Healthcare Claims Data Research Articles

Sources of Geographic Variation in Health Care Spending Among Individuals With Employer Sponsored Insurance.

We use health care claims data from the Health Care Cost Institute to estimate the share of geographic variation in health care spending attributable to person-specific (demand) and place-specific (supply) factors. We exploit patient migration across 112 metropolitan areas between 2012 and 2016. Using an event study approach, we find that moving to an area with 10% higher (lower) spending leads to a 4.2% increase (decrease) in individual medical spending. Our estimate implies that 42% of variation in health care spending among the commercially insured is attributable to place-specific factors. We show that variation in both price and utilization jointly determine the place-specific impact on individual spending. All else equal, we find that moving to an area with 10% higher (lower) prices, on average leads to a 5% increase (decrease) in spending, while moving to an area with 10% higher (lower) utilization leads to a 3.6% increase (decrease).

1155-P: Durable vs. Disposable Insulin Pumps for Type 1 Diabetes: Health Care Costs and Utilization

Objectives: To compare the economic impact of using a durable versus disposable insulin pump over a 4-year period (i.e., the typical warranty period) in individuals with type 1 diabetes (T1D). Methods: This population-based, longitudinal analysis used a large repository of healthcare claims data were used to compare healthcare costs (in 2017 U.S. dollars) and utilization in individuals with T1D using durable pumps (N=2,013) to those using disposable pumps (N=642), before and after pump adoption. T-test was used to examine the between-group difference in pump-related costs, and a generalized linear model regression was used to estimate the difference-in-difference effect of pump type on other outcomes. Results: Mean pump-related costs over the 4-year period were $6,606 less for durable pump users versus disposable pump users (Figure; p<0.0001). Mean out-of-pocket pump-related costs for the same time period were $1,037 less for durable pump versus disposable pump users (p<0.0001). There were no statistically significant differences between durable and disposable pump users for any other variables studied (continuous glucose monitoring costs, insulin costs, inpatient costs, emergency room costs, outpatient costs, hospital admissions, or emergency room visits). Conclusions: Use of a durable versus disposable insulin pump is associated with significantly lower pump-related costs. Disclosure M. Shah: Employee; Self; Medtronic. C. Zhu: Employee; Self; Medtronic.

Sulfonylureas and Metformin Were Not Associated With an Increased Rate of Serious Bleeding in Warfarin Users: A Self-Controlled Case Series Study.

Drug interactions between warfarin and sulfonylureas are suggested by pharmacokinetic information and prior studies. However, clinical evidence on the association of such interactions and the risk of bleeding is lacking. Using healthcare claims data from 5 US Medicaid programs from 1999-2011 and a self-controlled case series design with warfarin as an object drug, we calculated confounder-adjusted rate ratios (RRs) for concomitant use of sulfonylureas and metformin for 3 outcomes separately: (i) serious bleeding as a composite outcome of gastrointestinal bleeding (GIB) and nontraumatic intracranial hemorrhage (ICH); (ii) GIB; and (iii) ICH. In 6,463 warfarin users experiencing serious bleeding, an increased rate of serious bleeding was not associated with concomitant use of glimepiride (RR: 0.93; 95% confidence interval (CI) 0.75-1.15), glipizide (RR: 0.97; 95% CI 0.84-1.13), glyburide (RR: 0.89; 95% CI 0.76-1.06), or metformin (RR: 0.85; 95% CI 0.76-0.96), nor was the occurrence of the component outcomes of GIB or ICH. These results suggest that use of sulfonylureas or metformin was not associated with an increased rate of serious bleeding in warfarin users.

Predicting 30-day hospital readmissions using artificial neural networks with medical code embedding.

Reducing unplanned readmissions is a major focus of current hospital quality efforts. In order to avoid unfair penalization, administrators and policymakers use prediction models to adjust for the performance of hospitals from healthcare claims data. Regression-based models are a commonly utilized method for such risk-standardization across hospitals; however, these models often suffer in accuracy. In this study we, compare four prediction models for unplanned patient readmission for patients hospitalized with acute myocardial infarction (AMI), congestive health failure (HF), and pneumonia (PNA) within the Nationwide Readmissions Database in 2014. We evaluated hierarchical logistic regression and compared its performance with gradient boosting and two models that utilize artificial neural networks. We show that unsupervised Global Vector for Word Representations embedding representations of administrative claims data combined with artificial neural network classification models improves prediction of 30-day readmission. Our best models increased the AUC for prediction of 30-day readmissions from 0.68 to 0.72 for AMI, 0.60 to 0.64 for HF, and 0.63 to 0.68 for PNA compared to hierarchical logistic regression. Furthermore, risk-standardized hospital readmission rates calculated from our artificial neural network model that employed embeddings led to reclassification of approximately 10% of hospitals across categories of hospital performance. This finding suggests that prediction models that incorporate new methods classify hospitals differently than traditional regression-based approaches and that their role in assessing hospital performance warrants further investigation.

Identifying the Prescribing Physician in US Healthcare Claims Data.

Identifying the prescribing physician may be essential in studies that use health care claims databases but often there is no single consistent physician identifier (ID) available that spans an entire database and no standard methods have emerged to address this issue. Unique Drug Enforcement Administration (DEA) numbers (N=36,721) were identified from the pharmacy claims of a cohort of postmenopausal (55+) initiators of osteoporosis medications (2008-2011) in the United Healthcare database. The proportion of times a Provider ID appeared at least once on the outpatient medical service claims (OMSCs) in the 14 days before a new prescription (Rx) out of the total number of new Rxs for each DEA was calculated and the Provider ID with the highest proportion was considered the most likely match to the DEA. We used regression models to evaluate how characteristics of OMSCs (N=36,721) related to the probability that the provider would be the prescribing physician on a proximal pharmacy claim for each patient (N=20,058). A total of 37,448,056 new Rxs and 132,135,673 OMSCs were associated with the DEAs. Family and general practitioner, hospitalist, and osteoporosis diagnoses were strong predictors of a match. Our models had sensitivities ranging from 81.9% to 82.8% and specificities ranging from 52.4% to 53.5%. Our algorithms were good at identifying matches for the prescribing physician according to our reference standard but may have included some false positives. Predicted probabilities from our models could be used in other databases where physician IDs are unavailable to help identify the prescriber.

Concordance and Discordance in the Geographic Distribution of Childhood Obesity and Pediatric Type 2 Diabetes in New York City.

As rates of childhood obesity and pediatric type 2 diabetes (T2D) increase, a better understanding is needed of how these 2 conditions relate and which subgroups of children are more likely to develop diabetes with and without obesity. To compare hotspots of childhood obesity and pediatric T2D in New York City, we performed geospatial clustering analyses on obesity estimates obtained from surveys of school-aged children and diabetes estimates obtained from health care claims data, from 2009 to 2013. Analyses were performed at the Census tract level. We then used multivariable regression analysis to identify sociodemographic and environmental factors associated with these hotspots. We identified obesity hotspots in Census tracts with a higher proportion of Black or Hispanic residents, with low median household income, or located in a food swamp. Total 51.1% of pediatric T2D hotspots overlapped with obesity hotspots. For pediatric T2D, hotspots were identified in Census tracts with a higher proportion of Black residents and a lower proportion of Hispanic residents. Non-Hispanic Black neighborhoods had a higher probability of being hotspots of both childhood obesity and pediatric T2D. However, we identified a discordance between hotspots of childhood obesity and pediatric diabetes in Hispanic neighborhoods, suggesting either under-detection or under-diagnosis of diabetes, or that obesity may influence diabetes risk differently in these 2 populations. These findings warrant further investigation of the relationship between childhood obesity and pediatric diabetes among different racial and ethnic groups, and may help guide pediatric public health interventions to specific neighborhoods.

Independent Validation in a Large Privately Insured Population of the Risk Index for Serious Prescription Opioid-Induced Respiratory Depression or Overdose.

To assess the generalizability of the overdose or serious opioid-induced respiratory depression risk index (VHA-RIOSORD), created by Zedler et al., using claims data from a large private insurer. A retrospective nested case-control analysis of health care claims data. Commercially insured individuals with a claim for an opioid prescription between October 1, 2014, and September 30, 2016 (N = 1,431,737). An overdose or serious opioid-induced respiratory depression (OSORD) occurred in 1,097 patients. Ten controls were selected per case (N = 10,970). Items and the assignment of point values to predictors were consistent with those determined by Zedler et al. Modeling of risk index scores produced predicted probabilities of OSORD; risk classes were defined by the predicted probability distribution. All 15 items of the VHA-RIOSORD were used to determine a member's risk of OSORD. The average predicted probability of experiencing OSORD ranged from 3% in the lowest risk decile to 90% in the highest, with excellent agreement between predicted and observed incidence across risk classes. The model's C-statistic was 0.88. Consistent with the findings of its developers, the VHA-RIOSORD performed well in identifying members of a large private insurance company who were medical users of prescription opioids at elevated risk of overdose or life-threatening respiratory depression, those most likely to benefit from preventive interventions.

Association of preconception paternal health on perinatal outcomes: analysis of U.S. claims data

To assess whether paternal health is associated with maternal peripartum and neonatal outcomes. Retrospective cohort study. University research departments. Analytic sample of children born to paired fathers and mothers covering live births within the United States between 2009-2016. Paternal health status (e.g., metabolic syndrome diagnoses, individual chronic disease diagnoses). Primary outcome of preterm birth (i.e., live birth before 37 weeks), and secondary outcomes of low birth weight, neonatal intensive care unit (NICU) stay, gestational diabetes, preeclampsia, eclampsia, and length of maternal stay. The IBM Marketscan Research database covers reimbursed health care claims data on inpatient and outpatient encounters who are privately insured through employment-sponsored health insurance. We assessed 785,809 singleton live births, with 6.6% born preterm. The presence of paternal comorbidities was associated with higher odds of preterm birth, low birth weight (LBW), and NICU stay. After adjusting for maternal factors, fathers with most or all components of the metabolic syndrome had 19% higher odds of having a child born preterm (95% CI 1.11-1.28), 23% higher odds of LBW (95% CI 1.01-1.51), and 28% higher odds of NICU stay (95% CI 1.08-1.52). Maternal morbidity (e.g., gestational diabetes or preeclampsia) was also positively associated with preconception paternal health. Increased preconception paternal comorbidity may be associated with negative infant and maternal outcomes. Although the paternal effect remains modest, these findings highlight the importance of the health of both parents, particularly the mother, on healthy pregnancy.

A Supervised Approach for Detection of Outliers in Healthcare Claims Data

A Supervised Approach for Detection of Outliers in Healthcare Claims Data

Replication of the TRITON-TIMI Antiplatelet Trial in Healthcare Claims Data

Replication of the TRITON-TIMI Antiplatelet Trial in Healthcare Claims Data

Replication of the PLATO Antiplatelet Trial in Healthcare Claims Data

Replication of the PLATO Antiplatelet Trial in Healthcare Claims Data

Treatment Patterns, Health Care Resource Utilization, and Health Care Cost Associated with Atypical Antipsychotics or Guanfacine Extended Release in Children and Adolescents with Attention-Deficit/Hyperactivity Disorder in Quebec, Canada.

Objective: To assess treatment patterns, health care resource utilization, and health care costs associated with use of atypical antipsychotics (AAPs) or the nonstimulant guanfacine extended release (GXR) after stimulant therapy for attention-deficit/hyperactivity disorder (ADHD). In Canada, GXR is approved as a monotherapy for children and adolescents with ADHD or as an adjunct to stimulants, and AAPs are commonly used off-label as an adjunct to stimulants.Methods: Health care claims data (January 1, 2007 to March 31, 2016) from Quebec's provincial health plan were assessed for individuals with ADHD, 6–17 years of age, who received ≥1 stimulant followed by a first AAP or GXR prescription (index medication), without a diagnosis for which AAPs are indicated.Results: Overall, 1327 individuals were included (AAPs, 1098; GXR, 229). Rates of discontinuation, augmentation, or switching of the index medication did not differ between AAPs and GXR during the first follow-up year. Discontinuation rates were significantly lower with GXR than with AAPs during the second year (22.0% vs. 35.9%; p = 0.03). GXR and AAPs resulted in similar increases in total health care cost. In GXR users, the increase in prescription drug cost after 6 months was higher than in AAP users, whereas the increase in overall medical cost was higher with AAPs than GXR, owing to more psychiatric department visits.Conclusions: In children and adolescents with ADHD who used AAPs or GXR after stimulants, secondary treatment changes were similar with both treatments after 1 year, but discontinuation rates were significantly lower with GXR than with AAPs in the second year. The greater increase in prescription cost with GXR was balanced by a greater increase in overall medical costs with AAPs, resulting in no overall difference in total health care cost between the two treatments.

The DREAM Initiative: study protocol for a randomized controlled trial testing an integrated electronic health record and community health worker intervention to promote weight loss among South Asian patients at risk for diabetes

BackgroundElectronic health record (EHR)-based interventions that use registries and alerts can improve chronic disease care in primary care settings. Community health worker (CHW) interventions also have been shown to improve chronic disease outcomes, especially in minority communities. Despite their potential, these two approaches have not been tested together, including in small primary care practice (PCP) settings. This paper presents the protocol of Diabetes Research, Education, and Action for Minorities (DREAM) Initiative, a 5-year randomized controlled trial integrating both EHR and CHW approaches into a network of PCPs in New York City (NYC) in order to support weight loss efforts among South Asian patients at risk for diabetes.Methods/designThe DREAM Initiative was funded by the National Institute of Diabetes and Digestive and Kidney Diseases (National Institutes of Health). A total of 480 individuals at risk for type 2 diabetes will be enrolled into the intervention group, and an equal number will be included in a matched control group. The EHR intervention components include the provision of technical assistance to participating PCPs regarding prediabetes-related registry reports, alerts, and order sets. The CHW intervention components entail group education sessions on diabetes prevention, including weight loss and nutrition. A mixed-methods approach will be used to evaluate the feasibility, adoption, and impact (≥ 5% weight loss) of the integrated study components. Additionally, a cost effectiveness analysis will be conducted using outcomes, implementation costs, and healthcare claims data to determine the incremental cost per person achieving 5% weight loss.DiscussionThis study will be the first to test the efficacy of an integrated EHR–CHW intervention within an underserved, minority population and in a practical setting via a network of small PCPs in NYC. The study’s implementation is enhanced through cross-sector partnerships, including the local health department, a healthcare payer, and EHR vendors. Through use of a software platform, the study will also systematically track and monitor CHW referrals to social service organizations. Study findings, including those resulting from cost-effectiveness analyses, will have important implications for translating similar strategies to other minority communities in sustainable ways.Trial registrationThis study protocol has been approved and is made available on ClinicalTrials.gov by NCT 03188094 as of 15 June 2017.

Validation of claims-based diagnoses of adult and pediatric neuromyelitis optica spectrum disorder and variations in diagnostic evaluation and treatment initiation

BackgroundNeuromyelitis optica spectrum disorder (NMOSD) is a rare demyelinating disease in need of more studies to determine effective treatment regimens. The rarity of the disorder, however, makes large randomized-controlled trials challenging. Validation of the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) code for NMO could facilitate the use of large healthcare claims data for future research. We aimed 1) to determine the positive predictive value (PPV) of the ICD-9-CM code for NMO as well as evaluate case-finding algorithms for the identification of patients with NMO/NMOSD and 2) to compare the evaluation of and treatment for pediatric versus adult patients. MethodsThis was a multicenter retrospective cohort study of patients with ≥ 1 ICD-9 code for NMO seen at 3 pediatric and 2 adult United States medical centers from 2001–2016. Using a standardized data entry form, pediatric and adult neurologists and rheumatologists reviewed patients’ medical records to determine whether patients fulfilled the 2006 criteria for NMO and/or the 2015 criteria for NMOSD in order to determine the positive predictive value (PPV) for the ICD-9-CM code. Demographic and clinical information was abstracted from patient medical records to ascertain variables then evaluated in case-based finding algorithms for further identification of patients with true NMO/NMOSD. We also evaluated differences in clinical characteristics between pediatric and adult patients using chi-squared or Fisher's exact tests, as appropriate, to assess for treatment variation. ResultsA single code for NMO had a PPV of 47% across all sites, with significant site variation (0–77%). The best case-finding algorithm included at least 5 codes as well as a documented hospitalization (PPV = =90% for children and PPV = 92% for adults). Children were more likely to be evaluated by a rheumatologist or ophthalmologist, undergo magnetic resonance imaging of the orbits, and receive immunosuppressive and biologic agents than their adult counterparts. Rituximab was administered similarly among the two groups. ConclusionThe ICD-9 code for neuromyelitis optica (NMO) is inaccurate for identification of NMO/NMOSD. Using case-finding algorithms increases the PPV. The initial diagnostic evaluation and treatment of NMOSD differs significantly between children and adults.

Nonrandomized Real-World Evidence to Support Regulatory Decision Making: Process for a Randomized Trial Replication Project.

Recent legislation mandates that the US Food and Drug Administration issue guidance regarding when real-world evidence (RWE) could be used to support regulatory decision making. Although RWE could come from randomized or nonrandomized designs, there are significant concerns about the validity of RWE assessing medication effectiveness based on nonrandomized designs. We propose an initiative using healthcare claims data to assess the ability of nonrandomized RWE to provide results that are comparable with those from randomized controlled trials (RCTs). We selected 40 RCTs, and we estimate that approximately30 attempted replications will be completed after feasibility analyses. We designed an implementation process to ensure that each attempted replication is consistent, transparent, and reproducible. This initiative is the first to systematically evaluate the ability of nonrandomized RWE to replicate multiple RCTs using a structured process. Results from this study should provide insight on the strengths and limitations of using nonrandomized RWE from claims for regulatory decision making.

Age-related difference in health care use and costs of patients with chronic kidney disease and matched controls: analysis of Dutch health care claims data.

BackgroundThe financial burden of chronic kidney disease (CKD) is increasing due to the ageing population and increased prevalence of comorbid diseases. Our aim was to evaluate age-related differences in health care use and costs in Stage G4/G5 CKD without renal replacement therapy (RRT), dialysis and kidney transplant patients and compare them to the general population.MethodsUsing Dutch health care claims, we identified CKD patients and divided them into three groups: CKD Stage G4/G5 without RRT, dialysis and kidney transplantation. We matched them with two controls per patient. Total health care costs and hospital costs unrelated to CKD treatment are presented in four age categories (19–44, 45–64, 65–74 and ≥75 years).ResultsOverall, health care costs of CKD patients ≥75 years of age were lower than costs of patients 65–74 years of age. In dialysis patients, costs were highest in patients 45–64 years of age. Since costs of controls increased gradually with age, the cost ratio of patients versus controls was highest in young patients (19–44 years). CKD patients were in greater need of additional specialist care than the general population, which was already evident in young patients.ConclusionAlready at a young age and in the earlier stages of CKD, patients are in need of additional care with corresponding health care costs far exceeding those of the general population. In contrast to the general population, the oldest patients (≥75 years) of all CKD patient groups have lower costs than patients 65–74 years of age, which is largely explained by lower hospital and medication costs.

Adherence to heart failure management medications following cardiac resynchronization therapy

Objective: The purpose of this study is to assess the real-world impact of cardiac resynchronization therapy (CRT) on adherence to heart failure (HF) medications.Methods: MarketScan administrative health care claims data from 2008 to 2014 among patients with HF were used. The date of first CRT implantation served as the index date. Adherence to guideline-directed medical therapy (GDMT) classes were compared during pre- and post-index periods using proportion of days covered (PDC). Comparisons between the two periods were made using the Wilcoxon sign-rank test for continuous PDC and McNemar's test for dichotomized PDC.Results: Increases in medication adherence were observed for major classes of HF GDMT medications. Specifically, adherence to angiotensin-converting enzyme inhibitors (ACE-I), angiotensin receptor blockers (ARB), beta blockers (BB), and furosemide increased by 22, 24, 32, and 28% (all p < .001), respectively, in the 12 months pre to 12 months post-CRT. Large increases between the pre- and post-CRT period were also observed when considering adherence as dichotomized PDC ≥0.80 in the 12 months pre- versus post-CRT.Conclusion: Adherence to HF medications significantly improved among HF patients post-CRT implantation. Further research is needed to better understand the underlying determinants of this effect, including whether the effect is attributable to factors such as enhanced patient monitoring and improved access to high-quality specialized HF care among patients receiving CRT.

Characteristics, treatment patterns, and survival from three cohorts of advanced or metastatic cancer patients using health care claims data in the United States.

129 Background: Our main objective was to identify, in healthcare claims data, patients with advanced or metastatic: urothelial carcinoma (amUC), gastric cancer (amGC) and non-small cell lung cancer (amNSCLC) and to report on their characteristics, treatments, and survival rates using contemporaneous real-world data. Methods: This cohort study was conducted in the HealthCore Integrated Research Database (HIRD), from January 2010 to January 2018, which contains healthcare claims data from commercial health plans across the US (60 million lives). We applied algorithms, previously validated on registry data, to the HIRD to define 3 cohorts of advanced stage cancer. Cohort characteristics and treatment patterns were described. Patient vital status was captured through probabilistic linkage with the National Death Index (NDI) and survival was assessed using the Kaplan-Meier method. Results: Algorithms to predict advanced stage cancer resulted in the following cohorts: 1,501 amUC, 6,253 amGC and 38,451 amNSCLC cases. Most patients in each cohort were de novo advanced or metastatic, but subsets were diagnosed at early stage and progressed to advanced stage (ranging from 15.1% for amNSCLC to 23.1% for amUC). Patient characteristics, treatments and survival outcomes are described in Table 1. Not all received systemic treatment; Immune Checkpoint inhibitors (ICI) were used in 5.3%, 2.2% and 10.8% of treated amUC, amGC and amNSCLC patients, respectively. Conclusions: In these cohorts of advanced or metastatic cancer patients, median survival time was limited despite most receiving treatment: radiation, systemic therapy or surgery. Treatment with ICI was low despite recent data in amUC and amNSCLC.[Table: see text]

Primary Care Provider Encounter Cadence and HbA1c Control in Older Patients With Diabetes

Primary care provider encounters are associated with health and well-being; however, limited evidence guides optimal primary care provider rate of visit, referred to as encounter cadence. This study measures associations between primary care provider encounter cadence and diabetes outcomes among individuals newly diagnosed with type 2 diabetes mellitus. In this retrospective cohort study, 7,106 people enrolled in Medicare Advantage and newly diagnosed with type 2 diabetes mellitus between July 1, 2012 and June 30, 2013 were identified and followed for 36 months. Two methods measured primary care provider encounter cadence: total primary care provider encounters (frequency) and quarters with primary care provider encounter (regularity). Logistic regression measured relationships between primary care provider encounter cadence and non-insulin diabetes medication adherence, HbA1c control, emergency department visits, and inpatient admissions. Non-insulin diabetes medication adherence was defined according to the National Committee for Quality Assurance, Healthcare Effectiveness Data and Information Set specifications and measured using healthcare claims data. Post-hoc models examined adherence and diabetes control among those nonadherent (n=5,212) and with noncontrolled HbA1c (n=326) during the encounter/cadence period. Data were extracted and analyzed in 2017. Adjusted models indicated that both frequency (AOR=1.08, 95% CI=1.06, 1.10) and regularity (AOR=1.18, 95% CI=1.13, 1.22) of primary care provider encounters were associated with increased odds of adherence. Post-hoc analyses indicated that more frequent (AOR=1.12, 95% CI=1.10, 1.15) and regular (AOR=1.27, 95% CI=1.22, 1.33) primary care provider encounters were associated significantly with adherence and were associated directionally with HbA1c control. More frequent and regular primary care provider encounters are associated with an increased likelihood of non-insulin diabetes medication adherence. These findings contribute to data needed to establish evidence-based guidelines for primary care provider encounter cadence for those newly diagnosed with type 2 diabetes mellitus.

Comparison of healthcare utilization among managed Medicaid individuals diagnosed with multiple sclerosis treated with emergent versus established disease modifying therapy

Background: Emerging disease-modifying therapies (DMTs) have evolved as an alternative treatment for patients with multiple sclerosis (MS). The efficacy and safety of established DMTs (interferons, glatiramer acetate, natalizumab, fingolimod and mitoxantrone) have been well studied and clinical trials with small sample sizes have suggested that emerging DMTs (iteriflunomide, dimethyl fumarate/BG-12, alemtuzumab and pegylated IFN) may have distinct advantages relative to established DMTs including better outcomes and reduced healthcare resource utilization. However, there is limited real-world information regarding which DMTs (established vs. emerging) provide the best clinical response and outcomes in managed care populations of patients with MS. Aims: To compare MS related healthcare use within one year of initiating emergent and established DMTs among Managed Medicaid individuals diagnosed with MS in the US. Methods: A large national sample of patient-level administrative healthcare claims data was used for this analysis. MS patients aged 18 years and over with a new prescription fill for an established or emergent DMT between 2013 and 2016 were evaluated. Patients were eligible if they were continuously enrolled in a health plan with pharmacy and medical coverage for at least 6 months before and 1 year after initiation of therapy. Four types of healthcare use were examined: MS-related hospitalizations, emergency room (ER) visits and relapse events (inpatient and outpatient). Multivariate negative binomial models with robust standard errors were used to estimate the association between MS related healthcare use and type of DMT. All models adjusted for age, gender, Charlson index and geographic region. Results: During the study period, 6981 Managed Medicaid individuals with a MS diagnosis initiated a DMT. Of those, 79.8% were female, 50.4% were aged 40–64 years and 21.5% were on emergent DMTs. Emergent DMT users had fewer hospitalizations compared to first generation DMT users within one year of initiating therapy (adjusted risk ratio [ARR] = 0.64, 95% confidence interval [CI]: 0.46–0.88) and fewer outpatient relapses (ARR = 0.86%, CI: 0.79–0.95). Differences in inpatient relapses and ER visits were not observed by DMT type. Conclusions: This study suggests emergent DMTs are associated with reduced MS-related hospitalizations and outpatient relapses within one year of initiating therapy. Studies examining a longer treatment time frame and additional outcomes are warranted to confirm these findings.