Transfer learning for clustering single-cell RNA-seq data crossing-species and batch, case on uterine fibroids.

Due to the high dimensionality and sparsity of the gene expression matrix in single-cell RNA-sequencing (scRNA-seq) data, coupled with significant noise generated by shallow sequencing, it poses a great challenge for cell clustering methods. While numerous computational methods have been proposed, the majority of existing approaches center on processing the target dataset itself. This approach disregards the wealth of knowledge present within other species and batches of scRNA-seq data. In light of this, our paper proposes a novel method named graph-based deep embedding clustering (GDEC) that leverages transfer learning across species and batches. GDEC integrates graph convolutional networks, effectively overcoming the challenges posed by sparse gene expression matrices. Additionally, the incorporation of DEC in GDEC enables the partitioning of cell clusters within a lower-dimensional space, thereby mitigating the adverse effects of noise on clustering outcomes. GDEC constructs a model based on existing scRNA-seq datasets and then applying transfer learning techniques to fine-tune the model using a limited amount of prior knowledge gleaned from the target dataset. This empowers GDEC to adeptly cluster scRNA-seq data cross different species and batches. Through cross-species and cross-batch clustering experiments, we conducted a comparative analysis between GDEC and conventional packages. Furthermore, we implemented GDEC on the scRNA-seq data of uterine fibroids. Compared results obtained from the Seurat package, GDEC unveiled a novel cell type (epithelial cells) and identified a notable number of new pathways among various cell types, thus underscoring the enhanced analytical capabilities of GDEC. Availability and implementation: https://github.com/YuzhiSun/GDEC/tree/main.

SAM: Self-adapting mixture prior to dynamically borrow information from historical data in clinical trials.

Mixture priors provide an intuitive way to incorporate historical data while accounting for potential prior-data conflict by combining an informative prior with a noninformative prior. However, prespecifying the mixing weight for each component remains a crucial challenge. Ideally, the mixing weight should reflect the degree of prior-data conflict, which is often unknown beforehand, posing a significant obstacle to the application and acceptance of mixture priors. To address this challenge, we introduce self-adapting mixture (SAM) priors that determine the mixing weight using likelihood ratio test statistics or Bayes factors. SAM priors are data-driven and self-adapting, favoring the informative (noninformative) prior component when there is little (substantial) evidence of prior-data conflict. Consequently, SAM priors achieve dynamic information borrowing. We demonstrate that SAM priors exhibit desirable properties in both finite and large samples and achieve information-borrowing consistency. Moreover, SAM priors are easy to compute, data-driven, and calibration-free, mitigating the risk of data dredging. Numerical studies show that SAM priors outperform existing methods in adopting prior-data conflicts effectively. We developed R package "SAMprior" and web application that are freely available at CRAN and www.trialdesign.org to facilitate the use of SAM priors.

Neighborhood Socioeconomic Disadvantage and Mortality Among Medicare Beneficiaries Hospitalized for Acute Myocardial Infarction, Heart Failure, and Pneumonia.

BACKGROUND: The Centers for Medicare and Medicaid Services' Hospital Value-Based Purchasing program uses 30-day mortality rates for acute myocardial infarction, heart failure, and pneumonia to evaluate US hospitals, but does not account for neighborhood socioeconomic disadvantage when comparing their performance. OBJECTIVE: To determine if neighborhood socioeconomic disadvantage is associated with worse 30-day mortality rates after a hospitalization for acute myocardial infarction (AMI), heart failure (HF), or pneumonia in the USA, as well as within the subset of counties with a high proportion of Black individuals. DESIGN AND PARTICIPANTS: This retrospective, population-based study included all Medicare fee-for-service beneficiaries aged 65 years or older hospitalized for acute myocardial infarction, heart failure, or pneumonia between 2012 and 2015. EXPOSURE: Residence in most socioeconomically disadvantaged vs. less socioeconomically disadvantaged neighborhoods as measured by the area deprivation index (ADI). MAIN MEASURE(S): All-cause mortality within 30 days of admission. KEY RESULTS: The study included 3,471,592 Medicare patients. Of these patients, 333,472 resided in most disadvantaged neighborhoods and 3,138,120 in less disadvantaged neighborhoods. Patients living in the most disadvantaged neighborhoods were younger (78.4 vs. 80.0 years) and more likely to be Black adults (24.6% vs. 7.5%) and dually enrolled in Medicaid (39.4% vs. 21.8%). After adjustment for demographics (age, sex, race/ethnicity), poverty, and clinical comorbidities, 30-day mortality was higher among beneficiaries residing in most disadvantaged neighborhoods for AMI (adjusted odds ratio 1.08, 95% CI 1.06-1.11) and pneumonia (aOR 1.05, 1.03-1.07), but not for HF (aOR 1.02, 1.00-1.04). These patterns were similar within the subset of US counties with a high proportion of Black adults (AMI, aOR 1.07, 1.03-1.11; HF 1.02, 0.99-1.05; pneumonia 1.03, 1.00-1.07). CONCLUSIONS: Neighborhood socioeconomic disadvantage is associated with higher 30-day mortality for some conditions targeted by value-based programs, even after accounting for individual-level demographics, clinical comorbidities, and poverty. These findings may have implications as policymakers weigh strategies to advance health equity under value-based programs.

Use of Administrative Claims to Assess Outcomes and Treatment Effect in Randomized Clinical Trials for Transcatheter Aortic Valve Replacement: Findings From the EXTEND Study.

BACKGROUND: Whether passively collected data can substitute for adjudicated outcomes to reproduce the magnitude and direction of treatment effect observed in cardiovascular clinical trials is not well known. METHODS: We linked adults ≥65 years of age in the HiR (US CoreValve Pivotal High Risk) and SURTAVI trials (Surgical or Transcatheter Aortic Valve Replacement in Intermediate-Risk Patients) to 100% Medicare inpatient claims, January 1, 2011, to December 31, 2016. Primary (eg, death and stroke) and secondary trial end points were compared across treatment arms (eg, transcatheter aortic valve replacement [TAVR] versus surgical aortic valve replacement [SAVR]) using trial-adjudicated outcomes versus outcomes derived from claims at 1 year (HiR) or 2 years (SURTAVI). RESULTS: Among 600 linked HiR participants (linkage rate, 80.0%), the rate of the trial's primary end point of all-cause mortality occurred in 13.7% of patients receiving TAVR and 16.4% of patients receiving SAVR at 1 year by using both trial data (hazard ratio, 0.84 [95% CI, 0.65-1.09]; P=0.33) and claims data (hazard ratio, 0.86 [95% CI, 0.66-1.11]; P=0.34; interaction P value=0.80). Noninferiority of TAVR relative to SAVR was seen by using both trial- and claims-based outcomes (Pnoninferiority<0.001 for both). Among 1005 linked SURTAVI trial participants (linkage rate, 60.5%), the trial's primary end point was 12.9% for TAVR and 13.1% for SAVR using trial data (hazard ratio, 1.08 [95% CI, 0.79-1.48]; P=0.90), and 11.3% for TAVR and 12.5% for SAVR patients using claims data (hazard ratio, 1.02 [95% CI, 0.73-1.41]; P=0.58; interaction P value=0.89). TAVR was noninferior to SAVR when compared using both trial and claims (Pnoninferiority<0.001 for both). Rates of procedural secondary outcomes (eg, aortic valve reintervention, pacemaker rates) were more closely concordant between trial and claims data than nonprocedural outcomes (eg, stroke, bleeding, cardiogenic shock). CONCLUSIONS: In the HiR and SURTAVI trials, ascertainment of trial primary end points using claims reproduced both the magnitude and direction of treatment effect in comparison with adjudicated event data, but nonfatal and nonprocedural secondary outcomes were not as well reproduced. Use of claims to substitute for adjudicated outcomes in traditional trial treatment comparisons may be valid and feasible for all-cause mortality and certain procedural outcomes but may be less suitable for other end points.

Use of etomidate in patients with heart failure undergoing noncardiac surgery.

BACKGROUND: Patients with symptomatic and asymptomatic heart failure undergoing noncardiac surgery may benefit from the haemodynamic profile of etomidate. However, the safety of etomidate in this population is unknown. We examined anaesthesiologist variation in etomidate use and assessed its safety using an instrumental variable approach to account for differences in treatment selection. METHODS: A retrospective cohort study of 19 714 patients with heart failure undergoing noncardiac surgery at two tertiary care institutions from January 2006 to December 2017 was performed. The proportion of etomidate use among 294 anaesthesiologists was examined and adjusted risk differences (aRD) for in-hospital and 30-day mortality were calculated using physician preference for etomidate as an instrumental variable. RESULTS: Etomidate was used in 14.3% (2821/19 714) of patients. Preference for etomidate varied substantially among individual anaesthesiologists with the lowest and highest quartile users using etomidate in 0-4.7% and 20.4-66.7% of their own heart failure patients, respectively. The adjusted instrumental variable analysis showed no significant differences in the risk of in-hospital (aRD -0.2%; 95% confidence interval, -2.4%-1.9%; P=0.83) or 30 day mortality (aRD 0.2%; 95% confidence interval, -2.5%-2.9%; P=0.90). Anaesthesiologists with higher preferences for etomidate were more experienced (greater heart failure and total case volume) than anaesthesiologists with lower preferences for etomidate. CONCLUSIONS: We found substantial variability in anaesthesiologists' preference for etomidate for use in patients with heart failure undergoing noncardiac surgery. There was no association between etomidate use and in-hospital or 30-day mortality. Etomidate is not inferior to other currently used options for induction of general anaesthesia in patients with heart failure.

Validating the use of registries and claims data to support randomized trials: Rationale and design of the Extending Trial-Based Evaluations of Medical Therapies Using Novel Sources of Data (EXTEND) Study.

BACKGROUND: Randomized controlled trials are the "gold standard" for comparing the safety and efficacy of therapies but may be limited due to high costs, lack of feasibility, and difficulty enrolling "real-world" patient populations. The Extending Trial-Based Evaluations of Medical Therapies Using Novel Sources of Data (EXTEND) Study seeks to evaluate whether data collected within procedural registries and claims databases can reproduce trial results by substituting surrogate non-trial-based variables for exposures and outcomes. METHODS AND RESULTS: Patient-level data from 2 clinical trial programs-the Dual Antiplatelet Therapy Study and the United States CoreValve Studies-will be linked to a combination of national registry, administrative claims, and health system data. The concordance between baseline and outcomes data collected within nontrial data sets and trial information, including adjudicated end point events, will be assessed. We will compare the study results obtained using these alternative data sources to those derived using trial-ascertained variables and end points using trial-adjudicated end points and covariates. CONCLUSIONS: Linkage of trials to registries and claims data represents an opportunity to use alternative data sources in place of and as adjuncts to randomized clinical trial data but requires further validation. The results of this research will help determine how these data sources can be used to improve our present and future understanding of new medical treatments.

Extracorporeal Membrane Oxygenation Use in Cardiogenic Shock: Impact of Age on In-Hospital Mortality, Length of Stay, and Costs.

OBJECTIVES: Increasing age is a well-recognized risk factor for in-hospital mortality in patients receiving extracorporeal membrane oxygenation for cardiogenic shock, but the shape of this relationship is unknown. In addition, the impact of age on hospital length of stay, patterns of patient disposition, and costs has been incompletely characterized. DESIGN: Retrospective analysis of the National Inpatient Sample. SETTING: U.S. nonfederal hospitals, years 2004-2016. PATIENTS: Adults with cardiogenic shock treated with extracorporeal membrane oxygenation (3,094; weighted national estimate: 15,415). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The mean age of extracorporeal membrane oxygenation recipients was 54.8 ± 15.4 years (range, 18-90 yr). Crude in-hospital mortality was 57.7%. Median time-to-death was 8 days (interquartile range, 3-17 d). A linear relationship between age and in-hospital mortality was observed with a 14% increase in the adjusted odds of in-hospital mortality for every 10-year increase in age (adjusted odds ratio, 1.14; 95% CI, 1.08-1.21; p < 0.0001). Thirty-four percent of patients were discharged alive at a median time of 30 days (interquartile range, 19-48 d). The median length of stay and total hospitalization costs were 14 days (interquartile range, [5-29 d]) and $134,573 ($71,782-$239,439), respectively, both of which differed significantly by age group (length of stay range from 17 d [18-49 yr] to 9 d [80-90 yr]; p < 0.0001 and cost range $147,548 [18-49 yr] to $105,350 [80-90 yr]; p < 0.0001). CONCLUSIONS: Age is linearly associated with increasing in-hospital mortality in individuals receiving extracorporeal membrane oxygenation for cardiogenic shock without evidence of a threshold effect. Median time-to-death is approximately 1 week. One third of patients are discharged from the hospital alive, but the median time-to-discharge is 1 month. Median length of stay ranges from 9 to 17 days depending on age. Hospitalization costs exceed $100,000 in all age groups.

Interactions between Environmental Factors and Glutathione S-Transferase (GST) Genes with Respect to Detectable Blood Aluminum Concentrations in Jamaican Children.

Aluminum (Al) is a metallic toxicant at high concentrations following natural or unnatural exposures. Dietary intake is considered as the main source of aluminum exposure in children. We used data from 366 typically developing (TD) children (ages 2−8 years) who participated as controls in an age- and sex-matched case−control study in Jamaica. We investigated additive and interactive associations among environmental factors and children's genotypes for glutathione S-transferase (GST) genes (GSTT1, GSTM1, GSTP1), in relation to having a detectable blood aluminum concentration (BAlC) of >5.0 µg/L, using multivariable logistic regression models. Findings from interactive models revealed that the odds of having a detectable BAlC was significantly higher among children who ate string beans (p ≤ 0.01), whereas about 40% lower odds of having a detectable BAlC was observed in children with higher parental education level, (p = 0.02). A significant interaction between consumption of saltwater fish and GSTP1 in relation to having a detectable BAlC using either co-dominant or dominant genetic models (overall interaction p = 0.02 for both models) indicated that consumption of saltwater fish was associated with higher odds of having a detectable BAlC only among children with the GSTP1 Ile105Val Ile/Ile genotype using either co-dominant or dominant models [OR (95% CI) = 2.73 (1.07, 6.96), p = 0.04; and OR (95% CI) = 2.74 (1.08, 6.99), p = 0.03]. Since this is the first study from Jamaica that reports such findings, replication in other populations is warranted.

Additive and Interactive Associations of Environmental and Sociodemographic Factors with the Genotypes of Three Glutathione S-Transferase Genes in Relation to the Blood Arsenic Concentrations of Children in Jamaica.

Arsenic (As) is a metalloid that has been classified as a xenobiotic with toxic effects on human beings, especially on children. Since the soil in Jamaica contains As, dietary intake is considered the main source of As exposure in Jamaicans. In addition, glutathione S-transferase (GST) genes, including GSTT1, GSTP1, and GSTM1, play an important role in the metabolism of xenobiotics including As in humans. Using data from 375 typically developing children (2-8 years) in Jamaica, we investigated the environmental and sociodemographic factors, as well as their possible interactions with the children's genotype for GST genes in relation to having a detectable level of blood As concentration (i.e., >1.3 µg/L). Using multivariable logistic regression, we have identified environmental factors significantly associated with blood As concentrations that include a child's age, parental education levels, and the consumption of saltwater fish, cabbage, broad beans, and avocado (all p < 0.01). Based on the multivariable analysis including gene x environment interactions, we found that among children with the Ile/Ile genotype for GSTP1 Ile105Val, children who consumed avocado had higher odds of having a detectable blood As concentration compared to children who did not eat avocado.

Development and validation of an echocardiographic algorithm to predict long-term mitral and tricuspid regurgitation progression.

AIMS: Prediction of mitral (MR) and tricuspid (TR) regurgitation progression on transthoracic echocardiography (TTE) is needed to personalize valvular surveillance intervals and prognostication. METHODS AND RESULTS: Structured TTE report data at Beth Israel Deaconess Medical Center, 26 January 2000-31 December 2017, were used to determine time to progression (≥1+ increase in severity). TTE predictors of progression were used to create a progression score, externally validated at Massachusetts General Hospital, 1 January 2002-31 December 2019. In the derivation sample (MR, N = 34 933; TR, N = 27 526), only 5379 (15.4%) individuals with MR and 3630 (13.2%) with TR had progression during a median interquartile range) 9.0 (4.1-13.4) years of follow-up. Despite wide inter-individual variability in progression rates, a score based solely on demographics and TTE variables identified individuals with a five- to six-fold higher rate of MR/TR progression over 10 years (high- vs. low-score tertile, rate of progression; MR 20.1% vs. 3.3%; TR 21.2% vs. 4.4%). Compared to those in the lowest score tertile, those in the highest tertile of progression had a four-fold increased risk of mortality. On external validation, the score demonstrated similar performance to other algorithms commonly in use. CONCLUSION: Four-fifths of individuals had no progression of MR or TR over two decades. Despite wide interindividual variability in progression rates, a score, based solely on TTE parameters, identified individuals with a five- to six-fold higher rate of MR/TR progression. Compared to the lowest tertile, individuals in the highest score tertile had a four-fold increased risk of mortality. Prediction of long-term MR/TR progression is not only feasible but prognostically important.

Validation of Administrative Claims to Ascertain Outcomes in Pivotal Trials of Transcatheter Aortic Valve Replacement.

OBJECTIVES: The aim of this study was to evaluate the performance of administrative claims in ascertaining trial clinical events committee-adjudicated outcomes in the U.S. CoreValve studies. BACKGROUND: Real-world data offer tremendous opportunity to improve outcome ascertainment in clinical trials. However, little is known about the validity of outcomes ascertained using real-world data to capture trial endpoints. METHODS: Patients enrolled in 3 pivotal trials and 2 pre-market continued-access studies evaluating transcatheter aortic valve replacement were linked to Medicare fee-for-service inpatient claims. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and kappa agreement statistic of claims to detect clinical endpoints and procedural complications in trial patients were calculated. RESULTS: Claims accurately identified trial-adjudicated deaths (sensitivity, specificity, PPV, and NPV all >99.6%; kappa 1.00). Claims had good performance in identifying trial-adjudicated permanent pacemaker implantation (sensitivity 92.2%, specificity 99.1%, PPV 96.1%, NPV 98.2%, kappa 0.93) and aortic valve reintervention (sensitivity 84.4%, specificity 99.6%, PPV 69.1%, NPV 99.8%, kappa 0.76). Claims had more modest performance in ascertaining trial-adjudicated myocardial infarction (sensitivity 63.6%, specificity 97.2%, PPV 29.9%, NPV 99.3%, kappa 0.39) and acute kidney injury (sensitivity 70.2%, specificity 85.4%, PPV 38.2%, NPV 95.7%, kappa 0.41) and the poorest performance for identifying trial-adjudicated bleeding events (sensitivity 86.4%, specificity 36.8%, PPV 35.0%, NPV 86.3%, kappa 0.16). CONCLUSIONS: Compared with trial-adjudicated outcomes, claims data performed well in ascertaining death and outcomes with procedural billing codes and more modestly in identifying other outcomes. Claims may be cautiously and selectively used to augment data collection in future cardiovascular device trials.

Comparison of Clinical Trials and Administrative Claims to Identify Stroke Among Patients Undergoing Aortic Valve Replacement: Findings From the EXTEND Study.

BACKGROUND: Cerebrovascular events (CVEs) are devastating complications after aortic valve replacement. We assessed whether billing claims accurately identify CVEs in place of clinical event adjudication in structural heart disease trials. METHODS: Adult participants in the US CoreValve High Risk and SURTAVI trials (Surgical or Transcatheter Aortic-Valve Replacement in Intermediate-Risk Patients) were linked to Medicare inpatient claims from January 1, 2006 to December 31, 2016. Claims consistent with CVEs within 14 days of a similar trial-adjudicated CVE were considered a match. The sensitivity, specificity, and positive and negative predictive values of International Classification of Diseases,Ninth andTenthRevisions, Clinical Modification billing codes for cerebrovascular disease were determined against trial-defined CVEs as the criterion standard. Kaplan-Meier estimates of claims-defined versus trial-defined CVEs were compared. RESULTS: Among 4230 linked trial participants (linkage rate 79.8%), 550 (13.0%) sustained 630 adjudicated CVEs over a 5-year follow-up period. Linked and nonlinked individuals were similar. An algorithm using 4 International Classification of Diseases, Ninth Revision, Clinical Modification codes (434.91, 434.11, 433.11, and 997.02) had a sensitivity of 60.9%, specificity of 99.0%, positive predictive value of 86.5%, and negative predictive value of 95.8% for identifying a trial-adjudicated ischemic stroke. An algorithm using 3 International Classification of Diseases, Tenth Revision, Clinical Modification codes (I63.9, I63.40, I63.49) had a sensitivity of 66.7%, specificity of 99.4%, positive predictive value of 88.9%, and negative predictive value of 97.6%. CONCLUSIONS: In linked clinical trial and Medicare claims data, 4 International Classification of Diseases, Ninth Revision, Clinical Modification and 3 International Classification of Diseases, Tenth Revision, Clinical Modification billing codes identified half of trial-adjudicated CVEs during follow-up with high specificity and predictive value, but imperfect sensitivity. Although low sensitivity may limit the use of claims to substitute for traditional trial outcomes to identify CVEs, high specificity suggests claims could be used to trigger evaluation of neurological events, potentially improving the efficiency of the evaluation of techniques and devices designed to reduce such events.

Hospital Variation in the Utilization of Short-Term Nondurable Mechanical Circulatory Support in Myocardial Infarction Complicated by Cardiogenic Shock.

BACKGROUND: Limited knowledge exists on inter-hospital variation in the utilization of short-term, nondurable mechanical circulatory support (MCS) for myocardial infarction (MI) complicated by cardiogenic shock (CS). METHODS AND RESULTS: Hospitalizations for MI with CS in 2014 in a nationally representative all-payer database were included. The proportion of hospitalizations for MI with CS using MCS (MCS ratio) and in-hospital mortality were evaluated. Hospital characteristics and outcomes were compared across quartiles of MCS usage. Of 1813 hospitals evaluated, 1440 (79.4%) performed ≥10 percutaneous coronary interventions annually. Of these, 1064 (73.9%) had at least one code for MCS. Forty-one percent of hospitals did not use MCS. The median (interquartile range) proportion of MCS use among admissions for MI with CS was 33.3% (0.0%-50.0%). High MCS utilizing hospitals were larger ( P<0.001). Eighty-five percent (2808/3301) of MCS use was intra-aortic balloon pump. There was significant variation in receipt of MCS at different hospitals (median odds ratio of receiving MCS at 2 random hospitals: 1.58; 95% CI, 1.45-1.70). Adjusted in-hospital mortality was not different across quartiles of MCS use (Q4 versus Q1; odds ratio, 0.95; 95% CI, 0.77-1.16; P=0.58). CONCLUSIONS: Wide variation exists in hospital use of MCS for MI with CS, unexplained by patient characteristics. The predominant form of MCS use is intra-aortic balloon pump. Risk-adjusted mortality rates were not different between higher and lower MCS-utilizing hospitals.

National trends, predictors of use, and in-hospital outcomes in mechanical circulatory support for cardiogenic shock.

AIMS: Despite rising rates of cardiogenic shock (CS), data on trends and in-hospital outcomes of short-term non-durable mechanical circulatory support (MCS) are limited. Thus, we aimed to identify recent national trends in MCS utilisation in the USA, patient-level predictors of MCS use, and in-hospital outcomes in CS inclusive of extracorporeal membrane oxygenation (ECMO). METHODS AND RESULTS: Hospitalisations of US adults with a discharge diagnosis of CS, from January 2004 to December 2014, in the National Inpatient Sample were included. Rates of MCS were stratified by device type and clinical presentation. Outcomes included in-hospital mortality, hospitalisation costs, and number of procedures. A total of 183,516 hospitalisations with CS (47,636 [25.9%] involving MCS) were included. MCS recipients were younger, less frequently female, received more procedures, had higher costs, and more frequently presented with MI (MCS vs. non-MCS: 71.6% vs. 42.9%; p<0.0001). Growth in CS hospitalisations (214.4%) outpaced annual MCS use (160.0%), with relative declines in intra-aortic balloon pump use starting in 2008. Right heart catheterisation rates for both groups remained low (MCS vs. non-MCS: 5.9% vs. 3.3%; p<0.0001). In-hospital mortality declined but remained high in both groups (MCS vs. non-MCS [2014]: 32.7% vs. 41.5%; p<0.0001). CONCLUSIONS: In-hospital mortality for CS has declined but remains high. Rates of CS have outpaced MCS utilisation which remains uncommon in non-MI hospitalisations with shock. MCS is associated with utilisation of other procedures during hospitalisation.

A Survey of Interventional Cardiologists' Attitudes and Beliefs About Public Reporting of Percutaneous Coronary Intervention.

Importance: Public reporting of procedural outcomes has been associated with lower rates of percutaneous coronary intervention (PCI) and worse outcomes after myocardial infarction. Contemporary data are limited on the influence of public reporting on interventional cardiologists' clinical decision making. Objective: To survey a contemporary cohort of interventional cardiologists in Massachusetts and New York about how public reporting of PCI outcomes influences clinical decision making. Design, Setting, and Participants: An online survey was developed with public reporting experts and administered electronically to eligible physicians in Massachusetts and New York who were identified by Doximity (an online physician networking site) and 2014 Medicare fee-for-service claims for PCI procedures. The personal and hospital characteristics of participants were ascertained via a comprehensive database from Doximity and the American Hospital Association annual surveys of US hospitals (2012 and 2013) and linked to survey responses. Associations between survey responses and characteristics of participants were evaluated in univariable and multivariable analyses. Main Outcomes and Measures: Reported rate of avoidance of performing PCIs in high-risk patients and of perception of pressure from colleagues to avoid performing PCIs. Results: Of the 456 physicians approached, 149 (32.7%) responded, including 67 of 129 (51.9%) in Massachusetts and 82 of 327 (25.1%) in New York. The mean (SD) age was 49 (9.2) years; 141 of 149 participants (94.6%) were men. Most participants reported practicing at medium to large, nonprofit hospitals with high-volume cardiac catheterization laboratories and cardiothoracic surgery capabilities. In 2014, participants had higher annual PCI volumes among Medicare patients than nonparticipants did (median, 31; interquartile range [IQR], 13-47 vs median, 17; IQR, 0-41; P < .001). Among participants, 65% reported avoiding PCIs on at least 2 occasions becase of concern that a bad outcome would negatively impact their publicly reported outcomes; 59% reported sometimes or often being pressured by colleagues to avoid performing PCIs because of a concern about the patient's risk of death. After multivariable adjustment, more years of experience practicing interventional cardiology was associated with lower odds of PCI avoidance. The state of practice was not associated with survey responses. Conclusions and Relevance: Current PCI public reporting programs can foster risk-averse clinical practice patterns, which do not vary significantly between interventional cardiologists in New York and Massachusetts. Coordinated efforts by policy makers, health systems leadership, and the interventional cardiology community are needed to mitigate these unintended consequences.