Median Household Income Quartile Research Articles

Correlation of socioeconomic status and lead concentration in tap water in Missouri

Lead exposure can have devastating health consequences. This effect is especially detrimental to the developing brain of children. Tap water is the most common source of lead in our diet. There is evidence that lower socioeconomic populations are at higher risk of exposure to tap water with a higher concentration of contaminants. We hypothesized that lead concentration is higher in the tap water of communities with lower per capita income and lower median household income. Using publicly available data from the annual water quality reports, we analyzed the 90th percentile lead levels for municipalities in the state of Missouri. Our results demonstrated that cities in the highest per capita income quartile had significantly lower 90th percentile lead levels than those in the lowest per capita income quartile (2.62±0.28 ppb vs. 4.74±0.83 ppb). Similarly, the highest median household income cities had significantly lower 90th percentile lead levels than those in the lowest median household income quartile (2.44±0.29 ppb vs. 4.62±0.88 ppb). In view of the long-lasting and irreversible health consequences of lead exposure, these findings have important implications when it comes to allocation of resources for mitigating effects of lead on health of populations.

Utilization of epilepsy surgery in the United States: A study of the National Inpatient Sample investigating the roles of race, socioeconomic status, and insurance.

Background: Epilepsy is estimated to affect 70 million people worldwide and is medically refractory in 30% of cases.Methods: This is a retrospective cross-sectional study using a US database from 2012 to 2014 to identify patients aged ≥18 years admitted to the hospital with epilepsy as the primary diagnosis. The sampled population was weighted using Healthcare Cost and Utilization Project guidelines. Procedural ICD-9 codes were utilized to stratify the sampled population into two cohorts: resective surgery and implantation or stimulation procedure.Results: Query of the database yielded 152,925 inpatients, of which 8535 patients underwent surgical intervention. The nonprocedural group consisted of 76,000 White patients (52.6%) and 28,390 Black patients (19.7%) while the procedural group comprised 5550 White patients (64%) and 730 Black patients (8.6%) (P < 0.001). Patients with Medicare were half as likely to receive a surgical procedure (14.8% vs. 28.4%) while patients with private insurance were twice as likely to receive a procedure (53.4% vs. 29.3%), both were statistically significant (P < 0.01). Those in the lowest median household income quartile by zip code (<$40,000) were 68% less likely to receive a procedure (21.5% vs. 31.4%) while the highest income quartile was 133% more likely to receive a procedure (26.1% vs. 19.5%). Patients from rural and urban nonteaching hospitals were, by a wide margin, less likely to receive a surgical procedure.Conclusion: We demonstrate an area of need and significant improvement at institutions that have the resources and capability to perform epilepsy surgery. The data show that institutions may not be performing enough epilepsy surgery as a result of racial and socioeconomic bias. Admissions for epilepsy continue to increase without a similar trend for epilepsy surgery despite its documented effectiveness. Race, socioeconomic status, and insurance all represent significant barriers in access to epilepsy surgery. The barriers can be remedied by improving referral patterns and implementing cost-effective measures to improve inpatient epilepsy services in rural and nonteaching hospitals.

Impact of Insurance and Income levels on Survival Rates in Prolactinomas

<h3>Purpose/Objective(s)</h3> Prolactinoma is the most common type (∼40%) of functional pituitary adenoma. Patients with prolactinoma generally have a good prognosis with appropriate treatment. However, disparities in insurance and socioeconomic status can influence the outcomes of illnesses that have good prognoses. This study aims to analyze the impact of insurance and socioeconomic status on the survival rates of prolactinoma patients across the United States. <h3>Materials/Methods</h3> Patients with prolactinomas diagnosed from 2004 to 2016 were identified from the National Cancer Database (NCDB). Patients with the histology code 8271 in the target area C75.1 (pituitary gland) were included. Patients with unknown status for age, sex, race, insurance status, income quartile, Charlson-Deyo comorbidity score, treatment modality, or overall survival (OS) were excluded. OS was evaluated with the Kaplan-Meier and Cox proportional hazards methods. <h3>Results</h3> A total of 2391 patients with prolactinomas were identified. The median age was 38 years. 8.1% of patients were uninsured, 64.5% of patients had private insurance or managed care. 13.5% of patients had Medicaid, 12.0% of patients had Medicare, and 1.8% of patients had other government insurance. 72.9% of patients were white, and 19.8% of patients were from households in the lowest income quartile based on 2016 US census data. Using univariate analysis, only median household income quartile was statistically associated with OS: patients in the lowest income quartile (less than $40,277 annually) were associated with poorer OS (HR 1.596, 95% CI 1.03-2.48, <i>P</i> < 0.05) compared to patients in higher income quartiles. With multivariate analysis adjusting for age, sex, race, Charlson-Deyo comorbidity score, and insurance status, statistically significant differences in OS were seen for patients in the lowest income quartile (HR 1.993, 95% CI 1.24-3.21, <i>P</i> < 0.05) compared to patients in higher income quartiles. Poorer outcomes were also associated with older age at diagnosis (HR 1.063, 95% CI 1.05-1.08, <i>P</i> < 0.05) and female sex (HR 1.600, 95% CI 1.12-2.31, <i>P</i> < 0.05). <h3>Conclusion</h3> This NCDB study demonstrates the statistically significant disparity in outcomes for prolactinoma patients in the lower income quartile levels. Since prolactinomas represent the most common form of pituitary tumors, it is important to recognize the impact of income disparity on outcomes and develop strategies to improve clinical outcome for these patients.

National Disparities in Autologous Breast Reconstruction

PURPOSE: Autologous breast reconstruction has evolved from more morbid procedures that sacrificed the patient’s abdominal muscle (the TRAM or transverse rectus abdominus muscle flap) to more elegant autologous reconstructions termed “perforator” flaps that spare fascia and muscle to harvest only the adipose tissue. Commercial insurers have recognized the higher technical demand for perforator flaps relative to other autologous reconstructions by creating separate procedural codes with significantly higher professional fees. This study examined whether a perforator flap procedure code unavailable with Federally issued Medicare or Medicaid disproportionally incentivizes perforator flaps among the commercially insured and, subsequently, patients from a higher socioeconomic status. METHODS: Autologous reconstructions performed between 2008 and 2014 were reviewed from the National Inpatient Sample using the ICD-9-CM procedure codes 85.72, 85.73, 85.74, 85.75, and 85.76. Extracted variables included age, race, comorbidities, hospital type, hospital region, insurance payer type, and median household income quartile. Autologous breast reconstruction was subdivided into microvascular perforator flaps (85.74, 85.75, 85.76), microvascular TRAM flaps (85.73), and pedicled TRAM flaps (85.72). Demographics, comorbidities, and access to care were compared between cohorts by chi-squared and ANOVA tests. A logistic regression comparing microvascular reconstructions only was created to predict the effects of insurance, geography, income quartile, and race on the likelihood of perforator flap reconstruction while controlling for age and comorbidities. RESULTS: After querying and weighting National Inpatient Sample data, 33,246 microvascular perforator flap breast reconstructions, 16,804 microvascular TRAM flap reconstructions, and 16,918 pedicled TRAM flap reconstructions were compared. The majority of patients undergoing autologous reconstruction were Whites (64.1%), with a mean age of 51.1 years. Patients receiving microvascular perforator flaps had fewer total comorbidities than patients receiving microvascular TRAM (P < 0.001) or pedicled TRAM (P = 0.003) flaps. Perforator flaps were significantly more likely among the commercially insured (perforator flap: 85.8% versus microvascular TRAM: 75.9% versus pedicled TRAM: 75.0%, P < 0.001), while TRAM flaps were more likely among patients with Medicare or Medicaid (perforator flap: 14.2% versus microvascular TRAM: 24.1% versus pedicled TRAM: 25.0%, P < 0.001). Patients of higher income quartiles were significantly more likely to receive perforator flap autologous reconstruction (P < 0.001). When comparing microvascular reconstruction, logistic regression revealed an odds ratio of 1.72 (P < 0.001) for perforator flaps among the commercially insured when compared with patients with Medicare or Medicaid. Income trends paralleled insurance status. When compared with the lowest income quartile, the second quartile had an odds ratio of 1.11 (P = 0.003) for perforator flap reconstruction, the third 1.07 (P = 0.029), and the fourth 1.36 (P < 0.001). Compared with rural locations, urban nonteaching hospitals had an odds ratio of 2.42 (P < 0.001) for perforator flap reconstruction and urban teaching hospitals had an odds ratio of 4.08 (P < 0.001). Asian patients had a higher odds ratio of receiving perforator flaps than White patients, with an odds ratio of 1.16 (P = 0.010). Black and Hispanic patients had comparable rates as White patients, with odds ratios of 0.97 and 0.95, respectively (P = 0.342, P = 0.192). CONCLUSIONS: Reimbursement incentives disproportionally favor perforator flap autologous reconstruction among the commercially insured. Differences across insurance status exaggerate already existing disparities in breast reconstruction across socioeconomic status.

Patient Characteristics Associated with Delays to Benign Gynecologic Surgery: Impact of the COVID-19 Pandemic

Study ObjectiveTo examine how demographic and socioeconomic characteristics impact timing of minimally invasive gynecologic surgery (MIGS) before and during the COVID-19 pandemic.DesignRetrospective cohort study using electronic medical record data. Primary outcome was interval between referral to MIGS and date of surgery.SettingTertiary-level MIGS division in the southeast US.Patients or ParticipantsHistorical cohort undergoing surgery with MIGS 2014-2016 (n=377) and cohort in 2020 referred during the pandemic (n=191).InterventionsLaparoscopic hysterectomy, myomectomy, adnexal surgery, or excision of endometriosis.Measurements and Main ResultsPatient demographics (race, age, marital status, language, insurance, and socioeconomic factors) were evaluated for significant associations with surgical delay. Patients with fibroids and abnormal uterine bleeding had a shorter interval to surgery (median 95 days, range 66-133) compared to patients with chronic pelvic pain (median 127 days, range 73-274). Our model adjusting for surgical indication revealed that single patients were 2.13 times as likely to wait >90 days (95% CI 1.35-3.36) compared to partnered patients prior to the pandemic. Additionally, those in the lowest quartile of median household income (<$42,572 vs > $75,020; OR 2.42, 95% CI 1.32, 4.46) and those from zip codes with the highest proportion of population in poverty (≥ 0.20 vs <0.07; OR 1.93, 95% CI 1.04, 3.6) were more likely to wait > 90 days. However, all of these differences disappeared during the pandemic. There were no differences in time to surgery by race, ethnicity, language, population density, markers of education by zip code, or insurance before or during the pandemic.ConclusionHistorically, race and socioeconomic factors are associated with decreased access to MIGS and vulnerable populations were disproportionately affected by the COVID-19 pandemic. Despite this, we found decreased time to surgery at our institution, and previous socioeconomic disparities associated with scheduling delays were improved during the pandemic, suggesting improved equitable access to tertiary-level MIGS.

The Association of Socioeconomic Status and Discharge Destination with 30-Day Readmission after Ischemic Stroke

ObjectivesThis study aimed to explore the association of socioeconomic status and discharge destination with 30-day readmission after ischemic stroke. Materials and MethodsWe examined 30-day all-cause readmission among patients hospitalized for ischemic stroke in states of Arkansas, Iowa, and Wisconsin in 2016 and 2017 and New York in 2016 using Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases. ResultsAmong the 52301 patients included, 51.1% were female. The 30-day readmission rates were 10.2%, 8.2%, 9.3%, 10.4%, 11.6%, and 11.2% for age group 18-34, 35-44, 45-54, 55-64, 65-74, and ≥75 years, respectively (p<0.001). In Generalized Estimating Equation analysis, patients with Medicare and Medicaid insurance were more likely to be readmitted, compared with private insurance, (adjusted Odds Ratio [aOR] 1.37, 95% CI 1.23-1.53; and aOR 1.26, 95% CI 1.09-1.45, respectively). Patients in the bottom quartile of zip code level median household income had higher 30-day readmission rate (12.4%) than those in the 2nd, 3rd and 4th quartile (10.3%, 10.1%, and 10.7%, respectively, p<0.001). Compared with those discharged home with self-care which had the lowest readmission rate (8.4%), patients who left against medical advice had the highest readmission rate (18.6%; aOR 2.23, 95% CI 1.75-2.83), followed by rehabilitation and skilled nursing facilities (13.2%; aOR 1.33, 95% CI 1.22-1.46), and home with home health care (11.3%, aOR 1.18, 95% CI 1.08-1.28). ConclusionsSocioeconomic status and discharged destination affect readmission after stroke. These results provide evidence to inform vulnerable patient population as targets for readmission prevention.

MP28-06 THE ASSOCIATION OF ANXIETY IN YOUNGER VS OLDER MEN WITH BENIGN PROSTATIC HYPERPLASIA

You have accessJournal of UrologyBenign Prostatic Hyperplasia: Epidemiology & Evaluation (MP28)1 Sep 2021MP28-06 THE ASSOCIATION OF ANXIETY IN YOUNGER VS OLDER MEN WITH BENIGN PROSTATIC HYPERPLASIA Kevin A. George, Sarah S. Makhani, Adam Nolte, George Wayne, and Alan Nieder Kevin A. GeorgeKevin A. George More articles by this author , Sarah S. MakhaniSarah S. Makhani More articles by this author , Adam NolteAdam Nolte More articles by this author , George WayneGeorge Wayne More articles by this author , and Alan NiederAlan Nieder More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002025.06AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Men with benign prostatic hyperplasia (BPH) commonly have lower urinary tract symptoms (LUTS), including frequency, urgency, and hesitancy, which cause significant distress. Recent evidence has shown a bidirectional relationship between LUTS and mental health. This study aims to assess the association of anxiety in younger men with BPH as compared to older men. METHODS: Data were obtained from the 2016 National Inpatient Sample (NIS). NIS is the largest all-payer inpatient database for United States hospitalizations. For analysis, we selected a subset of this sample, which included male hospitalizations over the age of 18 with BPH. Men between the ages of 18-64 were compared with men 65 years or older. Univariate and multivariate logistic regressions were used to determine the association between age and anxiety, reporting odds ratios and their corresponding 95% confidence intervals. Multivariate analysis controlled for race, major comorbidities (diabetes, hypertension, coronary artery disease), hospital location, insurance status, median household income, and overall disease burden. Sampling weights were used to generate nationally representative estimates. RESULTS: Among the 1,501,714 men with BPH, 142,959 (9.5%) had comorbid anxiety while 1,358,754 (90.5%) did not. Of the 262,124 men 18-64 years old, 37,725 (14.4%) had anxiety, while of the remaining 1,239,589 men 65+ years old, 105,234 (8.5%) had anxiety (p <0.001). Accordingly, upon multivariate analysis, the odds of anxiety were 1.5 times greater in the younger cohort as compared to the older cohort (adjusted OR=0.55, p <0.001, 95% CI [0.54, 0.57]). Additionally, multivariate analysis revealed significantly greater odds of anxiety in Black, Hispanic, and Asian men, men with Medicare or Medicaid, men in the first, second, and third quartiles of median household income, and men with hypertension and diabetes mellitus. CONCLUSIONS: This study shows a higher prevalence of anxiety in younger individuals (18-64 yrs) with BPH as compared to older individuals (65+ yrs). To our knowledge, this is the first study to examine the association between anxiety and patient age in individuals with BPH in a nationally representative sample. These findings may have implications for the biopsychosocial approach to the management of BPH. Source of Funding: None © 2021 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 206Issue Supplement 3September 2021Page: e482-e483 Advertisement Copyright & Permissions© 2021 by American Urological Association Education and Research, Inc.MetricsAuthor Information Kevin A. George More articles by this author Sarah S. Makhani More articles by this author Adam Nolte More articles by this author George Wayne More articles by this author Alan Nieder More articles by this author Expand All Advertisement Loading ...

Admission trends in pediatric isolated linear skull fracture across the United States.

Pediatric isolated linear skull fractures commonly result from head trauma and rarely require surgery, yet patients are often admitted to the hospital-a costly care plan. In this study, the authors utilized a national database to investigate trends in admission for skull fractures across the United States. Children younger than 18 years with isolated linear skull fracture, according to ICD-9 diagnosis codes in the Kids' Inpatient Database of the Healthcare and Utilization Project (HCUP), who presented between 2003 and 2016 were included. HCUP collected data in 2003, 2006, 2009, 2012, and 2016. Children with a depressed skull fracture, multiple traumatic injuries, and acute intracranial findings were excluded. Sample-level data were translated into population-level data by using an HCUP-specific discharge weight. Overall, 11,355 patients (64% males) were admitted to 1605 hospitals. National admissions decreased from 3053 patients in 2003 to 1203 in 2016. The mean ± SD age at admission also decreased from 6.3 ± 5.9 years to 1.2 ± 3.0 years (p < 0.001). The proportion of patients in the lowest quartile of median household income increased by 9%, while that in the highest income quartile decreased by 7% (p < 0.001). Admission was generally more common in the summer months (June, July, and August) and on weekdays (68%). The mean ± SD hospital length of stay decreased from 2.0 ± 3.1 days to 1.4 ± 1.4 days between 2003 and 2012, and then increased to 2.1 ± 6.8 days in 2016 (p < 0.001). When adjusted for inflation, the mean total hospital charges increased from $13,099 to $21,204 (p < 0.001). The greatest proportion of admissions was in the South (35%), and the lowest was in the Northeast (17%). The proportion of patients admitted to large hospitals increased (59% to 72%, p < 0.001), which corresponded to a decrease in patients admitted to small hospitals (16% to 9%, p < 0.001). Overall, the total proportion of admissions to rural hospitals decreased by 6%, and that to urban teaching centers increased by 15% (p < 0.001). Since 2003, no child has undergone a neurosurgical procedure or died as an inpatient. This study identified a general nationwide decrease in admissions for pediatric linear isolated skull fracture, but associated costs increased. Admissions became less common at smaller rural hospitals and more common at larger urban teaching hospitals. This patient population required no inpatient neurosurgical intervention after 2003.

The Role of Upfront Autologous Hematopoietic Cell Transplantation in the Treatment of Mantle Cell Lymphoma, a Population Based Study Using the National Cancer Data Base (NCDB)

Introduction: There are no prospective randomized data supporting the use of upfront consolidative autologous stem cell transplant (ASCT) in newly diagnosed mantle cell lymphoma (MCL) patients (pts). The aim of this population-based study in MCL pts is to better define the impact of frontline ASCT on overall survival (OS), and to identify factors influencing treatment selection.Methods: Using the National Cancer Data Base (NCDB), we identified pts with newly diagnosed MCL treated from 2004-2012 with multi-agent chemo alone or in combination with consolidative ASCT as part of first line treatment. Univariable and multivariable Cox proportional hazards regression analyses were used to identify predictors for treatment selection and OS, including age, gender, race, stage, Charlson-Deyo comorbidity index (CDCI), income, education, insurance, treatment facility type, distance to treatment facility and year of diagnosis. Income was recorded as quartiles of median annual household income and education as percentage of persons with less than high school education in pts' residence census. Including these same factors, a propensity score matched analysis (PSMA) was performed to further account for confounding variables.Results: We identified 16,035 MCL pts with known survival times. The following groups were excluded: no chemo (n=3,193), chemo status/type unknown (n=1,056), chemo contraindicated (n=199), single-agent chemo (n=983), received allogeneic SCT (n=79) or ASCT data missing (n=235). The final analysis included 10,290 pts. Median age was 65 years (range 20-90); 66% were > 60 years, 73% male, 93% Caucasian, and 78% had stage III-IV. CDCI was 0 in 79%, 1 in 16% and ≥2 in 5%. Forty percent of pts were treated at academic institutions and 49% travelled >11.5 miles to receive treatment. Medicare was the most common insurance type (48%), followed by private (43%) and Medicaid (4%). Median follow up was 34 months (interquartile range (IQR) 16-60). Median OS for the entire cohort was 64 months; 5- and 10-year OS rates were 52% and 33%, respectively. Seventeen percent of the pts (n=1774) underwent upfront ASCT, and there was a trend of increased use from 2004-2012 (11.9% in 2004 and 21.2% in 2012) that resulted in an annual increase of 1.2% (p<0.001). 31% of pts ≤60 and 25% of pts ≤70 years received ASCT. For the ASCT group, median age was 58 years (IQR 52-64; range 23-77); 89 pts (5%) were ≥70 years of age. Predictors for ASCT use included (p<0.001 for each): stage III-IV (odds ratio (OR) 2.4), having private insurance (OR 2.71), higher education (OR 2.1), no transition of care (diagnosis and treatment done at the same facility) (OR 3.15), longer distance (> 11.5 miles) to treatment facility (OR 1.73) and year of diagnosis ≥2009 (OR 1.62) (Figure 1). Median follow up was 44 months (IQR 28-69) for ASCT group and 31 months (IQR 14-58) for chemo only group. Five and 10-year OS rates for ASCT and chemo group were 77%, 57%, 47% and 28%, respectively. After adjusting for other covariates, the use of ASCT was associated with superior OS (HR: 0.46; 95% CI 0.41-0.52; p<0.001). Other factors associated with superior OS were: female gender (HR 0.92, p=0.01), income (HR 0.92, p<0.001), private insurance (HR 0.71, p<0.001), no transition of care (HR 0.82, p<0.001) and diagnosis year ≥2009 (HR 0.91, p=0.005). Predictors for inferior OS were: age >60 (HR 1.62, p<0.001), African American race (HR 1.23, p=0.009), stage III-IV (HR 1.58, p<0.001) and CDCI ≥2 (HR 2.0, p<0.001). Using a subset of 1582 pts treated with ASCT and 1582 matched controls treated with chemo only, PSMA confirmed the OS benefit associated with ASCT (HR 0.5; CI 0.44-0.57, p<0.001). This OS benefit was observed in both younger (≤60 years: HR 0.44, CI 0.37-0.53, p<0.001) and older pts (>60 years: HR 0.59, CI 0.46-0.68; p<0.001) (Figure 2).Conclusion: In this large MCL cohort, the use of upfront consolidative ASCT was associated with superior OS when compared to chemo alone. Less than one third of young MCL pts treated with multi-agent chemo received consolidative ASCT, and this was influenced by sociodemographic factors with patterns favoring pts with higher education and private insurance. While one caveat in interpreting this retrospective study is that ASCT may have been preferentially applied to pts who responded to induction therapy, these data suggest clinical efficacy of ASCT in MCL. Prospective randomized trials are needed for further validation. [Display omitted] DisclosuresMajhail:Sanofi: Honoraria; Anthem, Inc.: Consultancy.

Race, Income, and Survival in Stage III Colon Cancer: CALGB 89803 (Alliance).

BackgroundDisparities in colon cancer outcomes have been reported across race and socioeconomic status, which may reflect, in part, access to care. We sought to assess the influences of race and median household income (MHI) on outcomes among colon cancer patients with similar access to care.MethodsWe conducted a prospective, observational study of 1206 stage III colon cancer patients enrolled in the CALGB 89803 randomized adjuvant chemotherapy trial. Race was self-reported by 1116 White and 90 Black patients at study enrollment; MHI was determined by matching 973 patients’ home zip codes with publicly available US Census 2000 data. Multivariate analyses were adjusted for baseline sociodemographic, clinical, dietary, and lifestyle factors. All statistical tests were 2-sided.ResultsOver a median follow-up of 7.7 years, the adjusted hazard ratios for Blacks (compared with Whites) were 0.94 (95% confidence interval [CI] = 0.66 to 1.35, P = .75) for disease-free survival, 0.91 (95% CI = 0.62 to 1.35, P = .65) for recurrence-free survival, and 1.07 (95% CI = 0.73 to 1.57, P = .73) for overall survival. Relative to patients in the highest MHI quartile, the adjusted hazard ratios for patients in the lowest quartile were 0.90 (95% CI = 0.67 to 1.19, Ptrend = .18) for disease-free survival, 0.89 (95% CI = 0.66 to 1.22, Ptrend = .14) for recurrence-free survival, and 0.87 (95% CI = 0.63 to 1.19, Ptrend = .23) for overall survival.ConclusionsIn this study of patients with similar health-care access, no statistically significant differences in outcomes were found by race or MHI. The substantial gaps in outcomes previously observed by race and MHI may not be rooted in differences in tumor biology but rather in access to quality care.

Association of Patient Sex and Pregnancy Status With Naloxone Administration During Emergency Department Visits.

To evaluate the association of sex and pregnancy status with rates of naloxone administration during opioid overdose-related emergency department (ED) visits by using the Nationwide Emergency Department Sample. A retrospective cohort study was conducted using the Nationwide Emergency Department Sample 2016 and 2017 data sets. Eligible records included men and women, 15-49 years of age, with an opioid overdose-related ED visit; records for women were stratified by pregnancy status (International Classification of Diseases, Tenth Revision O codes). A multivariable logistic regression model was used to assess the primary outcome of naloxone administration (Current Procedural Terminology code: J2310). Secondary outcomes included subsequent admission and mortality. A subgroup analysis compared pregnant women who did receive naloxone compared with those who did not receive naloxone. Records from 443,714 men, 304,364 nonpregnant women, and 25,056 pregnant women were included. Nonpregnant women had lower odds for naloxone administration (1.70% vs 2.10%; adjusted odds ratio [aOR] 0.86 [95% CI 0.83-0.89]) and mortality (2.21% vs 2.99%; aOR 0.71 [95% CI 0.69-0.73]) but higher odds of subsequent admission (30.22% vs 27.18%; aOR 1.04 [95% CI 1.03-1.06]) compared with men. Pregnant women had lower odds for naloxone administration (0.27% vs 1.70%; aOR 0.16 [95% CI 0.13-0.21]) and mortality (0.41% vs 2.21%; aOR 0.28 [95% CI 0.23-0.35]) but higher odds of subsequent admission (40.50% vs 30.22%; aOR 2.04 [95% CI 2.00-2.10]) compared with nonpregnant women. Pregnant women who received naloxone had higher odds of mortality (14% vs 0.39%; aOR 6.30 [95% CI 2.11-18.78]) compared with pregnant women who did not receive naloxone. Pregnant women who did not receive naloxone were more likely to have Medicaid as their expected insurance payer, be in the lowest quartile of median household income for residence ZIP codes, and have a concurrent mental health diagnosis compared with pregnant women who did receive naloxone. Reproductive-aged women who are nonpregnant and pregnant were less likely to receive naloxone during opioid overdose-related ED visits compared with reproductive-aged men. Naloxone administration for reproductive-aged women should be prioritized in the efforts to reduce opioid- and pregnancy-related morbidity and mortality in the United States.

Abstract MP26: Medicaid Expansion Did Not Reduce Stroke Readmissions After One Year

Introduction: Up to 21% of stroke survivors are re-hospitalized within 30 days. Health insurance promotes access to follow-up care that can mitigate the risk of readmission, but 12 states do not participate in the Affordable Care Act’s Medicaid expansion. Hypothesis: The probability of 30-day hospital readmission after acute ischemic stroke was lower in Medicaid expansion states than in non-expansion states. Methods: A retrospective, quasi-experimental study using six inpatient databases from AHRQ’s Healthcare Cost and Utilization Project: four from expansion states (AR, MD, NM and WA) and two non-expansion (FL and GA). The sample comprised all patients hospitalized in 2012-14 with a principal diagnosis of ischemic stroke (ICD-9-CM 433.x1, 434.x1 or 436) who were aged 19-64; resided in the state where admitted; had a primary payer of Medicaid, self-pay or no charge; and were discharged alive ( N =18,766). Mixed effects logit models with a time-by-treatment interaction were built to test if the probability of readmission changed differentially between expansion and non-expansion states from 2012-13 (before expansion) to 2014. Any in-state hospitalization within 30 days of discharge (except for rehabilitation, psychiatry, or cancer treatment) was considered a readmission. A secondary analysis of unplanned, potentially preventable readmissions (adapted from the AHRQ Prevention Quality Indicators) was also conducted. Models included race, sex, age, number of diagnoses, median household income quartile of patient ZIP code, and metropolitan residence as fixed effects, with random intercepts for hospital and state. Results: In 2012-13, 8.9% of the expansion state patients were readmitted compared to 9.0% in non-expansion states; in 2014, 11.1% were readmitted in expansion states versus 10.5% in non-expansion states. In multivariable models, the time-by-treatment interaction was not statistically significant: β=0.072, p= .541, for all readmissions, β=0.168, p =.683, for unplanned, potentially preventable readmissions. Conclusions: Medicaid expansion did not reduce 30-day readmissions after stroke in the first year of implementation in four diverse states. Stroke readmissions among non-elderly adults require more targeted interventions.

Socioeconomic Disparities in Patient Use of Telehealth During the Coronavirus Disease 2019 Surge

The coronavirus disease 2019 (COVID-19) pandemic required the rapid transition to telehealth with the aim of providing patients with medical access and supporting clinicians while abiding by the stay-at-home orders. To assess demographic and socioeconomic factors associated with patient participation in telehealth during the COVID-19 pandemic. This cohort study included all pediatric and adult patient encounters at the Department of Otolaryngology-Head & Neck Surgery in a tertiary care, academic, multisubspecialty, multisite practice located in an early hot spot for the COVID-19 pandemic from March 17 to May 1, 2020. Encounters included completed synchronous virtual, telephone, and in-person visits as well as visit no-shows. Patient demographic characteristics, insurance status, and 2010 Census block level data as a proxy for socioeconomic status were extracted. Univariate and multivariate logistic regression models were created for patient-level comparisons. Of the 1162 patients (604 females [52.0%]; median age, 55 [range, 0-97] years) included, 990 completed visits; of these, 437 (44.1%) completed a virtual visit. After multivariate adjustment, females (odds ratio [OR], 1.71; 95% CI, 1.11-2.63) and patients with preferred provider organization insurance (OR, 2.70; 95% CI, 1.40-5.20) were more likely to complete a virtual visit compared with a telephone visit. Increasing age (OR per year, 0.98; 95% CI, 0.98-0.99) and being in the lowest median household income quartile (OR, 0.60; 95% CI, 0.42-0.86) were associated with lower odds of completing a virtual visit overall. Those patients within the second (OR, 0.53; 95% CI, 0.28-0.99) and lowest (OR, 0.33; 95% CI, 0.17-0.62) quartiles of median household income by census block and those with Medicaid, no insurance, or other public insurance (OR, 0.47; 95% CI, 0.23-0.94) were more likely to complete a telephone visit. Finally, being within the lower 2 quartiles of proportion being married (OR for third quartile, 0.49 [95% CI, 0.29-0.86]; OR for lowest quartile, 0.39 [95% CI, 0.23-0.67]) was associated with higher likelihood of a no-show visit. These findings suggest that age, sex, median household income, insurance status, and marital status are associated with patient participation in telehealth. These findings identify vulnerable patient populations who may not engage with telehealth, yet still require medical care in a changing health care delivery landscape.

Trends in Incidence of Nicotine Use Disorder Among Adolescents in the Pediatric Hospital, 2012-2019.

To assess trends in the incidence of nicotine use disorder (NUD) and describe associated factors among adolescents in the pediatric emergency department (ED) and inpatient settings. We conducted a retrospective cohort study of all adolescents (11-18 years) with a hospital encounter (inpatient, observation, or ED) in the Pediatric Health Information System between January 1, 2012, and September 30, 2019. After excluding adolescents with a previous International Classification of Diseases, Ninth Revision, and International Classification of Diseases, Tenth Revision, NUD diagnosis in the past 2 years, adolescents with new NUD diagnosis (ie, NUD incidence) were identified. A multivariable generalized liner mixed model was used to assess adjusted NUD incidence and investigate the relationship of NUD with patient characteristics and any interactions between characteristics and time. Spearman's correlation coefficient was used to assess the correlation between NUD incidence and e-cigarette use reported among youth. Of 3 963 754 adolescents, 15 376 (0.4%) had a new diagnosis of NUD. Between 2012 and 2019, NUD incidence increased from 0.3% to 0.4% (P < .001). Findings from the time interaction effect analysis revealed increasing NUD incidence among certain subpopulations, including boys, those with a commercial or other insurance type, adolescents seen in the ED, those from the lowest and highest median household income quartile, and those in the South and West US Census regions. The correlation between NUD incidence and e-cigarette use among high school students was ρ = 0.884 (P = .006). The incidence of NUD among adolescents is increasing. Efforts to increase the screening and treatment of NUD among adolescents in the hospital, particularly among the at-risk populations identified, are needed.

Patients leaving without being seen from the emergency department: A prediction model using machine learning on a nationwide database.

ObjectiveThe objective of this study was to develop a US‐representative prediction model identifying factors with a greater likelihood of patients leaving without being seen.MethodsWe conducted a retrospective cohort analysis using a 2016 nationwide emergency department (ED) sample. Patient factors considered for analysis were the following: age, sex, acuity, chronic diseases, weekend visit, quarter of presentation, median household income quartile for patient's zip code, primary/secondary insurance, total charges for the visit, and urban/rural household. Hospital factors considered were urban/rural location, trauma center/teaching hospital, and annual ED volume. Multivariable logistic regression was used to find significant predictors and their interactions. A random forest algorithm was used to determine the order of importance of factors.ResultsA total of 32,680,232 hospital‐based ED visits with 466,047 incidences of leaving without being seen were included. The cohort comprised 55.5% females, with a median (IQR) age of 37 (21–58) years. Positively associating factors were male sex (odds ratio [OR], 1.22; 99% confidence interval [CI], 1.17–1.26), lower acuity (P < 0.001), and annual ED visits ≥60,000 (OR, 1.44; 99% CI, 1.21–1.7) versus <20,000. Negatively associating factors were primary insurance being Medicare/Tricare or private insurance (P < 0.001); weekend presentations (OR, 0.87; 99% CI, 0.85–0.89); age >64 or <18 years (P < 0.001); and higher median household income for patient's zip code second (OR, 0.86; 99% CI, 0.77–0.97), third (OR, 0.8; 99% CI, 0.7–0.91), and fourth (OR, 0.7; 99% CI, 0.6–0.8) quartiles versus the first quartile. Significant interactions existed between age, acuity, primary insurance, and chronic conditions. Primary insurance was the most predictive.ConclusionOur derivation model reiterated several modifiable and non‐modifiable risk factors for leaving without being seen established previously while rejecting the importance of others.

Health Disparities Among People Infected With Influenza, Rhode Island, 2013-2018.

Health disparities are associated with poor outcomes related to public health. The objective of this study was to assess health disparities associated with influenza infection based on median household income and educational attainment. We geocoded people with documented confirmed influenza infection by home address to identify the US Census 2010 tract in which they lived during 4 influenza surveillance seasons (2013-2014, 2015-2016, 2016-2017, and 2017-2018) in Rhode Island. We dichotomized influenza as severe if the person with influenza infection was hospitalized (ie, inpatient) or as nonsevere if the person was not hospitalized (ie, outpatient). We examined 2 socioeconomic factors: median household income (defined as low, medium low, medium high, and high) and educational attainment (defined as a ratio among people who completed <high school, high school, some college, or ≥bachelor's degree). We calculated relative rates (RRs) to determine the associated level of risk for each socioeconomic factor. The incidence of influenza per 100 000 person-years was significantly higher in populations with low vs high median household income (620 vs 303; P < .001) and in populations with low vs high educational attainment (583 vs 323; P < .001). The RR of a severe infection in the quartile with the lowest educational attainment (0.57) was significantly higher than the RR in the other 3 quartiles of educational attainment (range, 0.36-0.39; P = .01). However, the RR of a severe infection was higher in the 3 quartiles of median household income (range, 0.38-0.40) than in the quartile with the lowest median household income (0.29). People in Rhode Island with a lower socioeconomic status are at greater risk of an influenza infection than people with higher socioeconomic status. The reasons for these disparities require further investigation.

National Trends and Outcomes of Nonautoimmune Hemolytic Anemia in Alcoholic Liver Disease: Analysis of the Nationwide Inpatient Sample.

The aim of this study was to determine the burden of nonautoimmune hemolytic anemia (NAHA) in hospitalized patients with coexisting alcoholic liver disease (ALD), identify risk factors for NAHA in ALD and describe the hospitalization outcomes. ALD can result in structural and metabolic alterations in the red-blood cell membrane leading to premature destruction of erythrocytes and hemolytic anemia of varying severity. Hospitalized ALD patients with concomitant NAHA were identified in the Nationwide Inpatient Sample database using International Classification of Diseases-9 codes from 2009 to 2014. The primary outcome was to determine the nationwide prevalence and risk factors of NAHA in patients hospitalized with ALD. The prevalence of NAHA was 0.17% (n=3585) among all ALD patients (n=2,125,311) that were hospitalized. Multivariate analysis indicated higher odds of NAHA in ALD patients in the following groups: female gender [adjusted odds ratio (AOR) AOR 1.80, P<0.0001]; highest quartile of median household income (AOR 1.88, P<0.0001); increasing Charlson-Deyo Comorbidity Index (3 to 4 vs. 0, AOR 2.16, P=0.0042) and cirrhosis (AOR 2.74, P<0.0001). Discharges of ALD with anemia had a significantly longer average length of stay (8.8 vs. 6.0 d, P<0.0001), increased hospital charges ($38,961 vs. $25,244, P<0.0001) and higher mortality (9.0% vs. 5.6%, P<0.0001) when compared with ALD with no anemia. NAHA in patients with ALD is an important prognostic marker, predicting a longer, costlier hospitalization and increased inpatient mortality in ALD.

Assessment of social factors influencing hospitalization cost of US preterm newborns, 2016

Background and objective Preterm newborn (PTN) care is costly in the US, and the preterm birth (PTB) rate has not seen substantial improvement. PTB remains a significant public health issue because of risks of adverse health outcomes and the associated economic burden. Assessing factors leading to high hospital cost is imperative in order to ease economic burden associated with PTB. Social factors influencing hospitalization cost for PTN have not been studied extensively. This study aims to examine the contributions of hospital type, race/ethnicity, and median household income quartile to the cost of the PTN’s initial hospitalization after birth. Materials and methods Data used in this study originated from the 2016 Healthcare Cost and Utilization Project Kids’ Inpatient Database (KID) which provided the most recent national records of pediatric hospitalization in the US. International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes P07.2x and P07.3x were used to identify birth gestational age (GA) in complete weeks. Q00-Q99 codes were used to identify congenital anomalies. Data records reflecting hospitalization of newborns without congenital anomalies were extracted to estimate total cost in 2016. Analysis of factors influencing total cost was performed on the subset of newborns that were in-born and with birth GA from 24 to 36 weeks. Total hospital cost was calculated by multiplying total charges and cost-to-charge ratios. Per-diem cost was calculated by dividing total hospital cost by length of stay. Linear regression modeling was performed using weighted data, and the best model was selected using the Akaike information criteria. Results Hospitalization cost for PTNs, who constituted 7.6% of all studied newborns, comprised 45% of total newborn hospitalization cost. PTNs in freestanding children’s hospitals had significantly higher total cost compared to non-children’s teaching and non-teaching hospitals. The cost difference was GA-dependent. The latter two hospital types reported similar cost for PTNs. Although race/ethnicity and median household income quartile were related, the two covariates had independent effects on total hospitalization cost. A model built to assess the roles of covariates in the association between GA and total hospitalization cost found that all three covariates independently affected total cost when controlling for one another, with hospital type also showing an interaction effect with GA. Conclusions Hospitalization cost is highly influenced by hospital type, race/ethnicity, and median household income quartile. The modifiable aspects of these factors may be further dissected and targeted in order to ease the burden of high hospitalization cost associated with PTB.

Pre-operative headaches and obstructive hydrocephalus predict an extended length of stay following suboccipital decompression for pediatric Chiari I malformation.

For young children and adolescents with Chiari malformation type I (CM-I), the determinants of extended length of hospital stay (LOS) after neurosurgical suboccipital decompression are obscure. Here, we investigate the impact of patient- and hospital-level risk factors on extended LOS following surgical decompression for CM-I in young children to adolescents. The Kids' Inpatient Database year 2012 was queried. Pediatric CM-I patients (6-18 years) undergoing surgical decompression were identified. Weighted patient demographics, comorbidities, complications, LOS, disposition, and total cost were recorded. A multivariate logistic regression was used to determine the odds ratio for risk-adjusted LOS. The primary outcome was the degree patient comorbidities or post-operative complications correlated with extended LOS. A total of 1592 pediatric CM-I patients were identified for which 328 (20.6%) patients had extended LOS (normal LOS, 1264; extended LOS, 328). Age, gender, race, median household income quartile, and healthcare coverage distributions were similar between the two cohorts. Patients with extended LOS had significantly greater admission comorbidities including headache symptoms, nausea and vomiting, obstructive hydrocephalus, lack of coordination, deficiency anemias, and fluid and electrolyte disorders. On multivariate logistic regression, several risk factors were associated with extended LOS, including headache symptoms, obstructive hydrocephalus, and fluid and electrolyte disorders. Our study using the Kids' Inpatient Database demonstrates that presenting symptoms and signs, including headaches and obstructive hydrocephalus, respectively, are significantly associated with extended LOS following decompression for pediatric CM-I.

Unequal social vulnerability to Hurricane Sandy flood exposure.

Disparities exist in post-disaster flooding exposure and vulnerable populations bear a disproportionate impact of this exposure. We describe the unequal burden of flooding in a cohort of New York residents following Hurricane Sandy and assess whether the likelihood of flooding was distributed equally according to socioeconomic demographics, and whether this likelihood differed when analyzing self-reported or FEMA flood exposure measures. Residents of New York City and Long Island completed a self-administered survey 1.5–4.0 years after the storm. Multivariable logistic regressions were performed to determine the relationship between sociodemographic characteristics and flood exposure. Participants (n=1231) residing in areas of the lowest two quartiles of median household income experienced flooding the most often (FEMA/self-reported: < $40,298: 65.3%/42.0%, $40,298 - $67,188: 43.3%/32.1%) and these areas contained the highest proportions of non-white participants (< $40,298: 39.1%, $40,298 - $67,188: 36.6%) and those with < high school education (< $40,298: 35.5%, $40,298 - $67,188: 33.6%). Both self-report (p < 0.05) and FEMA (p < 0.05) flood measures indicated that older participants were more likely to live in a household exposed to flooding, while those living in higher income areas had decreased likelihood of flooding (p < 0.0001). Socioeconomic and age disparities were present in exposure to flooding during Hurricane Sandy. Future disaster preparedness responses must understand flooding from an environmental justice perspective to create effective strategies that minimize disproportionate exposure and its adverse outcomes.