Veterans receiving a second course of cognitive processing therapy or prolonged exposure therapy: is it better to switch or stay the same?

Cognitive processing therapy (CPT) and prolonged exposure therapy (PE) are effective psychotherapies for post-traumatic stress disorder (PTSD). However, these treatments also have high rates of dropout and non-response. Therefore, patients may need a second course of treatment. We compared outcomes for patients who switched between CPT/PE and those who repeated CPT/PE during a second course of treatment. We collected data from Iraq and Afghanistan war veterans (n = 2,958) who received a second course of CPT/PE in the Veterans Health Administration from 2001 to 2017 and had symptom outcomes (PTSD checklist; PCL). We measured the association between treatment sequence and change in PCL score over the second course of treatment using hierarchical Bayesian regression, adjusted for sociodemographic and clinical characteristics. All treatment sequences showed a significant reduction in PCL score over time (ß = -4.80; HDI95: -5.74, -3.86). Veterans who switched from CPT to PE had modestly greater PCL reductions during the second course than those who repeated CPT. However, no significant difference in PCL change during the second course was observed between veterans who repeated PE and those who switched from PE to CPT. Veterans participating in a second course of CPT/PE can benefit, and switching treatment may be slightly more beneficial following CPT.

Patient-centered medical home implementation and primary care provider turnover.

BACKGROUND: The Veterans Health Administration (VHA) began implementing a patient-centered medical home (PCMH) model of care delivery in April 2010 through its Patient Aligned Care Team (PACT) initiative. PACT represents a substantial system reengineering of VHA primary care and its potential effect on primary care provider (PCP) turnover is an important but unexplored relationship. This study examined the association between a system-wide PCMH implementation and PCP turnover. METHODS: This was a retrospective, longitudinal study of VHA-employed PCPs spanning 29 calendar quarters before PACT and eight quarters of PACT implementation. PCP employment periods were identified from administrative data and turnover was defined by an indicator on the last quarter of each uncensored period. An interrupted time series model was used to estimate the association between PACT and turnover, adjusting for secular trend and seasonality, provider and job characteristics, and local unemployment. We calculated average marginal effects (AME), which reflected the change in turnover probability associated with PACT implementation. RESULTS: The quarterly rate of PCP turnover was 3.06% before PACT and 3.38% after initiation of PACT. In adjusted analysis, PACT was associated with a modest increase in turnover (AME=4.0 additional PCPs per 1000 PCPs per quarter, P=0.004). Models with interaction terms suggested that the PACT-related change in turnover was increasing in provider age and experience. CONCLUSIONS: PACT was associated with a modest increase in PCP turnover, concentrated among older and more experienced providers, during initial implementation. Our findings suggest that policymakers should evaluate potential workforce effects when implementing PCMH.

Development of a novel metric of timely care access to primary care services.

OBJECTIVE: To develop a model for identifying clinic performance at fulfilling next-day and walk-in requests after adjusting for patient demographics and risk. DATA SOURCE: Using Department of Veterans Affairs (VA) administrative data from 160 VA primary care clinics from 2014 to 2017. STUDY DESIGN: Using a retrospective cohort design, we applied Bayesian hierarchical regression models to predict provision of timely care, with clinic-level random intercept and slope while adjusting for patient demographics and risk status. Timely care was defined as the provision of an appointment within 48 hours of any patient requesting the clinic's next available appointment or walking in to receive care. DATA COLLECTION/EXTRACTION METHODS: We extracted 1 841 210 timely care requests from 613 263 patients. PRINCIPAL FINDINGS: Across 160 primary care clinics, requests for timely care were fulfilled 86 percent of the time (range 83 percent-88 percent). Our model of timely care fit the data well, with a Bayesian R2 of .8. Over the four years of observation, we identified 25 clinics (16 percent) that were either struggling or excelling at providing timely care. CONCLUSION: Statistical models of timely care allow for identification of clinics in need of improvement after adjusting for patient demographics and risk status. VA primary care clinics fulfilled 86 percent of timely care requests.

A Resampling Based Grid Search Method to Improve Reliability and Robustness of Mixture-Item Response Theory Models of Multimorbid High-Risk Patients.

There are many statistics available to the applied statistician for assessing model fit and even more methods for assessing internal and external validity. We detail a useful approach using a grid search technique that balances the internal model consistency with generalizability and can be used with models that naturally lend themselves to multiple assessment techniques. Our method relies on resampling and a simple grid search method over 3 commonly used statistics that are simple to calculate. We apply this method in a latent traits framework using a mixture Item Response Theory (MIXIRT) model of common chronic health conditions. Model fit is assessed using Akaike's Information Criteria (AIC), latent class similarity is measured with the Variance of Information (VI), and the consistency of condition complexity and prevalence across latent classes is compared using Kendall's τ rank order statistic. From two patient cohorts at high risk for hospitalization in 2014 and 2018, we generated 19 MIXIRT models (allowing 2-20 latent classes) on 21 common comorbid conditions identified via healthcare encounter diagnosis codes. We ran these models on 100 bootstrap samples of size 10% for each cohort. Among the resulting models, combined AIC and VI statistics identified 5-7 latent classes, but the rank order correlation of condition complexity revealed that only the 5 class solutions had consistent condition complexity. The 5 class solutions were combined to produce a single parsimonious MIXIRT solution that balanced clinical significance with model fit, cluster similarity, and consistency of condition complexity.

Association of High-Cost Health Care Utilization With Longitudinal Changes in Patient-Centered Medical Home Implementation.

Importance: In 2010, the US Veterans Health Administration (VHA) implemented one of the largest patient-centered medical home (PCMH) models in the United States, the Patient Aligned Care Team initiative. Early evaluations demonstrated promising associations with improved patient outcomes, but limited evidence exists on the longitudinal association of PCMH implementation with changes in health care utilization. Objective: To determine whether a change in PCMH implementation is associated with changes in emergency department (ED) visits, hospitalizations for ambulatory care-sensitive conditions (ACSCs), or all-cause hospitalizations. Design, Setting, and Participants: This cohort study used national patient-level data from the VHA and Centers for Medicare & Medicaid Services between October 1, 2012, and September 30, 2015. A total of 1 650 976 patients from 897 included clinics were divided into 2 cohorts: patients younger than 65 years who received primary care at VHA sites affiliated with a VHA ED and patients 65 years or older who were enrolled in both VHA and Medicare services. Exposures: Clinics were categorized on improvement or decline in PCMH implementation based on their Patient Aligned Care Team implementation progress index (Pi2) score. Main Outcomes and Measures: Change in the number of ED visits, ACSC hospitalizations, and all-cause hospitalizations among patients at each clinic site. Results: The study included a total of 1 650 976 patients, of whom 581 167 (35.20%) were younger than 65 years (mean [SD] age, 49.03 [10.28] years; 495 247 [85.22%] men) and 1 069 809 (64.80%) were 65 years or older (mean [SD] age, 74.64 [7.41] years; 1 050 110 [98.16%] men). Among patients younger than 65 years, there were fewer ED visits among patients seen at clinics that had improved PCMH implementation (110.8 fewer visits per 1000 patients; P < .001) and clinics that had somewhat worse implementation (69.0 fewer visits per 1000 patients; P < .001) compared with clinics that had no change in Pi2 score. There were no associations of change in Pi2 scores with all-cause hospitalizations or ACSC hospitalizations among patients younger than 65 years. In patients 65 years or older, those seen at clinics that had somewhat worse PCMH implementation experienced fewer ED visits (20.1 fewer visits per 1000 patients; P = .002) and all-cause hospitalizations (12.4 fewer hospitalizations per 1000 patients; P = .007) compared with clinics with no change in Pi2 score. There was no association between change in Pi2 score with ACSC hospitalizations among patients 65 years or older. Conclusions and Relevance: There were no consistent associations of change in Pi2 score with high-cost health care utilization. This finding highlights the key differences in measuring PCMH implementation longitudinally compared with cross-sectional study designs.

Developing an Algorithm for Combining Race and Ethnicity Data Sources in the Veterans Health Administration.

INTRODUCTION: Racial/ethnic disparities exist in the Veterans Health Administration (VHA), despite financial barriers to care being largely mitigated and Veterans Administration's (VA) organizational commitment to health equity. Accurately identifying minority veterans is critical to monitoring progress toward equity as the VHA treats an increasingly racially and ethnically diverse veteran population. Although the VHA's completeness of race and ethnicity data is generally better than its public sector and private counterparts, the accuracy of the race and ethnicity in the various databases available to VHA is variable, as is the accuracy in identifying specific minority groups. The purpose of this article was to develop an algorithm for constructing race and ethnicity variables from data sources available to VHA researchers, to present demographic differences cross the data sources, and to apply the algorithm to one study year. MATERIALS AND METHODS: We used existing VHA survey data from the Survey of Healthcare Experiences of Patients (SHEP) and three commonly used administrative databases from 2003 to 2015: the VA Corporate Data Warehouse (CDW), VA Defense Identity Repository (VADIR), and Medicare. Using measures of agreement such as sensitivity, specificity, positive and negative predictive values, and Cohen kappa, we compared self-reported race and ethnicity from the SHEP and each of the other data sources. Based on these results, we propose an algorithm for combining data on race and ethnicity from these datasets. We included VHA patients who completed a SHEP and had race/ethnicity recorded in CDW, VADIR, and/or Medicare. RESULTS: Agreement between SHEP and other sources was high for Whites and Blacks and substantially lower for other minority groups. The CDW demonstrated better agreement than VADIR or Medicare. CONCLUSIONS: We developed an algorithm of data source precedence in the VHA that improves the accuracy of the identification of historically under-identified minorities: (1) SHEP, (2) CDW, (3) Department of Defense's VADIR, and (4) Medicare.

The Association Between Neighborhood Socioeconomic and Housing Characteristics with Hospitalization: Results of a National Study of Veterans.

BACKGROUND: Social determinants of health (SDOH) have an inextricable impact on health. If remained unaddressed, poor SDOH can contribute to increased health care utilization and costs. We aimed to determine if geographically derived neighborhood level SDOH had an impact on hospitalization rates of patients receiving care at the Veterans Health Administration's (VHA) primary care clinics. METHODS: In a 1-year observational cohort of veterans enrolled in VHA's primary care medical home program during 2015, we abstracted data on individual veterans (age, sex, race, Gagne comorbidity score) from the VHA Corporate Data Warehouse and linked those data to data on neighborhood socioeconomic status (NSES) and housing characteristics from the US Census Bureau on census tract level. We used generalized estimating equation modeling and spatial-based analysis to assess the potential impact of patient-level demographic and clinical factors, NSES, and local housing stock (ie, housing instability, home vacancy rate, percentage of houses with no plumbing, and percentage of houses with no heating) on hospitalization. We defined hospitalization as an overnight stay in a VHA hospital only and reported the risk of hospitalization for veterans enrolled in the VHA's primary care medical home clinics, both across the nation and within 1 specific case study region of the country: King County, WA. RESULTS: Nationally, 6.63% of our veteran population was hospitalized within the VHA system. After accounting for patient-level characteristics, veterans residing in census tracts with a higher NSES index had decreased odds of hospitalization. After controlling all other factors, veterans residing in census tracts with higher percentage of houses without heating had 9% (Odds Ratio, 1.09%; 95% CI, 1.04 to 1.14) increase in the likelihood of hospitalization in our regional Washington State analysis, though not our national level analyses. CONCLUSIONS: Our results present the impact of neighborhood characteristics such as NSES and lack of proper heating system on the likelihood of hospitalization. The application of placed-based data at the geographic level is a powerful tool for identification of patients at high risk of health care utilization.