Healthcare Claims Data Research Articles

Continuation rates of newly prescribed donepezil among patients with Alzheimer’s disease in Japan: Life study

AbstractBackgroundAlzheimer’s disease (AD) is a major public health problem. Although donepezil is used as a first‐line pharmaceutical treatment for AD, studies from various countries have reported low continuation rates. However, the only available data from Japan are from post‐marketing observational study reports. This study investigated the continuation rates of newly prescribed donepezil among AD patients in Japan.MethodThe study was conducted using health care claims data and long‐term care claims data from 17 Japanese municipalities participating in the Longevity Improvement & Fair Evidence (LIFE) Study. The study period was from April 2014 to October 2021, and the study population comprised patients who had been newly prescribed donepezil after receiving a diagnosis of AD. The treatment discontinuation was defined as a gap of ≥61 days between the last donepezil prescription and subsequent donepezil, cholinesterase inhibitors, or memantine prescription. Kaplan‐Meier curves were plotted to estimate the duration (days) from the start of a new prescription of donepezil until discontinuation, and patients were censored at the end of the observation period or death. The treatment continuation rates were also analyzed according to the patients’ certified long‐term care needs levels (ranging from levels 1‐5) at the time of the initial prescription.ResultWe analyzed 20,474 patients (mean age: 82.2 years, women: 65.7%) with a mean follow‐up duration of 483 days. The treatment continuation rates at 30 days, 90 days, 360 days, and 540 days after initial prescription were 89.1%, 79.4%, 64.5%, and 58.3%, respectively. By excluding patients with long‐term care needs levels 3‐5, we analyzed 18,924 patients with lower care needs levels (mean age: 82.0 years, women: 65.4%); their treatment continuation rates were almost identical to those of all patients (<1% difference). Within 60 days of donepezil discontinuation, only 2.9% and 3.1% of patients had been switched to galantamine and rivastigmine, respectively.ConclusionThe AD patients showed a rapid discontinuation of donepezil and subsequent AD treatments by 90 days, and approximately 40% had ceased this therapy after 1.5 years. In view of the limited pharmaceutical options available for AD in Japan, this finding emphasizes the need for AD preventive measures in the pre‐clinical stage.

Work absenteeism and disability days after diagnosis among patients with AML and caregivers.

e19002 Background: Treatment for acute myeloid leukemia (AML) includes intensive chemotherapy and/or hematopoietic stem cell transplant, and patients usually require extended hospital stays and absence from work. Caregivers may also take workplace absence to provide care during and after treatment. As data on the indirect burden of AML are underreported, we aimed to understand the impact of work absenteeism (ABS) and disability days among patients with AML and caregivers. Methods: This non-interventional, retrospective analysis of healthcare claims data was conducted using MarketScan Databases containing ABS data. Two cohorts were analyzed: patients newly diagnosed with AML between January 1, 2009 and December 1, 2019 (index period), and caregivers with an adult family member newly diagnosed with AML during the index period. All participants were full-time employees, aged 18–64 years, and had ≥12 months of continuous enrollment prior to index date and ≥30 days of continuous enrollment after index date. The index date was the date of first AML diagnosis for patients, or the date of the linked family member’s first AML diagnosis for caregivers. Participants were followed for ≥30 days up to 3 years, until the end of continuous enrollment, end of database eligibility, or end of the study period. Participants who were pregnant or in families with > 1 member with AML were excluded. Days of work loss and associated wage loss were calculated for each work loss type: ABS, short-term disability (STD), and long-term disability (LTD). Results: This analysis included 1,037 patients with AML and 781 caregivers. From baseline to follow-up (FU) period, the proportion of patients reporting ABS numerically decreased from 72.8% to 62.7% (p = 0.054), and changes in ABS days per patient per month (PPPM) (2.21– 2.10, p = 0.735) or ABS-related wage loss PPPM ($562–$511, p = 0.516) were not statistically significant. However, the proportion of patients reporting STD and LTD significantly increased (24.6%–50.0%, p < 0.001, and 4.0%–20.0%, p < 0.001, respectively). The number of days PPPM lost due to STD and LTD significantly increased (0.85–4.94, p < 0.001, and 0.24–1.40, p < 0.001), as did STD- and LTD-related wage loss PPPM ($203–$1194, p < 0.001, and $57–$330, p < 0.001). Among caregivers, the proportion reporting ABS and ABS days PPPM were similar between baseline and FU periods (83.2%–83.9%, p = 0.871, and 2.41–2.64, p = 0.315). The proportion of caregivers with STD claims significantly increased (5.4%–11.8%, p < 0.001), and days PPPM lost due to STD and associated wage loss PPPM numerically increased (0.15–0.25, p = 0.057 and $38–$60, p = 0.087). Conclusions: Following AML diagnosis, the proportion of patients with STD and LTD leave increased significantly, while absenteeism did not change significantly. Days of STD leave also increased among caregivers. Future research is needed to determine how work loss varies by AML treatment type and/or sequencing.

The Impact of Supplemental Nutrition Assistance Program (SNAP) Enrollment on Health and Cost Outcomes

Summary In the United States, food insecurity is prevalent in low-income households and is associated with elevated risk of chronic disease, poor health outcomes, and excess avoidable health care utilization and costs. Individuals dually eligible for Medicare and Medicaid are a unique low-income group with significant social challenges, complex health conditions, and higher-than-average health care costs. Interventions to address social determinants of health (SDOH) — such as food insecurity — have the potential to reduce avoidable health care utilization and cost, and the health insurers are increasingly integrating them into care delivery. Using Supplemental Nutrition Assistance Program (SNAP) data from the Pennsylvania Department of Human Services and health care claims data from an integrated health care delivery and finance system, this study reports findings of an innovative collaboration between a health care payer and a nonprofit organization to enroll dual-eligible individuals in SNAP. Data were collected prospectively from 661 SNAP-enrolled individuals and 1,320 individuals in a matched comparison group to explore the effect of SNAP enrollment on health care utilization (inpatient hospitalization, ED visits, and unplanned care) and health care costs (medical, pharmacy, and total cost of care) during the first and second years after SNAP enrollment. Data analysis was designed to answer questions that are important to payer and health care stakeholders, namely, when (i.e., how long after the start of SNAP benefits), how (i.e., what health outcome), and for whom (i.e., what population subgroups) is SNAP enrollment associated with health and cost outcomes. Compared with a propensity score–matched group not enrolled in SNAP, the SNAP enrollees had 16% and 21% lower total health care costs and pharmacy costs, respectively, within the first year postenrollment in SNAP. These group differences remained significant during the second year postenrollment, when SNAP enrollees had 16% and 20% lower total health care costs and pharmacy costs, respectively. Specifically, lower costs were seen among individuals who do not qualify for skilled nursing care. This study contributes to the growing evidence that novel payer-led interventions to address SDOH can influence positive changes in health care utilization and cost reduction, especially among disadvantaged populations.

Development of a COVID-19 vaccine effectiveness and safety assessment system in Japan: The VENUS study.

There are currently no COVID-19 vaccine assessment systems in Japan that allow for the active surveillance of both vaccinated and unvaccinated persons. Herein, we describe the development of Japan's first COVID-19 vaccine effectiveness and safety assessment system with active surveillance capabilities. The Vaccine Effectiveness, Networking, and Universal Safety (VENUS) Study was developed as a multi-source database that links four data types at the individual resident level: Basic Resident Register (base population information), Vaccination Record System (vaccination-related information), Health Center Real-time Information-sharing System on COVID-19 (HER-SYS; information on COVID-19 occurrence), and health care claims data (information on diagnoses, hospitalizations, diagnostic tests, and treatments). These data were obtained from four municipalities. Individual residents were linked across the data types using five matching algorithms based on names, birth dates, and sex; the data were anonymized after linkage. To ascertain the viability of the VENUS Study's database for COVID-19 vaccine assessments, we examined the trends in COVID-19 vaccinations, COVID-19 cases, and polymerase chain reaction (PCR) test numbers. We also evaluated the linkage rates across the data types. Our multi-source database was able to monitor COVID-19 vaccinations, COVID-19 cases, and PCR test numbers throughout the pandemic. Using the five algorithms, the data linkage rates between the COVID-19 occurrence information in the HER-SYS and the Basic Resident Register ranged from 85·4% to 91·7%. If used judiciously with an understanding of each data source's characteristics, the VENUS Study can provide a viable data platform that facilitates active surveillance and comparative analyses for population-based research on COVID-19 vaccine effectiveness and safety in Japan.

Using Healthcare Claims Data and Machine Learning to Identify Health Disparities for Individuals with Diagnosed and Undiagnosed Familial Hypercholesterolemia

Using Healthcare Claims Data and Machine Learning to Identify Health Disparities for Individuals with Diagnosed and Undiagnosed Familial Hypercholesterolemia

Identification of pregnancies and their outcomes in healthcare claims data, 2008-2019: An algorithm.

Pregnancy is a condition of broad interest across many medical and health services research domains, but one not easily identified in healthcare claims data. Our objective was to establish an algorithm to identify pregnant women and their pregnancies in claims data. We identified pregnancy-related diagnosis, procedure, and diagnosis-related group codes, accounting for the transition to International Statistical Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis and procedure codes, in health encounter reporting on 10/1/2015. We selected women in Merative MarketScan commercial databases aged 15-49 years with pregnancy-related claims, and their infants, during 2008-2019. Pregnancies, pregnancy outcomes, and gestational ages were assigned using the constellation of service dates, code types, pregnancy outcomes, and linkage to infant records. We describe pregnancy outcomes and gestational ages, as well as maternal age, census region, and health plan type. In a sensitivity analysis, we compared our algorithm-assigned date of last menstrual period (LMP) to fertility procedure-based LMP (date of procedure + 14 days) among women with embryo transfer or insemination procedures. Among 5,812,699 identified pregnancies, most (77.9%) were livebirths, followed by spontaneous abortions (16.2%); 3,274,353 (72.2%) livebirths could be linked to infants. Most pregnancies were among women 25-34 years (59.1%), living in the South (39.1%) and Midwest (22.4%), with large employer-sponsored insurance (52.0%). Outcome distributions were similar across ICD-9 and ICD-10 eras, with some variation in gestational age distribution observed. Sensitivity analyses supported our algorithm's framework; algorithm- and fertility procedure-derived LMP estimates were within a week of each other (mean difference: -4 days [IQR: -13 to 6 days]; n = 107,870). We have developed an algorithm to identify pregnancies, their gestational age, and outcomes, across ICD-9 and ICD-10 eras using administrative data. This algorithm may be useful to reproductive health researchers investigating a broad range of pregnancy and infant outcomes.

Severe Maternal Morbidity and Postpartum Readmission Through 1 Year.

To conduct a cohort study to estimate risk for readmission through 1 year postpartum and the most common readmission diagnoses for individuals with and without severe maternal morbidity (SMM) at delivery. Using national health care claims data from IBM MarketScan Commercial Research Databases (now known as Merative), we identified all delivery hospitalizations for continuously enrolled individuals 15-49 years of age that occurred between January 1, 2016, and December 31, 2018. Severe maternal morbidity at delivery was identified using diagnosis and procedure codes. Individuals were followed for 365 days after delivery discharge, and cumulative readmission rates were calculated for up to 42 days, up to 90 days, up to 180 days, and up to 365 days. We used multivariable generalized linear models to estimate adjusted relative risks (aRR), adjusted risk differences, and 95% CIs for the association between readmission and SMM at each of the timepoints. The study population included 459,872 deliveries; 5,146 (1.1%) individuals had SMM during the delivery hospitalization, and 11,603 (2.5%) were readmitted within 365 days. The cumulative incidence of readmission was higher in individuals with SMM than those without at all timepoints (within 42 days: 3.5% vs 1.2%, aRR 1.44, 95% CI 1.23-1.68; within 90 days: 4.1% vs 1.4%, aRR 1.46, 95% CI 1.26-1.69); within 180 days: 5.0% vs 1.8%, aRR 1.48, 95% CI 1.30-1.69; within 365 days: 6.4% vs 2.5%, aRR 1.44, 95% CI 1.28-1.61). Sepsis and hypertensive disorders were the most common reason for readmission within 42 and 365 days for individuals with SMM (35.2% and 25.8%, respectively). Severe maternal morbidity at delivery was associated with increased risk for readmission throughout the year after delivery, a finding that underscores the need for heightened awareness of risk for complications beyond the traditional 6-week postpartum period.

Association of prescription drug monitoring programs with benzodiazepine prescription dispensation and overdose in adolescents and young adults

Introduction Prescription drug monitoring programs are state-run databases designed to support safe prescribing of controlled substances and reduce prescription drug misuse. We analyzed healthcare claims data to determine the association between prescription drug monitoring programs with mandated provider review and adolescent and young adult benzodiazepine prescription dispensing and overdose. Methods We performed a state-level retrospective cohort study to evaluate the association between implementation of prescription drug monitoring programs with mandated provider review and benzodiazepine prescription dispensing and benzodiazepine-related overdoses among adolescents (13–18 years) and young adults (19–25 years) between 1 January 2008 and 31 December 2019. Data were obtained from a United States commercial health insurance company. Results There were 74,539 (1.8%) adolescents and 246,760 (4.0%) young adults with at least one benzodiazepine prescription dispensed. Benzodiazepine overdoses occurred among 1,569 (0.04%) and 3,202 (0.05%) adolescents and young adults, respectively. Implementation of a prescription drug monitoring program with mandated provider review was associated with a 6.8% (95% CI, 1.6–11.8) yearly reduction in benzodiazepine prescription dispensing among adolescents and a 12.5% (95% CI, 9.3–15.5) yearly reduction among young adults. There was no decrease in benzodiazepine overdoses in either age group (-15.4% [95% CI, −21.5 to 3.0] and −8.0% [95% CI, −18.0 to 3.2] yearly change in adolescents and young adults, respectively). Discussion Consistent with prior work, our study did not find an association between prescription drug monitoring program implementation and reduction in benzodiazepine-related overdoses among adolescents and young adults. However, the substantial reduction in benzodiazepine prescription dispensing is encouraging. Conclusion Prescription drug monitoring programs were associated with decreases in benzodiazepine prescription dispensing, but not benzodiazepine-related overdoses in this cohort of adolescents and young adults. These findings serve to inform development of further policies to address rising rates of benzodiazepine misuse and overdose in this patient population.

Assessing the impact of transplant case management on clinical outcomes.

Case management is commonly used by health plans to attempt to improve the care received by their members who have complex needs, such as those who undergo transplantation. There are few observational studies evaluating the effects that transplant case management programs have on clinical outcomes following a solid organ transplant. This limits the understanding of the quantitative effectiveness of such programs. This retrospective cohort study of solid organ transplant recipients with access to a transplant case management program used a case-control study design. Propensity score 1:1 matching was used to balance the comparison groups on demographic and pretransplant clinical characteristics. Health care claims data were used to determine whether program participation affected clinical outcomes following the transplant. A cohort of 1756 adults 18 years and older (878 cases and 878 controls) who had a solid organ transplant between 2018 and 2020 was followed beginning at the time of referral to transplant until 90 days following the transplant procedure. Transplant recipients who participated in the case management program had significantly lower 30-day and 90-day rejection rates, fewer 90-day readmissions, lower discharge mortality and 90-day mortality, and fewer bed days post transplant compared with those who did not participate in case management. Patients undergoing a solid organ transplant had improved clinical outcomes when they participated in a specialized case management program sponsored by their health plan.

Abstract P1-07-04: Claims-derived comorbidities associations with real-world overall survival (rwOS) among patients with metastatic breast cancer in an electronic health record-claims linked dataset

Abstract Introduction: Comorbid conditions can significantly impact various aspects of cancer care and therefore are often utilized as covariates to control for selection bias in observational studies. Oncology based electronic health record (EHR) data is often used as a source for comorbidity information but may be subject to missingness due to differences in clinician documentation. Utilizing additional data sources, such as health care claims data, may allow for greater capture of comorbidities. This study used a claims data linked to oncology EHR to identify comorbidities and related patient characteristics, as well as the association between a claims-derived comorbidity index and real-world overall survival (rwOS) among outpatient metastatic breast cancer (mBC) patients, to characterize the comorbidities data from this data source. Methods: We selected patients diagnosed with mBC between 2013 and 2021 from a linkage of mBC de-identified EHR-derived Flatiron Health Research Database (FHRD) and Komodo Health claims data. Charlson Comorbidity Index (CCI) were identified based on the presence of a pre-specified International Classification of Diseases (ICD) codes within 12 months preceding cancer diagnosis. To understand comorbidity data within this linkage, patient characteristics were assessed after the study cohort was stratified by the presence of any comorbidity assessed as part of CCI. We calculated the CCI for the study cohort as well as the prevalence of comorbid conditions that are used as part of the CCI. Kaplan-Meier estimates and Cox proportional hazards model were used to estimate median survival time and hazard ratios (HRs) of rwOS among patients with a CCI score of 1 and 2+. Results: The study cohort had 3,213 patients with a median age of 63 [IQR(52, 76)] and 98.7% were female. White patients accounted for 60.1% of the study population while Black and Hispanic/Latino patients accounted for 12.5% and 7.4% of the study cohort, respectively. Patients with comorbidities were older (mean age of 65 vs 61), had a higher (2+) ECOG performance status (PS) score (14.0% vs 9.7%), and were more likely to be Black (16.0% vs 11.0%) or Hispanic/Latino 10.8% vs 6.3%). Twenty percent of the patients had diabetes, 8.2% had peripheral vascular disease, 7.5% had mild liver disease, and 6.3% had congestive heart failure. According to index values, 21.0% of the study cohort had a CCI score of 1, and 18.0% had a score of 2 and above. The median survival times in years for patients with a CCI score of 1 and 2+ were 2.51 [95% CI 2.29,2.81] and 2.05 [95% CI 1.83,2.42], respectively. Based on univariate analysis, patients with a CCI score greater than zero had a higher risk of mortality (CCI 1: HR = 1.25 [95%CI 1.10, 1.41, p< 0.01], CCI 2+: HR = 1.54 [95%CI 1.35, 1.76, p< 0.01]). In the multivariate analysis, adjusted for age, ethnicity/race, and baseline ECOG PS score, the HR for the CCI 1 group was not statistically significant (HR=1.12 [95%CI 0.99, 1.27, p=0.08]). However, the risk of mortality was statistically significant for the CCI 2+ group (HR=1.31 [95%CI 1.14, 1.50], p < 0.01) in the adjusted analysis. Conclusion: Claims data linked to EHR can be used to identify comorbidities and describe patient characteristics. As anticipated, generally, the presence of more comorbid conditions was associated with worse rwOS for patients with mBC. This finding supports use of this linked dataset for similar future studies. Citation Format: Mustafa S. Ascha, Alemseged A. Asfaw, Prakirthi Yerram, Samantha N. Reiss, Sarah D. Brake, Niquelle B. Wadé. Claims-derived comorbidities associations with real-world overall survival (rwOS) among patients with metastatic breast cancer in an electronic health record-claims linked dataset [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P1-07-04.

TRICARE Extended Care Health Option Program: Prevalence of pediatric ECHO enrollees and healthcare service utilization in the Military Health System

BackgroundThe Extended Care Health Option (ECHO) Program is a TRICARE program aimed at reducing the disabling effects of chronic medical conditions for beneficiaries of the Department of Defense (DoD) healthcare program. However, little is known about military-connected children enrolled in the program. Objective/HypothesisThe aim of this study was to examine the demographic makeup of pediatric ECHO beneficiaries and their healthcare claims data. This is the first study to evaluate healthcare utilization of this subset of military dependents. MethodsA cross-sectional study was performed evaluating ECHO enrolled pediatric beneficiaries and their health service utilization during 2017–2019. TRICARE claims and military treatment facility (MTF) encounter data were utilized to evaluate health service utilization and identify the most frequently reported ICD-10-CM and CPT codes associated with care for this population. ResultsOf the 2,001,619 dependents aged 0–26 years who received medical care in the Military Health System (MHS) during 2017–2019, 21,588 individuals (1.1%) were enrolled in ECHO. The majority of encounters (65.4%) were provided in the MTFs. Inpatient visits, therapeutic services, and in-home nursing care were the top utilized private sector care services. Outpatient visits encompassed 94.8% of healthcare encounters, and neurodevelopmental disorders were the top principal diagnoses among ECHO beneficiaries. ConclusionsWith the increasing prevalence of children with medical complexity and developmental delay, the pediatric TRICARE beneficiaries eligible for ECHO will likely continue to rise. Improving services and supports for military children with special healthcare needs is needed to maximize their developmental trajectory.

Association Between Naloxone Coprescription Mandates and Postoperative Outcomes.

The opioid epidemic is a public health issue in the United States. The objective of this study was to evaluate the association between naloxone coprescription mandates and postoperative outcomes. Data on naloxone coprescription mandates show mixed evidence for fatal overdoses in the broader population. How these mandates have impacted surgical patients has not been fully explored. Healthcare claims data were used to identify all patients undergoing 1 of 50 common procedures between January 1, 2004, and June 30, 2019, and categorized as high risk for opioid overdose. The primary outcomes were an emergency department visit or hospital admission within 30 postoperative days. To reduce confounding, the association between this outcome and the implementation of naloxone coprescription mandates was estimated using a difference-in-differences approach. The study included 429,878 surgical patients with an average age of 54.8 years (SD=15.9 years) and with 257,728 females (60.0%). There was no significant association between naloxone prescribing mandates and the primary outcomes. After adjustment for potential confounders, the incidence of hospital admission was 3.26% after implementation of a naloxone coprescription mandate compared with 3.33% before (difference change: -0.08%, 95% CI: -0.44% to 0.29%, P =0.68). The incidence of an emergency department visit was 7.06% after implementation of a naloxone coprescription mandate compared with 7.73% before (difference: -0.67%, 95% CI: -1.39% to 0.05%, P =0.07). These results were robust to a variety of sensitivity and subgroup analyses. Naloxone coprescription mandates were not associated with a statistically or clinically significant change in emergency department visits or hospital admissions within 30 postoperative days.

Trends in Oral Antibiotic Use for Acne Treatment: A Retrospective, Population-Based Study in the United States, 2014 to 2016.

Antibiotic resistance related to prolonged antibiotic use is an emerging threat to public health. To evaluate recent trends in oral antibiotic use for acne treatment. A retrospective study was conducted from January 2014 through September 2016 using the IBM MarketScan® claims database. Patients were aged ≥9 years, prescribed an oral antibiotic, and diagnosed with acne vulgaris on 2 separate occasions. The primary outcome was the duration of oral antibiotic treatment over 12 months; continuous use was defined as ≤30-day gap between prescriptions. The most commonly prescribed antibiotic treatments (N=46,267) were doxycycline (36.7%) and minocycline (36.5%). Overall, 36%, 18%, 10%, and 5% of patients continuously used any oral antibiotic at 3, 6, 9, and 12 months, respectively. Among patients who continuously used tetracyclines, a similar percentage was prescribed minocycline (40.2%, 18.6%, 10.5%, and 5.1%) vs doxycycline (34.7%, 14.6%, 7.7%, and 3.9%) at 3, 6, 9, and 12 months, respectively. A greater percentage of patients continued use of tetracyclineclass antibiotics than other therapeutic classes. Retrospective analysis of health-care claims data. Relatively short study duration. Nearly 20% of patients continuously used oral antibiotics for ≥6 months, exceeding American Academy of Dermatology guideline recommendations of 3 to 4 months. J Drugs Dermatol. 2023;22(3):265-270. doi:10.36849/JDD.7345.

The application of machine learning to predict high-cost patients: A performance-comparison of different models using healthcare claims data.

Our aim was to predict future high-cost patients with machine learning using healthcare claims data. We applied a random forest (RF), a gradient boosting machine (GBM), an artificial neural network (ANN) and a logistic regression (LR) to predict high-cost patients in the following year. Therefore, we exploited routinely collected sickness funds claims and cost data of the years 2016, 2017 and 2018. Various specifications of each algorithm were trained and cross-validated on training data (n = 20,984) with claims and cost data from 2016 and outcomes from 2017. The best performing specifications of each algorithm were selected based on validation dataset performance. For performance comparison, selected models were applied to unforeseen data with features of the year 2017 and outcomes of the year 2018 (n = 21,146). The RF was the best performing algorithm measured by the area under the receiver operating curve (AUC) with a value of 0.883 (95% confidence interval (CI): 0.872-0.893) on test data, followed by the GBM (AUC = 0.878; 95% CI: 0.867-0.889). The ANN (AUC = 0.846; 95% CI: 0.834-0.857) and LR (AUC = 0.839; 95% CI: 0.826-0.852) were significantly outperformed by the GBM and the RF. All ML algorithms and the LR performed ´good´ (i.e. 0.9 > AUC ≥ 0.8). We were able to develop machine learning models that predict high-cost patients with 'good' performance facilitating routinely collected sickness fund claims and cost data. We found that tree-based models performed best and outperformed the ANN and LR.

Development and Validation of Coding Algorithms to Identify Patients with Incident Non-Small Cell Lung Cancer in United States Healthcare Claims Data.

We sought to develop and validate an incident non-small cell lung cancer (NSCLC) algorithm for United States (US) healthcare claims data. Diagnoses and procedures, but not medications, were incorporated to support longer-term relevance and reliability. Patients with newly diagnosed NSCLC per Surveillance, Epidemiology, and End Results (SEER) served as cases. Controls included newly diagnosed small-cell lung cancer and other lung cancers, and two 5% random samples for other cancer and without cancer. Algorithms derived from logistic regression and machine learning methods used the entire sample (Approach A) or started with a previous algorithm for those with lung cancer (Approach B). Sensitivity, specificity, positive predictive values (PPV), negative predictive values, and F-scores (compared for 1000 bootstrap samples) were calculated. Misclassification was evaluated by calculating the odds of selection by the algorithm among true positives and true negatives. The best performing algorithm utilized neural networks (Approach B). A 10-variable point-score algorithm was derived from logistic regression (Approach B); sensitivity was 77.69% and PPV = 67.61% (F-score = 72.30%). This algorithm was less sensitive for patients ≥80 years old, with Medicare follow-up time <3 months, or missing SEER data on stage, laterality, or site and less specific for patients with SEER primary site of main bronchus, SEER summary stage 2000 regional by direct extension only, or pre-index chronic pulmonary disease. Our study developed and validated a practical, 10-variable, point-based algorithm for identifying incident NSCLC cases in a US claims database based on a previously validated incident lung cancer algorithm.

Cost Savings and Reduced Health Care Utilization Associated with Participation in a Digital Diabetes Prevention Program in an Adult Workforce Population.

BackgroundThough in-person delivery of the Diabetes Prevention Program (DPP) has demonstrated medical cost savings, the economic impact of digital programs is not as well understood.ObjectiveThis study examines the impact of a digital DPP program on reducing all-cause health care costs and utilization among 2027 adult participants at 12 months.MethodsA longitudinal, observational analysis of health care claims data was conducted on a workforce population who participated in a digital diabetes prevention program. Differences in utilization and costs from the year prior to program delivery through 1 year after enrollment were calculated using medical claims data for digital DPP participants compared to a propensity matched cohort in a differences-in-differences model.ResultsAt 1 year, the digital DPP population had a reduction in all-cause health care spend of US$1169 per participant relative to the comparison group (P = 0.01), with US$699 of that savings coming from reduced inpatient spend (P = 0.001). Cost savings were driven by fewer hospital admissions and shorter length of stay (P < 0.001). No other significant results in cost differences were detected. There was a trend toward savings extending into the second year, but the savings did not reach statistical significance.ConclusionsThese results demonstrated significant short-term health care cost savings at 1 year associated with digital DPP program delivery.

P405: Project FIND-OUT (Fast Infant Neurogenetic Diagnosis via Outpatient Testing): Utilization of healthcare claims data to inform protocol design

P405: Project FIND-OUT (Fast Infant Neurogenetic Diagnosis via Outpatient Testing): Utilization of healthcare claims data to inform protocol design

Studies of Health Insurance Claims Data in Japan: A Scoping Review.

Health insurance claims data are used in various research fields; however, an overview on how they are used in healthcare research is scarce in Japan. Therefore, we conducted a scoping review to systematically map the relevant studies using Japanese claims data. MEDLINE, EMBASE, and Ichushi-Web were searched up to April 2021 for studies using Japanese healthcare claims data. We abstracted the data on study characteristics and summarized target diseases and research themes by the types of claims database. Moreover, we described the results of studies that aimed to compare health insurance claims data with other data sources narratively. A total of 1,493 studies were included. Overall, the most common disease classifications were "Diseases of the circulatory system" (18.8%, n = 281), "Endocrine, nutritional, and metabolic diseases" (11.5%, n = 171; mostly diabetes), and "Neoplasms" (10.9%, n = 162), and the most common research themes were "medical treatment status" (30.0%, n = 448), "intervention effect" (29.9%, n = 447), and "clinical epidemiology, course of diseases" (27.9%, n = 417). Frequent diseases and themes varied by type of claims databases. A total of 19 studies aimed to assess the validity of the claims-based definition, and 21 aimed to compare the results of claims data with other data sources. Most studies that assessed the validity of claims data compared to medical records were hospital-based, with a small number of institutions. Claims data are used in various research areas and will increasingly provide important evidence for healthcare policy in Japan. It is important to use previous claims database studies and share information on methodology among researchers, including validation studies, while informing policymakers about the applicability of claims data for healthcare planning and management.

LONELINESS DURING THE COVID-19 PANDEMIC AND THE EFFECT ON PSYCHOSOCIAL FACTORS AND HEALTH OUTCOMES

Abstract Loneliness and social isolation have harmful impacts on health and well-being; thus, the social distancing mandates during the COVID-19 pandemic, meant to protect our most vulnerable populations including older adults, may have had unintended consequences. The current study aimed to assess changes in loneliness (prior to and during the pandemic) and its impact on outcomes such as perceived stress, resilience, purpose in life, quality of life and health outcomes. An annual survey on healthy aging was mailed to a randomly selected national sample (age &amp;gt;65 years) in 2018/2019. Respondents completed the survey again in 2020 (N=3,564) to measure the impact of COVID-19. Measures included well-being and various psychosocial factors. Health care claims data were also used to derive diagnoses for health conditions. Respondents were 52% female and 44% between 65-74 years old. Loneliness symptoms increased for one-fifth of the sample during the pandemic. Respondents lonely at both time 1 and time 2 had lower resilience, purpose in life, and quality of life and higher perceived stress compared to those only lonely at one time point or not lonely. Respondents lonely at both time points had higher rates of health conditions compared to non-lonely respondents. In conclusion, older adults experienced increased levels of loneliness during COVID-19. Those with chronic loneliness had worse outcomes than those who were not lonely or inconsistently lonely and they exhibit lower resilience and purpose and higher stress and health complications. These findings suggest that timing and early intervention for loneliness is important for older adults.

Association Between the 2021 Heat Wave in Portland, Oregon, and Seattle, Washington, and Emergency Department Visits

This study used a health care claims data set of enrollees in commercial and Medicare Advantage insurance plans to assess the association between the June 2021 heat wave and the rates of emergency department visits in Portland, Oregon, and Seattle, Washington.