Background: Agreement between diagnosis codes for tuberculosis (TB) or latent tuberculosis infection (LTBI) in electronic health records (EHRs) and insurance claims for the same person has not been assessed.

Objective: To measure agreement levels between TB-related Systematized Nomenclature of Medicine–Clinical Terms (SNOMED-CT) diagnoses in patient records and International Statistical Classification of Diseases, 10^th rev., Clinical Modification (ICD-10-CM) diagnosis codes in insurance claims paid during 2015–2018 for a cohort of US patients, using a large nationwide database.

Methods: In a US population-based, retrospective cohort study of EHRs with commercial claims-linked data, Athena Software^® was used to match TB-related SNOMED-CT and ICD-10-CM codes. LTBI was identified by using a published ICD-based algorithm and all LTBI- and TB-related SNOMED-CT codes. Percentages of patients with SNOMED-CT and ICD-10-CM codes that matched exactly indicated agreement levels.

Results: Of persons with the 10 most frequent TB SNOMED-CT codes, 40% did not have an ICD-10-CM code that matched exactly. The code for positive tuberculin skin test was the most frequent unmatched SNOMED-CT code. Of persons with the 10 most frequent TB-related ICD-10-CM codes, 50% did not have a SNOMED-CT code that matched exactly. The most frequent unmatched ICD-10-CM code was for TB screening encounter. SNOMED-CT codes for LTBI matched to ICD-10-CM codes for TB testing; pulmonary TB; and nonspecific, positive, or adverse tuberculin reaction.

Conclusion: TB-related SNOMED-CT and ICD-10-CM diagnostic codes often disagree, and persons with ICD-10-CM codes for LTBI have unexpected SNOMED-CT codes, indicating a need for reconciling coding systems.

Medicare-Medicaid dual-eligible patients pose a special challenge, often requiring coordination-of-care between two separate programs for medical services and personal care services. Care management (CM) is the bridge for preventing adverse outcomes for patients, and reducing health care costs. Many patients don’t understand what care managers do for doctors and patients; and some providers don’t see the need for CM.

<i>(1) Methodology:</i> This study adapted core measures from Medicare CAHPS into an off-season survey (2019 n=3,114; 2018 n=1,853; 2017 n=316) with an extensive module of questions on patients’ perceptions about CM, and reasons for wanting or not wanting CM.

<i>(2) Setting:</i> Medicare-Medicaid patients at a large and diverse Medicaid health plan in the southwestern United States.

<i>(3) Analysis:</i> The analysis examined drivers of patients’ willingness to accept CM or not. 32.1% of patients reported having a care manager. 67.9% reported not having one (among whom 23.6% definitely want one; 26.3% might want; 18.5% “probably don’t want one” and 31.7% definitely don’t). Multiple choice reasons included: Not perceiving the need for CM (33.6% self-reliant, 47.5% get care from family/friends; 45.9% get care from other sources); not understanding what CM does (32.0%). Difference-of-proportions tests (a=0.05) revealed that older patients were least likely to want a care manager. Black patients were most likely to want care managers.

<i>(4) Implications:</i> Patients whose conditions might merit care management, but who don’t want a care manager, nevertheless noted services among 14 activities-of-daily-living (ADLs) and instrumental activities-of-daily-living (IADLs) that they would accept (<i>(when not described as care management</i>).

In 2017, the estimated total cost of diagnosed diabetes was $327 billion in the U.S., including $237 billion in direct costs and $90 billion in reduced productivity. Little is known of the incremental health cost by time period of diagnosis of chronic condition (CC) in diabetic patients.

Our study aims to estimate (i) treated prevalence of new vs. old diagnosis of chronic conditions in diabetics, and (ii) cost of each of the health services utilized, in addition to the total expenditure, using nationally representative Medical Expenditure Panel Survey data 2015 to 2017 for the U.S. civilian noninstitutionalized population.

The main independent variable would be presence of diabetes in adults. The outcome variables of utilization would be compared with three groups, namely diabetics only, diabetes + newly diagnosed (<=1year) CC, and diabetic + old (>1year) CC. Covariates includes body mass index, smoking status, usual source of care, Hb1Ac test, confident in in taking care of his/her diabetes, socio-demographic factors. To estimate the outcome variables related to cost, we would use two-part econometric model with logistic regression as first part and generalized linear regression for second part with log link function and gamma distribution. and generalized linear model would be used to estimate the medical expenditures.

The data analysis has not started. The preliminary results show higher prevalence of hypertension (76.86%) in diabetics.

The results of this study would help clinicians and policymakers to develop strategies to reduce the utilization, focus on making current inefficient services efficient and eventually reducing healthcare expenditure.