Implementing a Teaching Rural Mobile Health Clinic: Challenges and Adaptations.

In 2022, Penn State College of Medicine launched the LION Mobile Clinic, a teaching mobile health clinic offering preventive health services in rural Snow Shoe, Pennsylvania. We outline four challenges the clinic team faced in implementation, along with adaptations made to tailor the model to Snow Shoe's needs and opportunities.

Regional differences in fatal drug overdose deaths among Black and White individuals in the United States, 2012-2021.

BACKGROUND AND OBJECTIVES: The current study examines regional differences in Black/White fatal drug overdoses. METHODS: Black/White overdose mortality data (2012-2021; N = 537,085) were retrieved from CDC WONDER. We used death counts and corresponding Census Bureau population estimates by the decedent's age and race/ethnicity to calculate mortality rate ratios. RESULTS: From 2012 to 2021, there were 537,085 reported overdose deaths among White (85%) and Black (15%) individuals in the United States. In the South, Black individuals had lower fatal drug overdose deaths than their same-aged White counterparts. In the Northeast, Midwest, and West regions, Black individuals had around 10%-60% lower likelihood of overdoses among younger ages (15-24, 25-34, 35-44) but about 60%-300% higher likelihood of overdoses among older adults (55-64). Increases in overdose deaths during the pandemic (2020-2021) led to changes in Black/White overdose death patterns, whereas Black individuals of all ages in the Midwest and West regions had approximately 15%-425% higher likelihood of fatal overdoses than their same-aged White counterparts. Sex-stratified analysis suggests that Black females in the South had overdose death rates around 50% lower than same-aged White females, and overdose deaths were relatively equal for Black and White females in the Midwest-patterns not seen among males. CONCLUSIONS AND SCIENTIFIC SIGNIFICANCE: The findings indicate that the Black/White overdose mortality gap changed after the COVID-19 pandemic across all regions and age cohorts, with state and regional variations in magnitude. Behavioral interventions and policies to curb drug overdose deaths among populations most impacted should consider regional, sex, and age-related differences.

Relationship Between Availability of Urologists and Primary Care Providers and Stage of Diagnosis for Invasive Urinary Bladder Cancer.

INTRODUCTION: With no recommended screening approach, urinary bladder cancer patients rely on referral to urologists to ensure timely diagnosis of bladder cancer. This requires coordination between primary and specialty care. We provide estimates of the relative association between primary care physician and urologist density on stage of urinary bladder cancer diagnosis. METHODS: We used 2010 to 2016 Pennsylvania Cancer Registry data to identify all adult patients diagnosed with bladder cancer. Our primary outcome was locoregional stage of diagnosis, since treatment modality changes and prognosis worsens beyond this stage. Based on patient's residential location at the time of diagnosis we defined both density of urologists and number of primary care providers (defined as providers per population) within the patient's county. We used univariate and multivariate logistic regression to estimate the association between provider density and likelihood of locoregional stage of diagnosis. We also controlled for age, sex, race/ethnicity, insurance type, and year. RESULTS: Our sample included 11,771 urinary bladder cancer patients with 10,607 diagnosed at locoregional stage and 1164 at distant stage. Multivariate regression results show primary care density was associated with significantly higher odds of locoregional stage of diagnosis (odds ratio of 1.05 [95% CI: 1.02-1.08]) while urologist density was associated with significantly lower odds of locoregional stage (odds ratio of 0.65 [95% CI: 0.48-0.89]). CONCLUSIONS: We found primary care density but not urologist density was associated with earlier stage of diagnosis, highlighting the importance of access to primary care and need for timely referral to urologic care.

Prescription drug spending by payer: Implications for managed care.

Background: Brand-name prescription drugs are an important driver of prescription drug spending, but different payers may bear these costs differentially necessitating different policy goals for each payer. But little is known about how the top 10 selling drugs in the U.S. impact spending across payers. Objective: To estimate the differential spending burden of top prescription drugs on Medicaid, Medicare, commercial coverage, and out-of-pocket (OOP) spending. Methods: The percentage of total prescription drug spending, total spending, total prescriptions, and average cost per prescription overall and for each of the following payers - Medicaid, Medicare, private insurance, and OOP - was calculated for each of the top 10 selling prescription drugs using 2017-2019 Medical Expenditure Panel Survey data. Results: These 10 prescription drugs accounted for average annual spending of $83.4 billion and 19.0% of all prescription drug spending. Medicare tended to contribute the highest fraction of spending. The average annual cost per prescription ranged from $500 for Advair to $7400 for Tecfidera. Significant variation in the average annual number of prescriptions filled was observed, ranging from 1.4 million for Tecfidera to 13.6 million for Lantus. Conclusions: The findings highlight the significant impact of the top 10 selling prescription drugs on U.S. prescription drug spending. The wide variation in per prescription cost as well as contribution to each payer's prescription drug burden emphasizes how policies targeting top-selling drugs may differentially impact payers as well as how payer-specific policies may differ substantially even for top selling drugs.

Racial and Ethnic Differences in Heroin, Methamphetamine, and Cocaine Use, Treatment, and Mortality Trends in 3 National Data Sources-United States, 2010-2019.

BACKGROUND: As overdose deaths continue to rise, public health officials need comprehensive surveillance data to design effective prevention, harm reduction, and treatment strategies. Disparities across race and ethnicity groups, as well as trends in substance use, treatment, or overdose deaths, have been examined individually, but reports rarely compare findings across multiple substances or data sources. OBJECTIVE: To provide a broad assessment of the overdose crisis, we describe trends in substance use, treatment, and overdose mortality across racial and ethnic groups for multiple substances. RESEARCH DESIGN: We conducted a longitudinal, cross-sectional analysis comparing trends. SUBJECTS: We identified self-reported use from the National Survey on Drug Use and Health, substance use treatment admissions from the Treatment Episode Data Set-Admissions, and overdose deaths from the CDC's Multiple Cause of Death files. MEASURES: We measured rates of substance use, treatment, and deaths involving heroin, methamphetamine, and cocaine among United States adults from 2010 to 2019. RESULTS: Heroin, methamphetamine, and cocaine use increased, though not all changes were statistically significant. Treatment admissions indicating heroin and methamphetamine increased while admissions indicating cocaine decreased. Overdose deaths increased among all groups: methamphetamine (257%-1,115%), heroin (211%-577%), and cocaine (88%-259%). Changes in rates of use, treatment, and death for specific substances varied by racial and ethnic group. CONCLUSIONS: Substance use, treatment, and overdose mortality changed considerably, though not always equivalently. Identifying diverging trends in substance-related measures for specific substances and racial and ethnic groups can inform targeted investment in treatment to reduce disparities and respond to emerging changes in the overdose crisis.

Understanding Regional Patterns of Overdose Deaths Related to Opioids and Psychostimulants.

BACKGROUND: As overdose rates increase for multiple substances, policymakers need to identify geographic patterns of substance-specific deaths. In this study, we describe county-level opioid and psychostimulant overdose patterns and how they correlate with county-level social vulnerability measures. METHODS: A cross-sectional observational study, we used nationwide 2016-2018 restricted access Centers for Disease Prevention and Control county-level mortality files for 1,024 counties. We estimated quartiles of opioid and psychostimulant overdose mortality and provided estimates of their association with county-level Social Vulnerability Index (SVI) percentile. RESULTS: There was high opioid and psychostimulant overdose mortality in the Middle Atlantic, South Atlantic, East North Central, and Mountain regions. The Central US had the lowest opioid and psychostimulant overdose mortality rates. Counties with higher SVI scores (i.e. higher social vulnerability) were significantly more likely to experience high opioid and high psychostimulant overdose (high-high) mortality. A 10-percentile increase in SVI score was associated with a 3.1 percentage point increase in the likelihood of being a high-high county (p < 0.001) in unadjusted models and a 1.5 percentage point increase (p < 0.05) in models adjusting for region. CONCLUSION: Our results illustrated the heterogenous geographic distribution of the growing concurrent opioid and psychostimulant overdose crisis. The substantial regional variation we identified highlights the need for local data to guide policymaking and treatment planning. The association of opioid-psychostimulant overdose mortality with social vulnerability demonstrates the critical need in impacted counties for tailored treatment that addresses the complex medical and social needs of people who use both opioids and psychostimulants.

Increasing the adoption of evidence-based communication practices for HPV vaccination in primary care clinics: The HPV ECHO study protocol for a cluster randomized controlled trial.

BACKGROUND: The safe, highly-effective human papillomavirus (HPV) vaccine remains underused in the US. The Announcement Approach Training (AAT) has been shown to effectively increase HPV vaccine uptake by training providers to make strong vaccine recommendations and answer parents' common questions. Systems communications, like recall notices, can further improve HPV vaccination by reducing missed clinical opportunities for vaccination. Never tested in supporting HPV vaccination, the ECHO (Extension for Community Healthcare Outcomes) model is a proven implementation strategy to increase best practices among healthcare providers. This trial uses a hybrid effectiveness-implementation design (type II) to evaluate two ECHO-delivered interventions intended to increase HPV vaccination rates. METHODS: This 3-arm cluster randomized controlled trial will be conducted in 36 primary care clinics in Pennsylvania. Aim 1 evaluates the impact of HPV ECHO (AAT to providers) and HPV ECHO+ (AAT to providers plus recall notices to vaccine-declining parents) versus control on HPV vaccination (≥1 dose) among adolescents, ages 11-14, between baseline and 12-month follow-up (primary outcome). Using a convergent mixed-methods approach, Aim 2 evaluates the implementation of the HPV ECHO and HPV ECHO+ interventions. Aim 3 explores exposure to and impact of vaccine information from providers and other sources (e.g., social media) on secondary acceptance among 200 HPV vaccine-declining parents within 12 months. DISCUSSION: We expect to demonstrate the effectiveness and evaluate the implementation of two highly scalable interventions to increase HPV vaccination in primary care clinics. Our study seeks to address the communication needs of both providers and parents, increase HPV vaccination, and, eventually, prevent HPV-related cancers. TRIAL REGISTRATION: ClinicalTrials.govNCT04587167. Registered on October 14, 2020.

A scoping review of the opioid epidemic among U.S. Immigrants: Implications for treatment practices.

The purpose of this review was to identify knowledge gaps within the literature regarding the impact of opioid use disorder, specific to immigrants in the United States, by addressing the following questions: 1) What is presented in the literature about the impact of opioid use disorder (OUD) and the opioid epidemic on immigrants in the United States?; and 2) What role does culture play in the opioid use disorder experiences of immigrants in the United States? Nineteen research articles were uncovered that addressed immigrants in the U.S. and opioid use disorder. The following themes prevailed: 1) OUD comparisons, 2) OUD comorbidities, 3) disparate OUD treatment engagement, and 4) the role of country of origin. Limited review findings support the need for future research on the topic of opioid misuse among immigrants in the United States. The authors elaborated on additional issues that influence OUD rates and warrant further exploration. Matters related to the potential positive roles of religion and faith leaders, cultural perceptions and expectations about gender roles, immigration status, ethnically diverse needs among sub-groups of immigrants, the role of geographic location within the U.S., and the implications of COVID-19 on OUD among immigrants need to be addressed to alleviate the deleterious impact of opioid misuse among immigrants.

Association of emergency department crowding with inpatient outcomes.

OBJECTIVE: To examine the association of higher emergency department (ED) census with inpatient outcomes on the day of discharge (inpatient length of stay, in-hospital mortality, ED revisits, and readmissions). DATA SOURCES AND STUDY SETTING: All-payer ED and inpatient discharge data and hospital characteristics data from all non-federal, general, and acute care hospitals in the state of California from October 1, 2015 to December 31, 2017. STUDY DESIGN: In retrospective data analysis, we examined whether ED census was associated with inpatient outcomes for all inpatients, including those not admitted through the ED. The main predictor variable was ED census on day of discharge, categorized based on hospital year and day of week. Separate linear regression models with robust SEs and hospital fixed effects examined the association of ED census on inpatient outcomes (length of stay, 3-day ED revisit, 30-day all-cause readmission, in-hospital mortality), controlling for patient and visit-level factors. We stratified analyses by whether admission was elective or unscheduled. EXTRACTION METHODS: Inpatient discharges in non-federal, general medical hospitals with EDs. PRINCIPAL FINDINGS: We examined 5,784,253 discharges. The adjusted model showed that, compared to when the ED was below the median, higher ED census on the day of discharge was associated with longer inpatient length of stay, lower readmissions, and higher in-hospital mortality (90th percentile for length of stay: +0.8% [95% confidence interval, CI: +0.6% to +1.1%]; readmissions: -0.59 percentage points [or -5.6%] [95% CI: -0.0071 to -0.0048]; mortality: +0.14 percentage points [or +5.4%] [95% CI: +0.0009 to +0.0018]). [Correction added on 18 November 2022, after first online publication: '[odds rato, OR -5.6%]' and '[OR +5.4%]' of the preceding sentence have been corrected to '[or -5.6%]' and '[or +5.4%]', respectively, in this version.] Results for length of stay were primarily driven by patients with elective admissions, while results for readmissions and in-hospital mortality were primarily driven by patients with unscheduled admissions. CONCLUSIONS: This study suggests that ED crowding may affect inpatients throughout the hospital, even patients who are already admitted to the hospital.

Do-Not-Resuscitate Orders and Outcomes for Patients with Pancreatic Cancer.

Background: The impact of the do-not-resuscitate (DNR) order on patients with pancreatic cancer remains uncertain. In this study, we evaluated whether DNR status was associated with in-hospital mortality and costs for inpatient stay among patients hospitalized with pancreatic cancer. Methods: Data were obtained from the National Inpatient Sample, Healthcare Cost and Utilization Project, which represents ∼20% of all discharges from US community hospitals; 40,246 pancreatic cancer admissions between 2011 and 2016 were included. Mortality was modeled using a logistic regression model; costs for inpatient stay were modeled using a multivariable generalized linear regression model. Results: The sample included 6041 (15%) patients with a documented DNR order. After controlling for covariates, patients with a DNR order had approximately six times greater odds of mortality compared with patients without a DNR order (odds ratio 5.90, p < 0.0001). Compared with patients who survived without a DNR order during the hospital stay, patients who had a DNR order and died during the hospital stay had significantly lower costs (-US$983; p = 0.0270), and patients who died without a DNR order during the hospital stay had significantly higher costs (US$5638; p < 0.0001). Patients who survived with a DNR order had costs that were not significantly different from patients who survived without a DNR order. Conclusions: The presence of a DNR order among patients with pancreatic cancer was significantly associated with higher mortality risk as well as lower costs for patients who died during the hospital stay. However, DNR status was not significantly associated with costs for pancreatic cancer patients who were discharged alive.

Associations between prescription and illicit stimulant and opioid use in the United States, 2015-2020.

INTRODUCTION: Overdose deaths involving opioids and stimulants continue to reach unprecedented levels in the United States. Although significant attention has been paid to the relationship between prescription and illicit opioid use, little work has focused on the association between prescription and illicit stimulant use. Thus, this study explores characteristics of those who use or misuse prescription stimulants and/or opioids and associations with use of cocaine, methamphetamine, and heroin. METHODS: We used 2015-2020 data from the National Survey on Drug Use and Health. Using adjusted multivariable logistic regression, we estimated the associations between past year prescription stimulant or prescription opioid prescribed use and misuse; various demographic characteristics; and past-year cocaine, methamphetamine, or heroin use. RESULTS: From 2015 to 2020, 4.9 and 9.8 million US adults annually reported misusing prescription stimulants and opioids, respectively. Individuals who misused prescription stimulants were more likely to be ages 18-25 (45.8 %; 95 % CI: 44.0-47.5) than individuals who misused prescription opioids (21.7 %; 95 % CI: 20.7-22.7). We observed higher rates of cocaine use among individuals reporting prescription stimulant misuse (12.0 %; 95 % CI: 11.0-12.9) compared to those reporting prescription opioid misuse (5.7 %; 95 % CI: 5.1-6.3, p < 0.001). Heroin use was more common among individuals with prescription opioid misuse (2.1 %; 95 % CI: 1.7-2.2) than prescription stimulant misuse (0.6 %; 95 % CI: 0.4-0.7, p < 0.001). However, rates of methamphetamine use among individuals with prescription stimulant misuse (2.4 %; 95 % CI: 1.9-3.0) did not differ from individuals with prescription opioid misuse (2.1 %; 95 % CI: 1.7-2.5, p = 0.67). CONCLUSIONS: Prescription stimulant misuse, compared to prescription opioid misuse, was associated with higher levels of cocaine use but not methamphetamine use. Treatment providers should consider screening for other substance use disorders among people who report prescription stimulant use or misuse. Additional research should seek to understand the mechanism underlying the different associations between prescription stimulant misuse and cocaine or methamphetamine use.

Nonadherence to Multimodality Cancer Treatment Guidelines in the United States.

Purpose: Our purpose was to identify patients with cancer who do not receive guideline-concordant multimodality treatment and to identify factors that are associated with nonreceipt of guideline-concordant multimodality treatment. Methods and Materials: Five cancers for which the multimodal guideline-concordant treatment (with surgery, chemotherapy, and radiation therapy) is clearly defined in national guidelines were selected from the National Cancer Database: (1) nonmetastatic anal cancer, (2) locally advanced cervical cancer, (3) nonmetastatic nasopharynx cancer, (4) locally advanced rectal cancer, and (5) locally advanced non-small cell lung cancer. Multivariable logistic regression was used to determine the odds ratios (with 95% confidence intervals) of receiving the guideline-concordant treatment versus not, adjusting for common confounding variables. Results: 178,005 patients with cancer were included: 32,214 anal, 54,485 rectal, 13,179 cervical, 5061 nasopharyngeal, and 73,066 lung. Overall, 162,514 (91%) received guideline-concordant treatment and 15,491 (9%) did not. Twenty-one percent of patients with cervical cancer, 10% of patients with rectal cancer, 7% of patients with lung cancer, 5% of patients with anal cancer, and 3% of patients with nasopharynx cancer did not receive guideline-concordant treatment. In general, patients who were older, with comorbid conditions, and who were evaluated at low-volume facilities (odds ratios > 1 with P < .05) were less likely to receive guideline-concordant treatment. Conclusions: Nearly 1 in 10 patients in this cohort are not receiving appropriate multimodal cancer therapy. There appear to be significant disparities in receipt of guideline-concordant treatment based on primary tumor site, age, comorbidities, and reporting facility.

Effect of Out-of-Pocket Costs on Subsequent Mammography Screening.

OBJECTIVE: Although the Affordable Care Act eliminated cost sharing for screening mammography, a concern is that grandfathered plans, diagnostic mammograms, and follow-up testing may still lead to out-of-pocket (OOP) spending. Our study examines how OOP spending among women at their baseline screening mammogram may impact the decision to receive subsequent screening. METHODS: The study included commercially insured women aged 40 to 41 years with a screening mammogram between 2011 and 2014. We estimated multivariate linear probability models of the effect of OOP spending at the baseline mammogram on subsequent screening 12 to 36 months later. RESULTS: Having any OOP payments for the baseline screening mammogram significantly reduced the probability of screening in the subsequent 12 to 24 months by 3.0 percentage points (pp) (95% confidence interval [CI]: 1.1-4.8 pp decrease). For every $100 increase in the OOP expenses for the baseline mammogram, the likelihood of subsequent screening within 12 to 24 months decreased by 1.9 pp (95% CI: 0.8-3.1 pp decrease). Similarly, any OOP spending for follow-up tests resulting from the baseline screening led to a 2.7 pp lower probability of screening 12 to 24 months later (95% CI: 0.9-4.1 pp decrease). Higher OOP expenses were associated with significantly lower screening 24 to 36 months later (coefficient = -0.014, 95% CI: -0.025 to -0.003). DISCUSSION: Although cost sharing has been eliminated for screening mammograms, OOP costs may still arise, particularly for diagnostic and follow-up testing services, both of which may reduce rates of subsequent screening. For preventive services, reducing or eliminating cost sharing through policy and legislation may be important to ensuring continued adherence to screening guidelines.

Medical Service Use and Charges for Cancer Care in 2018 for Privately Insured Patients Younger Than 65 Years in the US.

Importance: Currently, there are limited published data regarding resource use and spending on cancer care in the US. Objective: To characterize the most frequent medical services provided and the associated spending for privately insured patients with cancer in the US. Design, Setting, and Participants: This cohort study used data from the MarketScan database for the calendar year 2018 from a sample of 27.1 million privately insured individuals, including patients with a diagnosis of the 15 most prevalent cancers, predominantly from large insurers and self-insured employers. Overall societal health care spending was estimated for each cancer type by multiplying the mean total spending per patient (estimated from MarketScan) by the number of privately insured patients living with that cancer in 2018, as reported by the National Cancer Institute's Surveillance, Epidemiology, and End Results program. Analyses were performed from February 1, 2018, to July 8, 2021. Exposures: Evaluation and management as prescribed by treating care team. Main Outcomes and Measures: Current Procedural Terminology and Healthcare Common Procedure Coding System codes based on cancer diagnosis code. Results: The estimated cost of cancer care in 2018 for 402 115 patients with the 15 most prevalent cancer types was approximately $156.2 billion for privately insured adults younger than 65 years in the US. There were a total of 38.4 million documented procedure codes for 15 cancers in the MarketScan database, totaling $10.8 billion. Patients with breast cancer contributed the greatest total number of services (10.9 million [28.4%]), followed by those with colorectal cancer (3.9 million [10.2%]) and prostate cancer (3.6 million [9.4%]). Pathology and laboratory tests contributed the highest number of services performed (11.7 million [30.5%]), followed by medical services (6.3 million [16.4%]) and medical supplies and nonphysician services (6.1 million [15.9%]). The costliest cancers were those of the breast ($3.4 billion [31.5%]), followed by lung ($1.1 billion [10.2%]) and colorectum ($1.1 billion [10.2%]). Medical supplies and nonphysician services contributed the highest total spent ($4.0 billion [37.0%]), followed by radiology ($2.1 billion [19.4%]) and surgery ($1.8 billion [16.7%]). Conclusions and Relevance: This analysis suggests that patients with breast, colorectal, and prostate cancers had the greatest number of services performed, particularly for pathology and laboratory tests, whereas patients with breast, lung, lymphoma, and colorectal cancer incurred the greatest costs, particularly for medical supplies and nonphysician services. The cost of cancer care in 2018 for the 15 most prevalent cancer types was estimated to be approximately $156.2 billion for privately insured adults younger than 65 years in the US.

Potential Winners and Losers: Understanding How the Oncology Care Model May Differentially Affect Hospitals.

PURPOSE: With the introduction of the Oncology Care Model and plans for the transition to Oncology Care First, alternative payment models (APMs) are an increasingly important piece of the oncology care landscape. Evidence is mixed on the Oncology Care Model's impact on utilization and costs, but as policymakers consider expansion of similar models, it is critical to understand the characteristics of hospitals that may be differentially affected. METHODS: We used 2007-2016 SEER-Medicare data to identify patients with breast and prostate cancer receiving chemotherapy, endocrine therapy (breast), or androgen deprivation therapy (prostate). For each hospital, we calculated 6-month expected mortality, emergency department (ED) visits, inpatient admissions, and costs, all commonly collected APM outcomes. After calculating observed-to-expected rates for each outcome by hospital, we estimated the association between observed-to-expected rates and characteristics of each hospital to understand hospital characteristics that might be associated with higher- or lower-than-expected rates of each outcome. RESULTS: Hospitals with > 15% rural patients had significantly higher-than-expected mortality (0.31 points higher, P < .001) and ED visit rates (0.10 points higher, P = .029) as well as significantly lower costs (0.06 points lower, P = .004). Hospitals unaffiliated with a medical school also experienced significantly higher-than-expected mortality and ED visits. Hospitals eligible for disproportionate share hospital payment experienced significantly higher ED visits but lower costs. For-profit hospitals experienced higher-than-expected mortality. CONCLUSION: Rural hospitals and those unaffiliated with a medical school may require special consideration as APMs expand in oncology care. Designated cancer centers and larger hospitals may be advantaged.

Health Savings Plans and Disparities in Access to Care by Race and Ethnicity.

INTRODUCTION: High-deductible health plans are often touted to motivate patients to become informed healthcare purchasers; however, racial/ethnic minorities report that high deductibles prevent them from seeking the needed care. One proposed way to mitigate the financial burden of high-deductible health plans is the use of health savings plans. This cross-sectional study investigates whether chronically ill Blacks and Hispanics enrolled in high-deductible health plans experience greater access to care difficulties than non-Hispanic Whites and whether racial/ethnic disparities are mitigated by the use of health savings plans. METHODS: Weighted, multivariate, linear probability regression models were estimated (analyses were conducted in December 2020), adjusting for individual attributes and contextual factors that may explain the variation in health care access. Chronically ill, U.S.-born Black, Hispanic, and White adults enrolled in a high-deductible health plan from the National Health Interview Survey in 2011-2018 were included. Associations were tested among 3 independent variables-being Black, being Hispanic, and health savings plan utilization (and their interaction)-and access to healthcare outcomes of interest, including affordability-related access, provider-related access, and delayed care. RESULTS: Blacks and Hispanics were less likely to use health savings plans, and Blacks were more likely to experience problems with access to health care. Although the use of health savings plans was found to have a minimal effect on reducing racial/ethnic disparities in affordability-related access, there was also evidence that health savings plans compounded racial/ethnic disparities in provider-related access. CONCLUSIONS: Understanding how health savings plans function to improve access to care within racial/ethnic minority groups may help to inform policy approaches related to their use.

Comparative Effectiveness of a Complex Care Program for High-Cost/High-Need Patients: a Retrospective Cohort Study.

BACKGROUND: High-cost/high-need (HCHN) adults and the healthcare systems that provide their care may benefit from a new patient-centered model of care involving a dedicated physician and nurse team who coordinate both clinical and social services for a small patient panel. OBJECTIVE: Evaluate the impact of a Complex Care Program (CCP) on likelihood of patient survival and hospital admission in 180 days following empanelment to the CCP. DESIGN: Retrospective cohort study using a quasi-experimental design with CCP patients propensity score matched to a concurrent control group of eligible but unempaneled patients. SETTING: Kaiser Permanente Mid-Atlantic States (KPMAS) during 2017-2018. PARTICIPANTS: Nine hundred twenty-nine CCP patients empaneled January 2017-June 2018, 929 matched control patients for the same period. INTERVENTIONS: The KPMAS CCP is a new program consisting of 8 teams each staffed by a physician and nurse who coordinate care across a continuum of specialty care, tertiary care, and community services for a panel of 200 patients with advanced clinical disease and recent hospitalizations. MAIN OUTCOMES: Time to death and time to first hospital admission in the 180 days following empanelment or eligibility. RESULTS: Compared to matched control patients, CCP patients had prolonged time to death (hazard ratio [HR]: 0.577, 95% CI: 0.474, 0.704), and CCP decedents had longer survival (median days 69.5 vs. 53.0, p=0.03). CCP patients had similar time to hospital admission (HR: 1.081, 95% CI: 0.930, 1.258), with similar results when adjusting for competing risk of death (HR: 1.062, 95% CI: 0.914, 1.084). LIMITATIONS: Non-randomized intervention; single healthcare system; patient eligibility limited to specific conditions. CONCLUSION: The KPMAS CCP was associated with significantly reduced short-term mortality risk for eligible patients who volunteered to participate in this intervention.