Health Services Research最新文献

Objective: To evaluate whether race- and sex-based biases are present in a predictive model of avoidable hospital (AH) events.

Study setting and design: We examined whether Medicare fee-for-service (FFS) beneficiaries in Maryland with similar risk scores differed in true AH event risk on the basis of race or sex (n = 324,834). This was operationalized as a logistic regression of true AH events on race or sex with fixed effects for risk score percentile.

Data sources and analytic sample: Beneficiary-level risk scores were derived from 36 months of Medicare FFS claims (April 2019-March 2022) and generated in May 2022. True AH events were observed in claims from June 2022.

Principal findings: Black patients had higher average risk scores than White patients; however, the likelihood of experiencing an AH event did not differ by race when controlling for predicted risk (Marginal Effect [ME] = 0.0003, 95%CI -0.0003 to 0.0009). AH event likelihood was lower in males when controlling for risk level; however, the effect was small (ME = -0.0008, 95% CI -0.0013 to -0.0003) and it did not differ by sex for the target group for intervention (ME = 0.0002, 95% CI -0.0031 to 0.0036).

Conclusions: We implemented a simple bias assessment methodology and found no evidence of meaningful race- or sex-based bias in this model. We encourage the incorporation of bias checks into predictive model development and monitoring processes.

Objective: To test quantitative process measures characterizing the work of social needs case managers as they assisted patients with diverse health-related needs-spanning both medical and social domains.

Study setting and design: The study analyzed secondary data on 7076 patients working with 147 case managers from the CommunityConnect social needs case management program in Contra Costa County, California from 2018 to 2021. The service-designed to be holistic with a focus on social determinants as root causes of health issues-helped patients navigate social services, health care, and mental health care.

Data sources and analytic sample: We used cross-sectional analyses to quantitatively characterize electronic health records (EHRs) derived measures of case management intensity (goal updates), duration (days goal was open), and outcomes for 19 different categories of health and social goals. Mixed-effects regression models were used to examine how work process measures varied according to goal categories. Models nested goals within patients within case managers and adjusted for patient-level covariates.

Principal findings: The most common goals were dental care (53%), food (40%), and housing (39%). In adjusted analyses, housing goals had significantly more case manager updates than any other type of goal with a marginal mean of 14.0 updates (95% CI: 13.4-14.7), were worked on for significantly longer (marginal mean of 417 days, 95% CI: 360-474) than any goal except dental care, and were least likely to be resolved. Utilities, insurance, and medication coordination goals were most likely to be resolved.

Conclusions: Case managers and patients repeatedly worked on goals over many months. Meeting housing needs and accessing dental care were issues that were not easily resolved and required extensive follow-up. One-time referral interventions may need follow-up systems to meaningfully support social and health needs.

Objective: To evaluate healthcare cost and utilization changes among Medicaid and dually eligible participants of a supportive housing program implemented by a managed care organization and community-based organization.

Study setting and design: Healthcare claims were reviewed retrospectively for 80 program participants in one urban Pennsylvania county between 1/1/2018 and 9/28/2023 who had ≥6 months of claims data in both pre- and post-housing periods. Eligibility included age >18 years, Medicaid/Special Needs Plan enrollment, and housing need. Due to limited housing units, potential participants were prioritized by medical need and history of unplanned care.

Data sources and analytic sample: Healthcare cost and utilization were compared during pre- (i.e., 12 months before housing initiation) and post-periods (i.e., 12 months after housing initiation).

Principal findings: Compared to the pre-period, significantly lower medical (-40.4%, p = 0.004), emergency department (-62.7%, p = 0.02), and total (-33.3%, p = 0.02) costs of care were observed in the post-period. Significantly lower primary care (-50.0%, p = 0.0003), specialist (-31.3%, p = 0.02), and emergency department (-50.0%, p = 0.03) utilization were also observed.

Conclusions: Healthcare cost and utilization among medically complex individuals were lower with supportive housing. Future evaluations with randomized designs can address the potential causal impact of supportive housing as a healthcare intervention on specific outcomes.

Objective: To examine the relationship between insurers' commercial market power and negotiated prices in Medicaid Managed Care (MMC) plans for hospital care.

Data sources: MMC prices from hospital-disclosed price transparency data as of July 2023 compiled by Turquoise Health, insurance enrollment information from the 2021 Clarivate InterStudy enrollment data.

Study design: Log-transformed linear regression with hospital and procedure fixed effects estimating the within-hospital MMC price variation as a function of insurers' commercial market share quartile and MMC market share for 15 common outpatient hospital services.

Data collection/extraction methods: A total of 39,049 MMC price samples measured at hospital-procedure-MMC insurer level are merged with county-insurer level market share data.

Principal findings: Around 25% of price variation in MMC plans are driven by within-hospital factors. Compared with MMC insurers from the lowest commercial market share quartile (<0.8%), those from the highest commercial market share quartile (>17%) are associated with negotiating 4.6% (95% confidence interval: [2.8%-6.4%], p < 0.001) lower MMC prices for outpatient hospital care, including 3.6% (p < 0.05) for medical/surgical procedures, 3.6% (p < 0.01) for radiology, and 6.7% (p < 0.001) for emergency department visits.

Conclusions: MMC insurers with substantial commercial market share negotiate lower MMC prices for multiple outpatient hospital services.

Objective: To investigate provider and administrators' perspectives about the impact of the Department of Veterans Affairs' (VA) Community Care program on acute and residential mental health treatment of rural Veterans.

Data sources and study setting: Primary data were collected from participants via interviews. Participants were employees of VA Healthcare Systems located in Northern New England, or employees of non-VA mental health treatment settings affiliated with VA in Northern New England.

Study design: This study was informed by the Consolidated Framework for Implementation Research (CFIR), with Community Care as the implemented program. Individual, semi-structured interviews were conducted.

Data collection/extraction methods: Individual interviews were transcribed, coded deductively using the CFIR, and inductively coded by locating themes.

Principal findings: Twenty-one people completed interviews. Commonly reported challenges included community programs not focused on Veterans' needs, poor coordination of care, communication challenges, and problems tracking Veteran care. Facilitators included increased access to care and strengthening coordination of care.

Conclusions: The VA's Community Care program can address the acute or residential mental health needs of Veterans in rural settings in some circumstances, however there are challenges to successful implementation.

Objectives: To examine heterogeneity in physician batch ordering practices and measure the associations between a physician's tendency to batch order imaging tests on patient outcomes and resource utilization.

Study setting and design: In this retrospective study, we used comprehensive EMR data from patients who visited the Mayo Clinic of Arizona Emergency Department (ED) between October 6, 2018 and December 31, 2019. Primary outcomes are patient length of stay (LOS) in the ED, number of diagnostic imaging tests ordered during a patient encounter, and patients' return with admission to the ED within 72 h. The association between outcomes and physician batch tendency was measured using a multivariable linear regression controlling for various covariates.

Data sources and analytic sample: The Mayo Clinic of Arizona Emergency Department recorded approximately 50,836 visits, all randomly assigned to physicians during the study period. After excluding rare complaints, we were left with an analytical sample of 43,299 patient encounters.

Principal findings: Findings show that having a physician with a batch tendency 1 standard deviation (SD) greater than the average physician was associated with a 4.5% increase in ED LOS (p < 0.001). It was also associated with a 14.8% (0.2 percentage points) decrease in the probability of a 72-h return with admission (p < 0.001), implying that batching may lead to more comprehensive evaluations, reducing the need for short-term revisits. A batch tendency 1SD greater than that of the average physician was also associated with an additional 8 imaging tests ordered per 100 patient encounters (p < 0.001), suggesting that batch ordering may be leading to tests that would not have been otherwise ordered had the physician waited for the results from one test before placing their next order.

Conclusions: This study highlights the considerable impact of physicians' diagnostic test ordering strategies on ED efficiency and patient care. The results also highlight the need to develop guidelines to optimize ED test ordering practices.

Objective: This study aimed to compare imputation approaches to identify the likely undocumented patient population in electronic health record (EHRs). EHR are a promising source of information on undocumented immigrants' medical needs and care utilization, but there is no verified way to identify immigration status in the data. Different approaches to approximating immigration status in EHR introduce unique biases, which in turn has major implications on our understanding of undocumented immigrant patients.

Study setting and design: We used a dataset of all emergency department (ED) visits from 2016 to 2019 in the Los Angeles Department of Health Services (LADHS) merged across patient medical records, demographic data, and claims data. We included all ED visits from our patient groups of interest and limited to patients at or over the age of 18 years at the time of their ED visit and excluded empty encounter records (n = 1,106,086 ED encounters).

Data sources and analytic sample: We created three patient groups: (1) US-born, (2) foreign-born documented, and (3) undocumented using two different imputation approaches: a logical approach versus statistical assignment. We compared predicted probabilities for two outcomes: an ED visit related to a behavioral health (BH) disorder and inpatient admission/transfer to another facility.

Principal findings: Both approaches provide comparable estimates among the three patient groups for ED encounters for a BH disorder and inpatient admission/transfer to another facility. Undocumented immigrants are less likely to have a BH diagnosis in the ED and are less likely to be admitted or transferred compared to the US-born.

Conclusions: Researchers should consider expanding EHR with administrative data when studying the undocumented patient population and may prefer a logical approach to estimate immigration status. Researchers who rely on payer status alone (i.e., restricted Medicaid) as a proxy for undocumented immigrants in EHR should consider how this may bias their results. As Medicaid expands for undocumented immigrants, statistical assignment may become the preferred method.

Objective: To examine how a novel payer-provider joint venture (JV) between one payer and multiple competitive delivery systems in New Hampshire (NH), which included value-based payment, care management, and non-financial supports, impacted healthcare value and payer and provider group experiences.

Study setting and design: We conducted a mixed-methods study. We used a quasi-experimental longitudinal difference-in-differences design to examine the impact of the JV (which started in January 2016 and ended in December 2020) on healthcare utilization, quality, and spending, using members in Maine (ME) as a control group. We also analyzed patient uptake of the JV's care management program using routinely collected administrative data and assessed payer and provider group leaders' perspectives about the JV via semi-structured interviews.

Data sources and analytic sample: We used administrative and claims data from 2013 to 2019 in a commercially insured population under 65 years in NH and ME. We also used administrative data on care management eligibility and uptake and conducted semi-structured interviews with payer and provider group leaders affiliated with the JV.

Principal findings: The JV was associated with no sustained change in medical utilization, quality, and spending throughout the study period. In the first year of the JV, there was a $142 (95% confidence interval: $41, $243) increase in pharmaceutical spending per member and a 13% (4.4%, 25%) relative increase in days covered for diabetes medications. Only 15% of eligible members engaged in care management, which was a key component of the JV's multi-pronged approach. In a disconnect from the empirical findings, payer and provider group leaders believed that the JV reduced healthcare costs and improved quality.

Conclusions: Our findings provide evidence for future payer-provider JVs and demonstrate the importance of having a valid control group when evaluating JVs and value-based payment arrangements.

Objective: To examine the response of community health center (CHC) quality to quality levels at neighboring CHCs in the presence of non-price competition.

Data setting and design: A quasi-experimental study of US community health centers. Outcome variables were indices that measured overall quality of CHC care. Using patient flow data, we constructed CHC-specific Hirschman-Herfindahl index (HHI) and competitors' composite quality measure. The plausibly exogenous change in characteristics of "competitors' competitors" was exploited to identify the relationship between competition and quality of care, using a generalized two-stage least square model with instrumental variables.

Data sources and analytic sample: Using the Health Center Program Uniform Data System (2014-2018), linked with American Community Survey and Medical Expenditure Panel Survey, we analyzed 1098 unique federally funded CHCs in 50 states and District of Columbia which had at least one neighboring CHC and had non-missing data for 2015-2018 (4226 CHC-years).

Principal findings: Most of CHCs served populations in overlapping geographic markets, with median market concentration decreasing during the study period. A one-percent increase in competitors' quality was associated with a 0.71-percent increase in an index CHC's composite quality (p < 0.01), consisting of a 0.59-percent increase in chronic condition control rates (p < 0.01); a 0.68-percent increase in the screening and assessment rates (p < 0.01); and a 0.78-percent increase in medication management rates (p < 0.01). The association was stronger at CHCs serving a smaller proportion of uninsured patients. No significant quality reaction was observed at CHCs with a percentage of uninsured patients larger than the 75th percentile. We observed no significant associations between HHI and quality.

Conclusions: Increasing competition does not harm quality of care at CHCs. A CHC appears to improve its quality if its competitors improved quality. The beneficial quality effect was less pronounced in CHCs providing a significant proportion of care to uninsured patients, suggesting lack of incentives faced by these CHCs.