = 43), trazodone (n = 42), or placebo (n = 41). The change in ISI scores from baseline to 7 weeks with CBT-I or trazodone was no different from placebo: CBT-I, -3.7 (95% CI, -5.5 to -1.9); trazodone, -4.2 (CI, -5.9 to -2.4); and placebo, -3.1 (CI, -4.9 to -1.3). There was no meaningful change in ISI scores from baseline to 25 weeks: CBT-I, -4.8 (CI, -7.0 to -2.7); trazodone, -4.0 (CI, -6.0 to -1.9); and placebo, -4.3 (CI, -6.4 to -2.2). Serious adverse events (SAEs), particularly serious cardiovascular events, were more frequent with trazodone (annualized cardiovascular SAE incidence rates: CBT-I, 0.05 [CI, 0.00 to 0.29]; trazodone, 0.64 [CI, 0.34 to 1.10]; and placebo, 0.21 [CI, 0.06 to 0.53]). LIMITATION: Modest sample size and most participants had mild or moderate insomnia. CONCLUSION: In patients undergoing hemodialysis with mild or moderate chronic insomnia, there was no difference in the effectiveness of 6 weeks of CBT-I or trazodone compared with placebo. The incidence of SAEs was higher with trazodone. PRIMARY FUNDING SOURCE: National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases.

Time to regain birthweight and association with neurodevelopmental outcomes among extremely preterm newborns.

OBJECTIVE: Determine association between time to regain birthweight and 2-year neurodevelopment among extremely preterm (EP) newborns. STUDY DESIGN: Secondary analysis of the Preterm Erythropoietin Neuroprotection Trial evaluating time to regain birthweight, time from birth to weight nadir, time from nadir to regain birthweight, and cumulative weight loss with 2-year corrected Bayley Scales of Infant and Toddler Development 3rd edition. RESULTS: Among n = 654 EP neonates, those with shorter nadir-to-regain had lower cognitive scores (≤1 day versus ≥8 days: -5.0 points, [CI -9.5, -0.6]) and lower motor scores (≤1 day versus ≥8 days: -4.6 points [CI -9.2, -0.03]) in adjusted stepwise forward regression modeling. Increasingly cumulative weight loss was associated with lower cognitive scores (≤-50 percent-days: -5.6, [CI -9.4, -1.8]), motor scores (≤-50 percent-days: -4.2, [CI -8.2, -0.2]); and language scores (≤-50 percent-days: -6.0, [CI -10.1, -1.9]). CONCLUSION: Faster nadir-to-regain and excessive cumulative weight loss are associated with adverse 2-year neurodevelopmental outcomes. TRIAL REGISTRATION: PENUT Trial Registration: NCT01378273. https://clinicaltrials.gov/ct2/show/NCT01378273 . CLINICAL TRIAL REGISTRATION: This study is a post-hoc secondary analysis of pre-existing data from the PENUT Trial (NCT #01378273).

Evolution of the Sarnat exam and association with 2-year outcomes in infants with moderate or severe hypoxic-ischaemic encephalopathy: a secondary analysis of the HEAL Trial.

OBJECTIVE: To study the association between the Sarnat exam (SE) performed before and after therapeutic hypothermia (TH) and outcomes at 2 years in infants with moderate or severe hypoxic-ischaemic encephalopathy (HIE). DESIGN: Secondary analysis of the High-dose Erythropoietin for Asphyxia and EncephaLopathy Trial. Adjusted ORs (aORs) for death or neurodevelopmental impairment (NDI) based on SE severity category and change in category were constructed, adjusting for sedation at time of exam. Absolute SE Score and its change were compared for association with risk for death or NDI using locally estimated scatterplot smoothing curves. SETTING: Randomised, double-blinded, placebo-controlled multicentre trial including 17 centres across the USA. PATIENTS: 479/500 enrolled neonates who had both a qualifying SE (qSE) before TH and a SE after rewarming (rSE). INTERVENTIONS: Standardised SE was used across sites before and after TH. All providers underwent standardised SE training. MAIN OUTCOME MEASURES: Primary outcome was defined as the composite outcome of death or any NDI at 22-36 months. RESULTS: Both qSE and rSE were associated with the primary outcome. Notably, an aOR for primary outcome of 6.2 (95% CI 3.1 to 12.6) and 50.3 (95% CI 13.3 to 190) was seen in those with moderate and severe encephalopathy on rSE, respectively. Persistent or worsened severity on rSE was associated with higher odds for primary outcome compared with those who improved, even when qSE was severe. CONCLUSION: Both rSE and change between qSE and rSE were strongly associated with the odds of death/NDI at 22-36 months in infants with moderate or severe HIE.

Time to Reaching Target Cooling Temperature and 2-year Outcomes in Infants with Hypoxic-Ischemic Encephalopathy.

OBJECTIVE: To determine if time to reaching target temperature (TT) is associated with death or neurodevelopmental impairment (NDI) at 2 years of age in infants with hypoxic-ischemic encephalopathy (HIE). STUDY DESIGN: Newborn infants ≥36 weeks of gestation diagnosed with moderate or severe HIE and treated with therapeutic hypothermia were stratified based on time at which TT was reached, defined as early (ie, ≤4 hours of age) or late (>4 hours of age). Primary outcomes were death or NDI. Secondary outcomes included neurodevelopmental assessment with Bayley Scales of Infant and Toddler Development, third edition (BSID-III) at age 2. RESULTS: Among 500 infants, the median time to reaching TT was 4.3 hours (IWR, 3.2-5.7 hours). Infants in early TT group (n = 211 [42%]) compared with the late TT group (n = 289 [58%]) were more likely to be inborn (23% vs 13%; P < .001) and have severe HIE (28% vs 19%; P = .03). The early and late TT groups did not differ in the primary outcome of death or any NDI (adjusted RR, 1.05; 95% CI, 0.85-0.30; P = .62). Among survivors, neurodevelopmental outcomes did not differ significantly in the 2 groups (adjusted mean difference in Bayley Scales of Infant Development-III scores: cognitive, -2.8 [95% CI, -6.1 to 0.5], language -3.3 [95% CI, -7.4 to 0.8], and motor -3.5 [95% CI, -7.3 to 0.3]). CONCLUSIONS: In infants with HIE, time to reach TT is not independently associated with risk of death or NDI at age 2 years. Among survivors, developmental outcomes are similar between those who reached TT at <4 and ≥4 hours of age. TRIAL REGISTRATION: High-dose Erythropoietin for Asphyxia and Encephalopathy (HEAL); NCT02811263; https://beta. CLINICALTRIALS: gov/study/NCT02811263.

Association Between Urinary Biomarkers and CKD in Extremely Low Gestational Age Neonates.

RATIONALE & OBJECTIVE: Children born before 28 weeks' gestation are at increased risk of chronic kidney disease (CKD). Urine biomarkers may shed light on mechanistic pathways and improve the ability to forecast CKD. We evaluated whether urinary biomarkers in neonates of low gestational age (GA) are associated with a reduced estimated glomerular filtration rate (eGFR) over time. STUDY DESIGN: A cohort study of neonates with an exploratory case-control study of a subset of the cohort. SETTING & PARTICIPANTS: 327 neonates born at 24-27 weeks' gestation with 2-year eGFR data from the PENUT (Preterm Erythropoietin Neuroprotection Trial) and the REPaIReD (Recombinant Erythropoietin for Prevention of Infant Renal Disease) study. EXPOSURES: 11 urinary biomarkers measured at 27, 30, and 34 weeks' postmenstrual age for the primary cohort study and 10 additional biomarkers for the exploratory case-control study. OUTCOMES: eGFR<90mL/min/1.73m2 at 2 years corrected for GA. ANALYTICAL APPROACH: Linear mixed models to assess differences in biomarker values between neonates in whom CKD did and did not develop, accounting for multiple comparisons using Bonferroni-Holm correction in the cohort study only. Cohort analyses were adjusted for sex, GA, and body mass index. Cases were matched to controls on these variables in the case-control study. RESULTS: After adjusting for weeks of GA, urinary levels of α-glutathione-S-transferase (log difference, 0.27; 95% CI, 0.12-0.43), albumin (log difference, 0.13; 95% CI, 0.02-0.25), and cystatin C (log difference, 0.19; 95% CI, 0.04-0.34) were higher in those in whom CKD developed than in those in whom it did not. Urinary albumin and cystatin C levels did not remain significantly different after Bonferroni-Holm correction. In the exploratory case-control analysis, there were no differences in any biomarkers between cases and controls. LIMITATIONS: Early deaths and a high number of subjects without eGFR at 2 years corrected for GA. CONCLUSIONS: Measurement of urinary biomarkers may assist in monitoring neonates who are at risk for CKD. Additional studies are needed to confirm these findings. FUNDING: Grants from government (National Institutes of Health). TRIAL REGISTRATION: Registered at ClinicalTrials.gov with study number NCT01378273. PLAIN-LANGUAGE SUMMARY: Approximately 15 million neonates worldwide are born prematurely, and 2 million are born before 28 weeks' gestation. Many of these children go on to experience chronic kidney disease. Urine biomarkers may allow for early recognition of those at risk for the development of kidney disease. In this study of more than 300 children born before 28 weeks' gestational age, we found higher mean urinary levels of α-glutathione-S-transferase at 27, 30, and 34 weeks in children whose estimated glomerular filtration rate was<90mL/min/1.73m2 at 2 years compared with children whose estimated glomerular filtration rate was>90mL/min/1.73m2 at 2 years. Measurement of urinary biomarkers may assist in monitoring neonates who are at risk for chronic kidney disease. Additional studies are needed to confirm our findings.

Planning stepped wedge cluster randomized trials to detect treatment effect heterogeneity.

Stepped wedge design is a popular research design that enables a rigorous evaluation of candidate interventions by using a staggered cluster randomization strategy. While analytical methods were developed for designing stepped wedge trials, the prior focus has been solely on testing for the average treatment effect. With a growing interest on formal evaluation of the heterogeneity of treatment effects across patient subpopulations, trial planning efforts need appropriate methods to accurately identify sample sizes or design configurations that can generate evidence for both the average treatment effect and variations in subgroup treatment effects. To fill in that important gap, this article derives novel variance formulas for confirmatory analyses of treatment effect heterogeneity, that are applicable to both cross-sectional and closed-cohort stepped wedge designs. We additionally point out that the same framework can be used for more efficient average treatment effect analyses via covariate adjustment, and allows the use of familiar power formulas for average treatment effect analyses to proceed. Our results further sheds light on optimal design allocations of clusters to maximize the weighted precision for assessing both the average and heterogeneous treatment effects. We apply the new methods to the Lumbar Imaging with Reporting of Epidemiology Trial, and carry out a simulation study to validate our new methods.

Frailty and Risk of Serious Infections in Patients With Rheumatoid Arthritis Treated With Biologic or Targeted-Synthetic Disease-Modifying Antirheumatic Drugs.

OBJECTIVE: It remains unknown whether frailty status portends an increased risk of adverse outcomes in patients with rheumatoid arthritis (RA) initiating biologic or targeted-synthetic (b/ts) disease-modifying antirheumatic drugs (DMARDs). The objective of our study was to evaluate the association between frailty and serious infections in a younger population of patients (<65 years old) with RA who initiated b/tsDMARDs. METHODS: Using MarketScan data, we identified new users of tumor necrosis factor inhibitors (TNFi), non-TNFi biologic DMARDs, or Janus kinase inhibitors (JAKi) between 2008 and 2019 among those with RA. Patients' baseline frailty risk score was calculated using a Claims-Based Frailty Index (≥0.2 defined as frail) 12 months prior to drug initiation. The primary outcome was time to serious infection; secondarily, we examined time-to-any infection and all-cause hospitalizations. We used Cox proportional hazards to estimate adjusted hazard ratios and 95% confidence intervals (95% CIs) and assessed the significance of interaction terms between frailty status and drug class. RESULTS: A total of 57,980 patients, mean (±SD) age 48.1 ± 10.1 were included; 48,139 (83%) started TNFi, 8,111 (14%) non-TNFi biologics, and 1,730 (3%) JAKi. Among these, 3,560 (6%) were categorized as frail. Frailty was associated with a 50% increased risk of serious infections (adjusted hazard ratio [95% CI] 1.5, 1.2-1.9) and 40% higher risk of inpatient admissions (1.4 [1.3-1.6]) compared with nonfrail patients among those who initiated TNFi. Frailty was also associated with a higher risk of any infection relative to nonfrail patients among those on TNFi (1.2 [1.1-1.3]) or non-TNFi (1.2 [1.0-1.4]) or JAKi (1.4 [1.0-2.0]). CONCLUSION: Frailty is an important predictor for the risk of adverse outcomes among patients with RA treated with biologic or targeted-synthetic DMARDs.

Using Machine Learning to Leverage Biomarker Change and Predict Colorectal Cancer Recurrence.

PURPOSE: The risk of colorectal cancer (CRC) recurrence after primary treatment varies across individuals and over time. Using patients' most up-to-date information, including carcinoembryonic antigen (CEA) biomarker profiles, to predict risk could improve personalized decision making. METHODS: We used electronic health record data from an integrated health system on a cohort of patients diagnosed with American Joint Committee on Cancer stage I-III CRC between 2008 and 2013 (N = 3,970) and monitored until recurrence or end of follow-up. We addressed missingness in recurrence outcomes and longitudinal CEA measures, and engineered CEA features using current and past biomarker values for inclusion in a risk prediction model. We used a discrete time Superlearner model to evaluate various algorithms for predicting recurrence. We evaluated the time-varying discrimination and calibration of the algorithms and assessed the role of individual predictors. RESULTS: Recurrence was documented in 448 (11.3%) patients. XGBoost with depth = 1 (XGB-D1) predicted recurrence substantially better than all other algorithms at all time points, with AUC ranging from 0.87 (95% CI, 0.86 to 0.88) at 6 months to 0.94 (95% CI, 0.92 to 0.96) at 54 months. The only variable used by XGB-D1 was 6-month change in log CEA. Predicted 1-year risk of recurrence was nearly zero for patients whose log CEA did not increase in the last 6 months, between 12.2% and 34.1% for patients whose log CEA increased between 0.10 and 0.40, and 43.6% for those with a log CEA increase >0.40. Compared with XGB, penalized regression approaches (lasso, ridge, and elastic net) performed poorly, with AUCs ranging from 0.58 to 0.69. CONCLUSION: A flexible, machine learning approach that incorporated longitudinal CEA information yielded a simple and high-performing model for predicting recurrence on the basis of 6-month change in log CEA.

Minimal important difference metrics and test-retest reliability of the PTSD Checklist for DSM-5 with a primary care sample.

The PTSD Checklist for DSM-5 (PCL-5) is a measure of posttraumatic stress disorder (PTSD) symptom severity that is widely used for clinical and research purposes. Although previous work has examined metrics of minimal important difference (MID) of the PCL-5 in veteran samples, no work has identified PCL-5 MID metrics among adults in primary care in the United States. In this secondary analysis, data were evaluated from primary care patients (N = 971) who screened positive for PTSD and participated in a large clinical trial in federally qualified health centers in three U.S. states. Participants primarily self-identified as women (70.2%) and White (70.3%). We calculated test-retest reliability using clinic registry data and multiple distribution- and anchor-based metrics of MID using baseline and follow-up survey data. Test-retest reliability (Pearson's r, Spearman's ρ, intraclass correlation coefficient) ranged from adequate to excellent (.79-.94), with the shortest time lag demonstrating the highest reliability estimate. The MID for the PCL-5 was estimated using multiple approaches. Distribution-based approaches indicated an MID range of 8.5-12.5, and anchor-based approaches indicated an MID range of 9.8-11.7. Taken together, the MID metrics indicate that PCL-5 change scores of 9-12 likely reflect real change in PTSD symptoms and indicate at least an MID for patients, whereas PCL-5 change scores of 5 or less likely are not reliable. These findings can help inform clinicians using the PCL-5 in similar populations to track patient responses to treatment and help researchers interpret PCL-5 score changes in clinical trials.