De-implementation of Axillary Dissection in Women With Breast Cancer is Largely Driven By Site-level Contextual Effects.

OBJECTIVE: To assess the contribution of unknown institutional factors (contextual effects) in the de-implementation of cALND in women with breast cancer. SUMMARY OF BACKGROUND DATA: Women included in the National Cancer Database with invasive breast carcinoma from 2012 to 2016 that underwent upfront lumpectomy and were found to have a positive sentinel node. METHODS: A multivariable mixed effects logistic regression model with a random intercept for site was used to determine the effect of patient, tumor, and institutional variables on the risk of cALND. Reference effect measureswere used to describe and compare the contribution of contextual effects to the variation in cALND use to that of measured variables. RESULTS: By 2016, cALND was still performed in at least 50% of the patients in a quarter of the institutions. Black race, younger women and those with larger or hormone negative tumors were more likely to undergo cALND. However, the width of the 90% reference effect measures range for the contextual effects exceeded that of the measured site, tumor, time, and patient demographics, suggesting institutional contextual effects were the major drivers of cALND de-implementation. For instance, a woman at an institution with low-risk of performing cALND would have 74% reduced odds of havinga cALND than if she was treated at a median-risk institution, while a patient at a high-risk institution had 3.91 times the odds. CONCLUSION: Compared to known patient, tumor, and institutional factors, contextual effects had a higher contribution to the variation in cALND use.

The Influence of the Rural Transitions Nurse Program for Veterans on Healthcare Utilization Costs.

BACKGROUND: The Veterans Affairs (VA) Healthcare System Rural Transitions Nurse Program (TNP) addresses barriers veterans face when transitioning from urban tertiary VA hospitals to home. Previous clinical evaluations of TNP have shown that enrolled veterans were more likely to follow up with their primary care provider within 14 days of discharge and experience a significant reduction in mortality within 30 days compared to propensity-score matched controls. OBJECTIVE: Examine changes from pre- to post-hospitalization in total, inpatient, and outpatient 30-day healthcare utilization costs for TNP enrollees compared to controls. DESIGN: Quantitative analyses modeling the changes in cost via multivariable linear mixed-effects models to determine the association between TNP enrollment and changes in these costs. PARTICIPANTS: Veterans meeting TNP eligibility criteria who were discharged home following an inpatient hospitalization at one of the 11 implementation sites from April 2017 to September 2019. INTERVENTION: The four-step TNP transitional care intervention. MAIN MEASURES: Changes in 30-day total, inpatient, and outpatient healthcare utilization costs were calculated for TNP enrollees and controls. KEY RESULTS: Among 3001 TNP enrollees and 6002 controls, no statistically significant difference in the change in total costs (p = 0.65, 95% CI: (- $675, $350)) was identified. However, on average, the increase in inpatient costs from pre- to post-hospitalization was approximately $549 less for TNP enrollees (p = 0.02, 95% CI: (- $856, - $246)). The average increase in outpatient costs from pre- to post-hospitalization was approximately $421 more for TNP enrollees compared to controls (p = 0.003, 95% CI: ($109, $671)). CONCLUSIONS: Although we found no difference in change in total costs between veterans enrolled in TNP and controls, TNP was associated with a smaller increase in direct inpatient medical costs and a larger increase in direct outpatient medical costs. This suggests a shifting of costs from the inpatient to outpatient setting.

Iso-osmolar versus low-osmolar contrast media and outcomes after percutaneous coronary intervention: Insights from the VA CART Program.

OBJECTIVES: To assess whether contrast media type is associated with outcomes in veterans undergoing percutaneous coronary intervention (PCI). BACKGROUND: There is uncertainty about the impact of iso-osmolar contrast medium (IOCM) versus low-osmolar contrast medium (LOCM) on acute kidney injury (AKI) and other major adverse renal or cardiovascular events (MARCE) after PCI. We assessed the association between contrast media type and MARCE in patients who underwent PCI within the Veterans Administration Healthcare System. METHODS: We reviewed PCIs performed between 2009 and 2019 using data from the Veterans Affairs Clinical Assessment, Reporting, and Tracking Program. The primary endpoint was MARCE, a composite of myocardial infarction, stroke, all-cause death, AKI, and dialysis onset at 30 days. RESULTS: The analysis cohort consisted of 50,389 patients of whom 25,555 received LOCM and 24,834 received IOCM. There was significant variation in contrast type across sites. After adjustment for comorbidities, no significant association between contrast media type and MARCE was observed in both site-unadjusted (odds ratio [OR] for IOCM: 0.99; 95% confidence interval [CI]: 0.92-1.08; p = 0.97) and site-adjusted (OR: 1.06; 95% CI: 0.95-1.18; p = 0.30) analyses. Similar results were obtained when contrast volume was imputed or the data was subset to individuals with available contrast volume. CONCLUSION: In a large cohort of veterans undergoing PCI, we found considerable site variation in the type of contrast media used but no significant association between contrast media type and the incidence of MARCE, both before and after adjustment for the site.

Radial versus femoral access for left main percutaneous coronary intervention: An analysis from the Veterans Affairs Clinical, Reporting, and Tracking Program.

OBJECTIVES: We aimed to compare clinical characteristics and procedural outcomes of left main percutaneous interventions (LM-PCI) by transradial (TRA) versus transfemoral (TFA) approach in the VA healthcare system. BACKGROUND: TRA for percutaneous coronary intervention (PCI) is steadily increasing. However, the frequency and efficacy of TRA for LM-PCI remain less studied. METHODS: All LM-PCIs performed in the VA healthcare system were identified for fiscal year 2008 through 2018. Patients' baseline characteristics and procedure-related variables were compared by access site. Both short- and long-term clinical outcomes were analyzed using propensity score matching. RESULTS: A total of 4004 LM-PCI were performed in the VA via either radial or femoral access from 2008 to 2018. Among these, 596 (14.9%) LM PCIs were performed via TRA. Use of TRA for LM-PCI increased from 2.2% to 31.5% over the study period. Propensity matched outcome analysis, comparing TRA versus TFA, showed a similar procedural success (98.4% for TRA vs. 97.8% for TFA; RR: 1.01 [0.98, 1.03]) and 1-year major adverse cardiovascular events (MACE) (25.9% for TRA vs. 26.8% TFA; RR: 0.96 [0.74, 1.25]). There were no statistically significant differences among secondary outcomes analyses including major bleeding. CONCLUSION: Use of TRA for LM-PCI has been steadily increasing in the VA healthcare system. These findings demonstrate similar procedural success and 1-year MACE across access strategies, suggesting an opportunity to continue increasing TRA use for LM-PCI.

Noninvasive Ventilation Use in Critically Ill Patients with Acute Asthma Exacerbations.

Rationale: Noninvasive ventilation decreases the need for invasive mechanical ventilation and mortality among patients with chronic obstructive pulmonary disease but has not been well studied in asthma.Objectives: To assess the association between noninvasive ventilation and subsequent need for invasive mechanical ventilation and in-hospital mortality among patients admitted with asthma exacerbation to the ICU.Methods: We performed a retrospective cohort study using administrative data collected during 2010-2017 from 682 hospitals in the United States. Outcomes included receipt of invasive mechanical ventilation and in-hospital mortality. Generalized estimating equations, propensity-matched models, and marginal structural models were used to assess the association between noninvasive ventilation and outcomes.Measurements and Main Results: The study population included 53,654 participants with asthma exacerbation. During the study period, 13,540 patients received noninvasive ventilation (25.2%; 95% confidence interval [CI], 24.9-25.6%), 14,498 underwent invasive mechanical ventilation (27.0%; 95% CI, 26.7-27.4%), and 1,291 died (2.4%; 95% CI, 2.3-2.5%). Among those receiving noninvasive ventilation, 3,013 patients (22.3%; 95% CI, 21.6-23.0%) required invasive mechanical ventilation after first receiving noninvasive ventilation, 136 of whom died (4.5%; 95% CI, 3.8-5.3%). Across all models, the use of noninvasive ventilation was associated with a lower odds of receiving invasive mechanical ventilation (adjusted generalized estimating equation odds ratio, 0.36; 95% CI, 0.32-0.40) and in-hospital mortality (odds ratio, 0.48; 95% CI 0.40-0.58). Those who received noninvasive ventilation before invasive mechanical ventilation were more likely to have comorbid pneumonia and severe sepsis.Conclusions: Noninvasive ventilation use during asthma exacerbation was associated with improved outcomes but should be used cautiously with acute comorbid conditions.

Effect of environmental conditions on injury rates at a Colorado ski resort.

OBJECTIVE: To determine the effect of daily environmental conditions on skiing and snowboarding-related injury rates. METHODS: Injury information was collected from a mountainside clinic at a large Colorado ski resort for the 2012/2013 through 2016/2017 seasons. Daily environmental conditions including snowfall, snow base depth, temperature, open terrain and participant visits were obtained from historical resort records. Snowpack and visibility information were obtained for the 2013/2014 through 2014/2015 seasons and included in a subanalysis. Negative binomial regression was used to estimate injury rate ratios (IRRs) and 95% CIs. RESULTS: The overall injury rate among skiers and snowboarders was 1.37 per 1000 participant visits during 2012/2013 through 2016/2017. After adjustment for other environmental covariates, injury rates were 22% higher (IRR=1.22, 95% CI 1.14 to 1.29) on days with <2.5 compared with ≥2.5 cm of snowfall, and 14% higher on days with average temperature in the highest quartile (≥-3.1°C) compared with the lowest (<-10.6°C; IRR=1.14, 95% CI 1.03 to 1.26). Rates decreased by 8% for every 25 cm increase in snow base depth (IRR=0.92, 95% CI 0.88 to 0.95). In a subanalysis of the 2013/2014 and 2014/2015 seasons including the same covariates plus snowpack and visibility, only snowpack remained significantly associated with injury rates. Rates were 71% higher on hardpack compared with powder days (IRR=1.71, 95% CI 1.18 to 2.49) and 36% higher on packed powder compared with powder days (IRR 1.36, 95% CI 1.12 to 1.64). CONCLUSIONS: Environmental conditions, particularly snowfall and snowpack, have a significant impact on injury rates. Injury prevention efforts should consider environmental factors to decrease injury rates in skiers and snowboarders.

Bayesian profiling for cost with zeros to decompose total cost into probability of cost and mean nonzero cost.

Cost of health care can vary substantially across hospitals, centers, or providers. Data from electronic health records provide information for studying patterns of cost variation and identifying high or low cost centers. Cost data often include zero values when patients receive no care, and joint two-part models have been developed for clustered cost data with zeros. Standard methods for center comparisons, sometimes called profiling, can use these methods to incorporate zero values into total cost. However, zero costs also provide opportunities to further examine sources of cost variation and outliers. For example, a hospital may have high (or low) cost due to frequency of nonzero cost, amount of nonzero cost, or a combination of those. We give methods for decomposing hospital differences in total cost with zeros into components for probability of use (i.e., of nonzero cost) and for cost of use (mean of nonzero cost). The components multiply to total cost and quantify components on the same easily interpreted multiplicative scales. The methods are based on Bayesian hierarchical models and counterfactual arguments, with Markov chain Monte Carlo estimation. We used simulated data to illustrate use, interpretation, and visualization of the methods in diverse situations, and applied the methods to 30,024 patients at 57 US Veterans Administration hospitals to characterize outlier hospitals in one year cost of inpatient care following a cardiac procedure. Twenty eight percent of patients had zero cost. These methods are useful in providing insight into cost variation and outliers for planning future studies or interventions.

A comparison of recreational skiing- and snowboarding-related injuries at a Colorado ski resort, 2012/13-2016/17.

The purpose of this study was to compare injury patterns between recreational skiers and snowboarders. Injured skiers (n = 3,961) and snowboarders (n = 2,428) presented to a mountainside medical clinic, 2012/13-2016/17. Variables investigated for analysis included demographics/characteristics, injury event information, and injury information. Skiers were older than snowboarders (34.3 ± 19.3 vs. 23.2 ± 10.5 years, p < 0.001); a greater proportion of skiers were female (46.3% vs. 27.8%, p < 0.001). Most skiers (84.4%) and snowboarders (84.5%) were helmeted at the time of injury (p = 0.93). Snowboarders were most frequently beginners (38.9%), skiers were intermediates (37.8%). Falls to snow (skiers = 72.3%, snowboarders = 84.8%) and collisions with natural objects (skiers = 9.7%, snowboarders = 7.4%) were common injury mechanisms. Common skiing injuries were knee sprains (20.5%) and head trauma (8.9%); common snowboarding injuries were wrist fractures (25.7%), shoulder separations (9.1%), and head trauma (9.0%). Given that injury patterns significantly differ between sports, it is important for clinicians, ski patrollers, and resorts to develop and deliver sport-specific injury prevention interventions to most effectively decrease injury burden.

Two-part models for cost with zeros to decompose effects of covariates on probability of cost, mean nonzero cost, and mean total cost.

Health care cost data often contain many zero values, for patients who did not use any care. Two-part models with logistic models for part I, probability of use (ie, nonzero cost) and log-link models for part II, mean cost of use (ie, nonzero cost) are often used. Effects of exposures or covariates on total (marginal) cost are often of interest, and recent work has proposed useful methods. Factors that affect total cost do so through a combination of effects on probability of use and cost of use. Such a decomposition is needed to understand and act on factors that affect total cost, but little work has been done on this question. This paper presents two new methods for decomposing effects on total cost, namely, an adjusted approach based on log-binomial models for part I and a population average approach based on counterfactual arguments. Extensive simulations illustrate performance, interpretation, and robustness over a wide range of data features. The method is applied to risk-adjusted 30-day outpatient cardiac cost for patients following percutaneous coronary intervention from the Department of Veterans Affairs Clinical Assessment, Reporting, and Tracking Program. Results illustrate the simple decomposition of effects on total cost into components for probability of use and cost of use.

Comparison of a time-varying covariate model and a joint model of time-to-event outcomes in the presence of measurement error and interval censoring: application to kidney transplantation.

BACKGROUND: Tacrolimus (TAC) is an immunosuppressant drug given to kidney transplant recipients post-transplant to prevent antibody formation and kidney rejection. The optimal therapeutic dose for TAC is poorly defined and therapy requires frequent monitoring of drug trough levels. Analyzing the association between TAC levels over time and the development of potentially harmful de novo donor specific antibodies (dnDSA) is complex because TAC levels are subject to measurement error and dnDSA is assessed at discrete times, so it is an interval censored time-to-event outcome. METHODS: Using data from the University of Colorado Transplant Center, we investigated the association between TAC and dnDSA using a shared random effects (intercept and slope) model with longitudinal and interval censored survival sub-models (JM) and compared it with the more traditional interval censored survival model with a time-varying covariate (TVC). We carried out simulations to compare bias, level and power for the association parameter in the TVC and JM under varying conditions of measurement error and interval censoring. In addition, using Markov Chain Monte Carlo (MCMC) methods allowed us to calculate clinically relevant quantities along with credible intervals (CrI). RESULTS: The shared random effects model was a better fit and showed both the average TAC and the slope of TAC were associated with risk of dnDSA. The simulation studies demonstrated that, in the presence of heavy interval censoring and high measurement error, the TVC survival model underestimates the association between the survival and longitudinal measurement and has inflated type I error and considerably less power to detect associations. CONCLUSIONS: To avoid underestimating associations, shared random effects models should be used in analyses of data with interval censoring and measurement error.

Modeling associations between latent event processes governing time series of pulsing hormones.

This work is motivated by a desire to quantify relationships between two time series of pulsing hormone concentrations. The locations of pulses are not directly observed and may be considered latent event processes. The latent event processes of pulsing hormones are often associated. It is this joint relationship we model. Current approaches to jointly modeling pulsing hormone data generally assume that a pulse in one hormone is coupled with a pulse in another hormone (one-to-one association). However, pulse coupling is often imperfect. Existing joint models are not flexible enough for imperfect systems. In this article, we develop a more flexible class of pulse association models that incorporate parameters quantifying imperfect pulse associations. We propose a novel use of the Cox process model as a model of how pulse events co-occur in time. We embed the Cox process model into a hormone concentration model. Hormone concentration is the observed data. Spatial birth and death Markov chain Monte Carlo is used for estimation. Simulations show the joint model works well for quantifying both perfect and imperfect associations and offers estimation improvements over single hormone analyses. We apply this model to luteinizing hormone (LH) and follicle stimulating hormone (FSH), two reproductive hormones. Use of our joint model results in an ability to investigate novel hypotheses regarding associations between LH and FSH secretion in obese and non-obese women.

Reference effect measures for quantifying, comparing and visualizing variation from random and fixed effects in non-normal multilevel models, with applications to site variation in medical procedure use and outcomes.

BACKGROUND: Multilevel models for non-normal outcomes are widely used in medical and health sciences research. While methods for interpreting fixed effects are well-developed, methods to quantify and interpret random cluster variation and compare it with other sources of variation are less established. Random cluster variation, sometimes referred to as general contextual effects (GCE), may be the main focus of a study; therefore, easily interpretable methods are needed to quantify GCE. We propose a Reference Effect Measure (REM) approach to 1) quantify GCE and compare it to individual subject and cluster covariate effects, and 2) quantify relative magnitudes of GCE and variation from sets of measured factors. METHODS: To illustrate REM, we consider a two-level mixed logistic model with patients clustered within hospitals and a random intercept for hospitals. We compare patients at hospitals at given percentiles of the estimated random effect distribution to patients at a median or 'reference' hospital. These estimates are then compared numerically and graphically to individual fixed effects to quantify GCE in the context of effects of other measured variables (aim 1). We then extend this approach by comparing variation from the random effect distribution to variation from sets of fixed effects to understand their magnitudes relative to overall outcome variation (aim 2). RESULTS: Using an example of initiation of rhythm control treatment in atrial fibrillation (AF) patients within the Veterans Affairs (VA), we use REM to demonstrate that random variation across hospitals (GCE) in initiation of treatment is substantially greater than that due to most individual patient factors, and explains at least as much variation in treatment initiation as do all patient factors combined. These results are contrasted with a relatively small GCE compared with patient factors in 1 year mortality following hospitalization for AF patients. CONCLUSIONS: REM provides a means of quantifying random effect variation (GCE) with multilevel data and can be used to explore drivers of outcome variation. This method is easily interpretable and can be presented visually. REM offers a simple, interpretable approach for evaluating questions of growing importance in the study of health care systems.

Association of Borderline Pulmonary Hypertension With Mortality and Hospitalization in a Large Patient Cohort: Insights From the Veterans Affairs Clinical Assessment, Reporting, and Tracking Program.

BACKGROUND: Pulmonary hypertension (PH) is associated with increased morbidity across the cardiopulmonary disease spectrum. Based primarily on expert consensus opinion, PH is defined by a mean pulmonary artery pressure (mPAP) ≥25 mm Hg. Although mPAP levels below this threshold are common among populations at risk for PH, the relevance of mPAP <25 mm Hg to clinical outcome is unknown. METHODS AND RESULTS: We analyzed retrospectively all US veterans undergoing right heart catheterization (2007-2012) in the Veterans Affairs healthcare system (n=21,727; 908-day median follow-up). Cox proportional hazards models were used to evaluate the association between mPAP and outcomes of all-cause mortality and hospitalization, adjusted for clinical covariates. When treating mPAP as a continuous variable, the mortality hazard increased beginning at 19 mm Hg (hazard ratio [HR]=1.183; 95% confidence interval [CI], 1.004-1.393) relative to 10 mm Hg. Therefore, patients were stratified into 3 groups: (1) referent (≤18 mm Hg; n=4,207); (2) borderline PH (19-24 mm Hg; n=5,030); and (3) PH (≥25 mm Hg; n=12,490). The adjusted mortality hazard was increased for borderline PH (HR=1.23; 95% CI, 1.12-1.36; P<0.0001) and PH (HR=2.16; 95% CI, 1.96-2.38; P<0.0001) compared with the referent group. The adjusted hazard for hospitalization was also increased in borderline PH (HR=1.07; 95% CI, 1.01-1.12; P=0.0149) and PH (HR=1.15; 95% CI, 1.09-1.22; P<0.0001). The borderline PH cohort remained at increased risk for mortality after excluding the following high-risk subgroups: (1) patients with pulmonary artery wedge pressure >15 mm Hg; (2) pulmonary vascular resistance ≥3.0 Wood units; or (3) inpatient status at the time of right heart catheterization. CONCLUSIONS: These data illustrate a continuum of risk according to mPAP level and that borderline PH is associated with increased mortality and hospitalization. Future investigations are needed to test the generalizability of our findings to other populations and study the effect of treatment on outcome in borderline PH.

Using Cox cluster processes to model latent pulse location patterns in hormone concentration data.

Many hormones, including stress hormones, are intermittently secreted as pulses. The pulsatile location process, describing times when pulses occur, is a regulator of the entire stress system. Characterizing the pulse location process is particularly difficult because the pulse locations are latent; only hormone concentration at sampled times is observed. In addition, for stress hormones the process may change both over the day and relative to common external stimuli. This potentially results in clustering in pulse locations across subjects. Current approaches to characterizing the pulse location process do not capture subject-to-subject clustering in locations. Here we show how a Bayesian Cox cluster process may be adapted as a model of the pulse location process. We show that this novel model of pulse locations is capable of detecting circadian rhythms in pulse locations, clustering of pulse locations between subjects, and identifying exogenous controllers of pulse events. We integrate our pulse location process into a model of hormone concentration, the observed data. A spatial birth-and-death Markov chain Monte Carlo algorithm is used for estimation. We exhibit the strengths of this model on simulated data and adrenocorticotropic and cortisol data collected to study the stress axis in depressed and non-depressed women.

Factors associated with rhythm control treatment decisions in patients with atrial fibrillation-Insights from the NCDR PINNACLE registry.

BACKGROUND: Decisions to use rhythm control in atrial fibrillation (AF) should generally be dictated by patient factors, such as quality of life, heart failure, and other comorbidities. Whether or not other factors affect decisions about the use of rhythm control, and catheter ablation in particular, is unknown. METHODS: A cohort of all patients diagnosed with nonvalvular AF were identified from the National Cardiovascular Data Registry's Practice Innovation and Clinical Excellence (PINNACLE) AF registry of US outpatient cardiology practices during the study period from May 1, 2008, to December 31, 2014. Overall and practice-specific rates of rhythm control (cardioversion, antiarrhythmic drug therapy, or catheter ablation) were assessed. We assessed patient and practice factors associated with rhythm control and determined the relative contribution of patient, practice, and unmeasured practice factors with its use. RESULTS: Among 511,958 PINNACLE AF patients, 22.3% were treated with rhythm control and 2.9% underwent catheter ablation. Significant practice variation in rhythm control was present (median rate of rhythm control across practices 22.8%, range 0.2%-62.9%). Significant patient factors associated with rhythm control therapy included white (vs nonwhite) race (odds ratio [OR] 2.43, P<.001), private (vs nonprivate) insurance (OR 1.04, P<.001), and whether a patient was seen by an electrophysiologist (OR 1.77, P<.001). In an analysis of the relative contribution of patient, practice, and unmeasured practice factors with rhythm control, the contribution of unmeasured practice factors (95% range OR 0.29-3.44) exceeded that of either patient (95% range OR 0.46-2.30) or practice (95% range OR 0.15-2.77) factors. CONCLUSIONS: One in 5 AF patients in the PINNACLE registry received rhythm control, and 1 in 50 received catheter ablation, suggesting that rhythm control may be underused. A variety of measured and unmeasured practice factors unrelated to patient characteristics play a disproportionate role in the use of rhythm control treatment decisions. Understanding the drivers of these decisions may identify inappropriate treatment variation and better inform optimal use of these therapies.

Facility-Level Variation in Hospitalization, Mortality, and Costs in the 30 Days After Percutaneous Coronary Intervention: Insights on Short-Term Healthcare Value From the Veterans Affairs Clinical Assessment, Reporting, and Tracking System (VA CART) Program.

BACKGROUND: Policies to reduce unnecessary hospitalizations after percutaneous coronary intervention (PCI) are intended to improve healthcare value by reducing costs while maintaining patient outcomes. Whether facility-level hospitalization rates after PCI are associated with cost of care is unknown. METHODS AND RESULTS: We studied 32,080 patients who received PCI at any 1 of 62 Veterans Affairs hospitals from 2008 to 2011. We identified facility outliers for 30-day risk-standardized hospitalization, mortality, and cost. Compared with the risk-standardized average, 2 hospitals (3.2%) had a lower-than-expected hospitalization rate, and 2 hospitals (3.2%) had a higher-than-expected hospitalization rate. We observed no statistically significant variation in facility-level risk-standardized mortality. The facility-level unadjusted median per patient 30-day total cost was $23,820 (interquartile range, $19,604-$29,958). Compared with the risk-standardized average, 17 hospitals (27.4%) had lower-than-expected costs, and 14 hospitals (22.6%) had higher-than-expected costs. At the facility level, the index PCI accounted for 83.1% of the total cost (range, 60.3%-92.2%), whereas hospitalization after PCI accounted for only 5.8% (range, 2.0%-12.7%) of the 30-day total cost. Facilities with higher hospitalization rates were not more expensive (Spearman ρ=0.16; 95% confidence interval, -0.09 to 0.39; P=0.21). CONCLUSIONS: In this national study, hospitalizations in the 30 day after PCI accounted for only 5.8% of 30-day cost, and facility-level cost was not correlated with hospitalization rates. This challenges the focus on reducing hospitalizations after PCI as an effective means of improving healthcare value. Opportunities remain to improve PCI value by reducing the variation in total cost of PCI without compromising patient outcomes.

Mortality after presentation with stent thrombosis is associated with time from index percutaneous coronary intervention: a report from the VA CART program.

BACKGROUND: The risk of mortality for patients presenting to the cardiac catheterization laboratory with stent thrombosis (ST) may differ as a function of the timing from initial stent implantation. We hypothesized that the 30-day mortality would differ for angiographically defined early ST (EST), late ST (LST), and very late ST (VLST). METHODS: All patients undergoing angiography for diagnosis and treatment of ST were identified by the Department of Veterans Affairs (VA) Clinical Assessment, Reporting, and Tracking (CART) Program from 2006 to 2012. Stent thrombosis occurring ≤30 days after stent implantation were defined as EST; 31 to 365 days as LST; and >365 days as VLST. Log-rank test and Cox proportional hazard regression modeling were used to describe unadjusted and adjusted differences in mortality between groups. RESULTS: A total of 656 patients were diagnosed with angiographic definite ST with known timing. This cohort consisted of 129 (20%), 138 (21%), and 389 (59%) patients with EST, LST, and VLST, respectively. Over three fourths (76%) of VLST cases occurred >2 years after stent implantation. Stent thrombosis timing was significantly associated with 30-day mortality risk in unadjusted (P < .001) and adjusted (P = .04) analyses. Unadjusted mortality risk was 3% in VLST, 6% in LST, and 13% in EST. After multivariable adjustment, patients with LST had nonsignificantly lower 30-day mortality (hazard ratio [HR] for LST, 0.5; 95% CI, 0.2-1.2), whereas VLST had significantly lower 30-day mortality (HR for VLST, 0.38; 95% CI, 0.18-0.82) when compared with patients with EST. CONCLUSIONS: Thirty-day mortality after an angiographic definite ST presentation is associated with time from index percutaneous coronary intervention to ST. This relationship potentially reflects the differing mechanisms of ST that are postulated to predominate at different timeframes.

Clopidogrel prescription filling delays and cardiovascular outcomes in a pharmacy system integrating inpatient and outpatient care: insights from the Veterans Affairs CART Program.

BACKGROUND: Delays in filling clopidogrel prescriptions after percutaneous coronary intervention (PCI) have been demonstrated previously and associated with adverse outcomes. METHODS: This was a retrospective cohort study of 11,418 patients undergoing PCI with stent placement in Veterans Affairs (VA) hospitals between January 1, 2005, and September 30, 2010. Data were obtained from the national VA Clinical Assessment, Reporting, and Tracking Program, including post-PCI clopidogrel prescription fill date and outcomes of myocardial infarction and death within 90 days of discharge. Patients who did not fill a clopidogrel prescription on the day of discharge were considered to have a delay. Multivariable models assessed the association between clopidogrel delay and myocardial infarction/death using clopidogrel delay as a time-varying covariate. RESULTS: Of the patients, 7.2% had a delay in filling their clopidogrel prescription. Delay in filling clopidogrel was associated with increased risk of major adverse events (hazard ratio 2.34, 95% CI 1.66-3.29, P < .001). The percentage of patients who delayed filling varied by hospital, ranging from 0 to 43.5% with a median of 6.2% (P < .001, χ(2) for difference across hospitals) and a median odds ratio of 2.13 (95% CI 1.85-2.68) suggesting large site variation in clopidogrel delay across hospitals. CONCLUSIONS: In a health care system with integrated inpatient and outpatient pharmacy services, 1 in 14 patients delays filling a clopidogrel prescription. The large site variation suggests a need to identify best practices that allow hospitals to optimize prescription filling at discharge to potentially improve patient outcomes.

Adherence to dabigatran therapy and longitudinal patient outcomes: insights from the veterans health administration.

BACKGROUND: Dabigatran is a novel oral anti-coagulant (NOAC) that reduces risk of stroke in patients with non-valvular atrial fibrillation (NVAF). It does not require routine monitoring with laboratory testing which may have an adverse impact on adherence. We aimed to describe adherence to dabigatran in the first year after initiation and assess the association between non-adherence to dabigatran and clinical outcomes in a large integrated healthcare system. METHODS: We studied a national cohort of 5,376 patients with NVAF, initiated on dabigatran between October-2010 and September-2012 at all Veterans Affairs hospitals. Adherence to dabigatran was calculated as proportion of days covered (PDC) and association between PDC and outcomes was assessed using standard regression techniques. RESULTS: Mean age of the study cohort was 71.3 ± 9.7 years; 98.3% were men and mean CHADS2 score was 2.4 ± 1.2 (mean CHA2DS2VASc score 3.2 ± 1.4). Median PDC was 94% (IQR 76%-100%; mean PDC 84% ± 22%) over a median follow-up of 244 days (IQR 140-351). A total of 1,494 (27.8%) patients had a PDC <80% and were classified as non-adherent. After multivariable adjustment, lower adherence was associated with increased risk for combined all-cause mortality and stroke (HR 1.13, 95% CI 1.07-1.19 per 10% decrease in PDC). Adherence to dabigatran was not associated with non-fatal bleeding or myocardial infarction. CONCLUSIONS: In the year after initiation, adherence to dabigatran for a majority of patients is very good. However, 28% of patients in our cohort had poor adherence. Furthermore, lower adherence to dabigatran was associated with increased adverse outcomes. Concerted efforts are needed to optimize adherence to NOACs.

Regression calibration for models with two predictor variables measured with error and their interaction, using instrumental variables and longitudinal data.

Regression calibration provides a way to obtain unbiased estimators of fixed effects in regression models when one or more predictors are measured with error. Recent development of measurement error methods has focused on models that include interaction terms between measured-with-error predictors, and separately, methods for estimation in models that account for correlated data. In this work, we derive explicit and novel forms of regression calibration estimators and associated asymptotic variances for longitudinal models that include interaction terms, when data from instrumental and unbiased surrogate variables are available but not the actual predictors of interest. The longitudinal data are fit using linear mixed models that contain random intercepts and account for serial correlation and unequally spaced observations. The motivating application involves a longitudinal study of exposure to two pollutants (predictors) - outdoor fine particulate matter and cigarette smoke - and their association in interactive form with levels of a biomarker of inflammation, leukotriene E4 (LTE 4 , outcome) in asthmatic children. Because the exposure concentrations could not be directly observed, we used measurements from a fixed outdoor monitor and urinary cotinine concentrations as instrumental variables, and we used concentrations of fine ambient particulate matter and cigarette smoke measured with error by personal monitors as unbiased surrogate variables. We applied the derived regression calibration methods to estimate coefficients of the unobserved predictors and their interaction, allowing for direct comparison of toxicity of the different pollutants. We used simulations to verify accuracy of inferential methods based on asymptotic theory.