Symptoms as a Predictor of the Placebo-Controlled Efficacy of PCI in Stable Coronary Artery Disease.

BACKGROUND: Placebo-controlled evidence from ORBITA-2 found that percutaneous cutaneous intervention (PCI) in stable coronary artery disease with little or no antianginal medication relieved angina, but residual symptoms persisted in many. The reason for this was unclear. OBJECTIVES: ; This ORBITA-2 secondary analysis investigates the relationship between presenting symptoms and disease severity (anatomic, non-invasive, and invasive ischemia) and the ability of symptoms to predict the placebo-controlled efficacy of PCI. METHODS: Pre-randomization symptom severity and nature were assessed using the ORBITA smartphone application and symptom and quality of life questionnaires including the Rose angina questionnaire. Disease severity was assessed using quantitative coronary angiography (QCA), stress echocardiography, fractional flow reserve (FFR), and instantaneous wave-free ratio (iFR). Bayesian ordinal regression was used. RESULTS: At pre-randomization, the median number of daily angina episodes was 0.8 (0.4-1.6), 64% had Rose angina, QCA diameter stenosis 61 (49-74), stress echocardiography score 1.0 (0.0-2.7), FFR 0.63 (0.49-0.75), and iFR 0.78 (0.55-0.87). There was little relationship between symptom severity and nature and disease severity: angina symptom score with QCA ordinal correlation coefficient 0.06 (95% CrI 0.00 to 0.08); stress echocardiography 0.09 (95% CrI 0.02 to 0.10); FFR 0.04 (95% CrI -0.03 to 0.07); and iFR 0.04 (95% CrI -0.01 to 0.07). However, Rose angina and guideline-based typical angina were strong predictors of placebo-controlled PCI efficacy (angina symptom score: OR 1.9, 95% CrI 1.6 to 2.1, Pr(Interaction)=99.9% and OR 1.8, 95% CrI 1.6 to 2.1, Pr(Interaction)=99.9%, respectively). CONCLUSIONS: Although symptom severity and nature were poorly associated with disease severity, the nature of symptoms powerfully predicted the placebo-controlled efficacy of PCI.

N-of-1 Trial of Angina Verification Before Percutaneous Coronary Intervention.

BACKGROUND: In stable coronary artery disease, 30% to 60% of patients remain symptomatic despite successful revascularization. Perhaps not all symptoms reported by a patient with myocardial ischemia are, in fact, angina. OBJECTIVES: This study sought to determine whether independent symptom verification using a placebo-controlled ischemic stimulus could distinguish which patients achieve greatest symptom relief from percutaneous coronary intervention (PCI). METHODS: ORBITA-STAR was a multicenter, n-of-1, placebo-controlled study in patients undergoing single-vessel PCI for stable symptoms. Participants underwent 4 episodes (60 seconds each) of low-pressure balloon occlusion across their coronary stenosis, randomly paired with 4 episodes of placebo inflation. Following each episode, patients reported the similarity of the induced symptom in comparison with their usual symptom. The similarity score ranged from -10 (placebo replicated the symptom more than balloon occlusion) to +10 (balloon occlusion exactly replicated the symptom). The primary endpoint was the ability of the similarity score to predict symptom relief with PCI. RESULTS: Fifty-one patients were recruited, aged 62.9 ± 8.6 years. The median fractional flow reserve was 0.68 (Q1-Q3: 0.57-0.79), and the instantaneous wave-free ratio was 0.80 (Q1-Q3: 0.48-0.89). The median similarity score was 3 (Q1-Q3: 0.875-5.25). The similarity score was a strong predictor of symptom improvement following PCI: a patient with an upper quartile similarity score of 5.25 was significantly more likely to have lower angina frequency at follow-up (OR: 8.01; 95% credible interval: 2.39-15.86) than a patient with a lower quartile similarity score of 0.875 (OR: 1.31; 95% credible interval: 0.71-1.99), Pr(difference) >99.9%. CONCLUSIONS: Similarity score powerfully predicted symptom improvement from PCI. These data lay the foundation for independent symptom mapping to target PCI to those patients most likely to benefit. (Systematic Trial of Angina Assessment Before Revascularization [ORBITA-STAR]; NCT04280575).

Efficacy and Safety of Dapagliflozin in Patients With Acute Heart Failure.

BACKGROUND: The primary goals during acute heart failure (AHF) hospitalization are decongestion and guideline-directed medical therapy (GDMT) optimization. Unlike diuretics or other GDMT, early dapagliflozin initiation could achieve both AHF goals. OBJECTIVES: The authors aimed to assess the diuretic efficacy and safety of early dapagliflozin initiation in AHF. METHODS: In a multicenter, open-label study, 240 patients were randomized within 24 hours of hospital presentation for hypervolemic AHF to dapagliflozin 10 mg once daily or structured usual care with protocolized diuretic titration until day 5 or hospital discharge. The primary outcome, diuretic efficiency expressed as cumulative weight change per cumulative loop diuretic dose, was compared across treatment assignment using a proportional odds model adjusted for baseline weight. Secondary and safety outcomes were adjudicated by a blinded committee. RESULTS: For diuretic efficiency, there was no difference between dapagliflozin and usual care (OR: 0.65; 95% CI: 0.41-1.02; P = 0.06). Dapagliflozin was associated with reduced loop diuretic doses (560 mg [Q1-Q3: 260-1,150 mg] vs 800 mg [Q1-Q3: 380-1,715 mg]; P = 0.006) and fewer intravenous diuretic up-titrations (P ≤ 0.05) to achieve equivalent weight loss as usual care. Early dapagliflozin initiation did not increase diabetic, renal, or cardiovascular safety events. Dapagliflozin was associated with improved median 24-hour natriuresis (P = 0.03) and urine output (P = 0.005), expediting hospital discharge over the study period. CONCLUSIONS: Early dapagliflozin during AHF hospitalization is safe and fulfills a component of GDMT optimization. Dapagliflozin was not associated with a statistically significant reduction in weight-based diuretic efficiency but was associated with evidence for enhanced diuresis among patients with AHF. (Efficacy and Safety of Dapagliflozin in Acute Heart Failure [DICTATE-AHF]; NCT04298229).

Coronary sinus reducer for the treatment of refractory angina (ORBITA-COSMIC): a randomised, placebo-controlled trial.

BACKGROUND: The coronary sinus reducer (CSR) is proposed to reduce angina in patients with stable coronary artery disease by improving myocardial perfusion. We aimed to measure its efficacy, compared with placebo, on myocardial ischaemia reduction and symptom improvement. METHODS: ORBITA-COSMIC was a double-blind, randomised, placebo-controlled trial conducted at six UK hospitals. Patients aged 18 years or older with angina, stable coronary artery disease, ischaemia, and no further options for treatment were eligible. All patients completed a quantitative adenosine-stress perfusion cardiac magnetic resonance scan, symptom and quality-of-life questionnaires, and a treadmill exercise test before entering a 2-week symptom assessment phase, in which patients reported their angina symptoms using a smartphone application (ORBITA-app). Patients were randomly assigned (1:1) to receive either CSR or placebo. Both participants and investigators were masked to study assignment. After the CSR implantation or placebo procedure, patients entered a 6-month blinded follow-up phase in which they reported their daily symptoms in the ORBITA-app. At 6 months, all assessments were repeated. The primary outcome was myocardial blood flow in segments designated ischaemic at enrolment during the adenosine-stress perfusion cardiac magnetic resonance scan. The primary symptom outcome was the number of daily angina episodes. Analysis was done by intention-to-treat and followed Bayesian methodology. The study is registered with ClinicalTrials.gov, NCT04892537, and completed. FINDINGS: Between May 26, 2021, and June 28, 2023, 61 patients were enrolled, of whom 51 (44 [86%] male; seven [14%] female) were randomly assigned to either the CSR group (n=25) or the placebo group (n=26). Of these, 50 patients were included in the intention-to-treat analysis (24 in the CSR group and 26 in the placebo group). 454 (57%) of 800 imaged cardiac segments were ischaemic at enrolment, with a median stress myocardial blood flow of 1·08 mL/min per g (IQR 0·77-1·41). Myocardial blood flow in ischaemic segments did not improve with CSR compared with placebo (difference 0·06 mL/min per g [95% CrI -0·09 to 0·20]; Pr(Benefit)=78·8%). The number of daily angina episodes was reduced with CSR compared with placebo (OR 1·40 [95% CrI 1·08 to 1·83]; Pr(Benefit)=99·4%). There were two CSR embolisation events in the CSR group, and no acute coronary syndrome events or deaths in either group. INTERPRETATION: ORBITA-COSMIC found no evidence that the CSR improved transmural myocardial perfusion, but the CSR did improve angina compared with placebo. These findings provide evidence for the use of CSR as a further antianginal option for patients with stable coronary artery disease. FUNDING: Medical Research Council, Imperial College Healthcare Charity, National Institute for Health and Care Research Imperial Biomedical Research Centre, St Mary's Coronary Flow Trust, British Heart Foundation.

Impact of Insurance Status and Region on Angiotensin Receptor-Neprilysin Inhibitor Prescription During Heart Failure Hospitalizations.

BACKGROUND: An angiotensin receptor-neprilysin inhibitor (ARNI) is the preferred renin-angiotensin system (RAS) inhibitor for heart failure with reduced ejection fraction (HFrEF). Among eligible patients, insurance status and prescriber concern regarding out-of-pocket costs may constrain early initiation of ARNI and other new therapies. OBJECTIVES: In this study, the authors sought to evaluate the association of insurance and other social determinants of health with ARNI initiation at discharge from HFrEF hospitalization. METHODS: The authors analyzed ARNI initiation from January 2017 to June 2020 among patients with HFrEF eligible to receive RAS inhibitor at discharge from hospitals in the Get With The Guidelines-Heart Failure registry. The primary outcome was the proportion of ARNI prescription at discharge among those prescribed RAS inhibitor who were not on ARNI on admission. A logistic regression model was used to determine the association of insurance status, U.S. region, and their interaction, as well as self-reported race, with ARNI initiation at discharge. RESULTS: From 42,766 admissions, 24,904 were excluded for absolute or relative contraindications to RAS inhibitors. RAS inhibitors were prescribed for 16,817 (94.2%) of remaining discharges, for which ARNI was prescribed in 1,640 (9.8%). Self-reported Black patients were less likely to be initiated on ARNI compared to self-reported White patients (OR: 0.64; 95% CI: 0.50-0.81). Compared to Medicare beneficiaries, patients with third-party insurance, Medicaid, or no insurance were less likely to be initiated on ARNI (OR: 0.47 [95% CI: 0.31-0.72], OR: 0.41 [95% CI: 0.25-0.67], and OR: 0.20 [95% CI: 0.08-0.47], respectively). ARNI therapy varied by hospital region, with lowest utilization in the Mountain region. An interaction was demonstrated between the impact of insurance disparities and hospital region. CONCLUSIONS: Among patients hospitalized between 2017 and 2020 for HFrEF who were prescribed RAS inhibitor therapy at discharge, insurance status, geographic region, and self-reported race were associated with ARNI initiation.

Would You Rather: Quantifying TBI Survivor Perceptions of Functional Status via Their Surrogates.

OBJECTIVE: To quantify health utilities of the Glasgow Outcome Scale-Extended (GOSE) states after actual Traumatic Brain Injury (TBI). BACKGROUND: Recovery after TBI is measured using the GOSE, a validated clinical trial endpoint. A recent public survey quantified the health utilities of some GOSE states after hypothetical TBI as worse than death. However, no health utilities exist for disability after actual TBI. METHODS: This national computer-adaptive survey followed EQUATOR-CHERRIES guidelines and recruited adult TBI survivors (injury>1 y prior) via their available surrogates. Using a standard gamble approach in randomized order, participants gave preferences for post-TBI categorical health states ranging from GOSE 2-8. We calculated median [interquartile range, IQR] health utilities for each GOSE state, from -1 (worse than death) to 1 (full health), with 0 as reference (death, GOSE 1). RESULTS: Of 515 eligible, 298 surrogates (58%) consented and completed the scenarios on TBI survivors' behalf. TBI survivors had a current median GOSE 5 [3-7]. GOSE 2, GOSE 3, and GOSE 4 were rated worse than death by 89%, 64%, and 38%, respectively. The relationship was nonlinear, and intervals were unequal between states, with a bimodal distribution for GOSE 4. CONCLUSIONS: In this index study of actual post-TBI disability, poor neurologic outcomes represented by GOSE 2-4 were perceived as worse than death by at least one in three survivors. Similar to previously reported public perceptions after a hypothetical TBI, these long-term perceptions may inform earlier post-TBI shared decision making, as well as help shape value-based research and quality of care. LEVEL OF EVIDENCE: II, Economic & Value-based Evaluations.

A double-blind, randomised, placebo-controlled trial of the coronary sinus Reducer in refractory angina: design and rationale of the ORBITA-COSMIC trial.

The coronary sinus Reducer (CSR) is an hourglass-shaped device which creates an artificial stenosis in the coronary sinus. Whilst placebo-controlled data show an improvement in angina, these results are unreplicated and are the subject of further confirmatory research. The mechanism of action of this unintuitive therapy is unknown. The Coronary Sinus Reducer Objective Impact on Symptoms, MRI Ischaemia, and Microvascular Resistance (ORBITA-COSMIC) trial is a randomised, placebo-controlled, double-blind trial investigating the efficacy of the CSR. Patients with (i) established epicardial coronary artery disease, (ii) angina on maximally tolerated antianginal medication, (iii) evidence of myocardial ischaemia and (iv) no further options for percutaneous coronary intervention or coronary artery bypass grafting will be enrolled. Upon enrolment, angina and quality-of-life questionnaires, treadmill exercise testing and quantitative stress perfusion cardiac magnetic resonance (CMR) imaging will be performed. Participants will record their symptoms daily on a smartphone application throughout the trial. After a 2-week symptom assessment phase, participants will be randomised in the cardiac catheterisation laboratory to CSR or a placebo procedure. After 6 months of blinded follow-up, all prerandomisation tests will be repeated. A prespecified subgroup will undergo invasive coronary physiology assessment at prerandomisation and follow-up. The primary outcome is stress myocardial blood flow on CMR. Secondary outcomes include angina frequency, quality of life and treadmill exercise time. (ClinicalTrials.gov: NCT04892537).

Surrogate Perception of Disability after Hospitalization for Traumatic Brain Injury.

BACKGROUND: The Glasgow Outcome Scale Extended (GOSE) is a measure of recovery after traumatic brain injury (TBI). Public surveys rate some GOSE states as worse than death. Direct family experience caring for patients with TBI may impact views of post-TBI disability. STUDY DESIGN: We conducted a national cross-sectional computer-adaptive survey of surrogates of TBI dependents incurring injury more than 1 year earlier. Using a standard gamble approach in randomized order, surrogates evaluated preferences for post-TBI GOSE states from GOSE 2 (bedridden, unaware) to GOSE 8 (good recovery). We calculated median (interquartile range [IQR]) health utilities for each post-TBI state, ranging from -1 to 1, with 0 as reference (death = GOSE 1), and assessed sociodemographic associations using proportional odds logistic regression modeling. RESULTS: Of 515 eligible surrogates, 298 (58%) completed scenarios. Surrogates were median aged 46 (IQR 35 to 60), 54% married, with Santa Clara strength of faith 14 (10 to 18). TBI dependents had a median GOSE5 (3 to 7). Median (IQR) health utility ratings for GOSE 2, GOSE 3, and GOSE 4 were -0.06 (-0.50 to -0.01), -0.01 (-0.30 to 0.45), and 0.30 (-0.01 to 0.80), rated worse than death by 91%, 65%, and 40%, respectively. Surrogates rated GOSE 4 (daily partial help) worse than the general population. Married surrogates rated GOSE 4 higher (p < 0.01). Higher strength of faith was associated with higher utility scores across GOSE states (p = 0.034). CONCLUSIONS: In this index study of surrogate perceptions about disability after TBI, poor neurologic outcomes-vegetative, needing all-day or partial daily assistance-were perceived as worse than death by at least 1 in 3 surrogates. Surrogate perceptions differed from the unexposed public. Long-term perceptions about post-TBI disability may inform earlier, tailored shared decision-making after neurotrauma.

Evolution and optimization of clinical trial endpoints and design in pulmonary arterial hypertension.

Selection of endpoints for clinical trials in pulmonary arterial hypertension (PAH) is challenging because of the small numbers of patients and the changing expectations of patients, clinicians, and regulators in this evolving therapy area. The most commonly used primary endpoint in PAH trials has been 6-min walk distance (6MWD), leading to the approval of several targeted therapies. However, single surrogate endpoints such as 6MWD or hemodynamic parameters may not correlate with clinical outcomes. Composite endpoints of clinical worsening have been developed to reflect patients' overall condition more accurately, although there is no standard definition of worsening. Recently there has been a shift to composite endpoints assessing clinical improvement, and risk scores developed from registry data are increasingly being used. Biomarkers are another area of interest, although brain natriuretic peptide and its N-terminal prohormone are the only markers used for risk assessment or as endpoints in PAH. A range of other genetic, metabolic, and immunologic markers is currently under investigation, along with conventional and novel imaging modalities. Patient-reported outcomes are an increasingly important part of evaluating new therapies, and several PAH-specific tools are now available. In the future, alternative statistical techniques and trial designs, such as patient enrichment strategies, will play a role in evaluating PAH-targeted therapies. In addition, modern sequencing techniques, imaging analyses, and high-dimensional statistical modeling/machine learning may reveal novel markers that can play a role in the diagnosis and monitoring of PAH.

Impact of Tissue Expander Surface Texture on Two-Stage Breast Reconstruction Outcomes: A Combined Analysis of Incidence, Timing, and Severity.

BACKGROUND: With ongoing investigations of the impact of device texturing on breast implant-related anaplastic large cell lymphoma (BIA-ALCL), studies have begun comparing complications profiles of tissue expanders. However, there is a paucity of timing and severity data of complications. The aim of this study is to provide a comparative survival analysis of post-operative complications between smooth (STE) and textured tissue expanders (TTE) in breast reconstruction. METHODS: A single institution experience with tissue expander breast reconstruction was reviewed for complications up to 1 year post 2nd stage reconstruction from 2014-2020. Demographics, comorbidities, operation-related variables, and complications were evaluated. Kaplan-Meier curves, cox proportional hazard models, and a consensus based ordinal logistic regression model were used to compare complication profiles. RESULTS: Of 919 ttal patients, 65.3% (n=600) received TTEs and 34.7% (n=319) received STEs. There was increased risk of infection (p<0.0001), seroma (p=0.046), expander malposition (p<0.0001), and wound dehiscence (p=0.019) in STEs compared to TTEs. However, there were also decreased risk of capsular contracture (p=0.005) in STEs compared to TTEs. Failure of breast reconstruction (p<0.001) and wound dehiscence (p=0.018) occurred significantly earlier in STEs compared to TTEs. Predictors for significantly higher severity complications included: smooth tissue expander use (p=0.007), shorter time to complication (p<0.0001), higher BMI (p=0.005), smoking history (p=0.025), and nipple sparing mastectomy (p=0.012). CONCLUSIONS: Differences in the timing and severity of complications contribute to the safety profiles of tissue expanders. STEs are associated increased odds of higher severity and earlier complications. Therefore, tissue expander selection may depend on underlying risk factors and severity predictors.

Why are There not More Bayesian Clinical Trials? Ability to Interpret Bayesian and Conventional Statistics Among Medical Researchers.

OBJECTIVE AND BACKGROUND: We assessed current understandings in interpretation of Bayesian and traditional statistical results within the clinical researcher (non-statistician) community. METHODS: Within a 22-question survey, including demographics and experience and comfort levels with Bayesian analyses, we included questions on how to interpret both Bayesian and traditional statistical outputs. We also assessed whether Bayesian or traditional interpretations are considered more useful. RESULTS: Among the 323 respondent clinicians, 42.4% and 36.5% chose the correct interpretations of the posterior probability and 95% credible interval, respectively. Only 11.5% of respondents interpreted the p-value correctly and 23.5% interpreted the 95% confidence interval correctly. CONCLUSIONS: Based on these survey results, we conclude that most of these clinicians face uncertainty when attempting to interpret results from both Bayesian and traditional statistical outputs. When presented with accurate interpretations, clinicians generally conclude that Bayesian results are more useful than conventional ones. We believe there is a need for education of clinicians in statistical interpretation in ways that are customized to this audience.

Why are not There More Bayesian Clinical Trials? Perceived Barriers and Educational Preferences Among Medical Researchers Involved in Drug Development.

OBJECTIVE AND BACKGROUND: The clinical trials community has been hesitant to adopt Bayesian statistical methods, which are often more flexible and efficient with more naturally interpretable results than frequentist methods. We aimed to identify self-reported barriers to implementing Bayesian methods and preferences for becoming comfortable with them. METHODS: We developed a 22-question survey submitted to medical researchers (non-statisticians) from industry, academia, and regulatory agencies. Question areas included demographics, experience, comfort levels with Bayesian analyses, perceived barriers to these analyses, and preferences for increasing familiarity with Bayesian methods. RESULTS: Of the 323 respondents, most were affiliated with pharmaceutical companies (33.4%), clinical research organizations (29.7%), and regulatory agencies (18.6%). The rest represented academia, medical practice, or other. Over 56% of respondents expressed little to no comfort in interpreting Bayesian analyses. "Insufficient knowledge of Bayesian approaches" was ranked the most important perceived barrier to implementing Bayesian methods by a plurality (48%). Of the approaches listed, in-person training was the most preferred for gaining comfort with Bayesian methods. CONCLUSIONS: Based on these survey results, we recommend that introductory level training on Bayesian statistics be presented in an in-person workshop that could also be broadcast online with live Q&A. Other approaches such as online training or collaborative projects may be better suited for higher-level trainings where instructors may assume a baseline understanding of Bayesian statistics. Increased coverage of Bayesian methods at medical conferences and medical school trainings would help improve comfort and overcome the substantial knowledge barriers medical researchers face when implementing these methods.

Association of step counts over time with the risk of chronic disease in the All of Us Research Program.

The association between physical activity and human disease has not been examined using commercial devices linked to electronic health records. Using the electronic health records data from the All of Us Research Program, we show that step count volumes as captured by participants' own Fitbit devices were associated with risk of chronic disease across the entire human phenome. Of the 6,042 participants included in the study, 73% were female, 84% were white and 71% had a college degree, and participants had a median age of 56.7 (interquartile range 41.5-67.6) years and body mass index of 28.1 (24.3-32.9) kg m-2. Participants walked a median of 7,731.3 (5,866.8-9,826.8) steps per day over the median activity monitoring period of 4.0 (2.2-5.6) years with a total of 5.9 million person-days of monitoring. The relationship between steps per day and incident disease was inverse and linear for obesity (n = 368), sleep apnea (n = 348), gastroesophageal reflux disease (n = 432) and major depressive disorder (n = 467), with values above 8,200 daily steps associated with protection from incident disease. The relationships with incident diabetes (n = 156) and hypertension (n = 482) were nonlinear with no further risk reduction above 8,000-9,000 steps. Although validation in a more diverse sample is needed, these findings provide a real-world evidence-base for clinical guidance regarding activity levels that are necessary to reduce disease risk.

Securely sharing DSMB reports to speed decision making from multiple, concurrent, independent studies of similar treatments in COVID-19.

Introduction: As clinical trials were rapidly initiated in response to the COVID-19 pandemic, Data and Safety Monitoring Boards (DSMBs) faced unique challenges overseeing trials of therapies never tested in a disease not yet characterized. Traditionally, individual DSMBs do not interact or have the benefit of seeing data from other accruing trials for an aggregated analysis to meaningfully interpret safety signals of similar therapeutics. In response, we developed a compliant DSMB Coordination (DSMBc) framework to allow the DSMB from one study investigating the use of SARS-CoV-2 convalescent plasma to treat COVID-19 to review data from similar ongoing studies for the purpose of safety monitoring. Methods: The DSMBc process included engagement of DSMB chairs and board members, execution of contractual agreements, secure data acquisition, generation of harmonized reports utilizing statistical graphics, and secure report sharing with DSMB members. Detailed process maps, a secure portal for managing DSMB reports, and templates for data sharing and confidentiality agreements were developed. Results: Four trials participated. Data from one trial were successfully harmonized with that of an ongoing trial. Harmonized reports allowing for visualization and drill down into the data were presented to the ongoing trial's DSMB. While DSMB deliberations are confidential, the Chair confirmed successful review of the harmonized report. Conclusion: It is feasible to coordinate DSMB reviews of multiple independent studies of a similar therapeutic in similar patient cohorts. The materials presented mitigate challenges to DSMBc and will help expand these initiatives so DSMBs may make more informed decisions with all available information.

Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure.

Machine learning is increasingly being used to predict clinical outcomes. Most comparisons of different methods have been based on empirical analyses in specific datasets. We used Monte Carlo simulations to determine when machine learning methods perform better than statistical learning methods in a specific setting. We evaluated six learning methods: stochastic gradient boosting machines using trees as the base learners, random forests, artificial neural networks, the lasso, ridge regression, and linear regression estimated using ordinary least squares (OLS). Our simulations were informed by empirical analyses in patients with acute myocardial infarction (AMI) and congestive heart failure (CHF) and used six data-generating processes, each based on one of the six learning methods, to simulate continuous outcomes in the derivation and validation samples. The outcome was systolic blood pressure at hospital discharge, a continuous outcome. We applied the six learning methods in each of the simulated derivation samples and evaluated performance in the simulated validation samples. The primary observation was that neural networks tended to result in estimates with worse predictive accuracy than the other five methods in both disease samples and across all six data-generating processes. Boosted trees and OLS regression tended to perform well across a range of scenarios.

Development and Validation of a Model to Predict Postdischarge Opioid Use After Cesarean Birth.

OBJECTIVE: To develop and validate a prediction model for postdischarge opioid use in patients undergoing cesarean birth. METHODS: We conducted a prospective cohort study of patients undergoing cesarean birth. Patients were enrolled postoperatively, and they completed pain and opioid use questionnaires 14 days after cesarean birth. Clinical data were abstracted from the electronic health record (EHR). Participants were prescribed 30 tablets of hydrocodone 5 mg-acetaminophen 325 mg at discharge and were queried about postdischarge opioid use. The primary outcome was total morphine milligram equivalents used. We constructed three proportional odds predictive models of postdischarge opioid use: a full model with 34 predictors available before hospital discharge, an EHR model that excluded questionnaire data, and a reduced model. The reduced model used forward selection to sequentially add predictors until 90% of the full model performance was achieved. Predictors were ranked a priori based on data from the literature and prior research. Predictive accuracy was estimated using discrimination (concordance index). RESULTS: Between 2019 and 2020, 459 participants were enrolled and 279 filled the standardized study prescription. Of the 398 with outcome measurements, participants used a median of eight tablets (interquartile range 1-18 tablets) after discharge, 23.5% used no opioids, and 23.0% used all opioids. Each of the models demonstrated high accuracy predicting postdischarge opioid use (concordance index range 0.74-0.76 for all models). We selected the reduced model as our final model given its similar model performance with the fewest number of predictors, all obtained from the EHR (inpatient opioid use, tobacco use, and depression or anxiety). CONCLUSION: A model with three predictors readily found in the EHR-inpatient opioid use, tobacco use, and depression or anxiety-accurately estimated postdischarge opioid use. This represents an opportunity for individualizing opioid prescriptions after cesarean birth.