Gaia Pocobelli

Health insurance claims-based analyses inform many large epidemiologic studies but may have unmeasured confounding. Electronic health records (EHRs) have more detailed health information, but data may be missing on some individuals. Informed by recent work comparing approaches to address missing data, we selected and applied generalized raking (GR) and multiple imputation (MI) to illustrate how to efficiently combine data from these 2 sources. We considered a previously conducted claims-based study comparing 90-day arterial thromboembolism (ATE) risk among patients hospitalized with COVID-19 versus influenza, for which body mass index (BMI), a potential confounder, was unavailable. Using linked claims-EHR data, we adjusted for the same covariates as the original study and used GR and MI to additionally control for EHR-ascertained BMI, available for a subset. We included 912 hospitalized Kaiser Permanente Washington patients, 449 with COVID-19 and 463 with influenza (31.0% and 38.5% had EHR-ascertained BMI in the prior 90 days, respectively). Adjusted hazard ratios of ATE were similar with and without adjustment for BMI, in GR and MI analyses. In the setting of a claims-EHR–based study with missing data on a potential confounder, we describe our selection of the missing data approach, and its implementation, to demonstrate a broadly applicable process.

Purpose: The U.S. Food and Drug Administration (FDA) defined disease severity criteria to assist clinical development of medical products for management of COVID‐19. These definitions were translated to code‐based algorithms for use in real‐world data. We validated the algorithms' performance in ambulatory settings at three regional integrated healthcare delivery systems contributing data to FDA's Sentinel System. Methods: We identified cohorts of individuals ≥ 18 years that met the algorithms' criteria for mild, moderate, and severe COVID‐19 at incident COVID‐19 diagnosis or positive SARS‐CoV‐2 test, and separately, at incident COVID‐19 treatment, from January 2022 through April 2023. We validated the algorithms via chart review of a random sample to calculate positive predictive values (PPVs) and 95% CIs. Results: The algorithms identified 33 071 patients at COVID‐19 diagnosis or positive test; 26 985 mild (49 chart reviewed), 5180 moderate (55 reviewed), and 906 severe (56 reviewed). A total of 4512 patients were identified at COVID‐19 treatment; 3474 mild (56 reviewed), 848 moderate (60 reviewed), and 190 severe (46 reviewed). The PPVs (1) at COVID‐19 diagnosis or positive test were: mild 57% (95% CI: 43%–71%), moderate 58% (95% CI: 45%–71%), and severe 54% (95% CI: 41%–67%), and (2) at COVID‐19 treatment: mild 57% (95% CI: 44%–70%), moderate 70% (95% CI: 58%–82%) and severe 72% (95% CI: 59%–85%). Conclusion: The algorithms had low‐to‐moderate performance in classifying COVID‐19 severity in ambulatory settings, depending on assessment at diagnosis or treatment. Researchers should consider the performance of the algorithm when using real‐world data to assess COVID‐19 severity.

Cancer screening guidelines specify ages at which routine screening should be discontinued and, except for cervical cancer screening, do not require that specific screening history criteria be met for discontinuation. We estimated the prevalence of being up-to-date with average-risk screening guidelines for colorectal cancer (CRC), cervical cancer, and lung cancer as of the recommended ages for discontinuation of routine screening. We conducted a descriptive study among several US health-care systems during 2010-2019. Up-to-date screening prevalence, based on US Preventive Services Task Force guidelines, was ascertained before the 76th, 66th, and 81st birthdays among persons eligible for CRC screening (n = 316 756 persons), cervical cancer screening (n = 20 282 persons), and lung cancer screening (n = 1151 persons). Up-to-date screening prevalence was 84.4% for CRC, 58.9% for cervical cancer, and 6.3% for lung cancer screening. Up-to-date screening prevalence at the ages recommended for discontinuing routine CRC, cervical cancer, and lung cancer screening varied appreciably and was particularly low for lung cancer screening.

Not-yet-known not-yet-known not-yet-known unknown Purpose: The U.S. Food and Drug Administration (FDA) defined disease severity criteria to assist clinical development of medical products for management of COVID-19. These definitions were translated to code-based algorithms for use in real-world data. We validated the algorithms’ performance in ambulatory settings at three regional integrated healthcare delivery systems contributing data to FDA’s Sentinel System. Methods: We identified cohorts of individuals ≥18 years that met the algorithms’ criteria for mild, moderate, and severe COVID-19 at incident COVID-19 diagnosis or positive SARS-CoV-2 test, and separately, at incident COVID-19 treatment, from January 2022 through April 2023. We validated the algorithms via chart review of a random sample to calculate positive predictive values (PPVs) and 95% CIs. Results: The algorithms identified 33,071 patients at COVID-19 diagnosis or positive test; 26,985 mild (49 chart reviewed), 5,180 moderate (55 reviewed), and 906 severe (56 reviewed). A total of 4,512 patients were identified at COVID-19 treatment; 3,474 mild (56 reviewed), 848 moderate (60 reviewed), and 190 severe (46 reviewed). The PPVs 1) at COVID-19 diagnosis or positive test were: mild 57% (95% CI: 43–71%), moderate 58% (95% CI: 45–71%), and severe 54% (95% CI: 41–67%) and 2) at COVID-19 treatment: mild 57% (95% CI: 44–70%), moderate 70% (95% CI: 58%-82%) and severe 72% (95% CI: 59%-85%). Conclusion: The algorithms had low-to-moderate performance in classifying COVID-19 severity in ambulatory settings, depending on assessment at diagnosis or treatment. Researchers should consider performance of the algorithm when using real-world data to assess COVID-19 severity.

Purpose The positive predictive value (PPV) of cervical cancer screening was projected to decrease in the era of human papillomavirus (HPV) vaccination. Methods We conducted a retrospective cohort study at three U.S. healthcare systems during 2010–2018. Females aged 21–38 years with an abnormal cervical cancer screening test result for which colposcopy was guideline-recommended were included. We estimated age-specific PPVs of cervical intraepithelial neoplasia grade 2 or more severe diagnosis (≥ CIN 2) in HPV-vaccinated and unvaccinated females. Results The age-specific PPV point estimates were lower in vaccinated versus unvaccinated females in each age group (21–24; 25–29; and 30–38 years), but the difference was statistically significant only among 25–29-year-olds (PPV = 16.4% [95% CI (confidence interval), 14.6–18.4%] and PPV = 19.8% [95% CI 18.5–21.1%], respectively). Among vaccinated 25–29-year-olds, the PPV was lower in those who received their first dose at ≤ 20 versus > 20 years of age (PPV = 12.1% [95% CI 9.5–15.2%] and PPV = 18.8% [95% CI 16.4–21.4%], respectively). Among all age groups combined, the PPV was lower in vaccinated versus unvaccinated females. Conclusions Our findings suggest that among females with a cervical test result recommended for colposcopy, the PPV for ≥ CIN2 was lower in those vaccinated versus unvaccinated, and was relatively lower in those vaccinated at younger ages. Future studies will have greater ability to estimate the impact of vaccination at the recommended age (9–12 years) on the PPV as more recent birth cohorts, who had greater vaccine uptake, age-in to screening eligibility.

We evaluated neural tube defect (NTD) risk associated with prescription opioid analgesic use during early pregnancy. We conducted a cohort study of liveborn singletons during 2001-2014 among 9 U.S. health plans. Risks of medical chart-confirmed primary NTD (anencephaly and select spina bifida) and any NTD (primary and other NTD) were compared among singletons with maternal new use of opioids during 18-56 days after last menstrual period to those with no maternal use during this period or pre-pregnancy. In a sensitivity analysis, singletons exposed during the second or third trimesters only (“negative control exposure”) were compared to those never exposed. The main analysis adjusted odds ratio (OR) associated with exposure was 3.0 (95% CI, 0.7-12.7) for primary NTD and 2.3 (95% CI, 0.8-6.7) for any NTD. The sensitivity analysis ORs were similar 2.4 and 1.6, respectively. Our results do not support or refute the hypothesis that prescription opioid analgesic exposure in early pregnancy increases risk of NTD. Although ORs were increased in the main analysis, the sensitivity analysis suggested the presence of residual confounding. Future studies that are sufficiently powered and adjust for confounders beyond those in the present study are warranted but likely challenging.

Guidelines now recommend initiating colorectal cancer (CRC) screening at age 45 years rather than 50 years, but little is known about screening completion and yield among people aged 45 to 49 years. To evaluate fecal immunochemical test (FIT) completion and yield in patients aged 45 to 49 versus 50 years. Retrospective cohort study. Kaiser Permanente Northern California, Washington, and Colorado. Those distributed a FIT kit during January to September 2022. FIT completion within 3 months, FIT positivity, receipt of colonoscopy within 3 months after a positive FIT result, and colonoscopy yield. A total of 267 732 FIT kits were distributed: 213 928 (79.9%) to patients aged 45 to 49 years, and 53 804 (20.1%) to those aged 50 years. Overall, FIT completion was slightly higher in patients aged 45 to 49 years (38.9% vs. 37.5%; adjusted risk ratio [aRR], 1.05 [95% CI, 1.04 to 1.06]), although at Colorado, those aged 45 to 49 years were substantially less likely to complete a FIT (30.7% vs. 40.2%; aRR, 0.77 [CI, 0.73 to 0.80]). Overall, FIT positivity was lower in patients aged 45 to 49 years (3.6% vs. 4.0%; aRR, 0.91 [CI, 0.84 to 0.98]), and receipt of colonoscopy after a positive FIT result was similar between groups (64.9% vs. 67.4%; aRR, 1.00 [CI, 0.94 to 1.05]). Adenoma detection was lower in the younger group (58.8% vs. 67.7%; aRR, 0.88 [CI, 0.83 to 0.95]). Yields were similar for adenoma with advanced histology (13.2% vs. 15.9%; aRR, 0.86 [CI, 0.69 to 1.07]), polyp with high-grade dysplasia (3.4% vs. 5.1%; aRR, 0.68 [CI, 0.44 to 1.04]), sessile serrated lesion (10.3% vs. 11.7%; aRR, 0.92 [CI, 0.71 to 1.21]), and CRC (2.8% vs. 2.7%; aRR, 1.10 [CI, 0.62 to 1.96]). The small number of neoplasia events contributed to wide CIs. Similar FIT completion and yield rates in people aged 45 to 50 years support initiation of CRC screening at age 45 years. Kaiser Permanente Sidney R. Garfield Memorial Fund.

Purpose Females living with human immunodeficiency virus (FLWHIV) are at increased risk of cervical cancer and U.S. guidelines, first published in 2009 and updated since then, recommend more frequent screening in this population. We examined screening rates among FLWHIV in the U.S. during 2010–2019. Methods This cohort study included 18–89-year-old FLWHIV during 2010–2019 at three U.S. healthcare settings. Sociodemographics, comorbidities, and cervical cancer screening tests were ascertained from administrative and clinical databases. We reported cervical cancer screening rates overall and by modality. Generalized estimating equations with Poisson distribution were used to estimate screening rate ratios (SRRs) and 95% confidence intervals (CIs) for the associations between screening rates and calendar year, age, race and ethnicity, and comorbidity. Results Among 3,556 FLWHIV, a total of 7,704 cervical cancer screening tests were received over 18,605 person-years during 2010–2019 (screening rate = 41.4 per 100 person-years). Relatively lower screening rates were associated with later calendar years (SRR = 0.71 [95% CI 0.68–0.75] for 2017–2019 versus 2010–2013), older age (SRR = 0.82 [95% CI 0.74–0.89] for 50–65-year-olds versus 18–29-year-olds), non-Hispanic white race versus non-Hispanic Black race (SRR = 0.89 [95% CI 0.81–0.98]) and greater comorbidity burden (SRR = 0.89 [95% CI 0.82–0.98] for ≥ 9 versus 0–6 comorbidity score). Conclusion The decrease in cervical cancer screening rates during 2010–2019 in this large cohort of FLWHIV may be explained at least partly by guideline changes during the study period recommending longer screening intervals. Our findings of relatively lower screening rates in FLWHIV who were non-Hispanic white, older, and with greater comorbidity burden should be confirmed in other U.S. settings.

Background: Efficiently identifying patients with human immunodeficiency virus (HIV) using administrative health care data (e.g., claims) can facilitate research on their quality of care and health outcomes. No prior study has validated the use of only ICD-10-CM HIV diagnosis codes to identify patients with HIV. Methods: We validated HIV diagnosis codes among women enrolled in a large U.S. integrated health care system during 2010–2020. We examined HIV diagnosis code-based algorithms that varied by type, frequency, and timing of the codes in patients’ claims data. We calculated the positive predictive values (PPVs) and 95% confidence intervals (CIs) of the algorithms using a medical record-confirmed diagnosis of HIV as the gold standard. Results: A total of 272 women with ≥ 1 HIV diagnosis code in the administrative claims data were identified and medical records were reviewed for all 272 women. The PPV of an algorithm classifying women as having HIV as of the first HIV diagnosis code during the observation period was 80.5% (95% CI: 75.4–84.8%), and it was 93.9% (95% CI: 90.0-96.3%) as of the second. Little additional increase in PPV was observed when a third code was required. The PPV of an algorithm based on ICD-10-CM-era codes was similar to one based on ICD-9-CM-era codes. Conclusion: If the accuracy measure of greatest interest is PPV, our findings suggest that use of ≥ 2 HIV diagnosis codes to identify patients with HIV may perform well. However, health care coding practices may vary across settings, which may impact generalizability of our results.