Lower comorbidity scores and severity levels in Veterans Health Administration hospitals: a cross-sectional study

BACKGROUND

Previous studies found that documentation of comorbidities differed when Veterans received care within versus outside Veterans Health Administration (VHA). Changes to medical center funding, increased attention to performance reporting, and expansion of Clinical Documentation Improvement programs, however, may have caused coding in VHA to change.

METHODS

Using repeated cross-sectional data, we compared Elixhauser-van Walraven scores and Medicare Severity Diagnosis Related Group (DRG) severity levels for Veterans' admissions across settings and payers over time, utilizing a linkage of VHA and all-payer discharge data for 2012-2017 in seven US states. To minimize selection bias, we analyzed records for Veterans admitted to both VHA and non-VHA hospitals in the same year. Using generalized linear models, we adjusted for patient and hospital characteristics.

RESULTS

Following adjustment, VHA admissions consistently had the lowest predicted mean comorbidity scores (4.44 (95% CI 4.34-4.55)) and lowest probability of using the most severe DRG (22.1% (95% CI 21.4%-22.8%)). In contrast, Medicare-covered admissions had the highest predicted mean comorbidity score (5.71 (95% CI 5.56-5.85)) and highest probability of using the top DRG (35.3% (95% CI 34.2%-36.4%)).

CONCLUSIONS

More effective strategies may be needed to improve VHA documentation, and current risk-adjusted comparisons should account for differences in coding intensity.

Evaluating feedback reports to support documentation of veterans' care preferences in home based primary care

BACKGROUND

To evaluate the effectiveness of delivering feedback reports to increase completion of LST notes among VA Home Based Primary Care (HBPC) teams. The Life Sustaining Treatment Decisions Initiative (LSTDI) was implemented throughout the Veterans Health Administration (VHA) in the United States in 2017 to ensure that seriously ill Veterans have care goals and LST decisions elicited and documented.

METHODS

We distributed monthly feedback reports summarizing LST template completion rates to 13 HBPC intervention sites between October 2018 and February 2020 as the sole implementation strategy. We used principal component analyses to match intervention to 26 comparison sites and used interrupted time series/segmented regression analyses to evaluate the differences in LST template completion rates between intervention and comparison sites. Data were extracted from national databases for VA HBPC in addition to interviews and surveys in a mixed methods process evaluation.

RESULTS

LST template completion rose from 6.3 to 41.9% across both intervention and comparison HBPC teams between March 1, 2018, and February 26, 2020. There were no statistically significant differences for intervention sites that received feedback reports.

CONCLUSIONS

Feedback reports did not increase documentation of LST preferences for Veterans at intervention compared with comparison sites. Observed increases in completion rates across intervention and comparison sites can likely be attributed to implementation strategies used nationally as part of the national roll-out of the LSTDI. Our results suggest that feedback reports alone were not an effective implementation strategy to augment national implementation strategies in HBPC teams.

Behavioral weight management use in the Veterans Health Administration: Sociodemographic and health correlates

INTRODUCTION

Over 40 % of United States Veterans Health Administration (VHA) primary care patients have obesity. Few patients use VHA's flagship weight management program, MOVE! and there is little information on other behavioral weight management program use.

METHODS

The national United States cohort included over 1.5 million primary care patients with obesity, age 18-79, based on VHA administrative data. Gender stratified multivariable logistic regression identified correlates of weight management use in the year after a patient's first primary care appointment (alpha of 0.05). Weight management use was defined as MOVE! or nutrition clinic visits.

RESULTS

The cohort included 121,235 women and 1,521,547 men with 13 % and 7 % using weight management, respectively. Point estimates for specific correlates of use were similar between women and men, and across programs. Black patients were more likely to use weight management than White patients. Several physical and mental health diagnoses were also associated with increased use, such as sleep apnea and eating disorders. Age and distance from VHA were negatively associated with weight management use.

CONCLUSIONS

When assessing multiple types of weight management visits, weight management care in VHA appears to be used more often by some populations at higher risk for obesity. Other groups may need additional outreach, such as those living far from VHA. Future work should focus on outreach and prevention efforts to increase overall use rates. This work could also examine the benefits of tailoring care for populations in greatest need.

Validity of Birth Certificate Data Compared With Hospital Discharge Data in Reporting Maternal Morbidity and Disparities

A growing number of studies are using birth certificate data, despite data-quality concerns, to study maternal morbidity and associated disparities. We examined whether conclusions about the incidence of maternal morbidity, including Black-White disparities, differ between birth certificate data and hospitalization data. Using linked birth certificate and hospitalization data from California and Michigan for 2018 (N=543,469), we found that maternal morbidity measures using birth certificate data alone are substantially underreported and have poor validity. Furthermore, the degree of underreporting in birth certificate data differs between Black and White individuals and results in erroneous inferences about disparities. Overall, Black-White disparities were more modest in the birth certificate data compared with the hospitalization data. Birth certificate data alone are inadequate for studies of maternal morbidity and associated racial disparities.

Association between stillbirth and severe maternal morbidity

BACKGROUND

Severe maternal morbidity has been increasing in the past few decades. Few studies have examined the risk of severe maternal morbidity among individuals with stillbirths vs individuals with live-birth deliveries.

OBJECTIVE

This study aimed to examine the prevalence and risk of severe maternal morbidity among individuals with stillbirths vs individuals with live-birth deliveries during delivery hospitalization as a primary outcome and during the postpartum period as a secondary outcome.

STUDY DESIGN

This was a retrospective cohort study using birth and fetal death certificate data linked to hospital discharge records from California (2008-2018), Michigan (2008-2020), Missouri (2008-2014), Pennsylvania (2008-2014), and South Carolina (2008-2020). Relative risk regression analysis was used to examine the crude and adjusted relative risks of severe maternal morbidity along with 95% confidence intervals among individuals with stillbirths vs individuals with live-birth deliveries, adjusting for birth year, state of residence, maternal sociodemographic characteristics, and the obstetric comorbidity index.

RESULTS

Of the 8,694,912 deliveries, 35,012 (0.40%) were stillbirths. Compared with individuals with live-birth deliveries, those with stillbirths were more likely to be non-Hispanic Black (10.8% vs 20.5%); have Medicaid (46.5% vs 52.0%); have pregnancy complications, including preexisting diabetes mellitus (1.1% vs 4.3%), preexisting hypertension (2.3% vs 6.2%), and preeclampsia (4.4% vs 8.4%); have multiple pregnancies (1.6% vs 6.2%); and reside in South Carolina (7.4% vs 11.6%). During delivery hospitalization, the prevalence rates of severe maternal morbidity were 791 cases per 10,000 deliveries for stillbirths and 154 cases per 10,000 deliveries for live-birth deliveries, whereas the prevalence rates for nontransfusion severe maternal morbidity were 502 cases per 10,000 deliveries for stillbirths and 68 cases per 10,000 deliveries for live-birth deliveries. The crude relative risk for severe maternal morbidity was 5.1 (95% confidence interval, 4.9-5.3), whereas the adjusted relative risk was 1.6 (95% confidence interval, 1.5-1.8). For nontransfusion severe maternal morbidity among stillbirths vs live-birth deliveries, the crude relative risk was 7.4 (95% confidence interval, 7.0-7.7), whereas the adjusted relative risk was 2.0 (95% confidence interval, 1.8-2.3). This risk was not only elevated among individuals with stillbirth during the delivery hospitalization but also through 1 year after delivery (severe maternal morbidity adjusted relative risk, 1.3; 95% confidence interval, 1.1-1.4; nontransfusion severe maternal morbidity adjusted relative risk, 1.2; 95% confidence interval, 1.1-1.3).

CONCLUSION

Stillbirth was found to be an important contributor to severe maternal morbidity.

The impact of volume and neonatal level of care on outcomes of moderate and late preterm infants

OBJECTIVE

Evaluate the relationship of neonatal unit level of care (LOC) and volume with mortality or morbidity in moderate-late preterm (MLP) (32-36 weeks' gestation) infants.

DESIGN

Retrospective cohort study of 650,865 inborn MLP infants in 4976 hospitals-years using 2003-2015 linked administrative data from 4 states. Exposure was combined neonatal LOC and MLP annual volume. The primary outcome was death or morbidity (respiratory distress syndrome, severe intraventricular hemorrhage, necrotizing enterocolitis, sepsis, infection, pneumothorax, extreme length of stay) with components as secondary outcomes. Poisson regression models adjusted for patient characteristics with a random effect for unit were used.

RESULTS

In adjusted models, high-volume level 2 units had a lower risk of the primary outcome compared to low-volume level 3 units (aIRR 0.90 [95% CI 0.83-0.98] vs. aIRR 1.13 [95% CI 1.03-1.24], p < 0.001) CONCLUSION: MLP infants had improved outcomes in high-volume level 2 units compared to low-volume level 3 units in adjusted analysis.

Hospital and Patient Factors Affecting Veterans' Hospital Choice

Veterans enrolled in the Veterans Affairs (VA) health care system gained greater access to non-VA care beginning in 2014. We examined hospital and Veteran characteristics associated with hospital choice. We conducted a longitudinal study of elective hospitalizations 2011 to 2017 in 11 states and modeled patients' choice of VA hospital, large non-VA hospital, or small non-VA hospital in conditional logit models. Patients had higher odds of choosing a hospital with an academic affiliation, better patient experience rating, location closer to them, and a more common hospital type. Patients who were male, racial/ethnic minorities, had higher VA enrollment priority, and had a mental health comorbidity were more likely than other patients to choose a VA hospital than a non-VA hospital. Our findings suggest that patients respond to certain hospital attributes. VA hospitals may need to maintain or achieve high levels of quality and patient experience to attract or retain patients in the future.

Perinatal mental health and pregnancy-associated mortality: opportunities for change

Perinatal mental health conditions have been associated with adverse pregnancy outcomes, including maternal death. This quality improvement project analyzed pregnancy-associated death among veterans with mental health conditions in order to identify opportunities to improve healthcare and reduce maternal deaths. Pregnancy-associated deaths among veterans using Veterans Health Administration (VHA) maternity care benefits between fiscal year 2011 and 2020 were identified from national VHA databases. Deaths among individuals with active mental health conditions underwent individual chart review using a standardized abstraction template adapted from the Centers for Disease Control and Prevention (CDC). Thirty-two pregnancy-associated deaths were identified among 39,720 paid deliveries with 81% (n = 26) occurring among individuals with an active perinatal mental health condition. In the perinatal mental health cohort, most deaths (n = 16, 62%) occurred in the late postpartum period and 42% (n = 11) were due to suicide, homicide, or overdose. Opportunities to improve care included addressing (1) racial disparities, (2) mental health effects of perinatal loss, (3) late postpartum vulnerability, (4) lack of psychotropic medication continuity, (5) mental health conditions in intimate partners, (6) child custody loss, (7) lack of patient education or stigmatizing patient education, and (8) missed opportunities for addressing reproductive health concerns in mental health contexts. Pregnancy-associated deaths related to active perinatal mental health conditions can be reduced. Mental healthcare clinicians, clinical teams, and healthcare systems have opportunities to improve care for individuals with perinatal mental health conditions.

Improving the Maternity Care Safety Net: Establishing Maternal Mortality Surveillance for Non-Obstetric Providers and Institutions

The siloed nature of maternity care has been noted as a system-level factor negatively impacting maternal outcomes. Veterans Health Administration (VA) provides multi-specialty healthcare before, during, and after pregnancy but purchases obstetric care from community providers. VA providers may be unaware of perinatal complications, while community-based maternity care providers may be unaware of upstream factors affecting the pregnancy. To optimize maternal outcomes, the VA has initiated a system-level surveillance and review process designed to improve non-obstetric care for veterans experiencing a pregnancy. This quality improvement project aimed to describe the VA-based maternal mortality review process and to report maternal mortality (pregnancy-related death up to 42 days postpartum) and pregnancy-associated mortality (death from any cause up to 1 year postpartum) among veterans who use VA maternity care benefits. Pregnancies and pregnancy-associated deaths between fiscal year (FY) 2011-2020 were identified from national VA databases. All deaths underwent individual chart review and abstraction that focused on multi-specialty care received at the VA in the year prior to pregnancy until the time of death. Thirty-two pregnancy-associated deaths were confirmed among 39,720 pregnancies (PAMR = 80.6 per 100,000 live births). Fifty percent of deaths occurred among individuals who had experienced adverse social determinants of health. Mental health conditions affected 81%. Half (n = 16, 50%) of all deaths occurred in the late postpartum period (43-365 days postpartum) after maternity care had ended. More than half of these late postpartum deaths (n = 9, 56.2%) were related to suicide, homicide, or overdose. Integration of care delivered during the perinatal period (pregnancy through postpartum) from primary, mental health, emergency, and specialty care providers may be enhanced through a system-based approach to pregnancy-associated death surveillance and review. This quality improvement project has implications for all healthcare settings where coordination between obstetric and non-obstetric providers is needed.

Better performance for right-skewed data using an alternative gamma model

BACKGROUND

The Maximum Likelihood Estimator (MLE) for parameters of the gamma distribution is commonly used to estimate models of right-skewed variables such as costs, hospital length of stay, and appointment wait times in Economics and Healthcare research. The common specification for this estimator assumes the variance is proportional to the square of the mean, which underlies estimation and specification tests. We present a specification in which the variance is directly proportional to the mean.

METHODS

We used simulation experiments to investigate finite sample results, and we used United States Department of Veterans Affairs (VA) healthcare cost data as an empirical example comparing the fit and predictive ability of the models.

RESULTS

Simulation showed the MLE based on a correctly specified alternative has less parameter bias, lower standard errors, and less skewness in distribution than a misspecified standard model. The application to VA healthcare cost data showed the alternative specification can have better R square, smaller root mean squared error, and smaller mean residuals within deciles of predicted values.

CONCLUSIONS

The alternative gamma specification can be a useful alternative to the standard specification for estimating models of right-skewed continuous variables.

Outcomes of Veterans Treated in Veterans Affairs Hospitals vs Non-Veterans Affairs Hospitals

Importance

Many veterans enrolled in the Veterans Affairs (VA) health care system have access to non-VA care through insurance and VA-purchased community care. Prior comparisons of VA and non-VA hospital outcomes have been limited to subpopulations.

Objective

To compare outcomes for 6 acute conditions in VA and non-VA hospitals for younger and older veterans using VA and all-payer discharge data.

Design, Setting, and Participants

This cohort study used a repeated cross-sectional analysis of hospitalization records for acute myocardial infarction (AMI), coronary artery bypass graft (CABG), gastrointestinal (GI) hemorrhage, heart failure (HF), pneumonia, and stroke. Participants included VA enrollees from 11 states at VA and non-VA hospitals from 2012 to 2017. Analysis was conducted from July 1, 2022, to October 18, 2023.

Exposures

Treatment in VA or non-VA hospital.

Main Outcome and Measures

Thirty-day mortality, 30-day readmission, length of stay (LOS), and costs. Average treatment outcomes of VA hospitals were estimated using inverse probability weighted regression adjustment to account for selection into hospitals. Models were stratified by veterans' age (aged less than 65 years and aged 65 years and older).

Results

There was a total of 593 578 hospitalizations and 414 861 patients with mean (SD) age 75 (12) years, 405 602 males (98%), 442 297 hospitalizations of non-Hispanic White individuals (75%) and 73 155 hospitalizations of non-Hispanic Black individuals (12%) overall. VA hospitalizations had a lower probability of 30-day mortality for HF (age ≥65 years, -0.02 [95% CI, -0.03 to -0.01]) and stroke (age <65 years, -0.03 [95% CI, -0.05 to -0.02]; age ≥65 years, -0.05 [95% CI, -0.07 to -0.03]). VA hospitalizations had a lower probability of 30-day readmission for CABG (age <65 years, -0.04 [95% CI, -0.06 to -0.01]; age ≥65 years, -0.05 [95% CI, -0.07 to -0.02]), GI hemorrhage (age <65 years, -0.04 [95% CI, -0.06 to -0.03]), HF (age <65 years, -0.05 [95% CI, -0.07 to -0.03]), pneumonia (age <65 years, -0.04 [95% CI, -0.06 to -0.03]; age ≥65 years, -0.03 [95% CI, -0.04 to -0.02]), and stroke (age <65 years, -0.11 [95% CI, -0.13 to -0.09]; age ≥65 years, -0.13 [95% CI, -0.16 to -0.10]) but higher probability of readmission for AMI (age <65 years, 0.04 [95% CI, 0.01 to 0.06]). VA hospitalizations had a longer mean LOS and higher costs for all conditions, except AMI and stroke in younger patients.

Conclusions and Relevance

In this cohort study of veterans, VA hospitalizations had lower mortality for HF and stroke and lower readmissions, longer LOS, and higher costs for most conditions compared with non-VA hospitalizations with differences by condition and age group. There were tradeoffs between better outcomes and higher resource use in VA hospitals for some conditions.

Differences in use of Veterans Health Administration and non-Veterans Health Administration hospitals by rural and urban Veterans after access expansions

PURPOSE

To examine changes in rural and urban Veterans' utilization of acute inpatient care in Veterans Health Administration (VHA) and non-VHA hospitals following access expansion from the Veterans Choice Act, which expanded eligibility for VHA-paid community hospitalization.

METHODS

Using repeated cross-sectional data of VHA enrollees' hospitalizations in 9 states (AZ, CA, CT, FL, LA, MA, NY, PA, and SC) between 2012 and 2017, we compared rural and urban Veterans' probability of admission in VHA and non-VHA hospitals by payer over time for elective and nonelective hospitalizations using multinomial logistic regression to adjust for patient-level sociodemographic features. We also used generalized linear models to compare rural and urban Veterans' travel distances to hospitals.

FINDINGS

Over time, the probability of VHA-paid community hospitalization increased more for rural Veterans than urban Veterans. For elective inpatient care, rural Veterans' probability of VHA-paid admission increased from 2.9% (95% CI 2.6%-3.2%) in 2012 to 6.5% (95% CI 5.8%-7.1%) in 2017. These changes were associated with a temporal trend that preceded and continued after the implementation of the Veterans Choice Act. Overall travel distances to hospitalizations were similar over time; however, the mean distance traveled decreased from 39.2 miles (95% CI 35.1-43.3) in 2012 to 32.3 miles (95% CI 30.2-34.4) in 2017 for rural Veterans receiving elective inpatient care in VHA-paid hospitals.

CONCLUSIONS

Despite limited access to rural hospitals, these data demonstrate an increase in rural Veterans' use of non-VHA hospitals for acute inpatient care and a small reduction in distance traveled to elective inpatient services.

Chronic Hypertension in Pregnancy and Racial-Ethnic Disparities in Complications

OBJECTIVE

To evaluate whether there are individual- and population-level associations between chronic hypertension and pregnancy complications, and to assess differences across seven racial-ethnic groups.

METHODS

This population-based study used linked vital statistics and hospitalization discharge data from all live and stillbirths in California (2008-2018), Michigan (2008-2020), Oregon (2008-2020), Pennsylvania (2008-2014), and South Carolina (2008-2020). We used multivariable log-binomial regression models to estimate risk ratios (RRs) and population attributable risk (PAR) percentages with 95% CIs for associations between chronic hypertension and several obstetric and neonatal outcomes, selected based on prior evidence and pathologic pathways. We adjusted models for demographic factors (race and ethnicity, payment method, educational attainment), age, body mass index, obstetric history, delivery year, and state, and conducted analyses stratified across seven racial-ethnic groups.

RESULTS

The study included 7,955,713 pregnancies, of which 168,972 (2.1%) were complicated by chronic hypertension. Chronic hypertension was associated with several adverse obstetric and neonatal outcomes, with the largest adjusted PAR percentages observed for preeclampsia with severe features or eclampsia (22.4; 95% CI 22.2-22.6), acute renal failure (13.6; 95% CI 12.6-14.6), and pulmonary edema (10.7; 95% CI 8.9-12.6). Estimated RRs overall were similar across racial-ethnic groups, but PAR percentages varied. The adjusted PAR percentages (95% CI) for severe maternal morbidity-a widely used composite of acute severe events-for people who were American Indian or Alaska Native, Asian, Black, Latino, Native Hawaiian or Other Pacific Islander, White, and Multiracial or Other were 5.0 (1.1-8.8), 3.7 (3.0-4.3), 9.0 (8.2-9.8), 3.9 (3.6-4.3), 11.6 (6.4-16.5), 3.2 (2.9-3.5), and 5.5 (4.2-6.9), respectively.

CONCLUSION

Chronic hypertension accounts for a substantial fraction of obstetric and neonatal morbidity and contributes to higher complication rates, particularly for people who are Black or Native Hawaiian or Other Pacific Islander.

Perinatal Care Measures Are Incomplete If They Do Not Assess The Birth Parent-Infant Dyad As A Whole

Measures of perinatal care quality and outcomes often focus on either the birth parent or the infant. We used linked vital statistics and hospital discharge data to describe a dyadic measure (including both the birth parent and the infant) for perinatal care during the birth hospitalization. In this five-state cohort of 2010-18 births, 21.6 percent of birth parent-infant dyads experienced at least one complication, and 9.6 percent experienced a severe complication. Severe infant complications were eight times more prevalent than severe birth parent complications. Among birth parents with a severe complication, the co-occurrence of a severe infant complication ranged from 2 percent to 51 percent, whereas among infants with a severe complication, the co-occurrence of a severe birth parent complication was rare, ranging from 0.04 percent to 5 percent. These data suggest that measures, clinical interventions, public reporting, and policies focused on either the birth parent or the infant are incomplete in their assessment of a healthy dyad. Thus, clinicians, administrators, and policy makers should evaluate dyadic measures, incentivize positive outcomes for both patients (parent and infant), and create policies that support the health of the dyad.

Association of Sickle Cell Disease With Racial Disparities and Severe Maternal Morbidities in Black Individuals

Importance

Little is known about the association between sickle cell disease (SCD) and severe maternal morbidity (SMM).

Objective

To examine the association of SCD with racial disparities in SMM and with SMM among Black individuals.

Design, Setting, and Participants

This cohort study was a retrospective population-based investigation of individuals with and without SCD in 5 states (California [2008-2018], Michigan [2008-2020], Missouri [2008-2014], Pennsylvania [2008-2014], and South Carolina [2008-2020]) delivering a fetal death or live birth. Data were analyzed between July and December 2022.

Exposure

Sickle cell disease identified during the delivery admission by using International Classification of Diseases, Ninth Revision and Tenth Revision codes.

Main Outcomes and Measures

The primary outcomes were SMM including and excluding blood transfusions during the delivery hospitalization. Modified Poisson regression was used to estimate risk ratios (RRs) adjusted for birth year, state, insurance type, education, maternal age, Adequacy of Prenatal Care Utilization Index, and obstetric comorbidity index.

Results

From a sample of 8 693 616 patients (mean [SD] age, 28.5 [6.1] years), 956 951 were Black individuals (11.0%), of whom 3586 (0.37%) had SCD. Black individuals with SCD vs Black individuals without SCD were more likely to have Medicaid insurance (70.2% vs 64.6%), to have a cesarean delivery (44.6% vs 34.0%), and to reside in South Carolina (25.2% vs 21.5%). Sickle cell disease accounted for 8.9% and for 14.3% of the Black-White disparity in SMM and nontransfusion SMM, respectively. Among Black individuals, SCD complicated 0.37% of the pregnancies but contributed to 4.3% of the SMM cases and to 6.9% of the nontransfusion SMM cases. Among Black individuals with SCD compared with those without, the crude RRs of SMM and nontransfusion SMM during the delivery hospitalization were 11.9 (95% CI, 11.3-12.5) and 19.8 (95% CI, 18.5-21.2), respectively, while the adjusted RRs were 3.8 (95% CI, 3.3-4.5) and 6.5 (95% CI, 5.3-8.0), respectively. The SMM indicators that incurred the highest adjusted RRs included air and thrombotic embolism (4.8; 95% CI, 2.9-7.8), puerperal cerebrovascular disorders (4.7; 95% CI, 3.0-7.4), and blood transfusion (3.7; 95% CI, 3.2-4.3).

Conclusions and Relevance

In this retrospective cohort study, SCD was found to be an important contributor to racial disparities in SMM and was associated with an elevated risk of SMM among Black individuals. Efforts from the research community, policy makers, and funding agencies are needed to advance care among individuals with SCD.

Physical Fitness in Relationship to Depression and Post-Traumatic Stress Disorder During Pregnancy Among U.S. Army Soldiers

Background: Depression and post-traumatic stress disorder (PTSD) are prevalent in pregnancy, especially among military members. These conditions can lead to adverse birth outcomes, yet, there's a paucity of evidence for prevention strategies. Optimizing physical fitness is one understudied potential intervention. We explored associations between prepregnancy physical fitness and antenatal depression and PTSD in soldiers. Materials and Methods: This was a retrospective cohort study of active-duty U.S. Army soldiers with live births between 2011 and 2014, identified with diagnosis codes from inpatient and outpatient care. The exposure was each individual's mean Army physical fitness score from 10 to 24 months before childbirth. The primary outcome was a composite of active depression or PTSD during pregnancy, defined using the presence of a code within 10 months before childbirth. Demographic variables were compared across four quartiles of fitness scores. Multivariable logistic regression models were conducted adjusting for potential confounders selected a priori. A stratified analysis was conducted for depression and PTSD separately. Results: Among 4,583 eligible live births, 352 (7.7%) had active depression or PTSD during pregnancy. Soldiers with the highest fitness scores (Quartile 4) were less likely to have active depression or PTSD in pregnancy (Quartile 4 vs. Quartile 1 adjusted odds ratio 0.55, 95% confidence interval 0.39-0.79). Findings were similar in stratified analyses. Conclusion: In this cohort, the odds of active depression or PTSD during pregnancy were significantly reduced among soldiers with higher prepregnancy fitness scores. Optimizing physical fitness may be a useful tool to reduce mental health burden on pregnancy.

Obstetric Volume and Severe Maternal Morbidity Among Low-Risk and Higher-Risk Patients Giving Birth at Rural and Urban US Hospitals

Importance

Identifying hospital factors associated with severe maternal morbidity (SMM) is essential to clinical and policy efforts.

Objective

To assess associations between obstetric volume and SMM in rural and urban hospitals and examine whether these associations differ for low-risk and higher-risk patients.

Design, Setting, and Participants

This retrospective cross-sectional study of linked vital statistics and patient discharge data was conducted from 2022 to 2023. Live births and stillbirths (≥20 weeks' gestation) at hospitals in California (2004-2018), Michigan (2004-2020), Pennsylvania (2004-2014), and South Carolina (2004-2020) were included. Data were analyzed from December 2022 to May 2023.

Exposures

Annual birth volume categories (low, medium, medium-high, and high) for hospitals in urban (10-500, 501-1000, 1001-2000, and >2000) and rural (10-110, 111-240, 241-460, and >460) counties.

Main Outcome and Measures

The main outcome was SMM (excluding blood transfusion); covariates included age, payer status, educational attainment, race and ethnicity, and obstetric comorbidities. Analyses were stratified for low-risk and higher-risk obstetric patients based on presence of at least 1 clinical comorbidity.

Results

Among more than 11 million urban births and 519 953 rural births, rates of SMM ranged from 0.73% to 0.50% across urban hospital volume categories (high to low) and from 0.47% to 0.70% across rural hospital volume categories (high to low). Risk of SMM was elevated for patients who gave birth at rural hospitals with annual birth volume of 10 to 110 (adjusted risk ratio [ARR], 1.65; 95% CI, 1.14-2.39), 111 to 240 (ARR, 1.37; 95% CI, 1.10-1.70), and 241 to 460 (ARR, 1.26; 95% CI, 1.05-1.51), compared with rural hospitals with greater than 460 births. Increased risk of SMM occurred for low-risk and higher-risk obstetric patients who delivered at rural hospitals with lower birth volumes, with low-risk rural patients having notable discrepancies in SMM risk between low (ARR, 2.32; 95% CI, 1.32-4.07), medium (ARR, 1.66; 95% CI, 1.20-2.28), and medium-high (ARR, 1.68; 95% CI, 1.29-2.18) volume hospitals compared with high volume (>460 births) rural hospitals. Among hospitals in urban counties, there was no significant association between birth volume and SMM for low-risk or higher-risk obstetric patients.

Conclusions and Relevance

In this cross-sectional study of births in US rural and urban counties, risk of SMM was elevated for low-risk and higher-risk obstetric patients who gave birth in lower-volume hospitals in rural counties, compared with similar patients who gave birth at rural hospitals with greater than 460 annual births. These findings imply a need for tailored quality improvement strategies for lower volume hospitals in rural communities.

Trends in Resources for Neonatal Intensive Care at Delivery Hospitals for Infants Born Younger Than 30 Weeks' Gestation, 2009-2020

Importance

In an ideal regionalized system, all infants born very preterm would be delivered at a large tertiary hospital capable of providing all necessary care.

Objective

To examine whether the distribution of extremely preterm births changed between 2009 and 2020 based on neonatal intensive care resources at the delivery hospital.

Design, Setting, and Participants

This retrospective cohort study was conducted at 822 Vermont Oxford Network (VON) centers in the US between 2009 and 2020. Participants included infants born at 22 to 29 weeks' gestation, delivered at or transferred to centers participating in the VON. Data were analyzed from February to December 2022.

Exposures

Hospital of birth at 22 to 29 weeks' gestation.

Main Outcomes and Measures

Birthplace neonatal intensive care unit (NICU) level was classified as A, restriction on assisted ventilation or no surgery; B, major surgery; or C, cardiac surgery requiring bypass. Level B centers were further divided into low-volume (<50 inborn infants at 22 to 29 weeks' gestation per year) and high-volume (≥50 inborn infants at 22 to 29 weeks' gestation per year) centers. High-volume level B and level C centers were combined, resulting in 3 distinct NICU categories: level A, low-volume B, and high-volume B and C NICUs. The main outcome was the change in the percentage of births at hospitals with level A, low-volume B, and high-volume B or C NICUs overall and by US Census region.

Results

A total of 357 181 infants (mean [SD] gestational age, 26.4 [2.1] weeks; 188 761 [52.9%] male) were included in the analysis. Across regions, the Pacific (20 239 births [38.3%]) had the lowest while the South Atlantic (48 348 births [62.7%]) had the highest percentage of births at a hospital with a high-volume B- or C-level NICU. Births at hospitals with A-level NICUs increased by 5.6% (95% CI, 4.3% to 7.0%), and births at low-volume B-level NICUs increased by 3.6% (95% CI, 2.1% to 5.0%), while births at hospitals with high-volume B- or C-level NICUs decreased by 9.2% (95% CI, -10.3% to -8.1%). By 2020, less than half of the births for infants at 22 to 29 weeks' gestation occurred at hospitals with high-volume B- or C-level NICUs. Most US Census regions followed the nationwide trends; for example, births at hospitals with high-volume B- or C-level NICUs decreased by 10.9% [95% CI, -14.0% to -7.8%) in the East North Central region and by 21.1% (95% CI, -24.0% to -18.2%) in the West South Central region.

Conclusions and Relevance

This retrospective cohort study identified concerning deregionalization trends in birthplace hospital level of care for infants born at 22 to 29 weeks' gestation. These findings should serve to encourage policy makers to identify and enforce strategies to ensure that infants at the highest risk of adverse outcomes are born at the hospitals where they have the best chances to attain optimal outcomes.

Racial/ethnic disparities in costs, length of stay, and severity of severe maternal morbidity

BACKGROUND

In contrast to other high-resource countries, the United States has experienced increases in the rates of severe maternal morbidity. In addition, the United States has pronounced racial and ethnic disparities in severe maternal morbidity, especially for non-Hispanic Black people, who have twice the rate as non-Hispanic White people.

OBJECTIVE

This study aimed to examine whether the racial and ethnic disparities in severe maternal morbidity extended beyond the rates of these complications to include disparities in maternal costs and lengths of stay, which could indicate differences in the case severity.

STUDY DESIGN

This study used California's linkage of birth certificates to inpatient maternal and infant discharge data for 2009 to 2011. Of the 1.5 million linked records, 250,000 were excluded because of incomplete data, for a final sample of 1,262,862. Cost-to-charge ratios were used to estimate costs from charges (including readmissions) after adjusting for inflation to December 2017. Mean diagnosis-related group-specific reimbursement was used to estimate physician payments. We used the Centers for Disease Control and Prevention definition of severe maternal morbidity, including readmissions up to 42 days after delivery. Adjusted Poisson regression models estimated the differential risk of severe maternal morbidity for each racial or ethnic group, compared with the non-Hispanic White group. Generalized linear models estimated the associations of race and ethnicity with costs and length of stay.

RESULTS

Asian or Pacific Islander, Non-Hispanic Black, Hispanic, and other race or ethnicity patients all had higher rates of severe maternal morbidity than non-Hispanic White patients. The largest disparity was between non-Hispanic White and non-Hispanic Black patients, with unadjusted overall rates of severe maternal morbidity of 1.34% and 2.62%, respectively (adjusted risk ratio, 1.61; P<.001). Among patients with severe maternal morbidity, the adjusted regression estimates showed that non-Hispanic Black patients had 23% (P<.001) higher costs (marginal effect of $5023) and 24% (P<.001) longer hospital stays (marginal effect of 1.4 days) than non-Hispanic White patients. These effects changed when cases, such as cases where a blood transfusion was the only indication of severe maternal morbidity, were excluded, with 29% higher costs (P<.001) and 15% longer length of stay (P<.001). For other racial and ethnic groups, the increases in costs and length of stay were smaller than those observed for non-Hispanic Black patients, and many were not significantly different from non-Hispanic White patients. Hispanic patients had higher rates of severe maternal morbidity than non-Hispanic White patients; however, Hispanic patients had significantly lower costs and length of stay than non-Hispanic White patients.

CONCLUSION

There were racial and ethnic differences in the costs and length of stay among patients with severe maternal morbidity across the groupings that we examined. The differences were especially large for non-Hispanic Black patients compared with non-Hispanic White patients. Non-Hispanic Black patients experienced twice the rate of severe maternal morbidity; in addition, the higher relative costs and longer lengths of stay for non-Hispanic Black patients with severe maternal morbidity support greater case severity in that population. These findings suggest that efforts to address racial and ethnic inequities in maternal health need to consider differences in case severity in addition to the differences in the rates of severe maternal morbidity and that these differences in case severity merit additional investigation.

An exploratory analysis of factors associated with spontaneous preterm birth among pregnant veterans with post-traumatic stress disorder

BACKGROUND

Pregnant veterans with post-traumatic stress disorder (PTSD) are at increased risk for spontaneous preterm birth, yet the underlying reasons are unclear. We examined factors associated with spontaneous preterm birth among pregnant veterans with active PTSD.

METHODS

This was an observational study of births from administrative databases reimbursed by the Veterans Health Association (VA) between 2005 and 2015. Singleton livebirths among veterans with active PTSD within 12 months prior to childbirth were included. The primary outcome was spontaneous preterm birth. Maternal demographics, psychiatric history, and pregnancy complications were evaluated as exposures. Covariates significant on bivariate analysis, as well as age and race/ethnicity as a social construct, were included in multivariable logistic regression to identify factors associated with spontaneous preterm birth. Additional analyses stratified significant covariates by the presence of active concurrent depression and explored interactions between antidepressant use and preeclampsia.

RESULTS

Of 3,242 eligible births to veterans with active PTSD, 249 (7.7%) were spontaneous preterm births. The majority of veterans with active PTSD (79.1%) received some type of mental health treatment, and active concurrent depression was prevalent (61.4%). Preeclampsia/eclampsia (adjusted odds ratio [aOR] 3.30, 95% confidence interval [CI] 1.67-6.54) and ≥6 antidepressant medication dispensations within 12 months prior to childbirth (aOR 1.89, 95% CI 1.29-2.77) were associated with spontaneous preterm birth. No evidence of interaction was seen between antidepressant use and preeclampsia on spontaneous preterm birth (p=0.39). Findings were similar when stratified by active concurrent depression.

CONCLUSION

Among veterans with active PTSD, preeclampsia/eclampsia and ≥6 antidepressant dispensations were associated with spontaneous preterm birth. While the results do not imply that people should discontinue needed antidepressants during pregnancy in veterans with PTSD, research into these factors might inform preterm birth prevention strategies for this high-risk population.

Gynecologist Supply Deserts Across the VA and in the Community

BACKGROUND

The Veterans Health Administration (VA) refers patients to community providers for specialty services not available on-site. However, community-level specialist shortages may impede access to care.

OBJECTIVE

Compare gynecologist supply in veterans' county of residence versus at their VA site.

DESIGN

We identified women veteran VA patients from fiscal year (FY) 2017 administrative data and assessed availability of a VA gynecologist within 50 miles (hereafter called "local") of veterans' VA homesites (per national VA organizational survey data). For the same cohort, we then assessed community-level gynecologist availability; counties with < 2 gynecologists/10,000 women (per the Area Health Resource File) were "inadequate-supply" counties. We examined the proportion of women veterans with local VA gynecologist availability in counties with inadequate versus adequate gynecologist supply, stratified by individual and VA homesite characteristics. Chi-square tests assessed statistical differences.

PARTICIPANTS

All women veteran FY2017 VA primary care users nationally.

MAIN MEASURES

Availability of a VA gynecologist within 50 miles of a veteran's VA homesite; county-level "inadequate-supply" of gynecologists.

KEY RESULTS

Among 407,482 women, 9% were in gynecologist supply deserts (i.e., lacking local VA gynecologist and living in an inadequate-supply county). The sub-populations with the highest proportions in gynecologist supply deserts were rural residents (24%), those who got their primary care at non-VAMC satellite clinics (13%), those who got their care at a site without a women's clinic (13%), and those with American Indian or Alaska Native (12%), or white (12%) race. Among those in inadequate-supply counties, 59.9% had gynecologists at their local VA; however, 40.1% lacked a local VA gynecologist.

CONCLUSIONS

Most veterans living in inadequate-supply counties had local VA gynecology care, reflecting VA's critical role as a safety net provider. However, for those in gynecologist supply deserts, expanded transportation options, modified staffing models, or tele-gynecology hubs may offer solutions to extend VA gynecology capacity.

Successful Discharge of Short Stay Veterans from VA Community Living Centers

The Veterans Health Administration (VHA) long-term care rebalancing initiative encouraged VA Community Living Centers (CLCs) to shift from long-stay custodial-focused care to short-stay skilled and rehabilitative care. Using all VA CLC admissions during 2007-2010 categorized as needing short-stay rehabilitation or skilled nursing care, we assessed the patient and facility rates of successful discharge to the community (SDC) of these short-stay Veterans. We found large variation in inter- as well as intra- facility SDC rates across the rehabilitation and skilled nursing short-stay cohorts. We discuss how our results can help guide VHA policy directed at delivering high-quality short-stay CLC care for Veterans.

Evaluating implementation strategies to support documentation of veterans' care preferences

OBJECTIVE

To evaluate the effectiveness of feedback reports and feedback reports + external facilitation on completion of life-sustaining treatment (LST) note the template and durable medical orders. This quality improvement program supported the national roll-out of the Veterans Health Administration (VA) LST Decisions Initiative (LSTDI), which aims to ensure that seriously-ill veterans have care goals and LST decisions elicited and documented.

DATA SOURCES

Primary data from national databases for VA nursing homes (called Community Living Centers [CLCs]) from 2018 to 2020.

STUDY DESIGN

In one project, we distributed monthly feedback reports summarizing LST template completion rates to 12 sites as the sole implementation strategy. In the second involving five sites, we distributed similar feedback reports and provided robust external facilitation, which included coaching, education, and learning collaboratives. For each project, principal component analyses matched intervention to comparison sites, and interrupted time series/segmented regression analyses evaluated the differences in LSTDI template completion rates between intervention and comparison sites.

DATA COLLECTION METHODS

Data were extracted from national databases in addition to interviews and surveys in a mixed-methods process evaluation.

PRINCIPAL FINDINGS

LSTDI template completion rose from 0% to about 80% throughout the study period in both projects' intervention and comparison CLCs. There were small but statistically significant differences for feedback reports alone (comparison sites performed better, coefficient estimate 3.48, standard error 0.99 for the difference between groups in change in trend) and feedback reports + external facilitation (intervention sites performed better, coefficient estimate -2.38, standard error 0.72).

CONCLUSIONS

Feedback reports + external facilitation was associated with a small but statistically significant improvement in outcomes compared with comparison sites. The large increases in completion rates are likely due to the well-planned national roll-out of the LSTDI. This finding suggests that when dissemination and support for widespread implementation are present and system-mandated, significant enhancements in the adoption of evidence-based practices may require more intensive support.

Health Care Access Expansions and Use of Veterans Affairs and Other Hospitals by Veterans

This cohort study examines changes in the use of Veterans Affairs (VA) and non-VA hospitals by VA enrollees and mortality associated with these policies.

A data-driven health index for neonatal morbidities

Whereas prematurity is a major cause of neonatal mortality, morbidity, and lifelong impairment, the degree of prematurity is usually defined by the gestational age (GA) at delivery rather than by neonatal morbidity. Here we propose a multi-task deep neural network model that simultaneously predicts twelve neonatal morbidities, as the basis for a new data-driven approach to define prematurity. Maternal demographics, medical history, obstetrical complications, and prenatal fetal findings were obtained from linked birth certificates and maternal/infant hospitalization records for 11,594,786 livebirths in California from 1991 to 2012. Overall, our model outperformed traditional models to assess prematurity which are based on GA and/or birthweight (area under the precision-recall curve was 0.326 for our model, 0.229 for GA, and 0.156 for small for GA). These findings highlight the potential of using machine learning techniques to predict multiple prematurity phenotypes and inform clinical decisions to prevent, diagnose and treat neonatal morbidities.

•

Traditional definitions of prematurity based on gestational age need to be updated

•

Deep learning of maternal clinical data improves predictions of neonatal morbidity

•

Data-driven model leverages birthweight, type of delivery and maternal race

•

Accurate risk prediction can inform clinical decisions

Evaluating Care in Safety Net Hospitals: Clinical Outcomes and Neonatal Intensive Care Unit Quality of Care in California

OBJECTIVES

To examine the characteristics of safety net (sn) and non-sn neonatal intensive care units (NICUs) in California and evaluate whether the site of care is associated with clinical outcomes.

STUDY DESIGN

This population-based retrospective cohort study of 34 snNICUs and 104 non-snNICUs included 22 081 infants born between 2014 and 2018 with a birth weight of 401-1500 g or gestational age of 22-29 weeks. Quality of care as measured by the Baby-MONITOR score and rates of survival without major morbidity were compared between snNICUs and non-snNICUs.

RESULTS

Black and Hispanic infants were cared for disproportionately in snNICUs, where care and outcomes varied widely. We found no significant differences in Baby-Measure Of Neonatal InTensive care Outcomes Research (MONITOR) scores (z-score [SD]: snNICUs, -0.31 [1.3]; non-snNICUs, 0.03 [1.1]; P = .1). Among individual components, infants in snNICUs exhibited lower rates of human milk nutrition at discharge (-0.64 [1.0] vs 0.27 [0.9]), lower rates of no health care-associated infection (-0.27 [1.1] vs 0.14 [0.9]), and higher rates of no hypothermia on admission (0.39 [0.7] vs -0.25 [1.1]). We found small but significant differences in survival without major morbidity (adjusted rate, 65.9% [95% CI, 63.9%-67.9%] for snNICUs vs 68.3% [95% CI, 67.0%-69.6%] for non-snNICUs; P = .02) and in some of its components; snNICUs had higher rates of necrotizing enterocolitis (3.8% [3.4%-4.3%] vs 3.1% [95% CI, 2.8%-3.4%]) and mortality (95% CI, 7.1% [6.5%-7.7%] vs 6.6% [6.2%-7.0%]).

CONCLUSIONS

snNICUs achieved similar performance as non-snNICUs in quality of care except for small but significant differences in any human milk at discharge, infection, hypothermia, necrotizing enterocolitis, and mortality.

A Comprehensive Analysis of the Costs of Severe Maternal Morbidity

INTRODUCTION

The objectives of this study were to include readmissions and physician costs in the estimates of total costs of severe maternal morbidity (SMM), to consider the effect of SMM on maternal length of stay (LOS), and to examine these for the more restricted definition of SMM that excludes transfusion-only cases.

METHODS

California linked birth certificate-patient discharge data for 2009 through 2011 (n = 1,262,862) with complete costs and LOS were used in a secondary data analysis. Cost-to-charge ratios were used to estimate costs from charges, adjusting for inflation. Physician payments were estimated from the mean payments for specific diagnosis-related groups. Generalized linear models estimated the association between SMM and costs and LOS.

RESULTS

Excluding readmissions and physician costs, SMM was associated with a 60% increase in hospital costs (marginal effect [ME], $3,550) and a 33% increase in LOS (ME 0.9 days). These increased to 70% (ME $5,806) and 46% (ME 1.3 days) when physician costs and readmissions were included. The effects of SMM were roughly one-half as large for patients who only required a blood transfusion (49% [ME $4,056] and 31% [ME 0.9 days]) as for patients who had another indicator for SMM (93% [ME $7,664] and 62% [ME 1.7 days]).

CONCLUSIONS

Postpartum hospital readmissions and physician costs are important and previously unreported contributors to the costs of SMM. Excess costs and LOS associated with SMM vary considerably by indication. Cost effects were larger than the LOS effects, indicating that SMM increases treatment intensity beyond increasing LOS, and decreasing SMM may have broader health and cost benefits than previously understood.

Quantifying the variation in neonatal transport referral patterns using network analysis

OBJECTIVE

Regionalized care reduces neonatal morbidity and mortality. This study evaluated the association of patient characteristics with quantitative differences in neonatal transport networks.

STUDY DESIGN

We retrospectively analyzed prospectively collected data for infants <28 days of age acutely transported within California from 2008 to 2012. We generated graphs representing bidirectional transfers between hospitals, stratified by patient attribute, and compared standard network analysis metrics.

RESULT

We analyzed 34,708 acute transfers, representing 1594 unique transfer routes between 271 hospitals. Density, centralization, efficiency, and modularity differed significantly among networks drawn based on different infant attributes. Compared to term infants and to those transported for medical reasons, network metrics identify greater degrees of regionalization for preterm and surgical patients (more centralized and less dense, respectively [p < 0.001]).

CONCLUSION

Neonatal interhospital transport networks differ by patient attributes as reflected by differences in network metrics, suggesting that regionalization should be considered in the context of a multidimensional system.

Birth Volume and Geographic Distribution of US Hospitals With Obstetric Services From 2010 to 2018

IMPORTANCE

Timely access to clinically appropriate obstetric services is critical to the provision of high-quality perinatal care.

OBJECTIVE

To examine the geographic distribution, proximity, and urban adjacency of US obstetric hospitals by annual birth volume.

DESIGN, SETTING, AND PARTICIPANTS

This retrospective population-based cohort study identified US hospitals with obstetric services using the American Hospital Association (AHA) Annual Survey of Hospitals and Centers for Medicare & Medicaid provider of services data from 2010 to 2018. Obstetric hospitals with 10 or more births per year were included in the study. Data analysis was performed from November 6, 2020, to April 5, 2021.

EXPOSURE

Hospital birth volume, defined by annual birth volume categories of 10 to 500, 501 to 1000, 1001 to 2000, and more than 2000 births.

MAIN OUTCOMES AND MEASURES

Outcomes assessed by birth volume category were percentage of births (from annual AHA data), number of hospitals, geographic distribution of hospitals among states, proximity between obstetric hospitals, and urban adjacency defined by urban influence codes, which classify counties by population size and adjacency to a metropolitan area.

RESULTS

The study included 26 900 hospital-years of data from 3207 distinct US hospitals with obstetric services, reflecting 34 054 951 associated births. Most infants (19 327 487 [56.8%]) were born in hospitals with more than 2000 births/y, and 2 528 259 (7.4%) were born in low-volume (10-500 births/y) hospitals. More than one-third of obstetric hospitals (37.4%; 10 064 hospital-years) were low volume. A total of 46 states had obstetric hospitals in all volume categories. Among low-volume hospitals, 18.9% (1904 hospital-years) were not within 30 miles of any other obstetric hospital and 23.9% (2400 hospital-years) were within 30 miles of a hospital with more than 2000 deliveries/y. Isolated hospitals (those without another obstetric hospital within 30 miles) were more frequently low volume, with 58.4% (1112 hospital-years) located in noncore rural areas.

CONCLUSIONS AND RELEVANCE

In this cohort study, marked variations were found in birth volume, geographic distribution, proximity, and urban adjacency among US obstetric hospitals from 2010 to 2018. The findings related to geographic isolation and rural-urban distribution of low-volume obstetric hospitals suggest the need to balance proximity with volume to optimize effective referral and access to high-quality perinatal care.

The effect of data aggregation on estimations of nurse staffing and patient outcomes

OBJECTIVE

To examine how estimates of the association between nurse staffing and patient length of stay (LOS) change with data aggregation over varying time periods and settings, and statistical controls for unobserved heterogeneity.

DATA SOURCES/STUDY SETTING

Longitudinal secondary data from October 2002 to September 2006 for 215 intensive care units and 438 general acute care units at 143 facilities in the Veterans Affairs (VA) health care system.

RESEARCH DESIGN

This retrospective observational study used unit-level panel data to analyze the association between nurse staffing and LOS. This association was measured over both a month-long and a year-long period, with and without fixed effects.

DATA COLLECTION

We used VA administrative data to obtain patient data on the severity of illness and LOS, as well as labor hours and wages for each unit by month.

PRINCIPAL FINDINGS

Overall, shorter LOS was associated with higher nurse staffing hours and lower proportions of hours provided by licensed professional nurses (LPNs), unlicensed personnel, and contract staff. Estimates of the association between nurse staffing and LOS changed in magnitude when aggregating data over years instead of months, in different settings, and when controlling for unobserved heterogeneity.

CONCLUSIONS

Estimating the association between nurse staffing and LOS is contingent on the time period of analysis and specific methodology. In future studies, researchers should be aware of these differences when exploring nurse staffing and patient outcomes.

Postpartum Transition of Care: Racial/Ethnic Gaps in Veterans' Re-Engagement in VA Primary Care after Pregnancy

INTRODUCTION

Pregnancy presents an opportunity to engage veterans in health care. Guidelines recommend primary care follow-up in the year postpartum, but loss to follow-up is common, poorly quantified, and especially important for those with gestational diabetes (GDM) and hypertension. Racial maternal inequities are well-documented and might be exacerbated by differential postpartum care. This study explores variation in postpartum re-engagement in U.S. Department of Veteran Affairs health care system (VA) primary care to identify potential racial/ethnic inequities in this care transition.

METHODS

We conducted a complete case analysis of the 2005-2014 national VA birth cohort (n = 18,414), and subcohorts of veterans with GDM (n = 1,253), and hypertensive disorders of pregnancy (HDP; n = 2,052) using VA-reimbursed discharge claims and outpatient data. Outcomes included incidence of any VA primary care visit in the postpartum year; in age-adjusted logistic regression, we explored race/ethnicity as a primary predictor.

RESULTS

In the year after a VA-covered birth, the proportion of veterans with one or more primary care visit was 53.8% overall, and slightly higher in the GDM (56.0%) and HDP (57.4%) subcohorts. In adjusted models, the odds of VA primary care follow-up were significantly lower for Black/African American (odds ratio, 0.87; 95% confidence interval, 0.81-0.93), Asian (odds ratio, 0.76; 95% confidence interval, 0.61-0.95), and Hawaiian/other Pacific Islander (odds ratio, 0.73; 95% confidence interval, 0.55-0.96) veterans, compared with White veterans. Among the subcohorts with GDM or HDP, there were no significant associations between primary care and race/ethnicity.

CONCLUSIONS

One-half of veterans re-engage in VA primary care after childbirth, with significant racial differences in this care transition. Re-engagement for those with the common pregnancy complications of HDP and GDM is only slightly higher, and less than 60%. The potential for innovations like VA maternity care coordinators to address such gaps merits attention.

Access to risk-appropriate hospital care and disparities in neonatal outcomes in racial/ethnic groups and rural-urban populations

Variations in infant and neonatal mortality continue to persist in the United States and in other countries based on both socio-demographic characteristics, such as race and ethnicity, and geographic location. One potential driver of these differences is variations in access to risk-appropriate delivery care. The purpose of this article is to present the importance of delivery hospitals on neonatal outcomes, discuss variation in access to these hospitals for high-risk infants and their mothers, and to provide insight into drivers for differences in access to high-quality perinatal care using the available literature. This review also illustrates the lack of information on a number of topics that are crucial to the development of evidence-based interventions to improve access to appropriate delivery hospital services and thus optimize the outcomes of high-risk mothers and their newborns.

Association of Costs and Days at Home With Transfer Hospital in Home

IMPORTANCE

New Centers for Medicare & Medicaid Services waivers created a payment mechanism for hospital at home services. Although it is well established that direct admission to hospital at home from the community as a substitute for hospital care provides superior outcomes and lower cost, the effectiveness of transfer hospital at home-that is, completing hospitalization at home-is unclear.

OBJECTIVE

To evaluate the outcomes of the transfer component of a Veterans Affairs (VA) Hospital in Home program (T-HIH), taking advantage of natural geographical limitations in a program's service area.

DESIGN, SETTING, AND PARTICIPANTS

In this quality improvement study, T-HIH was offered to veterans residing in Philadelphia, Pennsylvania, and their outcomes were compared with those of propensity-matched veterans residing in adjacent Camden, New Jersey, who were admitted to the VA hospital from 2012 to 2018. Data analysis was performed from October 2019 to May 2020.

INTERVENTION

Enrollment in the T-HIH program.

MAIN OUTCOMES AND MEASURES

The main outcomes were hospital length of stay, 30-day and 90-day readmissions, VA direct costs, combined VA and Medicare costs, mortality, 90-day nursing home use, and days at home after hospital discharge. An intent-to-treat analysis of cost and utilization was performed.

RESULTS

A total of 405 veterans (mean [SD] age, 66.7 [0.83] years; 399 men [98.5%]) with medically complex conditions, primarily congestive heart failure and chronic obstructive pulmonary disease exacerbations (mean [SD] hierarchical condition categories score, 3.54 [0.16]), were enrolled. Ten participants could not be matched, so analyses were performed for 395 veterans (all of whom were men), 98 in the T-HIH group and 297 in the control group. For patients in the T-HIH group compared with the control group, length of stay was 20% lower (6.1 vs 7.7 days; difference, 1.6 days; 95% CI, -3.77 to 0.61 days), VA costs were 20% lower (-$5910; 95% CI, -$13 049 to $1229), combined VA and Medicare costs were 22% lower (-$7002; 95% CI, -$14 314 to $309), readmission rates were similar (23.7% vs 23.0%), the numbers of nursing home days were significantly fewer (0.92 vs 7.45 days; difference, -6.5 days; 95% CI, -12.1 to -0.96 days; P = .02), and the number of days at home was 18% higher (81.4 vs 68.8 days; difference, 12.6 days; 95% CI, 3.12 to 22.08 days; P = .01).

CONCLUSIONS AND RELEVANCE

In this study, T-HIH was significantly associated with increased days at home and less nursing home use but was not associated with increased health care system costs.

VA home-based primary care interdisciplinary team structure varies with Veterans' needs, aligns with PACE regulation

BACKGROUND

Interdisciplinary team (IDT) care is central to home-based primary care (HBPC) of frail elders. Traditionally, all HBPC disciplines managed a patient (Full IDT), a costly approach to maintain. The recent PACE (Program of All-inclusive Care for the Elderly) regulation provides for a flexible approach of annual assessments from a core team with involvement of additional disciplines dependent upon patient needs (Core+). Current Department of Veterans Affairs (VA) HBPC guidance specifies Full IDTs care for medically complex and functionally impaired Veterans similar to PACE participants. We evaluated whether VA HBPC has adopted the flexible structure of the PACE regulation, aligned to Veteran needs.

DESIGN

Cross-sectional analysis.

SETTING

All 139 VA HBPC programs administered across 379 sites.

PARTICIPANTS

About 55,173 Veterans enrolled in HBPC during fiscal year 2018.

MEASUREMENTS

Patients' HBPC physician, nurse, psychologist/psychiatrist, social worker, therapist, dietitian, and pharmacist visits were grouped into interdisciplinary team types. Patient frailty was classified using VA HNHR v2 (High-Need High-Risk version 2, a measure of high, medium, and low risk of long-term institutionalization). Medical complexity was measured by clusters of impairment in the JEN frailty index (JFI). JFI clusters were validated by VA's Nosos measure to project cost and Care Assessment Need (CAN) measure of hospitalization and mortality risk.

RESULTS

HBPC provided Full IDT care to 21%, Core+ care to 54%, and Home Health+ (HHA+) care (skilled home health services plus additional disciplines, without primary care) to 16% of Veterans. Team type was associated with medical complexity (X² 2863.5 [66 df], p < 0.0001). High-risk Veterans (72% of sample) were more likely to receive Full IDT care (X² 62.9, 1 df), p < 0.0001), while low-risk Veterans (28%) were more likely to receive HHA+ care (X² 314.8, 1 df, p < 0.0001).

CONCLUSION

There is a strong association between HBPC team patterns and patient frailty, indicating tailoring of care to match Veteran needs.

Short-Term and Long-Term Educational Outcomes of Infants Born Moderately and Late Preterm

OBJECTIVE

To assess the relationship of moderate and late preterm birth (32^0/7-36^6/7 weeks) to long-term educational outcomes.

STUDY DESIGN

We hypothesized that moderate and late preterm birth would be associated with adverse outcomes in elementary school. To test this, we linked vital statistics patient discharge data from the Office of Statewide Health Planning and Development including birth outcomes, to the 2015-2016 school year administrative data of a large, urban school district (n = 72 316). We compared the relative risk of moderate and late preterm and term infants for later adverse neurocognitive and behavioral outcomes in kindergarten through the 12th grade.

RESULTS

After adjusting for socioeconomic status, compared with term birth, moderate and late preterm birth was associated with an increased risk of low performance in mathematics and English language arts, chronic absenteeism, and suspension. These risks emerged in kindergarten through grade 2 and remained in grades 3-5, but seemed to wash out in later grades, with the exception of suspension, which remained through grades 9-12.

CONCLUSIONS

Confirming our hypothesis, moderate and late preterm birth was associated with adverse educational outcomes in late elementary school, indicating that it is a significant risk factor that school districts could leverage when targeting early intervention. Future studies will need to test these relations in geographically and socioeconomically diverse school districts, include a wider variety of outcomes, and consider how early interventions moderate associations between birth outcomes and educational outcomes.

Development and use of an adjusted nurse staffing metric in the neonatal intensive care unit

OBJECTIVE

To develop a nurse staffing prediction model and evaluate deviation from predicted nurse staffing as a contributor to patient outcomes.

DATA SOURCES

Secondary data collection conducted 2017-2018, using the California Office of Statewide Health Planning and Development and the California Perinatal Quality Care Collaborative databases. We included 276 054 infants born 2008-2016 and cared for in 99 California neonatal intensive care units (NICUs).

STUDY DESIGN

Repeated-measures observational study. We developed a nurse staffing prediction model using machine learning and hierarchical linear regression and then quantified deviation from predicted nurse staffing in relation to health care-associated infections, length of stay, and mortality using hierarchical logistic and linear regression.

DATA COLLECTION METHODS

We linked NICU-level nurse staffing and organizational data to patient-level risk factors and outcomes using unique identifiers for NICUs and patients.

PRINCIPAL FINDINGS

An 11-factor prediction model explained 35 percent of the nurse staffing variation among NICUs. Higher-than-predicted nurse staffing was associated with decreased risk-adjusted odds of health care-associated infection (OR: 0.79, 95% CI: 0.63-0.98), but not with length of stay or mortality.

CONCLUSIONS

Organizational and patient factors explain much of the variation in nurse staffing. Higher-than-predicted nurse staffing was associated with fewer infections. Prospective studies are needed to determine causality and to quantify the impact of staffing reforms on health outcomes.

Racial and Ethnic Differences Over Time in Outcomes of Infants Born Less Than 30 Weeks' Gestation

OBJECTIVES

To examine changes in care practices over time by race and ethnicity and whether the decrease in hospital mortality and severe morbidities has benefited infants of minority over infants of white mothers.

METHODS

Infants 22 to 29 weeks' gestation born between January 2006 and December 2017 at a Vermont Oxford Network center in the United States were studied. We examined mortality and morbidity rate differences and 95% confidence intervals for African American and Hispanic versus white infants by birth year. We tested temporal differences in mortality and morbidity rates between white and African American or Hispanic infants using a likelihood ratio test on nested binomial regression models.

RESULTS

Disparities for certain care practices such as antenatal corticosteroids and for some in-hospital outcomes have narrowed over time for minority infants. Compared with white infants, African American infants had a faster decline for mortality, hypothermia, necrotizing enterocolitis, and late-onset sepsis, whereas Hispanic infants had a faster decline for mortality, respiratory distress syndrome, and pneumothorax. Other morbidities showed a constant rate difference between African American and Hispanic versus white infants over time. Despite the improvements, outcomes including hypothermia, mortality, necrotizing enterocolitis, late-onset sepsis, and severe intraventricular hemorrhage remained elevated by the end of the study period, especially among African American infants.

CONCLUSIONS

Racial and ethnic disparities in vital care practices and certain outcomes have decreased. That the quality deficit among minority infants occurred for several care practice measures and potentially modifiable outcomes suggests a critical role for quality improvement initiatives tailored for minority-serving hospitals.

Racial Segregation and Inequality in the Neonatal Intensive Care Unit for Very Low-Birth-Weight and Very Preterm Infants

IMPORTANCE

Racial and ethnic minorities receive lower-quality health care than white non-Hispanic individuals in the United States. Where minority infants receive care and the role that may play in the quality of care received is unclear.

OBJECTIVE

To determine the extent of segregation and inequality of care of very low-birth-weight and very preterm infants across neonatal intensive care units (NICUs) in the United States.

DESIGN, SETTING, AND PARTICIPANTS

This cohort study of 743 NICUs in the Vermont Oxford Network included 117 982 black, Hispanic, Asian, and white infants born at 401 g to 1500 g or 22 to 29 weeks' gestation from January 2014 to December 2016. Analysis began January 2018.

MAIN OUTCOMES AND MEASURES

The NICU segregation index and NICU inequality index were calculated at the hospital level as the Gini coefficients associated with the Lorenz curves for black, Hispanic, and Asian infants compared with white infants, with NICUs ranked by proportion of white infants for the NICU segregation index and by composite Baby-MONITOR (Measure of Neonatal Intensive Care Outcomes Research) score for the NICU inequality index.

RESULTS

Infants (36 359 black [31%], 21 808 Hispanic [18%], 5920 Asian [5%], and 53 895 white [46%]) were segregated among the 743 NICUs by race and ethnicity (NICU segregation index: black: 0.50 [95% CI, 0.46-0.53], Hispanic: 0.58 [95% CI, 0.54-0.61], and Asian: 0.45 [95% CI, 0.40-0.50]). Compared with white infants, black infants were concentrated at NICUs with lower-quality scores, and Hispanic and Asian infants were concentrated at NICUs with higher-quality scores (NICU inequality index: black: 0.07 [95% CI, 0.02-0.13], Hispanic: -0.10 [95% CI, -0.17 to -0.04], and Asian: -0.26 [95% CI, -0.32 to -0.19]). There was marked variation among the census regions in weighted mean NICU quality scores (range: -0.69 to 0.85). Region of residence explained the observed inequality for Hispanic infants but not for black or Asian infants.

CONCLUSIONS AND RELEVANCE

Black, Hispanic, and Asian infants were segregated across NICUs, reflecting the racial segregation of minority populations in the United States. There were large differences between geographic regions in NICU quality. After accounting for these differences, compared with white infants, Asian infants received care at higher-quality NICUs and black infants, at lower-quality NICUs. Explaining these patterns will require understanding the effects of sociodemographic factors and public policies on hospital quality, access, and choice for minority women and their infants.

Birth Hospitalization Costs and Days of Care for Mothers and Neonates in California, 2009-2011

OBJECTIVE

To provide population-based estimates of the hospital-related costs of maternal and newborn care, and how these vary by gestational age and birth weight.

STUDY DESIGN

We conducted a retrospective analysis of 2009-2011 California in-hospital deliveries at nonfederal hospitals with the infant and maternal discharge data successfully (96%) linked to birth certificates. Cost-to-charge ratios were used to estimate costs from charges. Physician hospital payments were estimated by mean diagnosis related group-specific reimbursement and costs were adjusted for inflation to December 2017 values. After exclusions for incomplete or missing data, the final sample was 1 265 212.

RESULTS

The mean maternal costs for all in-hospital deliveries was $8204, increasing to $13 154 for late preterm (32-36 weeks) and $22 702 for very preterm (<32 weeks) mothers. The mean cost for all newborns was $6389: $2433 for term infants, $22 102 for late preterm, $223 931 for very preterm infants, and $317 982 for extremely preterm infants (<28 weeks). Preterm infants were 8.1% of cases but incurred 60.9% of costs; for very preterm and extremely preterm infants, these shares were 1.0% and 36.5%, and 0.4% and 20.0%, respectively. Overall, mothers incurred 56% of the total costs during the delivery hospitalization.

CONCLUSIONS

Both maternal and neonatal costs are skewed, with this being much more pronounced for infants. Preterm birth is much more expensive than term delivery, with the additional costs predominately incurred by the infants. The small share of infants who require extensive stays in neonatal intensive care incur a large share of neonatal costs and these costs have increased over time.

The Impact of Medicaid Enrollment on Veterans Health Administration Enrollees' Behavioral Health Services Use

OBJECTIVE

To examine Veterans Health Administration (VA) enrollees' use of VA services for treatment of behavioral health conditions (BHCs) after gaining Medicaid, and if VA reliance varies by complexity of BHCs.

DATA SOURCES/STUDY SETTING

VA and Medicaid Analytic eXtract utilization data from 31 states, 2006-2010.

STUDY DESIGN

A retrospective, longitudinal study of Veterans enrolled in VA care in the year before and year after enrollment in Medicaid among 7,249 nonelderly Veterans with serious mental illness (SMI), substance use disorder (SUD), posttraumatic stress disorder (PTSD), depression, or other BHCs.

DATA COLLECTION/EXTRACTION METHODS

Utilization and VA reliance (proportion of care received at VA) for BH outpatient and inpatient services in unadjusted and adjusted analyses.

PRINCIPAL FINDINGS

In adjusted analyses, we found that overall Veterans did not significantly change their use of VA outpatient BH services after Medicaid enrollment. In beta-binomial models predicting VA BH outpatient reliance, veterans with SMI (IRR = 1.38, p < .05), PTSD (IRR = 1.62, p < .01), and depression (IRR = 1.36, p < .05) had higher reliance than veterans with other BHCs after Medicaid enrollment.

CONCLUSIONS

While veterans did not change the amount of VA outpatient BH services they used after enrolling in Medicaid, the proportion of care they received through VA or Medicaid varied by BHC.

Validation of the JEN frailty index in the National Long-Term Care Survey community population: identifying functionally impaired older adults from claims data

Background

Use of a claims-based index to identify persons with physical function impairment and at risk for long-term institutionalization would facilitate population health and comparative effectiveness research. The JEN Frailty Index [JFI] is comprised of diagnosis domains representing impairments and multimorbid clusters with high long-term institutionalization [LTI] risk. We test the index’s discrimination of activities-of-daily-living [ADL] dependency and 1-year LTI and mortality in a nationally representative sample of over 12,000 Medicare beneficiaries, and compare long-term community survival stratified by ADL and JFI.

Methods

2004 U.S. National Long-Term Care Survey data were linked to Medicare, Minimum Data Set, Veterans Health Administration files and vital statistics. ADL dependencies, JFI score, age and sex were measured at baseline survey. ADL and JFI groups were cross-tabulated generating likelihood ratios and classification statistics. Logistic regression compared discrimination (areas under receiver operating characteristic curves), multivariable calibration and accuracy of the JFI and, separately, ADLs, in predicting 1-year outcomes. Hall-Wellner bands facilitated contrasts of JFI- and ADL-stratified 5-year community survival.

Results

Likelihood ratios rose evenly across JFI risk categories. Areas under the curves of functional dependency at ≥3 and ≥ 2 for JFI, age and sex models were 0.807 [95% c.i.: 0.795, 0.819] and 0.812 [0.801, 0.822], respectively. The area under the LTI curve for JFI and age (0.781 [0.747, 0.815]) discriminated less well than the ADL-based model (0.829 [0.799, 0.860]). Community survival separated by JFI strata was comparable to ADL strata.

Conclusions

The JEN Frailty Index with demographic covariates is a valid claims-based measure of concurrent activities-of-daily-living impairments and future long-term institutionalization risk in older populations lacking functional information.

Electronic supplementary material

The online version of this article (10.1186/s12913-018-3689-2) contains supplementary material, which is available to authorized users.

Attributable Cost of Dementia: Demonstrating Pitfalls of Ignoring Multiple Health Care System Utilization

OBJECTIVES

To determine dementia prevalence and costs attributable to dementia using Veterans Health Administration (VHA) data with and without Medicare data.

DATA SOURCES

VHA inpatient, outpatient, purchased care and other data and Medicare enrollment, claims, and assessments in fiscal year (FY) 2013.

STUDY DESIGN

Analyses were conducted with VHA data alone and with combined VHA and Medicare data. Dementia was identified from a VHA sanctioned list of ICD-9 diagnoses. Attributable cost of dementia was estimated using recycled predictions.

DATA COLLECTION

Veterans age 65 and older who used VHA and were enrolled in Traditional Medicare in FY 2013 (1.9 million).

PRINCIPAL FINDINGS

VHA records indicated the prevalence of dementia in FY 2013 was 4.8 percent while combined VHA and Medicare data indicated the prevalence was 7.4 percent. Attributable cost of dementia to VHA was, on average, $10,950 per veteran per year (pvpy) using VHA alone and $6,662 pvpy using combined VHA and Medicare data. Combined VHA and Medicare attributable cost of dementia was $11,285 pvpy. Utilization attributed to dementia using VHA data alone was lower for long-term institutionalization and higher for supportive care services than indicated in combined VHA and Medicare data.

CONCLUSIONS

Better planning for clinical and cost-efficient care requires VHA and Medicare to share data for veterans with dementia and likely more generally.

Selection of Higher Risk Pregnancies into Veterans Health Administration Programs: Discoveries from Linked Department of Veterans Affairs and California Birth Data

OBJECTIVE

To describe variation in payer and outcomes in Veterans' births.

DATA/SETTING

Secondary data analyses of deliveries in California, 2000-2012.

STUDY DESIGN

We performed a retrospective, population-based study of all live births to Veterans (confirmed via U.S. Department of Veterans Affairs (VA) enrollment records), to identify payer and variations in outcomes among: (1) Veterans using VA coverage and (2) Veteran vs. all other births. We calculated odds ratios (aOR) adjusted for age, race, ethnicity, education, and obstetric demographics.

METHODS

We anonymously linked VA administrative data for all female VA enrollees with California birth records.

PRINCIPAL FINDINGS

From 2000 to 2012, we identified 17,495 births to Veterans. VA covered 8.6 percent (1,508), Medicaid 17.3 percent, and Private insurance 47.6 percent. Veterans who relied on VA health coverage had more preeclampsia (aOR 1.4, CI 1.0-1.8) and more cesarean births (aOR 1.2, CI 1.0-1.3), and, despite similar prematurity, trended toward more neonatal intensive care (NICU) admissions (aOR 1.2, CI 1.0-1.4) compared to Veterans using other (non-Medicaid) coverage. Overall, Veterans' birth outcomes (all-payer) mirrored California's birth outcomes, with the exception of excess NICU care (aOR 1.15, CI 1.1-1.2).

CONCLUSIONS

VA covers a higher risk fraction of Veterans' births, justifying maternal care coordination and attention to the maternal-fetal impacts of Veterans' comorbidities.

Effect of Repetitive Transcranial Magnetic Stimulation on Treatment-Resistant Major Depression in US Veterans: A Randomized Clinical Trial

IMPORTANCE

Treatment-resistant major depression (TRMD) in veterans is a major clinical challenge given the high risk for suicidality in these patients. Repetitive transcranial magnetic stimulation (rTMS) offers the potential for a novel treatment modality for these veterans.

OBJECTIVE

To determine the efficacy of rTMS in the treatment of TRMD in veterans.

DESIGN, SETTING, AND PARTICIPANTS

A double-blind, sham-controlled randomized clinical trial was conducted from September 1, 2012, to December 31, 2016, in 9 Veterans Affairs medical centers. A total of 164 veterans with TRD participated.

INTERVENTIONS

Participants were randomized to either left prefrontal rTMS treatment (10 Hz, 120% motor threshold, 4000 pulses/session) or to sham (control) rTMS treatment for up to 30 treatment sessions.

MAIN OUTCOMES AND MEASURES

The primary dependent measure of the intention-to-treat analysis was remission rate (Hamilton Rating Scale for Depression score ≤10, indicating that depression is in remission and not a clinically significant burden), and secondary analyses were conducted on other indices of posttraumatic stress disorder, depression, hopelessness, suicidality, and quality of life.

RESULTS

The 164 participants had a mean (SD) age of 55.2 (12.4) years, 132 (80.5%) were men, and 126 (76.8%) were of white race. Of these, 81 were randomized to receive active rTMS and 83 to receive sham. For the primary analysis of remission, there was no significant effect of treatment (odds ratio, 1.16; 95% CI, 0.59-2.26; P = .67). At the end of the acute treatment phase, 33 of 81 (40.7%) of those in the active treatment group achieved remission of depressive symptoms compared with 31 of 83 (37.4%) of those in the sham treatment group. Overall, 64 of 164 (39.0%) of the participants achieved remission.

CONCLUSIONS AND RELEVANCE

A total of 39.0% of the veterans who participated in this trial experienced clinically significant improvement resulting in remission of depressive symptoms; however, there was no evidence of difference in remission rates between the active and sham treatments. These findings may reflect the importance of close clinical surveillance, rigorous monitoring of concomitant medication, and regular interaction with clinic staff in bringing about significant improvement in this treatment-resistant population.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT01191333.

Hospice Care of Veterans in Medicare Advantage and Traditional Medicare: A Risk-Adjusted Analysis

OBJECTIVES

To compare the quality of end-of-life care in Medicare Advantage (MA) and traditional Medicare (TM), specifically, receipt and length of hospice care.

DESIGN

Retrospective analysis of administrative data.

SETTING

Hospice care.

PARTICIPANTS

Veterans dually enrolled in the Veterans Health Administration (VHA) and MA or TM who died between 2008 and 2013 (N = 1,515,441).

MEASUREMENTS

Outcomes studied included use and duration of hospice care. Use of a VHA-enrolled population allowed for risk adjustment that is otherwise challenging when studying MA.

RESULTS

Adjusted analyses revealed that MA beneficiaries were more likely to receive hospice than TM beneficiaries; results corroborate published non-risk-adjusted analyses. MA beneficiaries had shorter hospice duration; this is an opposite direction of effect than non-risk-adjusted analyses. Results were robust to multiple sensitivity analyses limiting the cohort to individuals in MA and TM who had equal opportunity for their comorbidities to be captured. Removing risk adjustment resulted in results that mirrored those in the existing published literature.

CONCLUSION

Our work provides two important insights regarding MA that are important to consider as enrollment in this insurance mechanism grows. First, MA beneficiaries received poorer-quality end-of-life care than TM beneficiaries, as ascertained by exposure to hospice. Second, any comparisons made between MA and TM require proper risk adjustment to obtain correct directions of effect. We encourage the Centers for Medicare & Medicaid Services to make comorbidity data specific to MA enrollees available to researchers for these purposes.