Large Integrated Health Care System Research Articles

Reducing Endometrial Cancer Disparities to Advance Health Equity (REACH Study) [ID 2683434

INTRODUCTION: Racial disparities in endometrial cancer survival have been increasing over the past 10 years. While Black women are diagnosed more often with aggressive histologic subtypes, studies have found tumor characteristics only partially account for their lower rates of survival. The degree to which system factors such as insurance and access to subspecialty care contribute to disparate survival outcomes is unclear. METHODS: This was a retrospective cohort study of individuals diagnosed with endometrial cancer from 2005 to 2021 within a large integrated health care system with high levels of access to gynecologic oncologists. Data were extracted from tumor registry and electronic databases. Multivariable logistic regression was used to evaluate associations between 1- and 5-year survival rates and race. RESULTS: Among 11,249 women, 61.9% identified as White, 15.1% Asian, 7.8% Black, 13.6% Hispanic, and 0.6% mixed or unknown race, with 46.5% having public insurance through Medicare or Medi-Cal. In preliminary analysis, Black women had lower survival compared to White women at 1 year (87.4% versus 92.3%) and 5 years (67% versus 78.7%) (OR 0.57; 95% CI, 0.46–0.71; P<.001). However, when adjusted for histology and stage, survival differences between Black and White women at both 1 year (OR 0.83; 95% CI, 0.65–1.07; P=.15) and 5 years (OR 0.92; 95% CI, 0.73–1.15; P=.45) were no longer significant. CONCLUSION: Among a diverse group of insured individuals receiving care within an integrated health care system, tumor characteristics largely accounted for racial disparities in survival. These findings suggest that disparities in survival not explained by tumor characteristics can be addressed by closing gaps in insurance coverage and access to care.

Association between SARS-CoV-2 Infection and Adverse Perinatal Outcomes by Race/Ethnicity in a Large Integrated Health Care System.

Recent studies have reported associations between severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) infection during pregnancy and adverse perinatal outcomes but the extent to which these associations vary by race/ethnicity remains uncertain. Therefore, we examined how the association between prenatal SARS-CoV-2 infection and adverse perinatal outcomes may be modified by race/ethnicity. A retrospective cohort study was performed using data on 67,986 pregnant women extracted from the Kaiser Permanente Southern California electronic health records between April 6, 2020, and December 31, 2021. Upon admission to labor and delivery, all women were routinely tested for coronavirus disease 2019 (COVID-19) using real-time reverse-transcriptase polymerase chain reaction test. Adjusted odds ratios (aORs) were used to estimate associations. During the study period, COVID-19 was diagnosed in 4,960 (7%) of singleton pregnancies, with the highest rates observed among Hispanics (9.4%) and non-Hispanic Blacks (6.2%). Compared with non-Hispanic Whites, Hispanics (aOR: 1.12, 95% CI: 1.03, 1.21) with SARS-CoV-2 infection had the highest odds of a pregnancy associated with nonreassuring fetal heart rate tracing. Neonates of all races/ethnicities, except for non-Hispanic Blacks, showed significantly increased odds of SARS-CoV-2 infection, with the highest risk observed among Asians/Pacific Islanders (aOR: 10.88, 95% CI: 1.33, 89.04). Non-Hispanic White mothers who tested positive were admitted to intensive care unit (ICU) at a higher rate at delivery and within 7 days of delivery (aOR: 34.77, 95% CI: 11.3, 107.04; aOR: 26.48, 95% CI: 9.55, 73.46, respectively). Hispanics were also at a significantly higher odds of admission to ICU (aOR: 4.62, 95% CI: 2.69, 7.94; aOR: 4.42, 95% CI: 2.58, 7.56, respectively). Non-Hispanic Black, Hispanic, and Asian/Pacific Islander mothers who tested positive for SARS-CoV-2 prenatally, were at increased risk for preeclampsia/eclampsia, and preterm birth as compared to non-Hispanic White mothers. The findings highlight racial/ethnic disparities in the association between SARS-CoV-2 infection and adverse perinatal outcomes. The risk of neonatal SARS-CoV-2 infection was highest for Asian/Pacific Islanders. We also observed a remarkably high risk of ICU admission for non-Hispanic White mothers infected with SARS-CoV-2. · Race/ethnicity influences perinatal outcomes in pregnancies impacted by SARS-CoV-2.. · The risk of neonatal SARS-CoV-2 infection was highest for Asian/Pacific Islanders.. · White mothers had a notably high risk of ICU admission at delivery following SARS-CoV-2 infection..

Impact of the Coronavirus Disease 2019 pandemic on neoadjuvant chemotherapy use in patients diagnosed with epithelial type ovarian cancer.

The Coronavirus Disease 2019 (COVID-19) pandemic posed critical challenges in providing care to ovarian cancer (OC) patients, including delays in OC diagnosis and treatment initiation. To accommodate for delays in OC surgery, the Society of Gynecologic Oncology (SGO) recommended preferential use of neoadjuvant chemotherapy during the pandemic. The purpose of this study was to assess the association of the COVID-19 pandemic with neoadjuvant chemotherapy use in patients diagnosed with OC. This retrospective cohort study included patients diagnosed with stage II-IV ovarian cancer of epithelial subtype between 01/01/2017-06/30/2021 at Kaiser Permanente Southern California (KPSC), a large integrated healthcare system in the United States. Ovarian cancer patients diagnosed between 2017-2020 were identified from KPSC's Surveillance, Epidemiology, and End Results (SEER)-affiliated cancer registry. Patients diagnosed in 2021 were identified from the electronic medical records (EMR) using ICD-10 diagnosis codes, followed by medical chart review to validate diagnosis and extract information on histology and stage at diagnosis. March 4, 2020 was used as the cut-off to define pre-pandemic and pandemic periods. Patients diagnosed with COVID-19 between OC diagnosis and treatment completion were excluded. Data on neoadjuvant chemotherapy use were extracted from the cancer registry and EMR, supplemented by chart review. Modified Poisson regression was used to evaluate the association of the pandemic with neoadjuvant chemotherapy use. Of 566 OC patients, 160 (28.3%) were diagnosed in the pandemic period. Patients diagnosed in the pandemic period were slightly younger (mean age 62.7 vs 64.9 years, p=0.07) and had a higher burden of Charlson comorbidities (p=0.05) than patients diagnosed in pre-pandemic period. No differences in time to treatment initiation were observed by pandemic periods. Neoadjuvant chemotherapy use was documented in 58.7% patients during the pandemic period compared to 47.3% in pre-pandemic period (p=0.01). After adjusting for covariates, patients diagnosed in the pandemic period were 29% more likely to receive neoadjuvant chemotherapy than patients diagnosed in pre-pandemic period [RR(95%CI): 1.29(1.12-1.49)]. Ovarian cancer patients diagnosed in the COVID-19 pandemic were more likely to receive neoadjuvant chemotherapy than patients diagnosed before the pandemic. Future research on patient outcomes and trends in the post-pandemic period are warranted.

Association Between Cardiovascular Health Status and Healthcare Utilization in a Large Integrated Healthcare System

IntroductionThe American Heart Association (AHA) “Life's Simple 7” schema can be used to categorize patients’ cardiovascular health (CVH) status as Poor, Intermediate or Ideal based on smoking, body mass index (BMI), physical activity, dietary patterns, blood pressure, cholesterol, and fasting blood sugar. This study examined the association between CVH status and subsequent healthcare utilization. MethodsThis was an observational cohort study of adults from an integrated health care delivery system – Kaiser Permanente Northern California – that had outpatient care between 2013-2014. Patients were categorized by AHA CVH status: Poor, Intermediate or Ideal. Individual-level health care utilization and costs in 2015 were accumulated for each patient and compared across the 3 CVH categories and stratified by age groups. Results991,698 patients were included in the study. 194,003 (19.6%) were age 18-39, 554,129 (55.9%) age 40-64, and 243,566 (24.6%) age ≥65 years. 259,931 (26.2%) had Ideal CVH, 521,580 (52.6%) had Intermediate CVH, and 210,187 (21.2%) had Poor CVH. Healthcare utilization measured by average relative cost per patient increased monotonically across age categories (p<0.001). In addition, CVH category was inversely associated with lower cost in each age group (p<0.001). ConclusionAdults who were younger and had more ideal CVH had relatively lower health care costs across age groups. Interventions to promote better CVH may improve patient outcomes and reduce overall healthcare expenditures.

Abstract 77: A Regional Transfer Center Pilot Improves Door-in-Door-out Time in a Large Integrated Healthcare System

Background: Successful treatment of patients with large vessel occlusions (LVO) is time dependent, yet timely transfers between hospitals for endovascular therapy (EVT) can be challenging. At Kaiser Permanente Northern California (KPNC) a wide variety of ambulance and EMS agencies may be involved in Urgent interfacility transfers. We evaluated a region-wide pilot utilizing a central transfer center to facilitate these transfers within a large integrated healthcare system. Methods: KPNC implemented a regional transfer center (TC) in March 2023, which arranged transfers for 17 of 19 primary stroke centers (PSC) from 8am to 9pm. Two centers were excluded as they were utilizing 911 transport rather than interfacility ambulance hub transfers. The study cohort included all stroke patients arriving to a PSC with LVO who were transferred to one of our Comprehensive Stroke Centers for EVT from 3/6/23 through 8/4/23. Data were compared to a similar cohort 10/6/23 through 3/5/23 (pre-TC). All patients were evaluated by teleneurologists on arrival, who requested initiation of ambulance transfer. Before TC, the ambulance request was made by the local ED facility; after the implementation of the TC, ambulance was arranged by the TC and communicated with teleneurology and the local ED. Door in door out (DIDO) times were measured for all transfers. Results: The pre-TC group had 81 transfers. The post-TC group had 68 patients. After TC, median DIDO dropped from 105 minutes to 83.5 minutes (p=0.02). The percent of patients transferred within 90 minutes rose from 41% to 57% (p=0.04) and the percent transferred with 75 minutes rose from 25% to 35% (p=0.14). Similar improvements were seen regardless of sex or ethnicity [Table]. Conclusions: Preliminary results show a significant improvement in DIDO times with initiation of a centralized TC to facilitate LVO transfers in a large integrated delivery system. Systems wishing to improve DIDO performance should consider implementing a similar TC.

Remote delivery of cancer genetic testing in veterans with metastatic prostate cancer: A Million Veteran Program study.

60 Background: Germline pathogenic alterations are found in approximately 10% of men with metastatic prostate cancer (mPC) and can inform precision therapy, surveillance, and cancer prevention for family. National guidelines recommend germline genetic testing be offered to all men with mPC, yet uptake of testing in the community has been documented to be 6-12% with many barriers to testing. We conducted a study to determine uptake of testing using remote consenting and testing for veterans with mPC who had participated in the Veteran’s Administration Million Veteran Program (MVP). We wanted to know if remote testing could overcome limited point-of-care counseling and ordering to increase uptake of germline testing. Methods: This prospective study enrolled veterans who participated in MVP study with a diagnosis of mPC. Veterans were contacted by mail with an option to opt-out of further contact. Those who did not opt-out were mailed information regarding the study and received a follow-up phone call to establish interest in germline testing with a CLIA-level germline test (BROCA). Those expressing interest provided verbal informed consent and were mailed a saliva collection kit for a multigene cancer predisposition gene panel test (University of Washington). Results were disclosed by phone and mailed to the patient with genetic counseling support and was also sent to the oncology provider by email, phone or both. Two research coordinators and two part-time genetic counselors provided review of consenting, counseling and return of results. Results: As of September 17, 2023, 1952 veterans who were alive with an incident diagnosis of mPC were identified through MVP and did not opt out of further contact. Informational letters were sent to the home address of all eligible participants. 681 (22%) of veterans were reached and completed informed consent. All participants were then sent a saliva kit and 432 (63%) completed testing, with 23 tests currently pending. Among the participants, 70% were White, 21% were Black, 0.5% Asian and 8% unknown. Thirteen percent (53/409) were found to have a germline pathogenic variant. Conclusions: We evaluated uptake of germline testing using a remote, VA system-wide approach to identify and offer genetic testing for veterans with mPC with access and cost issues removed. We completed germline testing at rates significantly higher than those reported in the community with modest personnel requirements, while also reaching a more diverse population of patients. Remote genetic testing can improve uptake of testing in large integrated health care systems.

Mode of Detection of Second Breast Cancers in Patients Undergoing Surveillance After Treatment of Ductal Carcinoma in Situ.

For patients undergoing posttreatment surveillance after ductal carcinoma in situ (DCIS), the NCCN Guidelines for Breast Cancer recommend annual breast imaging and physical examination every 6 to 12 months for 5 years, and then annually. The aim of our study was to evaluate the modes of detection (imaging, patient reported, or physical examination) of second cancers in a cohort of patients undergoing surveillance after primary DCIS treatment to better inform surveillance recommendations. We performed a retrospective cohort study of patients with DCIS treated between January 1, 2008, and December 31, 2011, within a large integrated health care system. Information on patient demographics, index DCIS treatment, tumor characteristics, and mode of detection of second breast cancer was obtained from the electronic health record or chart review. Our study cohort consisted of 1,550 women, with a median age of 59 years at diagnosis. Surgical treatment of DCIS included lumpectomy (75.0%; n=1,162), unilateral mastectomy (21.1%; n=327), or bilateral mastectomy (3.9%; n=61), with or without sentinel lymph node biopsy. Additionally, 44.4% (n=688) and 28.3% (n=438) received radiation and endocrine therapies, respectively. Median follow-up was 10 years, during which 179 (11.5%) women were diagnosed with a second breast cancer. Of the second cancers, 43.0% (n=77) were ipsilateral and 54.8% (n=98) contralateral, and 2.2% (n=4) presented with distant metastases; 61.5% (n=110) were invasive, 36.3% (n=65) were DCIS, and 2.2% (n=4) were Paget's disease. Second breast cancers were imaging-detected in 74.3% (n=133) of cases, patient-detected in 20.1% (n=36), physician-detected in 2.2% (n=4), and detected incidentally on imaging or pathology from procedures unrelated to oncologic care in 3.4% (n=6). In our cohort of patients undergoing surveillance following diagnosis and treatment of DCIS, 2% of second breast cancers were detected by a clinical breast examination. This suggests that survivorship care should prioritize mammography and patient education regarding breast self-examination and symptoms that warrant evaluation to detect second breast cancers.

Abstract WP87: To Walk-In or Not to Walk-In: Effects of ED Mode of Arrival on Acute Stroke Treatment and Hospitalization Outcomes in a Large Integrated Healthcare System

Background: Prior studies reported that only half of stroke patients arrived at emergency departments (ED) via emergency medical services (EMS). We sought to assess predictors of ED mode of arrival and its effects on acute stroke care and inpatient outcomes in a large racial-ethnically diverse cohort. Methods: In 2016, we launched a system-wide standardized telestroke program for 21 certified stroke centers. Non-cancelled stroke alerts included those patients who were potential candidates for further workup for acute ischemic stroke treatment. This study cohort included non-cancelled stroke alerts seen by Telestroke in 2021. Assessment included demographics, ED arrival mode , initial NIH Stroke Scale (NIHSS), neuroimaging results, IV thrombolytic, door-to-needle (DTN) time, large vessel occlusion (LVO), thrombectomy referral, hospital length of stay and discharge disposition. Clustered logistic regression was used to evaluate ED arrival mode’s effects on short-term outcomes. Results: Of the 7744 stroke alerts seen by Telestroke in 2021, 3264 (42.2%) were non-canceled. Among these, 2014 (61.7%) were in the early window (0-4.5 hours from last seen well). More of those arriving in the early window came via EMS compared to those seen in the extended window [1625 (80.7%) versus 729 (58.3%) respectively]. Among non-canceled ED stroke alerts, compared to patients who arrived via EMS, walk-in patients were younger, more Hispanic, had lower initial NIHSS scores, longer times for ED to contact Telestroke, slower DTN times, but less mortality [Table]. In multivariate analysis, walk-in patients were less likely to have a DTN time within 30 minutes (OR=0.37, 95% CI 0.25-0.34; p=&lt;0.001). Conclusions: In our cohort, there were several differences between walk-in patients and those who arrived via EMS. Further research is needed to understand these differences, the reasons for choosing walk-in, and whether ED mode of arrival affects clinical outcomes.

Abstract WP216: Results of a Pediatric Stroke Program Utilizing Stroke Telemedicine Across a Fifteen Hospital Healthcare System in Southern California

Background: The recognition of pediatric stroke symptoms is challenging. A regionwide pediatric stroke program utilizing an educational initiative across 15 emergency departments (ED) coupled with telemedicine was developed to standardize the process for immediate access to stroke experts and decrease door-to-intervention times. This study evaluates the impact of this intervention. Methods: Data from the pre- (1/2020-12/2021) vs post-intervention (1/2022-7/2023) including baseline demographics, volumes, and process measure timeframes: Door-to-Telestroke Page, Door-to-CT Interpretation, Door-to-Telestroke Call Back, and Door-to-Needle Time were abstracted. Chi-Square and Wilcoxon Rank Sum were used for statistical analysis. Results: There were a total of 92 Telestroke consults, with more in the post group: 35(38%) pre vs 57(62%) post (p=0.04). 52% were female with mean age of 12 (±1.2) years. 78(85%) were stroke mimics: 74% had non-cerebrovascular or migrainous neurological disorders as primary diagnoses. 7(7.6%) patients had interventions: 4(4.3%) thrombolytics, 9(9.7%) arterial or venous thrombectomies. 51% were discharged home and 43% were transferred to an acute care facility for higher level of care. There was no difference between pre- and post-implementation median [IQR] Door-to-Telestroke Page (17.5[14-21] vs 15[11-18] mins); Door-to-CT Interpretation (29[25-32] vs 27[23-29] mins) and Door-to-Needle Time. Median Post Door-to-Telestroke Call Back was faster (pre: 20[17-22] vs post: 16.0[13-18] mins [p=0.03]). Conclusion: Systemwide ED education coupled with telemedicine in implementing pediatric stroke systems of care in a large integrated healthcare system improves access to stroke specialists and time-based metrics in this low-volume, high-risk patient population.

Coding of Childhood Psychiatric and Neurodevelopmental Disorders in Electronic Health Records of a Large Integrated Health Care System: Validation Study.

Mental, emotional, and behavioral disorders are chronic pediatric conditions, and their prevalence has been on the rise over recent decades. Affected children have long-term health sequelae and a decline in health-related quality of life. Due to the lack of a validated database for pharmacoepidemiological research on selected mental, emotional, and behavioral disorders, there is uncertainty in their reported prevalence in the literature. We aimed to evaluate the accuracy of coding related to pediatric mental, emotional, and behavioral disorders in a large integrated health care system's electronic health records (EHRs) and compare the coding quality before and after the implementation of the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) coding as well as before and after the COVID-19 pandemic. Medical records of 1200 member children aged 2-17 years with at least 1 clinical visit before the COVID-19 pandemic (January 1, 2012, to December 31, 2014, the ICD-9-CM coding period; and January 1, 2017, to December 31, 2019, the ICD-10-CM coding period) and after the COVID-19 pandemic (January 1, 2021, to December 31, 2022) were selected with stratified random sampling from EHRs for chart review. Two trained research associates reviewed the EHRs for all potential cases of autism spectrum disorder (ASD), attention-deficit hyperactivity disorder (ADHD), major depression disorder (MDD), anxiety disorder (AD), and disruptive behavior disorders (DBD) in children during the study period. Children were considered cases only if there was a mention of any one of the conditions (yes for diagnosis) in the electronic chart during the corresponding time period. The validity of diagnosis codes was evaluated by directly comparing them with the gold standard of chart abstraction using sensitivity, specificity, positive predictive value, negative predictive value, the summary statistics of the F-score, and Youden J statistic. κ statistic for interrater reliability among the 2 abstractors was calculated. The overall agreement between the identification of mental, behavioral, and emotional conditions using diagnosis codes compared to medical record abstraction was strong and similar across the ICD-9-CM and ICD-10-CM coding periods as well as during the prepandemic and pandemic time periods. The performance of AD coding, while strong, was relatively lower compared to the other conditions. The weighted sensitivity, specificity, positive predictive value, and negative predictive value for each of the 5 conditions were as follows: 100%, 100%, 99.2%, and 100%, respectively, for ASD; 100%, 99.9%, 99.2%, and 100%, respectively, for ADHD; 100%, 100%, 100%, and 100%, respectively for DBD; 87.7%, 100%, 100%, and 99.2%, respectively, for AD; and 100%, 100%, 99.2%, and 100%, respectively, for MDD. The F-score and Youden J statistic ranged between 87.7% and 100%. The overall agreement between abstractors was almost perfect (κ=95%). Diagnostic codes are quite reliable for identifying selected childhood mental, behavioral, and emotional conditions. The findings remained similar during the pandemic and after the implementation of the ICD-10-CM coding in the EHR system.

Racial Disparities in COVID-19 Severity Are Partially Mediated by Chronic Stress-Evidence from a Large Integrated Healthcare System.

Racial and ethnic minorities have experienced a disproportionate burden of severe COVID-19. Whether chronic stress, also disproportionately experienced by racial and ethnic minorities, explains this excess risk is unknown. We identified 9577 adults (≥ 18years) diagnosed with COVID-19 from January 1, 2020, through September 30, 2021, enrolled in Kaiser Permanente Georgia (KPGA) with complete biomarker data. Self-reported race (Black or White) was defined from electronic medical records. Chronic stress, defined as allostatic load (AL), a composite score (scale 0-7) based on seven cardio-metabolic biomarkers, was categorized as below (low AL) or above (high AL) the median. Severe COVID-19 was defined as hospitalization or mortality within 30days of COVID-19 diagnosis. The association between race, AL, and severe COVID-19 was assessed using multivariable Poisson regression. The mediating effect of AL was assessed using the Valeri and VanderWeele method. All results were expressed as risk ratios (RRs) with 95% confidence intervals. Overall, Black (vs. White) KPGA members had an 18% excess risk of AL (RR: 1.18, 95%CI: 1.14-1.23) and a 24% excess risk of severe COVID-19 (RR: 1.24, 95%CI: 1.12, 1.37). AL explained 23% of the Black-White disparities in severe COVID-19. In our study, chronic stress, characterized by AL, partially mediated Black-White disparities in severe COVID-19 outcomes.

Homelessness Among Acute Care Patients Within a Large Health Care System in Northern California.

The impacts of homelessness on health and health care access are detrimental. Intervention and efforts to improve outcomes and increase availability of affordable housing have mainly originated from the public health sector and government. The role that large community-based health systems may play has yet to be established. This study characterizes patients self-identified as homeless in acute care facilities in a large integrated health care system in Northern California to inform the development of collaborative interventions addressing unmet needs of this vulnerable population. The authors compared sociodemographic characteristics, clinical conditions, and health care utilization of individuals who did and did not self-identify as homeless and characterized their geographical distribution in relation to Sutter hospitals and homeless resources. Between July 1, 2019 and June 30, 2020, 5% (N = 20,259) of the acute care settings patients had evidence of homelessness, among which 51.1% age <45 years, 66.4% males, and 24% non-Hispanic Black. Patients experiencing homelessness had higher emergency department utilization and lower utilization of outpatient and urgent care services. Mental health conditions were more common among patients experiencing homelessness. More than half of the hospitals had >5% of patients who identified as homeless. Some hospitals with higher proportions of patients experiencing homelessness are not located near many shelter resources. By understanding patients who self-identify as homeless, it is possible to assess the role of the health system in addressing their unmet needs. Accurate identification is the first step for the health systems to develop and deliver better solutions through collaborations with nonprofit organizations, community partners, and government agencies.

Identifying Symptoms Prior to Pancreatic Ductal Adenocarcinoma Diagnosis in Real-World Care Settings: Natural Language Processing Approach

Background Pancreatic cancer is the third leading cause of cancer deaths in the United States. Pancreatic ductal adenocarcinoma (PDAC) is the most common form of pancreatic cancer, accounting for up to 90% of all cases. Patient-reported symptoms are often the triggers of cancer diagnosis and therefore, understanding the PDAC-associated symptoms and the timing of symptom onset could facilitate early detection of PDAC. Objective This paper aims to develop a natural language processing (NLP) algorithm to capture symptoms associated with PDAC from clinical notes within a large integrated health care system. Methods We used unstructured data within 2 years prior to PDAC diagnosis between 2010 and 2019 and among matched patients without PDAC to identify 17 PDAC-related symptoms. Related terms and phrases were first compiled from publicly available resources and then recursively reviewed and enriched with input from clinicians and chart review. A computerized NLP algorithm was iteratively developed and fine-trained via multiple rounds of chart review followed by adjudication. Finally, the developed algorithm was applied to the validation data set to assess performance and to the study implementation notes. Results A total of 408,147 and 709,789 notes were retrieved from 2611 patients with PDAC and 10,085 matched patients without PDAC, respectively. In descending order, the symptom distribution of the study implementation notes ranged from 4.98% for abdominal or epigastric pain to 0.05% for upper extremity deep vein thrombosis in the PDAC group, and from 1.75% for back pain to 0.01% for pale stool in the non-PDAC group. Validation of the NLP algorithm against adjudicated chart review results of 1000 notes showed that precision ranged from 98.9% (jaundice) to 84% (upper extremity deep vein thrombosis), recall ranged from 98.1% (weight loss) to 82.8% (epigastric bloating), and F1-scores ranged from 0.97 (jaundice) to 0.86 (depression). Conclusions The developed and validated NLP algorithm could be used for the early detection of PDAC.

362 Identifying hemolytic disease of the fetus and newborn (HDFN) in a large integrated healthcare system

362 Identifying hemolytic disease of the fetus and newborn (HDFN) in a large integrated healthcare system

Randomized Trial of Patient Outreach Approaches to De-implement Outdated Colonoscopy Surveillance Intervals

Guidelines now recommend patients with low-risk adenomas receive colonoscopy surveillance in 7-10 years and those with the previously recommended 5-year interval be reevaluated. We tested three outreach approaches for transitioning patients to the 10-year interval recommendation. Design: 3-arm pragmatic randomized trial comparing telephone, secure messaging, and mailed letter outreach. Kaiser Permanente Northern California, a large integrated healthcare system. Patients aged 54-70 years with 1-2 small (<10 mm) tubular adenomas at baseline colonoscopy, due for 5-year surveillance in 2022, without high-risk conditions, and with access to all three outreach modalities. Patients were randomly assigned to outreach arm (telephone [n=200], secure message [n=203], and mailed letter [n=201]) stratified by age, sex, race, and ethnicity. Outreach in each arm was performed by trained medical assistants (unblinded) communicating in English with 1 reminder attempt at 2-4 weeks. Participants could change their assigned interval to 10 years or continue their planned 5-year interval. 60-day response rates were higher for telephone (64.5%) and secure messaging outreach (51.7%) versus mailed letter (31.3%). Also, more patients adopted the 10-year surveillance interval in the telephone (37.0%) and secure messaging arms (32.0%) compared to mailed letter (18.9%) and rate differences were significant for telephone (18.1%; 97.5% CI: 8.3%, 27.9%) and secure message outreach (13.1%; 97.5% CI: 3.5%, 22.7%) versus mailed letter outreach. Telephone and secure messaging were more effective than mailed letter outreach for de-implementing outdated colonoscopy surveillance recommendations among individuals with a history of low-risk adenomas in an integrated healthcare setting.

The association between the area deprivation index and dementia in a Veteran Population

AbstractBackgroundResidence in a disadvantaged neighborhood is associated with adverse health exposures and outcomes, and may increase risk for cognitive impairment and dementia. We investigated the association of the area deprivation index (ADI) with dementia in in a sample of the national Veteran population.MethodsWe examined data from the Veterans Health Administration (VHA), the largest integrated health care system in the United States. We studied 1,613,337 patients who obtained VHA care from 10/1/1999 to 9/30/2021, (n = 1,395,872 White; n = 147,530 Black) aged ≥55 years (mean: 68.6, SD: 7.7), 98% male. We characterized neighborhood deprivation with the ADI, a validated measure that combines 17 specific indicators (e.g., income, education, employment and housing) into an index of disadvantage. Participant zip codes from residential addresses were geocoded and assigned an ADI national rank according to their census block group. We used these data to categorize participants into ADI rank quintiles, where higher ADI rank quintile corresponds with more deprivation. We estimated adjusted incidence rates of dementia (International Classification of Diseases, 9th and 10th Edition codes) using Cox proportional hazards regression with age as the time scale.ResultsDuring a mean follow‐up of up to 11 (SD 4.8) years, 208,909 participants (12.8%) developed dementia. Compared to the least disadvantaged neighborhood quintile, all other groups had an increased risk of developing dementia in models adjusted for sex, race/ ethnicity, and psychiatric and medical comorbidities. The association was dose‐dependent, with the most disadvantaged quintile demonstrating the greatest risk (Quintile 2 HR: 1.09 95% CI: 1.07, 1.11; Quintile 3 HR: 1.14, 95% CI: 1.12, 1.16; Quintile 4 HR: 1.17, 95% CI: 1.16, 1.19, Quintile 5 HR: 1.24, 95% CI: 1.22, 1.26, Quintile 1 reference group). Repeating our main analysis using competing risks regression led to similar results.ConclusionWe found a dose‐dependent association between ADI and dementia The ADI may be a valuable tool for identifying high‐risk communities for dementia intervention.

Adverse drug events associated with nortriptyline compared with paroxetine and alternative medications in an older adult population: a retrospective cohort study in Southern California

ObjectiveInvestigate risk for falls, fractures and syncope in older adult patients treated with nortriptyline compared with paroxetine and alternative medications.DesignRetrospective cohort study.SettingThe electronic medical record and prescription drug database of...

Trends in firearm injury in a southern California health care system from 2010 to 2020

BackgroundFirearm injury is a significant public health concern in the United States.MethodsData on fatal and nonfatal firearm injuries were obtained from a cohort of N = 7,473,650 members of Kaiser Permanente Southern California, a large integrated healthcare system between 2010 and 2020. Age-adjusted rates of combined fatal and nonfatal firearm injury per 100,000 members were calculated by year, with the 2010 US census as the reference population. Trends were evaluated using Poisson or negative binomial regression.ResultsThere was an increasing trend in overall firearm injuries between 2010 and 2020 among adults in this large integrated healthcare system (p < .0001), primarily driven by non-self-inflicted firearm injuries (p < .0001). Self-inflicted injuries decreased during this time (p = .01). Injuries among youth showed no significant change.ConclusionThere was an increasing trend in firearm injuries between 2010 and 2020 among adults in this large integrated healthcare system, primarily driven by non-self-inflicted firearm injuries; however, self-inflicted injuries decreased during this time. Injuries among youth showed no significant change.

Abstract 16693: Trends in the Use of Oral Anticoagulation for Nonvalvular Atrial Fibrillation Vary by Specialty: Real World Results From a Large Integrated Health Care System 2011-2022

Introduction: Direct Oral anticoagulants (DOACs) have demonstrated superiority or non inferiority to vitamin K antagonists (VKAs) for stroke/embolic prevention in patients with nonvalvular atrial fibrillation (NVAF). However, clinical inertia may delay widespread adoption of newer therapies in medicine. Hypothesis: We evaluated whether uptake of DOACs into clinical medicine varied by physician specialty. Methods: We performed a retrospective review of patients with NVAF within the Geisinger Health System from 2011-2022. Longitudinal yearly trends for new/initial prescriptions (Rx) of DOAC vs VKA by cardiology or primary care (PCP) specialties Results: We identified patients with NVAF initiated on an oral anticoagulant from the years 2011 through 2023. Figures 1 and 2 demonstrate a statistically significant linear yearly increase in new Rx for DOACs for both cardiology and PCP specialties.By 2022, DOACs currently account for 96% of all new oral anticoagulant Rx by cardiologists and 90% of Rx by PCP. New Rx for DOAC by PCP lagged cardiology at all time points. Cardiology initial Rx &gt;50% DOAC by 2015 while PCP did not achieve this mark until 2018. Conclusions: In this real-world setting, we observed a significant increase in DOAC use as the initial prescription for patients with NVAF by both cardiology and primary care specialties.Significant difference in the adoption of DOAC was noted between cardiology and primary care specialties.Primary care use of DOAC as first Rx for NVAF increased over time but lagged cardiology by 2-3 years.

Evaluation of the international Ki67 working group cut point recommendations for early breast cancer: comparison with 21-gene assay results in a large integrated health care system

PurposeThe International Ki67 Working Group (IKWG) has developed training for immunohistochemistry (IHC) scoring reproducibility and recommends cut points of ≤ 5% and ≥ 30% for prognosis in ER+, HER2−, stage I/II breast cancer. We examined scoring reproducibility following IKWG training and evaluated these cut points for selecting patients for further testing with the 21-gene Recurrence Score (RS) assay.MethodsWe included 307 women aged 50+ years with node-negative, ER+PR+HER2− breast cancer and with available RS results. Slides from the diagnostic biopsy were stained for Ki67 and scored using digital image analysis (IA). Two IHC pathologists underwent IKWG training and visually scored slides, blinded to each other and IA readings. Interobserver reproducibility was examined using intraclass correlation (ICC) and Kappa statistics.ResultsDepending on reader, 8.8–16.0% of our cohort had Ki67 ≤ 5% and 11.4–22.5% had scores ≥ 30%. The ICC for Ki67 scores by the two pathologists was 0.82 (95% CI 0.78–0.85); it was 0.79 (95% CI 0.74–0.83) for pathologist 1 and IA and 0.76 (95% CI 0.71–0.80) for pathologist 2 and IA. For Ki67 scores ≤ 5%, the percentages with RS < 26 were 92.6%, 91.8%, and 90.9% for pathologist 1, pathologist 2, and IA, respectively. For Ki67 scores ≥ 30%, the percentages with RS ≥ 26 were 41.5%, 51.4%, and 27.5%, respectively.ConclusionThe IKWG’s Ki67 training resulted in moderate to strong reproducibility across readers but cut points had only moderate overlap with RS cut points, especially for Ki67 ≥ 30% and RS ≥ 26; thus, their clinical utility for a 21-gene assay testing pathway remains unclear.