Claims-based Algorithm Research Articles

Identification of cytokine release syndrome and indicators of severity in retrospective databases among patients receiving immunotherapy.

Cytokine release syndrome (CRS) can occur following cancer immunotherapies, but is most often mild and of limited duration. International Classification of Diseases (ICD)-10 codes allowing identification of CRS were introduced in 2020 but may be underutilized. We evaluated the performance of a published claims-based algorithm to detect CRS (any grade) and high-grade CRS (HG, grades 2-5), as well as identified indicators of HG CRS in retrospective data. Adults with low-grade and HG CRS during an encounter coinciding with administrations of blinatumomab or chimeric antigen receptor-T therapy were identified in three types of retrospective databases (hospital chargemaster data, electronic health records, and administrative claims). The algorithm's sensitivity in detecting any CRS and HG CRS was reported. A least absolute shrinkage and selection operator (LASSO) regression model was developed to identify indicators of HG CRS. Performance of the model was evaluated using area under the curve (AUC). The sensitivity of the algorithm to detect any grade CRS ranged between 77%-100% and between 8%-80% for HG CRS, depending on the type of database. The LASSO model identified hypotension, positive pressure (including mechanical ventilation), tocilizumab, and vasopressors as indicators of HG CRS. AUC varied between 60% and 75%. The algorithm accurately detected any grade CRS for over three-quarters of instances, but was not as reliable for HG CRS. Results varied based on database attributes. Hypotension, vasopressors, positive pressure, and tocilizumab were associated with HG CRS and may be methodologically helpful signals of CRS severity in retrospective data.

Abstract 4124585: Identification of Low-Value Carotid Ultrasound Tests: A Comparison of Administrative Claims Versus Chart Review

Introduction/Background: Low-value care is defined as care where harms or costs outweigh patient benefits. Prior studies have suggested that approximately 1 of every 10 carotid ultrasound tests are low-value because they were performed among patients without neurologic symptoms or a history of stroke. However, these studies have used administrative claims data that may not account for all patient symptoms or include all guideline-recommended indications. We assessed the validity of a claims-based measure of low-value carotid ultrasound testing compared to a reference standard of chart review. Hypothesis: Claims-based measures overestimate the proportion of carotid ultrasound tests classified as low-value care. Methods: We conducted a retrospective cohort study of Medicare fee-for-service beneficiaries attributed to an accountable care organization with administrative claims data linked to electronic health record data. Using Current Procedural Technology codes, the first 150 patients that underwent carotid ultrasound in 2021 were identified. We applied a claims-based algorithm used by researchers and policymakers to identify if testing was low-value or appropriate. We used linked electronic health record data to assess appropriateness of testing using the 2012 American Heart Association clinical practice guidelines. We assessed agreement between claims and chart review using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and Cohen’s Kappa. Results: Among 150 patients (mean age 76, 55% male, 86% white) who underwent carotid ultrasound, 76 tests (51%) were considered low value by the claims measure and 63 (42%) by chart review. Claims based measures were 67% sensitive, 61% specific, with a PPV of 55% and NPV of 62%. Cohen’s Kappa coefficient was 0.27 (CI 0.11 - 0.42), indicating only fair agreement between claims and chart review. Conclusion: We found that an often-used claims-based measure to identify low-value carotid ultrasound testing misclassifies a substantial proportion of tests as low-value. Using claims-based measures alone for assessing provider quality and making payment decisions could harm provider reputations and erode trust in the quality reporting system.

Claims-Based Algorithm to Identify Pre-Exposure Prophylaxis Indications for Tenofovir Disoproxil Fumarate and Emtricitabine Prescriptions (2012-2014): Validation Study.

To monitor the use of tenofovir disoproxil fumarate and emtricitabine (TDF/FTC) and related medicines for pre-exposure prophylaxis (PrEP) as HIV prevention using commercial pharmacy data, it is necessary to determine whether TDF/FTC prescriptions are used for PrEP or for some other clinical indication. This study aimed to validate an algorithm to distinguish the use of TDF/FTC for HIV prevention or infectious disease treatment. An algorithm was developed to identify whether TDF/FTC prescriptions were for PrEP or for other indications from large-scale administrative databases. The algorithm identifies TDF/FTC prescriptions and then excludes patients with International Classification of Diseases (ICD)-9 diagnostic codes, medications, or procedures that suggest indications other than for PrEP (eg, documentation of HIV infection, chronic hepatitis B, or use of TDF/FTC for postexposure prophylaxis). For evaluation, we collected data by clinician assessment of medical records for patients with TDF/FTC prescriptions and compared the assessed indication identified by the clinician review with the assessed indication identified by the algorithm. The algorithm was then applied and evaluated in a large, urban, community-based sexual health clinic. The PrEP algorithm demonstrated high sensitivity and moderate specificity (99.6% and 49.6%) in the electronic medical record database and high sensitivity and specificity (99% and 87%) in data from the urban community health clinic. The PrEP algorithm classified the indication for PrEP in most patients treated with TDF/FTC with sufficient accuracy to be useful for surveillance purposes. The methods described can serve as a basis for developing a robust and evolving case definition for antiretroviral prescriptions for HIV prevention purposes.

Development and Validation of Claims-Based Algorithms for Conjunctivitis and Keratitis.

Ocular surface disorders have been reported among patients with various medical conditions, including atopic dermatitis (AD). Nonetheless, validated algorithms to identify conjunctivitis and keratitis in claims data are lacking. Develop validated, claims-based algorithms for conjunctivitis and keratitis among patients with AD using medical records. Patients with AD were identified in a claims database between March 2017 and November 2019. Among these patients, candidate algorithms were developed that included diagnosis codes for conjunctivitis or keratitis, alone and in combination with ophthalmic treatments. Among patients who met ≥ 1 candidate algorithms, a subset was randomly selected for medical record review. Additionally, records from a random sample of patients with AD were reviewed to assess sensitivity. Overall, 341 records were sought and 262 adjudicated by an expert ophthalmologist. The positive predictive value (PPV) of each algorithm was calculated and compared to a pre-specified threshold of ≥ 70%. For conjunctivitis, the final algorithm was ≥ 1 conjunctivitis diagnosis (PPV = 81%, 95% confidence interval [CI]: 73%-87%). For keratitis, the final algorithm combined the following 2 candidate algorithms: ≥ 1 keratitis diagnosis and ≥ 1 dispensing of a topical antibiotic or antibiotic-steroid combination (PPV = 91%); and ≥ 1 keratitis diagnosis and ≥ 1 dispensing of an ophthalmic corticosteroid, topical immune-modulator, or topical NSAID (PPV = 68%) for an overall PPV of 80% (95% CI: 55%-93%). The first claims-based algorithms to identify conjunctivitis and keratitis among AD patients were developed and validated. They are available for use in future studies, particularly to better understand conjunctivitis and keratitis occurrence among patients with AD.

Adherence to First-Line Intravesical Bacillus Calmette-Guérin Therapy in the Context of Guideline Recommendations for US Patients With High-Risk Non-muscle Invasive Bladder Cancer.

Background: Bacillus Calmette-Guérin (BCG) can reduce recurrence and delay progression among patients with high-risk non-muscle invasive bladder cancer (NMIBC), but is associated with a substantial emotional, physical, and social burden. Objectives: This study evaluated the adequacy of first-line intravesical BCG treatment among high-risk NMIBC patients in the United States, including the subgroup with carcinoma in situ (CIS) of the bladder. Methods: Adults with high-risk NMIBC treated with BCG were selected from de-identified MarketScan® Commercial, Medicare, and Medicaid Databases (1/1/2010-2/28/2021). Adequacy of BCG induction and maintenance was evaluated from the first BCG claim until the end of the patient's observation, using a previously published claims-based algorithm (induction: ≥5 instillations within 70 days; induction and maintenance: ≥7 instillations within 274 days of first instillation) and a definition based on the landmark Southwest Oncology Group (SWOG) trial (induction: ≥5 instillations without gaps >7 days; followed by ≥2 instillations at month 3, 6, and every 6 months thereafter). Proportions of patients with adequate BCG induction and maintenance were reported overall and compared between those with and without CIS. Results: Of 5803 high-risk NMIBC patients treated with first-line BCG (mean age, 67.3 years; 20.6% female), 930 (16.0%) had CIS. After first-line BCG, 56.6% received another treatment. Although 86.9% had adequate BCG induction based on the claims-based algorithm (SWOG, 73.6%), only 41.5% had adequate BCG induction and maintenance (SWOG, 1.6%). Similar trends were observed for patients with and without CIS, with higher adherence to guidelines for patients with CIS (adequate induction using claims-based algorithm: 90.3% vs 86.2%; adequate induction and maintenance: 50.8% vs 39.7%, all P < .001). A greater proportion of CIS patients than non-CIS patients had cystectomy (CIS, 14.4%, non-CIS, 8.5%; P < .001) after first-line BCG. Discussion: Among patients with NMIBC treated with first-line intravesical BCG, most received adequate BCG induction but less than half had adequate BCG maintenance. BCG treatment was also inadequate for patients with CIS, with only half of patients receiving adequate BCG maintenance and a higher proportion undergoing cystectomy following first-line BCG. Conclusions: Results emphasize the need for additional treatment options for patients with NMIBC.

Adherence to First-Line Intravesical Bacillus Calmette-Guérin Therapy in the Context of Guideline Recommendations for US Patients With High-Risk Non-muscle Invasive Bladder Cancer

Background: Bacillus Calmette-Guérin (BCG) can reduce recurrence and delay progression among patients with high-risk non–muscle invasive bladder cancer (NMIBC), but is associated with a substantial emotional, physical, and social burden. Objectives: This study evaluated the adequacy of first-line intravesical BCG treatment among high-risk NMIBC patients in the United States, including the subgroup with carcinoma in situ (CIS) of the bladder. Methods: Adults with high-risk NMIBC treated with BCG were selected from de-identified MarketScan® Commercial, Medicare, and Medicaid Databases (1/1/2010-2/28/2021). Adequacy of BCG induction and maintenance was evaluated from the first BCG claim until the end of the patient’s observation, using a previously published claims-based algorithm (induction: ≥5 instillations within 70 days; induction and maintenance: ≥7 instillations within 274 days of first instillation) and a definition based on the landmark Southwest Oncology Group (SWOG) trial (induction: ≥5 instillations without gaps &gt;7 days; followed by ≥2 instillations at month 3, 6, and every 6 months thereafter). Proportions of patients with adequate BCG induction and maintenance were reported overall and compared between those with and without CIS. Results: Of 5803 high-risk NMIBC patients treated with first-line BCG (mean age, 67.3 years; 20.6% female), 930 (16.0%) had CIS. After first-line BCG, 56.6% received another treatment. Although 86.9% had adequate BCG induction based on the claims-based algorithm (SWOG, 73.6%), only 41.5% had adequate BCG induction and maintenance (SWOG, 1.6%). Similar trends were observed for patients with and without CIS, with higher adherence to guidelines for patients with CIS (adequate induction using claims-based algorithm: 90.3% vs 86.2%; adequate induction and maintenance: 50.8% vs 39.7%, all P &lt; .001). A greater proportion of CIS patients than non-CIS patients had cystectomy (CIS, 14.4%, non-CIS, 8.5%; P &lt; .001) after first-line BCG. Discussion: Among patients with NMIBC treated with first-line intravesical BCG, most received adequate BCG induction but less than half had adequate BCG maintenance. BCG treatment was also inadequate for patients with CIS, with only half of patients receiving adequate BCG maintenance and a higher proportion undergoing cystectomy following first-line BCG. Conclusions: Results emphasize the need for additional treatment options for patients with NMIBC.

Validation Study of the Claims-Based Algorithm Using the International Classification of Diseases Codes to Identify Patients With Coronavirus Disease in Japan From 2020 to 2022: The VENUS Study.

We validated claims-based algorithms using the International Classification of Diseases, Tenth Revision (ICD-10) to identify patients with the first-ever coronavirus disease (COVID-19) onset between May 2020 and August 2022. The study cohort was comprised of residents of one municipality enrolled in a public insurance program. This study used data provided by the municipality, including residents' insurer-based medical claims data linked to the Health Center Real-time Information-Sharing System (HER-SYS). The HER-SYS data included positive results from COVID-19 tests and were used as reference standards. Claims-based algorithms #1 and #2 were U07.1, B34.2, with and without suspicious diagnoses, respectively. Claims-based algorithms #3 and #4 were U07.1 with and without suspicious diagnoses, respectively. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each algorithm. The study cohort included 165 038 residents, including 13 402 residents were the reference standard. For the entire period, the sensitivity, specificity, PPV, and NPV were 55.7% (95% confidence interval: 54.8%-56.5%), 65.4% (65.2%-65.6%), 11.5% (11.3%-11.8%), and 98.9% (98.8%-99.0%) for Algorithm #1, and 67.0% (66.2%-67.8%), 88.1% (87.9%-88.3%), 31.6% (31.1%-32.2%), and 97.8% (97.7%-97.8%) for Algorithm #2, and 52.9% (52.0%-53.7%), 67.1% (66.9%-67.3%), 11.5% (11.2%-11.8%), and 98.3% (98.3%-98.4%) for Algorithm #3, 62.6% (61.8%-63.4%), 88.5% (88.3%-88.7%), 30.9% (30.3%-31.4%), and 97.3% (97.2%-97.4%) for Algorithm #4, respectively. Our study showed that the validity of claims-based algorithms consisting of COVID-19-related ICD-10 codes to identify patients with first-onset COVID-19 is limited.

Real-world perioperative treatment patterns and burden of recurrent disease in patients with high-risk endometrial cancer: a SEER-Medicare study

Objectives To elucidate unmet needs in high-risk endometrial cancer (EC), this study described perioperative treatment patterns in Medicare beneficiaries with high-risk EC and quantified the impact of disease recurrence on clinical and economic outcomes among patients receiving adjuvant therapy. Methods Patients aged ≥66 years with high-risk EC (stage I/II EC of non-endometrioid histology or stage III/IVA EC of any histology) receiving hysterectomy with bilateral salpingo-oophorectomy from SEER-Medicare data (2007–2019) were identified; perioperative treatment patterns were described. Post-operative treatment patterns were described among patients receiving adjuvant therapy; overall survival (OS), all-cause and EC-related healthcare resource utilization and costs were evaluated from recurrence date (using a claims-based algorithm developed with clinical input) for recurrent patients and from a frequency-matched date for non-recurrent patients. Results Of 2,279 patients receiving EC surgery, 3.1% received neoadjuvant therapy and 55.3% received adjuvant therapy. Among 1,199 patients receiving adjuvant therapy, systemic adjuvant therapy with radiation (38.9%) was most common. Median OS was 1.4 years among 378 (31.5%) recurrent patients identified over a median follow-up of 3.7 years. Recurrent patients had significantly higher per-patient-year rates of all-cause outpatient visits (37.7 vs. 22.6), EC-related outpatient visits (14.5 vs. 3.0), and all-cause hospitalizations (1.3 vs. 0.4) than non-recurrent patients, and an excess of $84,562 and $62,128 in all-cause and EC-related annual costs, predominantly driven by hospitalizations. Conclusions Our findings highlight the considerable clinical and economic burden experienced by patients with high-risk EC experiencing recurrence and emphasize the unmet need for novel therapies in early settings to mitigate this burden.

Development and Validation of Claims-based Algorithms for Identifying Hospitalized Patients With COVID-19 and Their Severity in 2020 and 2021.

This study aimed to develop and validate claims-based algorithms for identifying hospitalized patients with coronavirus disease 2019 (COVID-19) and disease severity. We used claims data including all patients at the National Center for Global and Medicine Hospital between January 1, 2020, and December 31, 2021. The claims-based algorithms for three statuses with COVID-19 (hospitalizations, moderate or higher status, and severe status) were developed using diagnosis codes (International Classification of Diseases, 10th revision code: U07.1, B34.2) and relevant medical procedure code. True cases were determined using the COVID-19 inpatient registry and electronic health records. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each algorithm at 6-month intervals. Of the 75,711 total patients, the number of true cases was 1,192 for hospitalizations, 622 for moderate or higher status, and 55 for severe status. The diagnosis code-only algorithm for hospitalization had sensitivities 90.4% to 94.9% and PPVs 9.3% to 19.4%. Among the algorithms consisting of both diagnosis codes and procedure codes, high sensitivity and PPV were observed during the following periods: 93.9% and 97.1% for hospitalization (January-June 2021), 90.4% and 87.5% for moderate or higher status (July-December 2021), and 92.3% and 85.7% for severe status (July-December 2020), respectively. Almost all algorithms had specificities and NPVs of approximately 99%. The diagnosis code-only algorithm for COVID-19 hospitalization showed low validity throughout the study period. The algorithms for hospitalizations, moderate or higher status, and severe status with COVID-19, consisting of both diagnosis codes and procedure codes, showed high validity in some periods.

The economic burden of recurrence in elderly patients with completely resected, stage IIB/IIC or III melanoma: an analysis of the Surveillance, Epidemiology, and End Results-Medicare linked database

Aims To compare healthcare resource utilization (HRU) and costs between patients with or without melanoma recurrence and between patients with distant or locoregional melanoma recurrence. Methods Patients aged ≥65 years with completely resected, stage IIB/IIC or III melanoma were identified from Surveillance, Epidemiology, and End Results-Medicare data and stratified based on whether they experienced a recurrence, and whether it was distant or locoregional (separately for each stage). The index date was the date of recurrence (recurrence group) or a randomly assigned date (non-recurrence group). Patients in the recurrence and non-recurrence groups were propensity score-matched 1:1 based on patient characteristics; HRU and healthcare costs were compared between the 2 groups and between patients with distant or locoregional recurrence during the ≤24 months following index. Results After matching, 507 pairs of patients with recurrent or non-recurrent stage IIB/IIC melanoma (236 patients with distant recurrence, 271 with locoregional) and 141 pairs of patients with recurrent or non-recurrent stage III melanoma (50 patients with distant recurrence, 91 with locoregional) were included. During the first year following recurrence, unadjusted HRU was generally higher in patients with versus without recurrence and patients with distant versus locoregional recurrence among both stage IIB/IIC and III cohorts. Patients who experienced recurrence incurred $6,474 (stage IIB/IIC) or $6,112 (stage III) per patient per month (PPPM) more in unadjusted, all-cause, total healthcare costs than patients without recurrence (both p < 0.001). Patients with distant recurrence incurred $7,292 (stage IIB/IIC) or $5,436 (stage III) PPPM more in unadjusted, all-cause, total healthcare costs than patients with locoregional recurrence (both p < 0.05). Limitations Melanoma recurrence was identified using a claims-based algorithm. Conclusions Economic burden is higher in patients with versus without melanoma recurrence and patients with distant versus locoregional recurrence. There is a high unmet need for adjuvant therapies that may help to prevent or delay recurrence.

Financial burden within first year of diagnosis for privately-insured patients with early-onset breast or colorectal cancer.

9 Background: Rising incidence of early-onset (E-O) cancers has been observed globally. E-O breast cancer (BC) are patients (pts) diagnosed ages 18 to 44 and E-O colorectal cancer (CRC) are aged 18 to 49. Because many E-O cancers are among pts who do not meet screening eligibility criteria by age, they are often diagnosed at advanced stage, which is also more expensive to treat. Younger pts may also have lower incomes and fewer assets, increasing risk of financial burden for pts with E-O cancers. We estimated the financial burden for pts and private payors for E-O BC and CRC, compared to non-elderly pts with BC and CRC diagnosed between the age of 45 to 64 and 50 to 64 (older-onset), respectively. Methods: We applied a claims-based algorithm to identify pts with incident BC and CRC diagnosed between 2014 and 2021 from MarketScan Commercial Claims and Encounters Data and required 12-month continuous enrollment before and after the index diagnosis for completeness of claims. Financial burden was quantified as total and out-of-pocket (OOP) payments, calculated as the sum of deductibles, copayments, and coinsurance. We used generalized linear models and quantile regressions to compare differences in mean and median costs, respectively, between pts with E-O vs. older onset cancer, adjusting for comorbidity, region, insurance plan type, sex (for CRC) and year. Results: Study cohort included 58,419 pts with BC and 11,251 pts with CRC. The proportion of E-O increased from 15% in 2014 to 17% in 2022 (Ptrend&lt;0.01) in BC and from 23% to 30% (Ptrend&lt;0.01) in CRC. Higher percentage of E-O pts were enrolled in high-deductible plans (BC: 25% v22%; CRC: 24% vs. 21%) and received systemic therapy (BC: 61% v 49%; CRC: 57% vs. 43%). Higher percent of surgery (93% v 89%) was observed in E-O BC and radiation (21% v 16%) in E-O CRC. Pts with E-O BC and CRC had significantly higher adjusted mean and median total and OOP payments (Table). Conclusions: E-O BC and CRC result in substantial costs to payors and moderately higher OOP payments. OOP maximums in private insurance may offer some financial protection for pts. E-O Older onset Difference 95% CIBC: BC: Mean total pay $182,770 $128,640 $54,130 $50445, $57817 Median total pay 143,351 93,317 50,034 47691, 52377 Mean OOP 6,014 5,063 951 817, 1085 Median OOP 5,142 4,258 884 791, 978 CRC: Mean total pay $153,446 $123,449 $29,998 $24244, $35752 Median total pay 119,747 89,340 30,406 25216, 35597 Mean OOP 5,429 4,791 638 428, 848 Median OOP 4,452 4,015 437 266, 607

Validation of an algorithm to prioritize patients for comprehensive medication management in primary care settings.

Comprehensive medication management (CMM) programs optimize the effectiveness and safety of patients' medication regimens, but CMM may be underutilized. Whether healthcare claims data can identify patients appropriate for CMM is not well-studied. Determine the face validity of a claims-based algorithm to prioritize patients who likely need CMM. We used claims data to construct patient-level markers of "regimen complexity" and "high-risk for adverse effects," which were combined to define four categories of claims-based CMM-need (very likely, likely, unlikely, very unlikely) among 180 patient records. Three clinicians independently reviewed each record to assess CMM need. We assessed concordance between the claims-based and clinician-review CMM need by calculating percent agreement as well as kappa statistic. Most records identified as 'very likely' (90%) by claims-based markers were identified by clinician-reviewers as needing CMM. Few records within the 'very unlikely' group (5%) were identified by clinician-reviewers as needing CMM. Interrater agreement between CMM-based algorithm and clinician review was moderate in strength (kappa = 0.6, p < 0.001). Claims-based pharmacy measures may offer a valid approach to prioritize patients into CMM-need groups. Further testing of this algorithm is needed prior to implementation in clinic settings.

Validating claims-based algorithms for a systemic lupus erythematosus diagnosis in Medicare data for informed use of the Lupus Index: a tool for geospatial research

ObjectiveThis study aimed to validate claims-based algorithms for identifying SLE and lupus nephritis (LN) in Medicare data, enhancing the use of the Lupus Index for geospatial research on SLE prevalence...

Association between mammographic screening history and stage at diagnosis among women with screen-detected breast cancer.

145 Background: Although screening mammography is an effective means of detecting breast cancer at earlier stages, some women are still diagnosed with later stage disease despite screening mammography. We hypothesized that screening history prior to diagnosis may impact stage at diagnosis and breast cancer mortality among women with screen-detected breast cancer. Methods: Retrospective cohort study of women aged 70+ in the SEER-Medicare database, diagnosed with ER+/HER2- breast cancer from 2010-2017, and enrolled in fee-for service Medicare Parts A and B from 5-years prior to diagnosis through 1 year after. We used validated claims-based algorithms to identify cancers that were detected by screening mammogram. We quantified 5-year screening history by number of prior screens. Stage at diagnosis was dichotomized into very early (T1N0) vs. later stage (T2+ or N+) disease. Multivariable logistic regression was used to assess the association between screening history and stage at diagnosis and Cox proportional hazards used to examine breast cancer mortality. Models were adjusted for demographic and clinical covariates associated with later stage disease in bivariate analyses. Results: We included 13,029 women with screen-detected breast cancer; 3,813 (29.27%) women had later stage disease at diagnosis, and 10,095 (77.48%) had at least one prior negative screening mammogram within the past five years. During this time, 2,359 (18.11%) women had only one prior mammogram, 3,132 (24.03%) had two, and 4,604 (35.34%) had three or four. Having at least one prior mammogram was associated with 56% lower odds of being diagnosed with later stage disease (OR: 0.44, 95% CI=0.40-0.48). Sociodemographic factors associated with increased odds of later stage disease included living in an area with greater poverty (20-100% vs. 0-5% of residents, OR: 1.19, 1.00-1.40), non-Hispanic Black race (OR: 1.18, 1.00-1.38), and Medicare/Medicaid dual eligibility (OR: 1.20, 1.06-1.36). Factors associated with later stage disease included living in an area with a more people with a high-school education (30-40% vs. 0-30% of residents, OR: 0.87, 0.77-0.97). At total of 696 deaths from breast cancer were observed, including 273 deaths among 2,935 women with no prior screening (9.3%) and 423 deaths among 10,097 women who had undergone screening within the past five years (4.2%, P &lt; 0.001). Receiving at least one negative screening prior to diagnosis was associated with decreased breast cancer-specific Mortality in adjusted analysis (HR: 0.48, 0.40-0.57). Conclusions: The use of prior routine screening mammography was associated with lower risk of being diagnosed with later stage disease and dying from breast cancer in elderly women with screen-detected breast cancer. Future research is warranted regarding the impact of routine vs. sporadic screening on potential morbidity and mortality in this age group.

Development and validation of claims-based algorithms for estimating gestational age of spontaneous abortion and termination.

To study the risk of spontaneous abortion (SAB) or termination using healthcare utilization databases, algorithms to estimate the gestational age (GA) are needed. Using Medicaid data, we developed a hierarchical algorithm to classify pregnancy outcomes. We identified the subset of potential SAB and termination cases, and abstracted the GA from linked electronic medical records (gold standard). We developed three approaches: (1) assign median GA for SAB and termination cases in the US; (2) draw a random GA from the population distributions; (3) estimate GA based on regression models. Algorithm performance was assessed based on the proportion of pregnancies with estimated GA within 1-4 weeks of the gold standard, the mean squared error (MSE) and the R-squared. Approach 1 and Approach 3 had similar performance, though approach 3 using random forest models with variables selected via the Boruta algorithm had better MSE and R-squared. For SAB, 58.0% of pregnancies were correctly classified within 2 weeks of the gold standard (MSE: 8.7, R-squared: 0.09). For termination, the proportions were 66.3% (MSE: 11.7; R-squared: 0.35). SABs and terminations can be studied in healthcare utilization data with careful implementation of validated algorithms though higher level of GA misclassification is expected compared to live births.

Development and Validation of an Algorithm to Predict Stillbirth Gestational Age in Medicaid Billing Records

Abstract Introduction With Medicaid covering half of US pregnancies, Medicaid Analytic eXtract (MAX) provides a valuable data source to enrich understanding about stillbirth etiologies. Objective We developed and validated a claims-based algorithm to predict GA at stillbirth. Method We linked the stillbirths identified in MAX 1999-2013 to Florida Fetal Death Records (FDRs) to obtain clinical estimates of GA (N=825). We tested several algorithms including using a fixed median GA, median GA at the time of specific prenatal screening tests, and expanded versions considering additional predictors of stillbirth within including linear regression and random forest models. We estimated the proportion of pregnancies with differences of ± 1, 2, 3 and 4 weeks between the predicted and FDR GA and the model mean square error (MSE). We validated the selected algorithms in two external samples. Results The best performing algorithm was a random forest model (MSE of 12.67 weeks2) with 84% of GAs within ± 4 weeks. Assigning a fixed GA of 28 weeks resulted in an MSE of 60.21 weeks2 and proportions of GA within ± 4 weeks of 32%. We observed consistent results in the external samples. Discussion Our prediction algorithm for stillbirths can facilitate pregnancy research in the Medicaid population.

Development and Optimization of a Bladder Cancer Algorithm Using SEER-Medicare Claims Data.

Categorizing patients with cancer by their disease stage can be an important tool when conducting administrative claims-based studies. As claims databases frequently do not capture this information, algorithms are increasingly used to define disease stage. To our knowledge, to date, no study has used an algorithm to categorize patients with bladder cancer (BC) by disease stage (non-muscle-invasive BC [NMIBC], muscle-invasive BC [MIBC], or locally advanced/metastatic urothelial carcinoma [la/mUC]) in a US-based health care claims database. A claims-based algorithm was developed to categorize patients by disease stage on the basis of the administrative claims portion of the SEER-Medicare linked data. The algorithm was validated against a reference SEER registry, and the algorithm's parameters were iteratively modified to improve its performance. Patients were included if they had an initial diagnosis of BC between January 2016 and December 2017 recorded in SEER registry data. Medicare claims data were available for these patients until December 31, 2019. The algorithm was evaluated by assessing percentage agreement, Cohen's kappa (κ), specificity, positive predictive value (PPV), and negative predictive value (NPV) against the SEER categorization. A total of 15,484 patients with SEER-confirmed BC were included: 10,991 (71.0%) with NMIBC, 3,645 (23.5%) with MIBC, and 848 (5.5%) with la/mUC. After multiple rounds of algorithm optimization, the final algorithm had an agreement of 82.5% with SEER, with a κ of 0.58, a PPV of 87.0% for NMIBC, and 76.8% for MIBC and a high NPV for la/mUC of 98.0%. This claims-based algorithm could be a useful approach for researchers conducting claims-based studies categorizing patients with BC at diagnosis.

Identifying Dementia Severity Among People Living With Dementia Using Administrative Claims Data

ObjectivesThere is currently no reliable tool for classifying dementia severity level based on administrative claims data. We aimed to develop a claims-based model to identify patients with severe dementia among a cohort of patients with dementia. DesignRetrospective cohort study. Setting and ParticipantsWe identified people living with dementia (PLWD) in US Medicare claims data linked with the Minimum Data Set (MDS) and Outcome and Assessment Information Set (OASIS). MethodsSevere dementia was defined based on cognitive and functional status data available in the MDS and OASIS. The dataset was randomly divided into training (70%) and validation (30%) sets, and a logistic regression model was developed to predict severe dementia using baseline (assessed in the prior year) features selected by generalized linear mixed models (GLMMs) with least absolute shrinkage and selection operator (LASSO) regression. We assessed model performance by area under the receiver operating characteristic curve (AUROC), area under precision-recall curve (AUPRC), and precision and recall at various cutoff points, including Youden Index. We compared the model performance with and without using Synthetic Minority Oversampling Technique (SMOTE) to reduce the imbalance of the dataset. ResultsOur study cohort included 254,410 PLWD with 17,907 (7.0%) classified as having severe dementia. The AUROC of our primary model, without SMOTE, was 0.81 in the training and 0.80 in the validation set. In the validation set at the optimized Youden Index, the model had a sensitivity of 0.77 and specificity of 0.70. Using a SMOTE-balanced validation set, the model had an AUROC of 0.83, AUPRC of 0.80, sensitivity of 0.79, specificity of 0.74, positive predictive value of 0.75, and negative predictive value of 0.78 when at the optimized Youden Index. Conclusions and ImplicationsOur claims-based algorithm to identify patients living with severe dementia can be useful for claims-based pharmacoepidemiologic and health services research.

Comparison of Medicare Claims-based Clostridioides difficile infection classification to chart review using a linked cohort

Background: Medicare claims are frequently used to study Clostridioides difficile infection (CDI) epidemiology. Categorizing CDI based on location of onset and potential exposure is critical in understanding transmission patterns and prevention strategies. While claims data are well-suited for identifying prior healthcare utilization exposures, they lack specimen collection and diagnosis dates to assign likely location of onset. Algorithms to classify CDI onset and healthcare association using claims data have been published, but the degree of misclassification is unknown. Methods: We linked patients with laboratory-confirmed CDI reported to four Emerging Infections Program (EIP) sites from 2016-2020 to Medicare beneficiaries using residence, birth date, sex, and hospitalization and/or healthcare exposure dates. Uniquely linked patients with fee-for-service Medicare A/B coverage and complete EIP case report forms were included. Patients with a claims CDI diagnosis code within ±28 days of a positive CDI test reported to EIP were categorized as hospital-onset (HO), long-term care facility onset (LTCFO), or community-onset (CO, either healthcare facility-associated [COHCFA] or community-associated [CA]) using a previously published algorithm based on claim type, ICD-10-CM code position, and duration of hospitalization (if applicable). EIP classifies CDI into these categories using positive specimen collection date and other information from chart review (e.g. admit/discharge dates). We assessed concordance of EIP and claims case classifications using Cohen’s kappa. Results: Of 10,002 eligible EIP-identified CDI cases, 7,064 were linked to a unique beneficiary; 3,451 met Medicare A/B fee-for-service coverage inclusion criteria. Of these, 650 (19%) did not have a claims diagnosis code ±28 days of the EIP specimen collection date (Table); 48% (313/650) of those without a claims diagnosis code were categorized by EIP as CA CDI. Among those with a CDI diagnosis code, concurrence of claims-based and EIP CDI classification was 68% (κ=0.56). Concurrence was highest for HO and lowest for COHCFA CDI. A substantial number of EIP-classified CO CDIs (30%, Figure) were misclassified as HO using the claims-based algorithm; half of these had a primary ICD-10 diagnosis code of sepsis (226/454; 50%). Conclusions: Evidence of CDI in claims data was found for 81% of EIP-reported CDI cases. Medicare classification algorithms concurred with the EIP classification in 68% of cases. Discordance was most common for community-onset CDI patients, many of whom were hospitalized with a primary diagnosis of sepsis. Misclassification of CO-CDI as HO may bias findings of claims-based CDI studies.

Identifying chronic kidney disease stage 3 with excess disease burden.

Chronic kidney disease (CKD) is a widely prevalent disease with heterogeneous disease progression. Prior study findings suggest that early referral to nephrologists can improve health outcomes for patients with CKD. Current practice guidelines recommend nephrology referral when patients are diagnosed with CKD stage 4. We tested whether a subset of patients with CKD stage 3 and common medical comorbidities demonstrates disease progression, cost, and utilization patterns that would merit earlier referral. Retrospective study of Medicare fee-for-service beneficiaries with CKD stages 3 through 5 and end-stage kidney disease. We identified 7 comorbidities with high prevalence in patients with progressive CKD and segmented beneficiaries with CKD stage 3 based on the presence of these comorbidities. Outcomes including costs, utilization, and disease progression were then compared across beneficiaries with different stages of CKD. We identified that beneficiaries with CKD stage 3 and at least 1 of the selected comorbidities (CKD stage 3-plus) represented 35.4% of all beneficiaries with CKD stage 3. The CKD stage 3-plus cohort had cost and utilization patterns that were more similar to beneficiaries with CKD stages 4 and 5 than to beneficiaries with CKD stage 3 without the selected comorbidities. Our findings demonstrate the use of a claims-based algorithm to identify patients with CKD stage 3 who have high costs and are at risk of disease progression, highlighting a potential subset of patients who might benefit from earlier nephrology intervention.