Medical Care最新文献

Background: In 2022, Louisiana Medicaid began offering reimbursement for some community health worker (CHW) services ordered and billed by a supervising clinician.

Objectives: We analyzed the extent to which CHW services were reimbursed by Louisiana Medicaid during 2022-2023, including number of encounters, demographics of Medicaid members served, number of encounters per member, geographic distribution of CHW encounters, primary diagnosis of members receiving CHW services, and provider type billing for CHW services.

Research design: Retrospective cohort study of Louisiana Medicaid members receiving CHW services reimbursed by Medicaid. We included paid fee-for-service claims and managed care encounters for CPT codes 98960, 98961, or 98962 with dates of service from 1/1/2022 through 12/31/23.

Results: A total of 10,726 unique individuals received 17,373 reimbursed CHW services, with an estimated total reimbursement of $314,905.52. Medicaid members ranged from 0 to 88 years, were majority female (56.8%), and received between 1 and 13 services; mean: 1.6 (SD: 1.1), median: 1 (IQR 1-2 encounters). The highest rate of CHW services per Medicaid member occurred in urban areas. Nearly all (99.97%) were individual services. A total of 41.9% of services were for screening without a specified diagnosis, and health-related social needs were more common than medical conditions. Over half (60.6%) of CHW services were billed by family practice or internal medicine providers.

Discussion: Health care practices may be asking CHWs primarily to conduct screenings, rather than provide the longitudinal services CHWs traditionally offer.

Conclusion: Uptake of billing for CHW services was limited. Providers may need education about CHW roles and technical assistance to support CHW integration and billing.

Background: In 2021, the Agency for Health Care Research and Quality (AHRQ) updated its guidelines for using the Present-on-Admission (POA) indicator in the Elixhauser comorbidity index. This update helps distinguish pre-existing comorbidities from complications that arise after hospital admission, improving the validity of hospital performance assessments and more accurately measuring patients' severity of illness upon admission.

Objective: To evaluate differences in comorbidity prevalence and the predictive performance of the Elixhauser Comorbidity Index for in-hospital mortality at admission under 3 comorbidity coding guidelines, including one that ignores the POA indicator.

Research design: A retrospective analysis of inpatient administrative data on Medicare beneficiaries.

Subjects: The dataset included 1,810,106 adult Medicare inpatient admissions across 6 U.S. states between 2017 and 2019.

Methods: Elastic net models were applied to predict in-hospital mortality using 3 approaches to coding comorbidities: (1) No-POA (including all conditions as admission comorbidities), (2) Full-POA (including only POA conditions as comorbidities), and (3) the 2021 AHRQ Partial-POA (applying POA to a subset of conditions to code comorbidities). Results: C-statistics were 0.800 (0.797-0.804), 0.768 (0.763-0.771), and 0.786 (0.781-0.790) for No-POA, full-POA, and 2021 AHRQ partial-POA guidelines, respectively.

Conclusion: Ignoring the POA inflated model performance by misclassifying complications as admission comorbidities. The 2021 Partial-POA guidelines achieved intermediate C-statistics while ensuring internal validity by accurately measuring illness severity at admission. This supports improved hospital evaluations, care quality, resource allocation, tailored intervention, and reimbursement. The elastic net model shows promise as a standard for predicting in-hospital mortality with the Elixhauser comorbidity measure.

Objective: To better understand financial barriers to care facing American Indian and Alaska Native households, this study builds on previous findings that these communities have a higher likelihood of having medical debt and engaging in cost avoidance. This study aims to build on those findings by controlling for health status, insurance type, and Indian Health Service (IHS) eligibility.

Design: This study uses data from the National Health Information Survey in binomial logistic regression models to examine the likelihood of American Indian and Alaska Native households having medical debt and engaging in cost avoidance.

Results: The results of the logistic regression analysis found that while health status and IHS eligibility significantly contribute to the likelihood of having medical debt or engaging in cost avoidance, racial disparities remain for American Indian and Alaska Native communities.

Conclusions: Despite access to the Indian Health Service and Tribal care, American Indian and Alaska Native households face disparities in financial barriers to care. These results suggest that, rather than the proposed cuts to the Indian Health Service, additional funding is needed to address shortcomings in the IHS/Tribal system of care in American Indian and Alaska Native communities.

Background: Poor social health is linked to incident cardiovascular disease, but less is known about how loneliness affects health care utilization after an acute myocardial infarction (AMI).

Objective: Determine the association between loneliness and 30-day emergency department (ED) visit or readmission after AMI hospitalization.

Research design: The REasons for Geographic and Racial Differences in Stroke (REGARDS) study is a national prospective cohort of 30,239 U.S. adults aged 45 years or older.

Measures: We examined the association between loneliness and 30-day post-AMI ED visit or readmission.

Subjects: Seven hundred forty-nine Medicare fee-for-service beneficiaries in REGARDS were discharged alive after an adjudicated AMI.

Results: The mean age was 77 years. Twenty-eight percent self-identified as non-Hispanic Black and 39% as women. Twenty percent reported loneliness. Twenty-nine percent had a 30-day ED visit or readmission. Lonely individuals had 61% increased risk of 30-day ED visit or readmission (RR: 1.61; 95% CI: 1.27-2.04; P <0.001), which remained significant after adjustment for sociodemographic and clinical factors (aRR: 1.48; 95% CI: 1.12-1.95; P =0.006). Stratified analyses demonstrated significant association for those aged 65-74 (aRR 2.48; 95% CI, 1.57-3.91; P <0.001), White adults (aRR: 1.86; 95% CI: 1.35-2.58; P <0.001), and men (aRR: 2.19; 95% CI: 1.59-3.01; P <0.001) but not for those 75+ (aRR: 0.94; 95% CI: 0.63-1.40; P =0.75), Black adults (aRR: 0.89; 95% CI: 0.53-1.49; P =0.660), or women (aRR: 0.81; 95% CI: 0.51-1.30; P =0.380).

Conclusions: Loneliness, even measured years before AMI, was associated with an increased risk of 30-day ED visit or readmission, specifically for those aged 65-74, White participants, and men. These findings may inform discharge strategies to reduce readmissions.

Background: Patients with Alzheimer disease and related dementias (ADRD) face risks from medications labeled "potentially inappropriate in older adults" (risky); concurrent receipt of anti-dementia drugs may amplify risk. We studied adverse events among older adults concurrently receiving anti-dementia and risky medications.

Methods: Using 2016-2019 administrative data from a random 40% sample of fee-for-service Medicare beneficiaries receiving anti-dementia medications (acetylcholinesterase inhibitors (AChEI) and/or memantine), we identified days with concurrent receipt of select, risky medications (benzodiazepines, sedative hypnotics, opioids). We measured diagnosed falls, hip fractures, and deaths among person-days with anti-dementia drug receipt comparing person-days with versus without concurrent risky drug receipt. We stratified regression analyses on long-term care (LTC) residence.

Results: We studied 633,528 beneficiaries; 64.3% were women, 33.7% met LTC residence criteria. Mean (SD) age was 80.9 (7.6) years. Each beneficiary contributed a mean (SD) of 551.7 (449.2) anti-dementia drug receipt days. Overall, 4.5% of person-days involved receipt of AChEI plus benzodiazepines; 3.8% involved AChEI plus an opioid. Falls, the most common outcome, affected 22.5% of our beneficiaries. Concurrent receipt of AChEI and opioids was associated with the greatest fall risk increase. Among community-dwelling beneficiaries, AChEI and opioid receipt (vs. AChEI alone) was associated with a hazard ratio for falls of 2.25 (95% CI: 2.19, 2.32); among LTC residents the corresponding hazard ratio was 1.46 (95% CI: 1.42, 1.51).

Conclusions: Assessment and treatment of symptoms among people with ADRD is complex; concurrent receipt of opioids and dementia medications is uncommon but seems risky. Efforts to eliminate avoidable opioids may decrease adverse events and associated suffering in this population.

Background: Telehealth can improve health care access in underserved areas. Hub-and-spoke-models, wherein providers in regional hubs deliver care through telehealth to patients visiting local "spoke" clinics, can improve access. However, cost impacts of this model are unknown.

Objective: Evaluate the utilization and cost impacts of VA's Clinical Resource Hub program for primary care (CRH-PC), a hub-and-spoke-model.

Design: Adjusted difference-in-difference and event study analyses comparing patients at program-sites who used CRH-PC services with patients who never used CRH-PC services, prepost program adoption, fiscal years 2018-2021. We also compared all patients at CRH-PC sites versus at non-CRH-PC sites to assess site-wide impacts.

Participants: CRH-PC sites: 164 sites and 1,546,892 patients; Non-CRH-PC sites: 704 sites and 4,062,797 patients.

Measures: Costs and number of VA-provided and VA-purchased primary, emergency, and acute inpatient care visits.

Results: At CRH-PC sites, 64,973 patients (4%) used CRH-PC services. Rural patients, African-American patients, and patients with greater comorbidities had higher odds of receiving program services. Program exposure was associated with an 18% increase in primary care visits (+0.7) and $612 per program-user per year. Comparing all patients (users and nonusers) at program-sites versus nonprogram sites, we found no impact, except video-based care more often replaced in-person services at program-sites.

Conclusions: Among program-users, VA's CRH-PC increased mean primary care visits and VA costs, but as only 4% of patients at program-clinics were program-users, there were no differences in overall cost or utilization between program and nonprogram clinics. Findings suggest clinics can offer primary care telehealth services to high-need populations without affecting clinic-level costs, but costs should be monitored upon wider adoption.

Objective: We sought to examine changes in mild-to-moderate TBI-related symptoms among service members and veterans (SM/Vs) following participation in a 5-site inpatient rehabilitation program with the US Department of Veterans Affairs between 7/1/2022 and 5/30/2024.

Methods: Neurobehavioral outcomes, posttraumatic stress disorder (PTSD) symptoms, pain interference, and lifestyle behaviors related to brain injury were assessed at baseline, discharge, and a 6-month follow-up. Mixed effects linear regression models, adjusting for key patient characteristics, were estimated to determine changes in TBI-related outcomes across the 3 time points.

Results: Mean participant age, for those with complete data (n = 127), was 41.64 years (SD = 5.57), with a mean of 7.45 deployments (SD = 3.12) and 16.32 concussive events (SD = 7.21). Participants were predominantly White (73.23%) Special Operations personnel (82.68%). TBI-related outcomes, including neurobehavioral symptoms, pain interference, PTSD, and brain injury adaptability, decreased significantly from baseline to discharge (b = -14.36, SE = 1.03; b = -3.79, SE = 0.49; b = -11.14, SE = 1.27; b = -2.41, SE = 0.41), with Cohen's d effect sizes of 1.14, 0.71, 0.69, and 0.56, respectively. Six-month follow-up, TBI-related outcomes remained statistically and practically below baseline levels in all measures except adaptability.

Conclusions and relevance: Findings illustrate an interdisciplinary, inpatient rehabilitation program for mild-to-moderate TBI yields significant improvements in TBI-related symptoms that are common among SM/Vs and are sustained at 6 months postdischarge.

Introduction: The Surveillance, Epidemiology, and End Results (SEER) Program collects data on the first course of cancer treatment, but no and unknown receipt of treatment cannot be distinguished for radiation therapy (RT) and chemotherapy. As part of the Data Use Agreement (DUA), users must acknowledge that they understand the data limitations and agree to include a description of the limitations in any analyses published using the data. The objective of this review was to evaluate users' compliance with the recommendations of the DUA.

Methods: Publications from a PubMed search were matched with the names of SEER treatment data users, and keywords were applied to identify relevant studies. Five reviewers (with 2 per publication) independently assessed if the authors (a) conducted analyses supported by these data, (b) correctly labelled no/unknown treatment as "no/unknown", and (c) described the limitations of their use. Publications were classified as "followed recommendations", "partially followed recommendations", or "did not follow recommendations" of the DUA.

Results: Among a total of 120 studies included in the review, 106 (88.3%) studies did not follow recommendations, 11 (9.2%) partially followed recommendations, and 3 studies (2.5%) followed recommendations. Only 11.7% of publications correctly labelled the "no/unknown" category as "no/unknown", and described the limitations associated with the no/unknown issue.

Conclusions: In this review, we found substantial misuse of the SEER treatment data and limited acknowledgement of the limitations of the SEER treatment data in publications. Such findings highlight the need to think of effective ways of encouraging appropriate use of the treatment data.

Background: Prior work has shown that Medicaid coverage offset reductions in employer-sponsored insurance after COVID-19-related job loss in Medicaid expansion states. However, the effect of Medicaid expansion on health care access is not fully understood.

Objective: To estimate the association of unemployment during COVID-19 with health insurance coverage and health care access in Medicaid expansion versus nonexpansion states.

Study design: We used restricted, longitudinal National Health Interview Survey (NHIS) data from 2019 to 2020, focusing on working-age adults (N=5156). Using triple-difference models, we estimated changes in outcomes for respondents becoming unemployed between 2019 and 2020 (after COVID-19) relative to continuously employed respondents, in Medicaid expansion versus nonexpansion states.

Principal findings: Compared with continuously employed respondents, there was a statistically significant increase in Medicaid coverage among the newly unemployed in expansion states (6.1 percentage points (pp), 95% CI: 1.0 to 11.3, P=0.019) but not in nonexpansion states (3.9 pp, 95% CI: -3.9 to 11.8, P=0.324); however, the triple difference was nonsignificant. Uninsurance increased among the newly unemployed in expansion states by only 4.9 pp (95% CI: 0.9 to 8.9, P=0.016) versus 12.4 pp in nonexpansion states (95% CI: 0.2 to 24.6, P=0.047), but the triple difference was statistically nonsignificant. There was a significant increase in delaying or skipping medical care among newly unemployed respondents in nonexpansion states, but not in expansion states and the triple difference was statistically significant (-15.5 pp, 95% CI: -26.9 to -4.0, P=0.008).

Conclusion: Our results suggest that Medicaid expansion prevented disruptions in health care access for the newly unemployed during COVID-19.