Impact of the COVID-19 Pandemic on Health Care Utilization in the Vaccine Safety Datalink: Retrospective Cohort Study.

BACKGROUND: Understanding the long-term impact of the COVID-19 pandemic on health care utilization is important to health care organizations and policy makers for strategic planning, as well as to researchers when designing studies that use observational electronic health record data during the pandemic period. OBJECTIVE: This study aimed to evaluate the changes in health care utilization across all care settings among a large, diverse, and insured population in the United States during the COVID-19 pandemic. METHODS: We conducted a retrospective cohort study within 8 health care organizations participating in the Vaccine Safety Datalink Project using electronic health record data from members of all ages from January 1, 2017, to December 31, 2021. The visit rates per person-year were calculated monthly during the study period for 4 health care settings combined as well as by inpatient, emergency department (ED), outpatient, and telehealth settings, both among all members and members without COVID-19. Difference-in-difference analysis and interrupted time series analysis were performed to assess the changes in visit rates from the prepandemic period (January 2017 to February 2020) to the early pandemic period (April-December 2020) and the later pandemic period (July-December 2021), respectively. An exploratory analysis was also conducted to assess trends through June 2023 at one of the largest sites, Kaiser Permanente Southern California. RESULTS: The study included more than 11 million members from 2017 to 2021. Compared with the prepandemic period, we found reductions in visit rates during the early pandemic period for all in-person care settings. During the later pandemic period, overall use reached 8.36 visits per person-year, exceeding the prepandemic level of 7.49 visits per person-year in 2019 (adjusted percent change 5.1%, 95% CI 0.6%-9.9%); inpatient and ED visits returned to prepandemic levels among all members, although they remained low at 0.095 and 0.241 visits per person-year, indicating a 7.5% and 8% decrease compared to pre-pandemic levels among members without COVID-19, respectively. Telehealth visits, which were approximately 42% of the volume of outpatient visits during the later pandemic period, were increased by 97.5% (95% CI 86.0%-109.7%) from 0.865 visits per person-year in 2019 to 2.35 visits per person-year in the later pandemic period. The trends in Kaiser Permanente Southern California were similar to those of the entire study population. Visit rates from January 2022 to June 2023 were stable and appeared to be a continuation of the use levels observed at the end of 2021. CONCLUSIONS: Telehealth services became a mainstay of the health care system during the late COVID-19 pandemic period. Inpatient and ED visits returned to prepandemic levels, although they remained low among members without evidence of COVID-19. Our findings provide valuable information for strategic resource allocation for postpandemic patient care and for designing observational studies involving the pandemic period.

Longitudinal adherence to breast cancer surveillance following cancer genetic testing in an integrated health care system.

PURPOSE: Screening with mammography and breast magnetic resonance imaging (MRI) is an important risk management strategy for individuals with inherited pathogenic variants (PVs) in genes associated with increased breast cancer risk. We describe longitudinal screening adherence in individuals who underwent cancer genetic testing as part of usual care in a vertically integrated health system. METHODS: We determined the proportion time covered (PTC) by annual mammography and breast MRI for individuals with PVs in TP53, BRCA1, BRCA2, PALB2, NF1, CHEK2, and ATM. We determined time covered by biennial mammography beginning at age 50 years for individuals who received negative results, uncertain results, or with PVs in genes without specific breast cancer screening recommendations. RESULTS: One hundred and forty individuals had PVs in TP53, BRCA1, BRCA2, PALB2, NF1, CHEK2, or ATM. Among these individuals, average PTC was 48% (range 0-99%) for annual screening mammography and 34% (range 0-100%) for annual breast MRI. Average PTC was highest for individuals with PVs in CHEK2 (N = 14) and lowest for individuals with PVs in TP53 (N = 3). Average PTC for biennial mammography (N = 1,027) was 49% (0-100%). CONCLUSION: Longitudinal screening adherence in individuals with PVs in breast cancer associated genes, as measured by the proportion of time covered, is low; adherence to annual breast MRI falls below that of annual mammography. Additional research should examine screening behavior in individuals with PVs in breast cancer associated genes with a goal of developing interventions to improve adherence to recommended risk management.

A safety study evaluating non-COVID-19 mortality risk following COVID-19 vaccination.

BACKGROUND: The safety of COVID-19 vaccines plays an important role in addressing vaccine hesitancy. We conducted a large cohort study to evaluate the risk of non-COVID-19 mortality after COVID-19 vaccination while adjusting for confounders including individual-level demographics, clinical risk factors, health care utilization, and community-level socioeconomic risk factors. METHODS: The retrospective cohort study consisted of members from seven Vaccine Safety Datalink sites from December 14, 2020 through August 31, 2021. We conducted three separate analyses for each of the three COVID-19 vaccines used in the US. Crude non-COVID-19 mortality rates were reported by vaccine type, age, sex, and race/ethnicity. The counting process model for survival analyses was used to analyze non-COVID-19 mortality where a new observation period began when the vaccination status changed upon receipt of the first dose and the second dose. We used calendar time as the basic time scale in survival analyses to implicitly adjust for season and other temporal trend factors. A propensity score approach was used to adjust for the potential imbalance in confounders between the vaccinated and comparison groups. RESULTS: For each vaccine type and across age, sex, and race/ethnicity groups, crude non-COVID-19 mortality rates among COVID-19 vaccinees were lower than those among comparators. After adjusting for confounders with the propensity score approach, the adjusted hazard ratios (aHRs) were 0.46 (95% confidence interval [CI], 0.44-0.49) after dose 1 and 0.48 (95% CI, 0.46-0.50) after dose 2 of the BNT162b2 vaccine, 0.41 (95% CI, 0.39-0.44) after dose 1 and 0.38 (95% CI, 0.37-0.40) after dose 2 of the mRNA-1273 vaccine, and 0.55 (95% CI, 0.51-0.59) after receipt of Ad26.COV2.S. CONCLUSION: While residual confounding bias remained after adjusting for several individual-level and community-level risk factors, no increased risk was found for non-COVID-19 mortality among recipients of three COVID-19 vaccines used in the US.

ORCA, a values-based decision aid for selecting additional findings from genomic sequencing in adults: Efficacy results from a randomized trial.

PURPOSE: Individuals having genomic sequencing can choose to be notified about pathogenic variants in genes unrelated to the testing indication. A decision aid can facilitate weighing one's values before making a choice about these additional results. METHODS: We conducted a randomized trial (N = 231) comparing informed values-choice congruence among adults at risk for a hereditary cancer syndrome who viewed either the Optional Results Choice Aid (ORCA) or web-based additional findings information alone. ORCA is values-focused with a low-literacy design. RESULTS: Individuals in both arms had informed values-choice congruence (75% and 73% in the decision aid and web-based groups, respectively; odds ratio [OR] = 1.10, 95% CI = 0.58-2.08). Most participants had adequate knowledge (79% and 76% in the decision aid and web-based groups, respectively; OR = 1.20, 95% CI = 0.61-2.34), with no significant difference between groups. Most had information-seeking values (97% and 98% in the decision aid and web-based groups, respectively; OR = 0.59, 95% CI = 0.10-3.61) and chose to receive additional findings. CONCLUSION: The ORCA decision aid did not significantly improve informed values-choice congruence over web-based information in this cohort of adults deciding about genomic results. Both web-based approaches may be effective for adults to decide about receiving medically actionable additional results.

Changes in incidence rates of outcomes of interest in vaccine safety studies during the COVID-19 pandemic.

BACKGROUND: The COVID-19 pandemic caused an abrupt drop in in-person health care (inpatient, Emergency Department, outpatient) and an increase in telehealth care, which poses challenges in vaccine safety studies that identify outcomes from in-person encounters. We examined the changes in incidence rates of selected encounter-based outcomes during the COVID-19 pandemic. METHODS: We assembled a cohort of members from 8 Vaccine Safety Datalink sites from January 1, 2017 through December 31, 2020. Using ICD-10 diagnosis codes or laboratory criteria, we identified 21 incident outcomes in traditional in-person settings and all settings. We defined 4 periods in 2020: January-February (pre-pandemic), April-June (early pandemic), July-September (middle pandemic), and October-December (late pandemic). We defined four corresponding periods in each year during 2017-2019. We calculated incidence rates, conducted difference in difference (DiD) analyses, and reported ratios of incidence rate ratios (RRR) to examine changes in rates from pre-pandemic to early, middle, and late pandemic in 2020, after adjusting for changes across similar periods in 2017-2019. RESULTS: Among > 10 million members, regardless of setting and after adjusting for changes during 2017-2019, we found that incidence rates of acute disseminated encephalomyelitis, encephalitis/myelitis/encephalomyelitis/meningoencephalitis, and thrombotic thrombocytopenic purpura did not significantly change from the pre-pandemic to early, middle or late pandemic periods (p-values ≥ 0.05). Incidence rates decreased from the pre-pandemic to early pandemic period during 2020 for acute myocardial infarction, anaphylaxis, appendicitis, Bell's palsy, convulsions/seizures, Guillain-Barré syndrome, immune thrombocytopenia (ITP), narcolepsy/cataplexy, hemorrhagic stroke, ischemic stroke, and venous thromboembolism (p-values < 0.05). Incidence rates of Bell's palsy, ITP, and narcolepsy/cataplexy were higher in all settings than in traditional in-person settings during the three pandemic periods (p-values < 0.05). CONCLUSION: Rates of some clinical outcomes during the pandemic changed and should not be used as historical background rates in vaccine safety studies. Inclusion of telehealth visits should be considered for vaccine studies involving Bell's palsy, ITP, and narcolepsy/cataplexy.

A decade of data: Adolescent vaccination in the vaccine safety datalink, 2007 through 2016.

BACKGROUND: Between May 2005 and March 2007, three vaccines were recommended by the Advisory Committee on Immunization Practices for routine use in adolescents in the United States: quadrivalent meningococcal conjugate vaccine (MenACWY), tetanus, diphtheria and acellular pertussis vaccine (Tdap), and human papillomavirus vaccine (HPV). Understanding historical adolescent vaccination patterns may inform future vaccination coverage efforts for these and emerging adolescent vaccines, including COVID-19 vaccines. METHODS: This was a descriptive, retrospective cohort study. All vaccines administered to adolescents aged 11 through 18 years in the Vaccine Safety Datalink population between January 1, 2007 and December 31, 2016 were examined. Vaccination coverage was assessed by study year for ≥1 dose Tdap or Td, ≥1 dose Tdap, ≥1 dose MenACWY, ≥1 dose HPV, and ≥3 dose HPV. The proportion of vaccine visits with concurrent vaccination (≥2 vaccines administered at the same visit) was calculated by sex and study year. The most common vaccine combinations administered in the study population were described by sex for two time periods: 2007-2010 and 2011-2016. RESULTS: The number of 11-18-year-olds in the study population averaged 522,565 males and 503,112 females per study year. Between January 2007 and December 2016 there were 4,884,553 vaccine visits in this population (45% among males). The overall proportion of concurrent vaccine visits among males was 43% (33-61% by study year). Among females, 39% of all vaccine visits included concurrent vaccination (32-48% by study year). Vaccine coverage for Tdap, MenACWY, and 1- and 3-dose HPV increased across the study period. A wide variety of vaccine combinations were administered among both sexes and in both time periods. CONCLUSIONS: The high vaccine uptake and multitude of vaccine combinations administered concurrently in the adolescent population of the Vaccine Safety Datalink provide historical patterns with which to compare future adolescent vaccination campaigns.

COVID-19 Vaccination and Non-COVID-19 Mortality Risk - Seven Integrated Health Care Organizations, United States, December 14, 2020-July 31, 2021.

By September 21, 2021, an estimated 182 million persons in the United States were fully vaccinated against COVID-19.* Clinical trials indicate that Pfizer-BioNTech (BNT162b2), Moderna (mRNA-1273), and Janssen (Johnson & Johnson; Ad.26.COV2.S) vaccines are effective and generally well tolerated (1-3). However, daily vaccination rates have declined approximately 78% since April 13, 2021†; vaccine safety concerns have contributed to vaccine hesitancy (4). A cohort study of 19,625 nursing home residents found that those who received an mRNA vaccine (Pfizer-BioNTech or Moderna) had lower all-cause mortality than did unvaccinated residents (5), but no studies comparing mortality rates within the general population of vaccinated and unvaccinated persons have been conducted. To assess mortality not associated with COVID-19 (non-COVID-19 mortality) after COVID-19 vaccination in a general population setting, a cohort study was conducted during December 2020-July 2021 among approximately 11 million persons enrolled in seven Vaccine Safety Datalink (VSD) sites.§ After standardizing mortality rates by age and sex, this study found that COVID-19 vaccine recipients had lower non-COVID-19 mortality than did unvaccinated persons. After adjusting for demographic characteristics and VSD site, this study found that adjusted relative risk (aRR) of non-COVID-19 mortality for the Pfizer-BioNTech vaccine was 0.41 (95% confidence interval [CI] = 0.38-0.44) after dose 1 and 0.34 (95% CI = 0.33-0.36) after dose 2. The aRRs of non-COVID-19 mortality for the Moderna vaccine were 0.34 (95% CI = 0.32-0.37) after dose 1 and 0.31 (95% CI = 0.30-0.33) after dose 2. The aRR after receipt of the Janssen vaccine was 0.54 (95% CI = 0.49-0.59). There is no increased risk for mortality among COVID-19 vaccine recipients. This finding reinforces the safety profile of currently approved COVID-19 vaccines in the United States.

Incidence of pediatric inflammatory bowel disease within the Vaccine Safety Datalink network and evaluation of association with rotavirus vaccination.

BACKGROUND: Recent studies have reported an increase in Inflammatory bowel disease (IBD) incidence in young children, highlighting the need to better understand risk factors for the development of IBD. Licensed for use in infants in 2006, the oral, live-attenuated rotavirus vaccine has biologic plausibility for instigating inflammation of the gut mucosa as a pathway to immune dysregulation. METHODS: Over a ten-year period, we evaluated incidence of IBD within a cohort of children under the age of ten, enrolled in seven integrated healthcare delivery systems. We conducted a nested case-control study to evaluate the association between rotavirus vaccination and IBD using conditional logistic regression. Cases were confirmed via medical record review and matched to non-IBD controls on date of birth, sex, and study site. RESULTS: Among 2.4 million children under the age of 10 years, 333 cases of IBD were identified with onset between 2007 and 2016. The crude incidence of IBD increased slightly over the study period (p-value for trend = 0.046). Of the 333 cases, 227 (68%) were born prior to 2007. Forty-two cases born in 2007 or later, with continuous enrollment since birth were included in the case-control study and matched to 210 controls. The adjusted odds ratio for any rotavirus vaccination in IBD cases, compared to matched controls, was 0.72 (95% confidence interval 0.19-2.65). CONCLUSIONS: Data from this large pediatric cohort demonstrate a small overall increase in IBD incidence in young children over a ten-year period. The data suggest that rotavirus vaccination is not associated with development of IBD.

Cancer Health Assessments Reaching Many (CHARM): A clinical trial assessing a multimodal cancer genetics services delivery program and its impact on diverse populations.

Advances in the application of genomic technologies in clinical care have the potential to increase existing healthcare disparities. Studies have consistently shown that only a fraction of eligible patients with a family history of cancer receive recommended cancer genetic counseling and subsequent genetic testing. Care delivery models using pre-test and post-test counseling are not scalable, which contributes to barriers in accessing genetics services. These barriers are even more pronounced for patients in historically underserved populations. We have designed a multimodal intervention to improve subsequent cancer surveillance, by improving the identification of patients at risk for familial cancer syndromes, reducing barriers to genetic counseling/testing, and increasing patient understanding of complex genetic results. We are evaluating this intervention in two large, integrated healthcare systems that serve diverse patient populations (NCT03426878). The primary outcome is the number of diagnostic (hereditary cancer syndrome) findings. We are examining the clinical and personal utility of streamlined pathways to genetic testing using electronic medical record data, surveys, and qualitative interviews. We will assess downstream care utilization of individuals receiving usual clinical care vs. genetic testing through the study. We will evaluate the impacts of a literacy-focused genetic counseling approach versus usual care genetic counseling on care utilization and participant understanding, satisfaction, and family communication. By recruiting participants belonging to historically underserved populations, this study is uniquely positioned to evaluate the potential of a novel genetics care delivery program to reduce care disparities.

A decision aid for additional findings in genomic sequencing: Development and pilot testing.

OBJECTIVE: To describe the development of a web-based, patient-facing decision aid to support patients and research participants to make an informed, values-based decision about whether to receive additional results from genomic sequencing. METHODS: We developed the decision aid following the multi-step process described in the International Patient Decision Aids Standards. This utilized literature review, focus groups, and alpha testing with research participants undergoing clinical genomic sequencing. RESULTS: The decision aid, the Optional Results Choice Aid (ORCA), includes a seven-question "values clarification exercise," illustrative patient quotes, and summative guidance for the user. The decision aid was found to be highly readable, acceptable and relevant in alpha testing. CONCLUSION: We developed a decision aid to support informed, values-based decision making for patients and research participants considering whether to receive additional results from genomic sequencing. ORCA is being implemented in the NHGRI-funded Cancer Health Assessment Reaching Many (CHARM) study, where we are measuring informed values-choice congruence. PRACTICE IMPLICATIONS: ORCA was designed to support patients and research participants to make an informed, values-based decision about whether to receive additional results from genomic sequencing.

Incidence of herpes zoster among varicella-vaccinated children, by number of vaccine doses and simultaneous administration of measles, mumps, and rubella vaccine.

INTRODUCTION: Children may receive measles-mumps-rubella (MMR) and varicella (VAR) vaccines separately or as measles-mumps-rubella-varicella (MMRV). We examined whether pediatric herpes zoster (HZ) incidence varied by pattern of varicella vaccine administration. METHODS: In six integrated health systems, we examined HZ incidence among children turning 12 months old during 2003-2008. All received varicella and MMR vaccines on recommended schedules. Cases were identified through 2014 using ICD-9 codes. Incidence was examined by number of varicella vaccine doses and same-day MMR. RESULTS: Among 199,797 children, overall HZ incidence was 18.6/100,000 person-years in the first-dose MMR + VAR group, 17.9/100,000 person-years in the MMRV group, and 7.5/100,000 person-years in the VAR-alone group. HZ incidence was lower following the second dose than before the second dose in all first-dose groups. CONCLUSIONS: HZ incidence was not meaningfully different between the MMRV and MMR + VAR first-dose groups. Overall and within first-dose groups, HZ incidence was lower among children receiving two varicella vaccine doses.

Incidence of Herpes Zoster Among Children: 2003-2014.

BACKGROUND AND OBJECTIVES: After the 1996 introduction of routine varicella vaccination in the United States, most studies evaluating pediatric herpes zoster (HZ) incidence reported lower incidence over time, with varying degrees of decline. Using the combined databases of 6 integrated health care organizations, we examined HZ incidence in children over a 12-year period in the varicella vaccine era. METHODS: This study included children aged 0 through 17 years from 2003 through 2014. Using electronic medical records, we identified HZ cases through International Classification of Diseases, Ninth Revision diagnosis code 053. We calculated HZ incidence rates per 100 000 person years of health plan membership for all children and among children who were vaccinated versus unvaccinated. We calculated rates for the 12-year period and examined temporal trends. Among children who were vaccinated, we compared HZ rates by month and year of age at vaccination. RESULTS: The study included 6 372 067 children with ≥1 month of health plan membership. For the 12-year period, the crude HZ incidence rate for all subjects was 74 per 100 000 person years, and the rate among children who were vaccinated was 38 per 100 000 person years, which was 78% lower than that among children who were unvaccinated (170 per 100 000 person years; P < .0001). Overall HZ incidence declined by 72% (P < .0001) from 2003 through 2014. Annual rates in children who were vaccinated were consistently lower than in children who were unvaccinated. CONCLUSIONS: With this population-based study, we confirm the decline in pediatric HZ incidence and the significantly lower incidence among children who are vaccinated, reinforcing the benefit of routine varicella vaccination to prevent pediatric HZ.

Balancing Adherence and Expense: The Cost-Effectiveness of Two-Sample vs One-Sample Fecal Immunochemical Test.

Colorectal cancer (CRC) causes more than 50,000 deaths each year in the United States but early detection through screening yields survival gains; those diagnosed with early stage disease have a 5-year survival greater than 90%, compared to 12% for those diagnosed with late stage disease. Using data from a large integrated health system, this study evaluates the cost-effectiveness of fecal immunochemical testing (FIT), a common CRC screening tool. A probabilistic decision-analytic model was used to examine the costs and outcomes of positive test results from a 1-FIT regimen compared with a 2-FIT regimen. The authors compared 5 diagnostic cutoffs of hemoglobin concentration for each test (for a total of 10 screening options). The principal outcome from the analysis was the cost per additional advanced neoplasia (AN) detected. The authors also estimated the number of cancers detected and life-years gained from detecting AN. The following costs were included: program management of the screening program, patient identification, FIT kits and their processing, and diagnostic colonoscopy following a positive FIT. Per-person costs ranged from $33 (1-FIT at 150ng/ml) to $92 (2-FIT at 50ng/ml) across screening options. Depending on willingness to pay, the 1-FIT 50 ng/ml and the 2-FIT 50 ng/ml are the dominant strategies with cost-effectiveness of $11,198 and $28,389, respectively, for an additional AN detected. The estimates of cancers avoided per 1000 screens ranged from 1.46 to 4.86, depending on the strategy and the assumptions of AN to cancer progression.

Pneumococcal Conjugate Vaccine Safety in Elderly Adults.

BACKGROUND: The 13-valent pneumococcal conjugate vaccine (PCV13) and the 23-valent pneumococcal polysaccharide vaccine (PPSV23) were both recommended to adults aged ≥65 years. The study examines adults ≥65 years for risk of adverse events (AEs) requiring medical attention following vaccination with PCV13 as compared with vaccination with PPSV23, a long-standing vaccine with a satisfactory safety profile. METHODS: The cohort study included 6 Vaccine Safety Datalink sites. The exposed person-time included follow-up time of the first PCV13 received by subjects age ≥65 years from January 1 to August 15, 2015. The comparator person-time included follow-up time after the first PPSV23 received by subjects of the same age during Janaury 1 to August 15 of each year of 2011-2015. The prespecified AEs included cardiovascular events, Bell's palsy, Guillain-Barré syndrome, syncope, erythema multiforme, thrombocytopenia, cellulitis and infection, allergic reaction, and anaphylaxis. Inverse probability of treatment weighting-adjusted Poisson regression models was used to estimate the relative risk (RR) of each AE. RESULTS: A total of 313 136 doses of PCV13 and 232 591 doses of PPSV23 were included. The adjusted RRs comparing the incidence of AEs following PCV13 vs PPSV23 were all <1, except for anaphylaxis, which was insignificant with an RR of 1.32 (95% confidence interval, 0.30-5.79). Only 1 patient who received PCV13 and 4 other vaccines concomitantly was confirmed by medical chart review as having experienced anaphylaxis after vaccination. CONCLUSIONS: These data do not support an increased rate of adverse events following PCV13 administration in elders compared with PPSV23 and should provide reassurance regarding continued use of PCV13.

Performance of a quantitative fecal immunochemical test for detecting advanced colorectal neoplasia: a prospective cohort study.

BACKGROUND: The fecal immunochemical test (FIT) is easier to use and more sensitive than the guaiac fecal occult blood test, but it is unclear how to optimize FIT performance. We compared the sensitivity and specificity for detecting advanced colorectal neoplasia between single-sample (1-FIT) and two-sample (2-FIT) FIT protocols at a range of hemoglobin concentration cutoffs for a positive test. METHODS: We recruited 2,761 average-risk men and women ages 49-75 referred for colonoscopy within a large nonprofit, group-model health maintenance organization (HMO), and asked them to complete two separate single-sample FITs. We generated receiver-operating characteristic (ROC) curves to compare sensitivity and specificity estimates for 1-FIT and 2-FIT protocols among those who completed both FIT kits and colonoscopy. We similarly compared sensitivity and specificity between hemoglobin concentration cutoffs for a single-sample FIT. RESULTS: Differences in sensitivity and specificity between the 1-FIT and 2-FIT protocols were not statistically significant at any of the pre-specified hemoglobin concentration cutoffs (10, 15, 20, 25, and 30 µg/g). There was a significant difference in test performance of the one-sample FIT between 50 ng/ml (10 µg/g) and each of the higher pre-specified cutoffs. Disease prevalence was low. CONCLUSIONS: A two-sample FIT is not superior to a one-sample FIT in detection of advanced adenomas; the one-sample FIT at a hemoglobin concentration cutoff of 50 ng/ml (10 µg/g) is significantly more sensitive for advanced adenomas than at higher cutoffs. These findings apply to a population of younger, average-risk patients in a U.S. integrated care system with high rates of prior screening.

Correction to: Early Colorectal Cancer Detected by Machine Learning Model Using Gender, Age, and Complete Blood Count Data.

The article Early Colorectal Cancer Detected by Machine Learning Model Using Gender, Age, and Complete Blood Count Data, written by Mark C. Hornbrook, Ran Goshen, Eran Choman, Maureen O'Keeffe-Rosetti, Yaron Kinar, Elizabeth G. Liles, and Kristal C. Rust, was originally published Online First without open access.

Early Colorectal Cancer Detected by Machine Learning Model Using Gender, Age, and Complete Blood Count Data.

BACKGROUND: Machine learning tools identify patients with blood counts indicating greater likelihood of colorectal cancer and warranting colonoscopy referral. AIMS: To validate a machine learning colorectal cancer detection model on a US community-based insured adult population. METHODS: Eligible colorectal cancer cases (439 females, 461 males) with complete blood counts before diagnosis were identified from Kaiser Permanente Northwest Region's Tumor Registry. Control patients (n = 9108) were randomly selected from KPNW's population who had no cancers, received at ≥1 blood count, had continuous enrollment from 180 days prior to the blood count through 24 months after the count, and were aged 40-89. For each control, one blood count was randomly selected as the pseudo-colorectal cancer diagnosis date for matching to cases, and assigned a "calendar year" based on the count date. For each calendar year, 18 controls were randomly selected to match the general enrollment's 10-year age groups and lengths of continuous enrollment. Prediction performance was evaluated by area under the curve, specificity, and odds ratios. RESULTS: Area under the receiver operating characteristics curve for detecting colorectal cancer was 0.80 ± 0.01. At 99% specificity, the odds ratio for association of a high-risk detection score with colorectal cancer was 34.7 (95% CI 28.9-40.4). The detection model had the highest accuracy in identifying right-sided colorectal cancers. CONCLUSIONS: ColonFlag® identifies individuals with tenfold higher risk of undiagnosed colorectal cancer at curable stages (0/I/II), flags colorectal tumors 180-360 days prior to usual clinical diagnosis, and is more accurate at identifying right-sided (compared to left-sided) colorectal cancers.

Implementation challenges and successes of a population-based colorectal cancer screening program: a qualitative study of stakeholder perspectives.

BACKGROUND: Few studies describe system-level challenges or facilitators to implementing population-based colorectal cancer (CRC) screening outreach programs. Our qualitative study explored viewpoints of multilevel stakeholders before, during, and after implementation of a centralized outreach program. Program implementation was part of a broader quality-improvement initiative. METHODS: During 2008-2010, we conducted semi-structured, open-ended individual interviews and focus groups at Kaiser Permanente Northwest (KPNW), a not-for-profit group model health maintenance organization using the practical robust implementation and sustainability model to explore external and internal barriers to CRC screening. We interviewed 55 stakeholders: 8 health plan leaders, 20 primary care providers, 4 program managers, and 23 endoscopy specialists (15 gastroenterologists, 8 general surgeons), and analyzed interview transcripts to identify common as well as divergent opinions expressed by stakeholders. RESULTS: The majority of stakeholders at various levels consistently reported that an automated telephone-reminder system to contact patients and coordinate mailing fecal tests alleviated organizational constraints on staff's time and resources. Changing to a single-sample fecal immunochemical test (FIT) lessened patient and provider concerns about feasibility and accuracy of fecal testing. The centralized telephonic outreach program did, however, result in some screening duplication and overuse. Higher rates of FIT completion and a higher proportion of positive results with FIT required more colonoscopies. CONCLUSIONS: Addressing barriers at multiple levels of a health system by changing the delivery system design to add a centralized outreach program, switching to a more accurate and easier-to-use fecal test, and providing educational and electronic support had both benefits and problematic consequences. Other health care organizations can use our results to understand the complexities of implementing centralized screening programs.