Estimating treatment effect in randomized trial after control to treatment crossover using external controls.

In clinical trials, it is common to design a study that permits the administration of an experimental treatment to participants in the placebo or standard of care group post primary endpoint. This is often seen in the open-label extension phase of a phase III, pivotal study of the new medicine, where the focus is on assessing long-term safety and efficacy. With the availability of external controls, proper estimation and inference of long-term treatment effect during the open-label extension phase in the absence of placebo-controlled patients are now feasible. Within the framework of causal inference, we propose several difference-in-differences (DID) type methods and a synthetic control method (SCM) for the combination of randomized controlled trials and external controls. Our realistic simulation studies demonstrate the desirable performance of the proposed estimators in a variety of practical scenarios. In particular, DID methods outperform SCM and are the recommended methods of choice. An empirical application of the methods is demonstrated through a phase III clinical trial in rare disease.

Persistence and Drivers of High-Cost Status Among Dual-Eligible Medicare and Medicaid Beneficiaries: An Observational Study.

Background: Little is known about the persistence of high-cost status among dual-eligible Medicare and Medicaid beneficiaries, who account for a substantial proportion of expenditures in both programs. Objective: To determine what proportion of this population has persistently high costs. Design: Observational study. Setting: Medicare-Medicaid Linked Enrollee Analytic Data Source data for 2008 to 2010. Participants: 1 928 340 dual-eligible Medicare and Medicaid beneficiaries who were alive all 3 years. Measurements: Medicare and Medicaid payments for these beneficiaries were calculated for each year. Beneficiaries were categorized as high-cost for a given year if their spending was in the top 10% for that year. Differences in spending were then examined for those who were persistently high-cost (all 3 years) versus those who were transiently high-cost (2008 but not 2009 or 2010) and those who were non-high-cost in all 3 years. Results: In the first year, 192 835 patients were high-cost. More than half (54.8%) remained high-cost across all 3 years. These patients were younger than transiently high-cost patients, with fewer medical comorbidities and greater intellectual impairment. Persistently high-cost patients spent $161 224 per year compared with $86 333 per year for transiently high-cost patients and $22 352 per year for non-high-cost patients. Most of the spending among persistently high-cost patients (68.8%) was related to long-term care, and very little (<1%) was related to potentially preventable hospitalizations for ambulatory care-sensitive conditions. Limitation: Potential misclassification of preventable spending and lack of detailed clinical data in administrative claims. Conclusion: A substantial majority of high-cost dual-eligible beneficiaries had persistently high costs over 3 years, with most of the cost related to long-term care and very little related to potentially preventable hospitalizations. Primary Funding Source: Peterson Center on Healthcare.

Safety-net Hospitals Face More Barriers Yet Use Fewer Strategies to Reduce Readmissions.

OBJECTIVE: US hospitals that care for vulnerable populations, "safety-net hospitals" (SNHs), are more likely to incur penalties under the Hospital Readmissions Reduction Program, which penalizes hospitals with higher-than-expected readmissions. Understanding whether SNHs face unique barriers to reducing readmissions or whether they underuse readmission-prevention strategies is important. DESIGN: We surveyed leadership at 1600 US acute care hospitals, of whom 980 participated, between June 2013 and January 2014. Responses on 28 questions on readmission-related barriers and strategies were compared between SNHs and non-SNHs, adjusting for nonresponse and sampling strategy. We further compared responses between high-performing SNHs and low-performing SNHs. RESULTS: We achieved a 62% response rate. SNHs were more likely to report patient-related barriers, including lack of transportation, homelessness, and language barriers compared with non-SNHs (P-values<0.001). Despite reporting more barriers, SNHs were less likely to use e-tools to share discharge summaries (70.1% vs. 73.7%, P<0.04) or verbally communicate (31.5% vs. 39.8%, P<0.001) with outpatient providers, track readmissions by race/ethnicity (23.9% vs. 28.6%, P<0.001), or enroll patients in postdischarge programs (13.3% vs. 17.2%, P<0.001). SNHs were also less likely to use discharge coordinators, pharmacists, and postdischarge programs. When we examined the use of strategies within SNHs, we found trends to suggest that high-performing SNHs were more likely to use several readmission strategies. CONCLUSIONS: Despite reporting more barriers to reducing readmissions, SNHs were less likely to use readmission-reduction strategies. This combination of higher barriers and lower use of strategies may explain why SNHs have higher rates of readmissions and penalties under the Hospital Readmissions Reduction Program.

The dynamics of ideology drift among U.S. Supreme Court justices: A functional data analysis.

We study the U.S. Supreme Court dynamics by analyzing the temporal evolution of the underlying policy positions of the Supreme Court Justices as reflected by their actual voting data, using functional data analysis methods. The proposed fully flexible nonparametric method makes it possible to dissect the time-dynamics of policy positions at the level of individual Justices, as well as providing a comprehensive view of the ideology evolution over the history of Supreme Court since its establishment. In addition to quantifying individual Justice's policy positions, we uncover average changes over time and also the major patterns of change over time. Additionally, our approach allows for representing highly complex dynamic trajectories by a few principal components which complements other models of analyzing and predicting court behavior.

Detection of second primary lung cancers on surveillance imaging following stereotactic ablative radiotherapy for non-small cell lung cancer.

INTRODUCTION/BACKGROUND: Second primary lung cancers (SPLC) are common following non-small cell lung cancer (NSCLC) treatment. Development of SPLC following stereotactic ablative radiation therapy (SABR) may differ as compared to surgical cohorts. We report incidence of and outcomes for SPLC detected by surveillance imaging in a cohort of patients treated with SABR. MATERIALS/METHODS: Patients treated with SABR for node-negative NSCLC between February, 2007 to May, 2019 were retrospectively identified. Patient characteristics, frequency of surveillance imaging, development of SPLC, recurrence patterns, and survival were reviewed. Surveillance CT was performed Q3-4 month year 1, Q3-6 month year 2, Q6-12 month year 3-5, and Q12 month thereafter. Actuarial estimates of development of SPLC and overall survival (OS) were generated with competing risk analysis. RESULTS: We identified 134 patients treated with SABR with ≥6 months follow up. Eighteen (13.4%) developed a total of 21 SPLC at a median of 28.5 months (range 3.0-84.7 months) following SABR, 19 (90.5%) biopsy-proven. Twenty (95.2%) SPLC were detected by surveillance imaging. Three patients developed 2 metachronous SPLC. Three and 5 year SPLC estimates were 11.7% and 13.1%. Eighteen (85.7%) SPLC were treated with curative intent. Two and 3 year estimate of OS following detection of SPLC was 79.8% and 54.7%. CONCLUSION: SPLC were more common in our cohort than other published studies. Outcomes following surveillance detected SPLC are similar to those of de novo early stage lung cancers. The high frequency of SPLC in our cohort suggests further studies to refine overall surveillance in very high-risk populations are needed.

Metagenomic features of bioburden serve as outcome indicators in combat extremity wounds.

Battlefield injury management requires specialized care, and wound infection is a frequent complication. Challenges related to characterizing relevant pathogens further complicates treatment. Applying metagenomics to wounds offers a comprehensive path toward assessing microbial genomic fingerprints and could indicate prognostic variables for future decision support tools. Wound specimens from combat-injured U.S. service members, obtained during surgical debridements before delayed wound closure, were subjected to whole metagenome analysis and targeted enrichment of antimicrobial resistance genes. Results did not indicate a singular, common microbial metagenomic profile for wound failure, instead reflecting a complex microenvironment with varying bioburden diversity across outcomes. Genus-level Pseudomonas detection was associated with wound failure at all surgeries. A logistic regression model was fit to the presence and absence of antimicrobial resistance classes to assess associations with nosocomial pathogens. A. baumannii detection was associated with detection of genomic signatures for resistance to trimethoprim, aminoglycosides, bacitracin, and polymyxin. Machine learning classifiers were applied to identify wound and microbial variables associated with outcome. Feature importance rankings averaged across models indicated the variables with the largest effects on predicting wound outcome, including an increase in P. putida sequence reads. These results describe the microbial genomic determinants in combat wound bioburden and demonstrate metagenomic investigation as a comprehensive tool for providing information toward aiding treatment of combat-related injuries.

Time dynamics of COVID-19.

We apply tools from functional data analysis to model cumulative trajectories of COVID-19 cases across countries, establishing a framework for quantifying and comparing cases and deaths across countries longitudinally. It emerges that a country's trajectory during an initial first month "priming period" largely determines how the situation unfolds subsequently. We also propose a method for forecasting case counts, which takes advantage of the common, latent information in the entire sample of curves, instead of just the history of a single country. Our framework facilitates to quantify the effects of demographic covariates and social mobility on doubling rates and case fatality rates through a time-varying regression model. Decreased workplace mobility is associated with lower doubling rates with a roughly 2 week delay, and case fatality rates exhibit a positive feedback pattern.

Trends in opioid use disorder and overdose among opioid-naive individuals receiving an opioid prescription in Massachusetts from 2011 to 2014.

AIMS: To examine how the risks of incident opioid use disorder (OUD), non-fatal and fatal overdose have changed over time among opioid-naive individuals receiving an initial opioid prescription. DESIGN: Retrospective, longitudinal study using the Massachusetts Chapter 55 data set, which linked multiple administrative data sets to study the opioid epidemic. We identified the cumulative incidence of OUD, non-fatal and fatal overdose among the opioid-naive initiating opioid treatment in Massachusetts from 2011 to 2014 and estimated rates of these outcomes at 6 months and at 1, 2, 3 and 4 years to 2015. We used Cox regression to examine the association between characteristics of the initial prescription and risk of these outcomes. SETTING: Massachusetts, USA. PARTICIPANTS: Massachusetts residents aged ≥ 11 years in 2011-15 who were opioid-naive (no opioid prescriptions or evidence of OUD in the 6 months prior to the index prescription) (n = 2 154 426). The mean age was 49.1 years, 55.3% were female and 47.3% had commercial insurance. MEASUREMENTS: Opioid prescriptions were identified in the Prescription Monitoring Program (PMP) database, as were the characteristics of the initial prescription database. The outcomes of OUD and non-fatal overdose were identified from claims in the All Payer Claims Database (APCD) and hospital encounters in the acute hospital case mix files. Fatal overdoses were identified using Registry of Vital Records and Statistics (RVRS) death certificates and the Office of the Chief Medical Examiner (OCME) circumstances of death and toxicology reports. FINDINGS: Among opioid-naive individuals receiving an initial opioid prescription, the risk of incident OUD appears to have declined between 2011 and 2014, while rates of overdose were largely unchanged. For example, the 1-year OUD rate was 1.18% in 2011, 1.11% in 2012, 1.26% in 2013 and 0.94% in 2014. Longer therapy duration was associated with higher risk of OUD [hazard ratio (HR) = 2.24, 95% confidence interval (CI) = 2.19-2.29 for duration of 3 or more months], non-fatal (HR = 1.67, 95% CI = 1.53-1.82) and fatal opioid overdose (HR = 2.24, 95% CI = 1.91-2.61). Concurrent benzodiazepine treatment was also associated with higher risk of OUD (HR = 1.14, 95% CI = 1.12-1.17), non-fatal (HR = 1.20, 95% CI = 1.10-1.30) and fatal overdose (HR = 1.86, 95% CI = 1.61-2.16). CONCLUSIONS: Among opioid-naive individuals in Massachusetts receiving an initial opioid prescription, the risk of incident opioid use disorder appears to have declined between 2011 and 2014, while rates of overdose were largely unchanged. Longer therapy duration and concurrent benzodiazepines were associated with higher rates of opioid use disorder and opioid overdose.

Characteristics And Spending Patterns Of Persistently High-Cost Medicare Patients.

One strategy for reducing health care spending is to target the Medicare beneficiaries who remain persistently high cost over time. Using a 20 percent sample of Medicare fee-for-service beneficiaries in the period 2012-14, we sought to identify the proportion of patients who remained persistently high cost (that is, in the top 10 percent of spending each year) and determine the characteristics and spending patterns that differentiated them from other patients. We found that 28.1 percent of patients who were high cost in 2012 remained persistently high cost over the subsequent two years. On average, persistently high-cost patients were younger, more likely to be members of racial/ethnic minority groups, eligible for Medicare based on having end-stage renal disease, and dually eligible for Medicaid, compared to transiently and never high-cost patients. Persistently high-cost patients had greater relative spending on outpatient care and medications, while very little of their spending was related to preventable hospitalizations. Health care systems and policy makers can use this information to better target spending reductions and care improvements over time.

Hospital-level care coordination strategies associated with better patient experience.

BACKGROUND: Patient experience is a key measure of hospital quality and is increasingly contained in value-based payment programmes. Understanding whether strategies aimed at improving care transitions are associated with better patient experience could help clinical leaders and policymakers seeking to improve care across multiple dimensions. OBJECTIVE: To determine the association of specific hospital care coordination and transition strategies with patient experience. DESIGN: We surveyed leadership at 1600 acute care hospitals and categorised respondents into three groups based on the strategies used: low-strategy (bottom quartile of number of strategies), mid-strategy (quartiles 2 and 3) and high-strategy (highest quartile). We used linear regression models to examine the association between use of these strategies and performance on measures of patient experience from the Hospital Consumer Assessment of Healthcare Providers and Systems survey. RESULTS: We achieved a 62% response rate. High-strategy hospitals reported using 7.7 strategies on average usually or always on their patient populations, while mid-strategy and low-strategy hospitals reported using 5.0 and 2.3 strategies, respectively. Compared with low-strategy hospitals, high-strategy hospitals had a higher overall rating (+2.23 percentage points (pp), P<0.001), higher recommendation score (+2.5 pp, P<0.001), and higher satisfaction with discharge process (+1.35 pp, P=0.01) and medication communication (+1.44 pp, P=0.002). Mid-strategy hospitals had higher scores than low-strategy hospitals except for discharge satisfaction. Patient-facing strategies, like sharing discharge summaries with patients prior to discharge, using discharge coordinators and calling patients 48 hours after discharge, were each individually associated with a higher overall hospital rating, and higher satisfaction with discharge process and medication communication. CONCLUSIONS: Hospitals with greater reported use of care coordination and transition strategies have better patient experience than hospitals with fewer reported strategies. Strategies that most directly involve patients have the strongest association with better experience.

Characteristics and spending patterns of high cost, non-elderly adults in Massachusetts.

BACKGROUND: Given that health care costs in Massachusetts continue to grow despite great efforts to contain them, we seek to understand characteristics and spending patterns of the costliest non-elderly adults in Massachusetts based on type of insurance. METHODS: We used the Massachusetts All-Payer Claims Database (APCD) from 2012 and analyzed demographics, utilization patterns and spending patterns across payers (Medicaid, Medicaid managed care, and private insurers) for high cost patients (those in the top 10% of spending) and non-high cost patients (the remaining 90%). RESULTS: We identified 3,712,045 patients between the ages of 18-64 years in Massachusetts in 2012 who met our inclusion criteria. Of this group, 8.5% had Medicaid fee-for-service, 11.1% had Medicaid managed care, and 80.3% had private insurance. High cost patients accounted for 65% of total spending in our sample. We found that high cost patients were more likely to be older (median age 48 vs 40, p<0.001), female (60.2% vs. 51.2%, p<0.001), and have multiple chronic conditions (4.4 vs. 1.3, p<0.001) compared to non-high cost patient patients. Medicaid patients were the most likely to be designated high cost (18.1%) followed by Medicaid managed care (MCO) (13.9%) and private insurance (8.6%). High cost Medicaid patients also had the highest mean annual spending and incurred the most preventable spending compared to high cost MCO and high cost private insurance patients. CONCLUSIONS & IMPLICATIONS: We used 2012 claims data from Massachusetts to examine characteristics and spending patterns of the state's costliest patients based on type of insurance. Providers and policymakers seeking to reduce costs and increase value under delivery system reform may wish to target the Medicaid population.

Recent Growth In Medicare Advantage Enrollment Associated With Decreased Fee-For-Service Spending In Certain US Counties.

Recent increases in Medicare Advantage enrollment may have caused lower spending growth in the fee-for-service (FFS) Medicare population. We identified the counties of largest Medicare Advantage growth and determined if increased enrollment was associated with reduced FFS Medicare spending growth in those counties. We found that 73 percent of counties experienced at least a 5-percentage-point increase in Medicare Advantage penetration between 2007 and 2014, with the most growth occurring in larger and poorer counties in the Northeast and South. The association between Medicare Advantage growth and FFS Medicare costs varied depending on baseline Medicare Advantage penetration: In counties with low baseline penetration, Medicare Advantage growth did not have a significant effect on per capita FFS Medicare spending, whereas in counties in the highest quartile of baseline Medicare Advantage penetration, it was associated with a significant decrease in FFS Medicare spending growth ($154 annually per 10-percentage-point increase in Medicare Advantage). These findings suggest that Medicare Advantage growth may be playing a role in moderating FFS Medicare costs.