The Cancer Research Network: a platform for epidemiologic and health services research on cancer prevention, care, and outcomes in large, stable populations

Epidemiology of early vs late recurrence among women with early stage estrogen receptor-positive breast cancer in the Pathways Study

BACKGROUND

Relatively little is known about the differences in prognostic factors for early vs late recurrence among women with early stage estrogen receptor-positive breast cancer.

METHODS

We analyzed factors related to early (<5 years) vs late (≥5 years) recurrence in 2992 women with stage I-IIB estrogen receptor-positive breast cancer in the Pathways Study, a prospective cohort of women with breast cancer enrolled between 2006 and 2013, with ascertainment of recurrence and death through December 2021.

RESULTS

After a median follow-up of 13.3 years, 341 (13.8%) women had recurrences, including 181 (53.7%) with late recurrence. Higher stage and grade were associated with recurrence regardless of timing, whereas progesterone receptor negativity was associated with early but not late recurrence. Receipt of endocrine therapy was associated with reduced risk of overall recurrence, but the length of endocrine therapy was not statistically significant in multivariable models. Minoritized racial and ethnic groups, including Asian, Black, and Hispanic women, had higher risk of early but not late recurrence compared to non-Hispanic White women. The trend of higher risk of early recurrence among these groups remained after adjustment for clinical, demographic, and socioeconomic factors but was statistically significant only in Asian women.

CONCLUSIONS

Our study revealed potentially important distinctions for early vs late recurrence, including the associations with progesterone receptor negativity and self-identified race and ethnicity. Possible higher risk of early recurrence among Asian, Black, and Hispanic women provides novel evidence for the existence of disparities in cancer outcomes, even within the breast cancer subtype indicative of generally good prognosis.

Representativeness of Patients With Lung Cancer in an Integrated Health Care Delivery System

INTRODUCTION

Observational research is important for understanding the real-world benefits of advancements in lung cancer care. Integrated health care systems, such as Kaiser Permanente Northern California, have extensive electronic health records suitable for such research, but the generalizability of their populations is often questioned.

METHODS

Leveraging data from the California Cancer Registry, the authors compared distributions of demographic and clinical characteristics, in addition to neighborhood and environmental conditions, between patients diagnosed with lung cancer from 2015 through 2019 at Kaiser Permanente Northern California, National Cancer Institute-designated cancer centers (NCICCs), and all other non-NCICC hospitals within the same catchment area.

RESULTS

Of 20,178 included patients, 30% were from Kaiser Permanente Northern California, 8% from NCICCs, and 62% from other non-NCICC hospitals. Compared to NCICC patients, Kaiser Permanente Northern California patients were more similar to other non-NCICC patients on most characteristics. Compared to other non-NCICC patients, Kaiser Permanente Northern California patients were slightly older, more likely to be female, and less likely to be Hispanic or Asian/Pacific Islander and to reside in lower socioeconomic status (SES) neighborhoods. In contrast, NCICC patients were younger, less likely to be female or from non-Asian/Pacific Islander minoritized racial groups, and more likely to present with early-stage disease and adenocarcinoma and to reside in neighborhoods with higher SES and lower air pollution than Kaiser Permanente Northern California or other non-NCICC patients.

DISCUSSION

Patients from Kaiser Permanente Northern California, compared to NCICCs, are more broadly representative of the underlying patient population with lung cancer.

CONCLUSION

Research using electronic health record data from integrated health care systems can contribute generalizable real-world evidence to benchmark and improve lung cancer care.

Cancer Diagnosis During Pregnancy and Livebirth Outcomes in the Adolescent and Young Adult Horizon Study

Objective: To describe patterns of cancer treatment and live birth outcomes that followed a cancer diagnosis during pregnancy. Study Design: The Adolescent and Young Adult (AYA) Horizon Study is an observational study evaluating outcomes in survivors of the five most common types of cancer in this age group (15-39 years old). Of the 23,629 individuals identified diagnosed with breast, lymphoma, thyroid, melanoma, or gynecological cancer in North Carolina (2000-2015) and California (2004-2016), we identified 555 live births to individuals who experienced cancer diagnosis during pregnancy. Births to individuals diagnosed with cancer during pregnancy were matched ∼1:5 on maternal age and year of delivery to live births to individuals without a cancer diagnosis (N = 2,667). Multivariable Poisson regression was used to compare birth outcomes between pregnancies affected by a cancer diagnosis and unaffected matched pregnancies. Results: Cancer diagnosis during pregnancy was associated with an increased risk of preterm delivery (prevalence ratio [PR] 2.70; 95% confidence interval [CI] 2.24, 3.26); very preterm delivery (PR 1.74; 95% CI 1.12, 2.71); induction of labor (PR 1.48; 95% CI 1.27, 1.73); low birth weight (PR 1.97; 95% CI 1.55, 2.50); and cesarean delivery (PR 1.18; 95% CI 1.04, 1.34) but not associated with low Apgar score (PR 0.90; 95% CI 0.39, 2.06). In our sample, 41% of patients received chemotherapy, half of whom initiated chemotherapy during pregnancy, and 86% received surgery, 58% of whom had surgery during pregnancy. Of the 19% who received radiation, all received radiation treatment following pregnancy. Conclusion: We identified an increased risk of birth outcomes, including preterm and very preterm delivery, induction of labor, low birth weight, and cesarean delivery, to those experiencing a cancer diagnosis during pregnancy. This analysis contributes to the available evidence for those experiencing a cancer diagnosis during pregnancy.

Incidence and Survival for Patients Diagnosed With Breast, Colorectal, and Lung Cancer in an Integrated System

INTRODUCTION

Documenting trends in cancer incidence and survival is a national priority. This study estimated age- and sex-adjusted incidence and 5-year relative survival among patients with cancer diagnosed within Kaiser Permanente compared to Surveillance, Epidemiology, and End Results (SEER) estimates.

METHODS

The cohort included Kaiser Permanente health plan members diagnosed with breast (BC), colorectal (CRC), or lung cancer (LC) between January 1, 1999 and December 31, 2018. Incidence was computed as age-adjusted rates per 100,000 member-years. SEER*Stat was used to compute 5-year relative survival.

RESULTS

Kaiser Permanente BC incidence rates were persistently higher than SEER from 2004 (126.5 [95% confidence interval (CI) = 123.2-129.9] vs 122.6 [95% CI = 121.3-123.2]) through 2013 (132.06 [95% CI = 129.5-135.7] vs 126.7 [95% CI = 125.9-127.5]). Kaiser Permanente CRC and LC incidence rates were lower than SEER for all years except 2008, showing a spike in CRC incidence (51.5 [95% CI = 49.9-53.0] vs 46.1 [95% CI = 45.7-46.4]). Kaiser Permanente BC, CRC, and LC survival estimates for all stages were higher than SEER.

CONCLUSIONS

Incidence rates for all-stage and localized-stage BC were consistently higher for Kaiser Permanente than for SEER. CRC and LC rates were lower. Kaiser Permanente survival rates were consistently higher than for SEER. The strengths of these findings are associated with the ability to capture "gold-standard" cancer registry data on defined Kaiser Permanente populations. However, findings should be interpreted cautiously given differences in the underlying populations and secular and regional differences between Kaiser Permanente and SEER. The Kaiser Permanente population is younger and more racially diverse than SEER aggregate populations, and Kaiser Permanente members are insured with access to preventive care (eg, smoking cessation programs, cancer screening).

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.

Challenges and Opportunities of Epidemiological Studies to Reduce the Burden of Cancers in Young Adults

There are >1.9 million survivors of adolescent and young adult cancers (AYA, diagnosed at ages 15-39) living in the U.S. today. Epidemiologic studies to address the cancer burden in this group have been a relatively recent focus of the research community. In this article, we discuss approaches and data resources for cancer epidemiology and health services research in the AYA population. We consider research that uses data from cancer registries, vital records, healthcare utilization, and surveys, and the accompanying challenges and opportunities of each. To illustrate the strengths of each data source, we present example research questions or areas that are aligned with these data sources and salient to AYAs. Integrating the respective strengths of cancer registry, vital records, healthcare data, and survey-based studies sets the foundation for innovative and impactful research on AYA cancer treatment and survivorship to inform a comprehensive understanding of diverse AYA needs and experiences.

Risk management actions following genetic testing in the Cancer Health Assessments Reaching Many (CHARM) Study: A prospective cohort study

BACKGROUND

Genetic testing can identify cancer risk early, enabling prevention and early detection. We describe use of risk management interventions following genetic testing in the Cancer Health Assessment Reaching Many (CHARM) study. CHARM assessed risk and provided genetic testing to low income, low literacy, and other underserved populations that historically face barriers to accessing cancer genetic services.

METHODS

CHARM was implemented in Kaiser Permanente Northwest (KPNW) and Denver Health (DH) between 2018 and 2020. We identified post-testing screening (mammography, breast MRI, colonoscopy) and surgical (mastectomy, oophorectomy) procedures using electronic health records. We examined utilization in participants who did and did not receive actionable risk management recommendations from study genetic counselors following national guidelines.

RESULTS

CHARM participants were followed for an average of 15.4 months (range: 0.4-27.8 months) after results disclosure. Less than 2% (11/680) received actionable risk management recommendations (i.e., could be completed in the initial years following testing) based on their test result. Among those who received actionable recommendations, risk management utilization was moderate (54.5%, 6/11 completed any procedure) and varied by procedure (mammogram: 0/3; MRI: 2/4; colonoscopy: 4/5; mastectomy: 1/5; oophorectomy: 0/3). Cancer screening and surgery procedures were rare in participants without actionable recommendations.

CONCLUSION

Though the number of participants who received actionable risk management recommendations was small, our results suggest that implementing CHARM's risk assessment and testing model increased access to evidence-based genetic services and provided opportunities for patients to engage in recommended preventive care, without encouraging risk management overuse.

Risk-reducing surgery in unaffected individuals receiving cancer genetic testing in an integrated health care system

BACKGROUND

Germline genetic testing enables primary cancer prevention, including through prophylactic surgery. We examined risk-reducing surgeries in unaffected individuals tested for hereditary cancer susceptibly between 2010 and 2018 in the Kaiser Permanente Northwest health system.

METHODS

We used an internal genetic testing database to create a cohort of individuals who received tests including one or more high-penetrance hereditary cancer susceptibility gene. We then identified, after testing, bilateral mastectomy, bilateral salpingo-oophorectomy (BSO), and total hysterectomy procedures in electronic health record and claims data through 2019. We describe surgery utilization by genetic test results and National Comprehensive Cancer Network (NCCN) guidelines.

RESULTS

The cohort included 1020 individuals, 16% with pathogenic/likely pathogenic (P/LP) variants in one or more of the following genes: BRCA1, BRCA2, CHEK2, APC, MUTYH, ATM, MSH2, PALB2, BRIP1, MLH1, MSH6, EPCAM, FLCN, RAD51C, RAD51D, or TP53. Among individuals with P/LP variants making them candidates for mastectomy, BSO, or hysterectomy per NCCN guidelines, 34% (33/97), 24% (23/94), and 8% (1/12), respectively, underwent surgery during follow-up. Fifty-three percent (18/37) of hysterectomies were among APC, BRCA1, and BRCA2 P/LP variant heterozygotes, typically concurrent with BSO. Three individuals with variants of uncertain significance (only) and 22 with negative results had prophylactic surgery after genetic testing.

CONCLUSIONS

Uptake of risk-reducing surgery following usual care genetic testing appears to be lower than in studies that actively recruit high-risk patients and provide testing and follow-up care in specialized settings. Factors in addition to genetic test results and NCCN guidelines motivate prophylactic surgery use and deserve further study.

The Kaiser Permanente Research Bank Cancer Cohort: a collaborative resource to improve cancer care and survivorship

Background

The Kaiser Permanente Research Bank (KPRB) is collecting biospecimens and surveys linked to electronic health records (EHR) from approximately 400,000 adult KP members. Within the KPRB, we developed a Cancer Cohort to address issues related to cancer survival, and to understand how genetic, lifestyle and environmental factors impact cancer treatment, treatment sequelae, and prognosis. We describe the Cancer Cohort design and implementation, describe cohort characteristics after 5 years of enrollment, and discuss future directions.

Methods

Cancer cases are identified using rapid case ascertainment algorithms, linkage to regional or central tumor registries, and direct outreach to KP members with a history of cancer. Enrollment is primarily through email invitation. Participants complete a consent form, survey, and donate a blood or saliva sample. All cancer types are included.

Results

As of December 31, 2020, the cohort included 65,225 cases (56% female, 44% male) verified in tumor registries. The largest group was diagnosed between 60 and 69 years of age (31%) and are non-Hispanic White (83%); however, 10,076 (16%) were diagnosed at ages 18–49 years, 4208 (7%) are Hispanic, 3393 (5%) are Asian, and 2389 (4%) are Black. The median survival time is 14 years. Biospecimens are available on 98% of the cohort.

Conclusions

The KPRB Cancer Cohort is designed to improve our understanding of treatment efficacy and factors that contribute to long-term cancer survival. The cohort’s diversity - with respect to age, race/ethnicity and geographic location - will facilitate research on factors that contribute to cancer survival disparities.

Factors influencing genetic counseling and testing for hereditary breast and ovarian cancer syndrome in a large US health care system

Investigate whether disparities and other factors influence referral to genetic counseling and testing for hereditary breast and ovarian cancer syndrome (HBOC) in a large health care system. Examination of clinical, demographic, and socioeconomic factors from electronic health records associated with genetic referral and testing within 12 months after a new cancer diagnosed between August 1, 2013 and December 31, 2018. For patients meeting institutional criteria for HBOC testing, 60.6% were referred for genetic counseling, 88% of whom underwent germline testing; at least one pathogenic variant was found in 15.3%. Referral rates for patients with breast (69%) or ovarian cancer (65.7%) were much higher than for metastatic prostate (11.1%, p < 0.0001) or pancreatic cancer (22.3%, p < 0.0001); referral criteria were implemented more recently for the latter two cancers. Younger age, family history, and chemotherapy were associated with referral. Higher Elixhauser comorbidity score and prior cancer were associated with non-referral. No other factors were associated with genetic referral for all eligible cancers combined, although differences were seen in specific cancers. Race was a significant factor only for breast cancer, with fewer Asians than Whites referred. Health disparities in referral to genetics for HBOC cancers are mitigated in a comprehensive integrated health care system.

Plant-Based Dietary Patterns and Breast Cancer Recurrence and Survival in the Pathways Study

Plant-based diets are recommended for cancer survivors, but their relationship with breast cancer outcomes has not been examined. We evaluated whether long-term concordance with plant-based diets reduced the risk of recurrence and mortality among a prospective cohort of 3646 women diagnosed with breast cancer from 2005 to 2013. Participants completed food frequency questionnaires at diagnosis and 6-, 25-, and 72-month follow-up, from which we derived plant-based diet indices, including overall (PDI), healthful (hPDI), and unhealthful (uPDI). We observed 461 recurrences and 653 deaths over a median follow-up of 9.51 years. Using multivariable-adjusted Cox proportional hazards models, we estimated hazard ratios (HR) and 95% confidence intervals for breast cancer recurrence and all-cause, breast-cancer-specific, and non-breast-cancer mortality. Increased concordance with hPDI was associated with a reduced hazard of all-cause (HR 0.93, 95% CI: 0.83–1.05) and non-breast-cancer mortality (HR 0.83, 95% CI: 0.71–0.98), whereas increased concordance with uPDI was associated with increased hazards (HR 1.07, 95% CI: 0.96–1.2 and HR 1.20, 95% CI: 1.02–1.41, respectively). No associations with recurrence or breast-cancer-specific mortality were observed. In conclusion, healthful vs. unhealthful plant-based dietary patterns had differing associations with mortality. To enhance overall survival, dietary recommendations for breast cancer patients should emphasize healthful plant foods.

The Adolescent and Young Adult (AYA) Horizon Study: An AYA Cancer Survivorship Cohort

BACKGROUND

In the United States, >45,000 adolescent and young adult (AYA) women are diagnosed with cancer annually. Reproductive issues are critically important to AYA cancer survivors, but insufficient information is available to address their concerns. The AYA Horizon Study was initiated to contribute high-quality, contemporary evidence on reproductive outcomes for female cancer survivors in the United States.

METHODS

The study cohort includes women diagnosed with lymphoma, breast, melanoma, thyroid, or gynecologic cancer (the five most common cancers among women ages 15-39 years) at three study sites: the state of North Carolina and the Kaiser Permanente health systems in Northern and Southern California. Detailed information on cancer treatment, fertility procedures, and pregnancy (e.g., miscarriage, live birth) and birth (e.g., birth weight, gestational length) outcomes are leveraged from state cancer registries, health system databases and administrative insurance claims, national data on assisted reproductive technology procedures, vital records, and survey data.

RESULTS

We identified a cohort of 11,072 female AYA cancer survivors that includes >1,200 African American women, >1,400 Asian women, >1,600 Medicaid enrollees, and >2,500 Hispanic women using existing data sources. Active response to the survey component was low overall (N = 1,679), and notably lower among minority groups compared with non-Hispanic white women.

CONCLUSIONS

Passive data collection through linkage reduces participant burden and prevents systematic cohort attrition or potential selection biases that can occur with active participation requirements.

IMPACT

The AYA Horizon study will inform survivorship planning as fertility and parenthood gain increasing recognition as key aspects of high-quality cancer care.

Trends in Imaging for Suspected Pulmonary Embolism Across US Health Care Systems, 2004 to 2016

IMPORTANCE

In response to calls to reduce unnecessary diagnostic testing with computed tomographic pulmonary angiography (CTPA) for suspected pulmonary embolism (PE), there have been growing efforts to create and implement decision rules for PE testing. It is unclear if the use of advanced imaging tests for PE has diminished over time.

OBJECTIVE

To assess the use of advanced imaging tests, including chest computed tomography (CT) (ie, all chest CT except for CTPA), CTPA, and ventilation-perfusion (V/Q) scan, for PE from 2004 to 2016.

DESIGN, SETTING, AND PARTICIPANTS

Cohort study of adults by age group (18-64 years and ≥65 years) enrolled in 7 US integrated and mixed-model health care systems. Joinpoint regression analysis was used to identify years with statistically significant changes in imaging rates and to calculate average annual percentage change (growth) from 2004 to 2007, 2008 to 2011, and 2012 to 2016. Analyses were conducted between June 11, 2019, and March 18, 2020.

MAIN OUTCOMES AND MEASURES

Rates of chest CT, CTPA, and V/Q scan by year and age, as well as annual change in rates over time.

RESULTS

Overall, 3.6 to 4.8 million enrollees were included each year of the study, for a total of 52 343 517 person-years of follow-up data. Adults aged 18 to 64 years accounted for 42 223 712 person-years (80.7%) and those 65 years or older accounted for 10 119 805 person-years (19.3%). Female enrollees accounted for 27 712 571 person-years (52.9%). From 2004 and 2016, chest CT use increased by 66.3% (average annual growth, 4.4% per year), CTPA use increased by 450.0% (average annual growth, 16.3% per year), and V/Q scan use decreased by 47.1% (decreasing by 4.9% per year). The use of CTPA increased most rapidly from 2004 to 2006 (44.6% in those aged 18-64 years and 43.9% in those ≥65 years), with ongoing rapid growth from 2006 to 2010 (annual growth, 19.8% in those aged 18-64 years and 18.3% in those ≥65 years) and persistent but slower growth in the most recent years (annual growth, 4.3% in those aged 18-64 years and 3.0% in those ≥65 years from 2010 to 2016). The use of V/Q scanning decreased steadily since 2004.

CONCLUSIONS AND RELEVANCE

From 2004 to 2016, rates of chest CT and CTPA for suspected PE continued to increase among adults but at a slower pace in more contemporary years. Efforts to combat overuse have not been completely successful as reflected by ongoing growth, rather than decline, of chest CT use. Whether the observed imaging use was appropriate or was associated with improved patient outcomes is unknown.

Prostate-specific antigen testing after the US Preventive Services Task Force recommendation: a population-based analysis of electronic health data

PURPOSE

This study describes longitudinal trends in the use of prostate-specific antigen (PSA)-based testing in two geographically distinct healthcare systems following the 2011 US Preventive Services Task Force (USPSTF) recommendations against routine PSA screening.

METHODS

We analyzed population-based health claims data from 253,139 men aged 40-80 who were enrolled at two US healthcare systems. We assessed trends in the percentage of eligible men receiving ≥ 1 PSA test per year by time period (2000-2008, 2009-2011, 2012-2014), age (40-54, 55-69, 70-80), and race (white, black, other, unknown), and conducted a joinpoint regression analysis.

RESULTS

Men aged 55-69 and 70-80 years of all races had similar use of PSA testing between 2000 and 2011, ranging between 47 and 56% of eligible men by year, while only 22-26% of men aged 40-54 had a PSA test per year during this period. Overall, the percentage of men receiving at least one PSA test per year decreased by 26% between 2009-2011 and 2012-2014, with similar trends across race and age groups. PSA testing declined significantly after 2011 (annual percent change = - 11.28).

CONCLUSIONS

Following the 2011 USPSTF recommendations against routine PSA screening, declines in PSA testing were observed among men of all races and across all age groups in two large US healthcare systems.

Implementing Quality Colon Cancer Survivorship Care: A Practical Proposal for a Path Forward

There are currently close to 17 million survivors of cancer in the United States. This number is expected to grow as both an aging population and improved treatment increase the number of survivors. Consequently, the importance of quality survivorship care has been recognized, but implementing, measuring, and paying for this care in a highly fragmented health care system, across a broad spectrum of diseases, is difficult. Quality measurement tied to payment is one approach that has commonly been used to improve the quality of care in the US health care system, but the complexity of applying quality measurement metrics across the spectrum of cancer survivorship care had led to stalemate. In this article, we draw on prior work to develop a quality cancer survivorship framework and propose a practical path forward with a focus on the provision of colon cancer survivorship care within integrated health care delivery networks. With this narrowly defined approach, we hope that we can promote a practical solution that can be extended to other diseases and payment systems over time.

Long-Term Patterns of Oral Anticancer Agent Adoption, Duration, and Switching in Patients With CML

BACKGROUND

Oral tyrosine kinase inhibitors (TKIs) have been the standard of care for chronic myeloid leukemia (CML) since 2001. However, few studies have evaluated changes in the treatment landscape of CML over time. This study assessed the long-term treatment patterns of oral anticancer therapies among patients with CML.

METHODS

This retrospective cohort study included patients newly diagnosed with CML between January 1, 2000, and December 31, 2016, from 10 integrated healthcare systems. The proportion of patients treated with 5 FDA-approved oral TKI agents-bosutinib, dasatinib, imatinib, nilotinib, and ponatinib-in the 12 months after diagnosis were measured, overall and by year, between 2000 and 2017. We assessed the use of each oral agent through the fourth-line setting. Multivariable logistic regression estimated the odds of receiving any oral agent, adjusting for sociodemographic and clinical characteristics.

RESULTS

Among 853 patients with CML, 81% received an oral agent between 2000 and 2017. Use of non-oral therapies decreased from 100% in 2000 to 5% in 2005, coinciding with imatinib uptake from 65% in 2001 to 98% in 2005. Approximately 28% of patients switched to a second-line agent, 9% switched to a third-line agent, and 2% switched to a fourth-line agent. Adjusted analysis showed that age at diagnosis, year of diagnosis, and comorbidity burden were statistically significantly associated with odds of receiving an oral agent.

CONCLUSIONS

A dramatic shift was seen in CML treatments away from traditional, nonoral chemotherapy toward use of novel oral TKIs between 2000 and 2017. As the costs of oral anticancer agents reach new highs, studies assessing the long-term health and financial outcomes among patients with CML are warranted.

Performance of Cancer Recurrence Algorithms After Coding Scheme Switch From International Classification of Diseases 9th Revision to International Classification of Diseases 10th Revision

PURPOSE

We previously developed and validated informatic algorithms that used International Classification of Diseases 9th revision (ICD9)-based diagnostic and procedure codes to detect the presence and timing of cancer recurrence (the RECUR Algorithms). In 2015, ICD10 replaced ICD9 as the worldwide coding standard. To understand the impact of this transition, we evaluated the performance of the RECUR Algorithms after incorporating ICD10 codes.

METHODS

Using publicly available translation tables along with clinician and other expertise, we updated the algorithms to include ICD10 codes as additional input variables. We evaluated the performance of the algorithms using gold standard recurrence measures associated with a contemporary cohort of patients with stage I to III breast, colorectal, and lung (excluding IIIB) cancer and derived performance measures, including the area under the receiver operating curve, average absolute prediction error, and correct classification rate. These values were compared with the performance measures derived from the validation of the original algorithms.

RESULTS

A total of 659 colorectal, 280 lung, and 2,053 breast cancer cases were identified. Area under the receiver operating curve derived from the updated algorithms was 89.0% (95% CI, 82.3% to 95.7%), 88.9% (95% CI, 79.3% to 98.2%), and 80.5% (95% CI, 72.8% to 88.2%) for the colorectal, lung, and breast cancer algorithms, respectively. Average absolute prediction errors for recurrence timing were 2.7 (SE, 11.3%), 2.4 (SE, 10.4%), and 5.6 months (SE, 21.8%), respectively, and timing estimates were within 6 months of actual recurrence for more than 80% of colorectal, more than 90% of lung, and more than 50% of breast cancer cases using the updated algorithm.

CONCLUSION

Performance measures derived from the updated and original algorithms had overlapping confidence intervals, suggesting that the ICD9 to ICD10 transition did not affect the RECUR Algorithm performance.

Hospitalization Risk During Chemotherapy for Advanced Cancer: Development and Validation of Risk Stratification Models Using Real-World Data

PURPOSE

Hospitalizations are a common occurrence during chemotherapy for advanced cancer. Validated risk stratification tools could facilitate proactive approaches for reducing hospitalizations by identifying at-risk patients.

PATIENTS AND METHODS

We assembled two retrospective cohorts of patients receiving chemotherapy for advanced nonhematologic cancer; cohorts were drawn from three integrated health plans of the Cancer Research Network. We used these cohorts to develop and validate logistic regression models estimating 30-day hospitalization risk after chemotherapy initiation. The development cohort included patients in two health plans from 2005 to 2013. The validation cohort included patients in a third health plan from 2007 to 2016. Candidate predictor variables were derived from clinical data in institutional data warehouses. Models were validated based on the C-statistic, positive predictive value, and negative predictive value. Positive predictive value and negative predictive value were calculated in reference to a prespecified risk threshold (hospitalization risk ≥ 18.0%).

RESULTS

There were 3,606 patients in the development cohort (median age, 63 years) and 634 evaluable patients in the validation cohort (median age, 64 years). Lung cancer was the most common diagnosis in both cohorts (26% and 31%, respectively). The selected risk stratification model included two variables: albumin and sodium. The model C-statistic in the validation cohort was 0.69 (95% CI, 0.62 to 0.75); 39% of patients were classified as high risk according to the prespecified threshold; 30-day hospitalization risk was 24.2% (95% CI, 19.9% to 32.0%) in the high-risk group and 8.7% (95% CI, 6.1% to 12.0%) in the low-risk group.

CONCLUSION

A model based on data elements routinely collected during cancer treatment can reliably identify patients at high risk for hospitalization after chemotherapy initiation. Additional research is necessary to determine whether this model can be deployed to prevent chemotherapy-related hospitalizations.

Adjuvant endocrine therapy for breast cancer patients: impact of a health system outreach program to improve adherence

PURPOSE

Reports suggest that up to 50% of women with hormone receptor-positive (HR+) breast cancer (BC) do not complete the recommended 5 years of adjuvant endocrine therapy (AET). We examined the impact of an outreach program at Kaiser Permanente Northern California (KPNC) on adherence and discontinuation of AET among patients who initiated AET.

METHODS

We assembled a retrospective cohort of all KPNC patients diagnosed with HR+, stage I-III BC initiating AET before (n = 4287) and after (n = 3580) implementation of the outreach program. We compared adherence proportions and discontinuation rates before and after program implementation, both crude and adjusted for age, race/ethnicity, education, income, and stage. We conducted a pooled analysis of data from six Cancer Research Network (CRN) sites that had not implemented programs for improving AET adherence, using identical methods and time periods, to assess possible secular trends.

RESULTS

In the pre-outreach period, estimated adherence in years 1, 2, and 3 following AET initiation was 75.2%, 71.0%, and 67.3%; following the outreach program, the estimates were 79.4%, 75.6%, and 72.2% (p-values < .0001 for pairwise comparisons). Results were comparable after adjusting for clinical and demographic factors. The estimated cumulative incidence of discontinuation was 0.22 (0.21-0.24) and 0.18 (0.17-0.19) at 3 years for pre- and post-outreach groups (p-value < .0001). We found no evidence of an increase in adherence between the study periods at the CRN sites with no AET adherence program.

CONCLUSION

Adherence and discontinuation after AET initiation improved modestly following implementation of the outreach program.

Trends in Use of Medical Imaging in US Health Care Systems and in Ontario, Canada, 2000-2016

Importance

Medical imaging increased rapidly from 2000 to 2006, but trends in recent years have not been analyzed.

Objective

To evaluate recent trends in medical imaging.

Design, Setting, and Participants

Retrospective cohort study of patterns of medical imaging between 2000 and 2016 among 16 million to 21 million patients enrolled annually in 7 US integrated and mixed-model insurance health care systems and for individuals receiving care in Ontario, Canada.

Exposures

Calendar year and country (United States vs Canada).

Main Outcomes and Measures

Use of computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and nuclear medicine imaging. Annual and relative imaging rates by imaging modality, country, and age (children [<18 years], adults [18-64 years], and older adults [≥65 years]).

Results

Overall, 135 774 532 imaging examinations were included; 5 439 874 (4%) in children, 89 635 312 (66%) in adults, and 40 699 346 (30%) in older adults. Among adults and older adults, imaging rates were significantly higher in 2016 vs 2000 for all imaging modalities other than nuclear medicine. For example, among older adults, CT imaging rates were 428 per 1000 person-years in 2016 vs 204 per 1000 in 2000 in US health care systems and 409 per 1000 vs 161 per 1000 in Ontario; for MRI, 139 per 1000 vs 62 per 1000 in the United States and 89 per 1000 vs 13 per 1000 in Ontario; and for ultrasound, 495 per 1000 vs 324 per 1000 in the United States and 580 per 1000 vs 332 per 1000 in Ontario. Annual growth in imaging rates among US adults and older adults slowed over time for CT (from an 11.6% annual percentage increase among adults and 9.5% among older adults in 2000-2006 to 3.7% among adults in 2013-2016 and 5.2% among older adults in 2014-2016) and for MRI (from 11.4% in 2000-2004 in adults and 11.3% in 2000-2005 in older adults to 1.3% in 2007-2016 in adults and 2.2% in 2005-2016 in older adults). Patterns in Ontario were similar. Among children, annual growth for CT stabilized or declined (United States: from 10.1% in 2000-2005 to 0.8% in 2013-2016; Ontario: from 3.3% in 2000-2006 to -5.3% in 2006-2016), but patterns for MRI were similar to adults. Changes in annual growth in ultrasound were smaller among adults and children in the United States and Ontario compared with CT and MRI. Nuclear medicine imaging declined in adults and children after 2006.

Conclusions and Relevance

From 2000 to 2016 in 7 US integrated and mixed-model health care systems and in Ontario, rates of CT and MRI use continued to increase among adults, but at a slower pace in more recent years. In children, imaging rates continued to increase except for CT, which stabilized or declined in more recent periods. Whether the observed imaging utilization was appropriate or was associated with improved patient outcomes is unknown.

Addressing Challenges in Converting Grant-Funded Infrastructures to Broadly Used Research Resources

The NCI invests heavily in research resources to serve the research community, including datasets, biospecimen banks, and networks of institutions in which clinical trials and other human subjects research are conducted. These resources often begin as grant-funded infrastructure initiated by scientists based on their own scientific interests, with a subsequent recognition of additional scientific uses. Although converting existing project-specific research activities into research resources may appear efficient in terms of time and financial investment, challenges can arise that undermine this efficiency and jeopardize future use. Here, we describe three challenges in the conversion process: (i) project-based infrastructure versus a research resource for a broader research community; (ii) complexity versus ease of use; and (iii) individual professional goals versus research resource priorities. We use our experience with the NCI-funded Cancer Research Network, particularly the Virtual Data Warehouse, to illustrate each challenge, concluding with strategies to mitigate each one. As studies grow in size and complexity, an ever-increasing volume of data, biospecimens, and human subjects research networks will be available for conversion to resources for scientific questions beyond those originally proposed. Addressing likely challenges thoughtfully can result in a more efficient conversion process and ultimately greater scientific impact.

Towards more Accessible Precision Medicine: Building a more Transferable Machine Learning Model to Support Prognostic Decisions for Micro- and Macrovascular Complications of Type 2 Diabetes Mellitus

Although machine learning models are increasingly being developed for clinical decision support for patients with type 2 diabetes, the adoption of these models into clinical practice remains limited. Currently, machine learning (ML) models are being constructed on local healthcare systems and are validated internally with no expectation that they would validate externally and thus, are rarely transferrable to a different healthcare system. In this work, we aim to demonstrate that (1) even a complex ML model built on a national cohort can be transferred to two local healthcare systems, (2) while a model constructed on a local healthcare system's cohort is difficult to transfer; (3) we examine the impact of training cohort size on the transferability; and (4) we discuss criteria for external validity. We built a model using our previously published Multi-Task Learning-based methodology on a national cohort extracted from OptumLabs® Data Warehouse and transferred the model to two local healthcare systems (i.e., University of Minnesota Medical Center and Mayo Clinic) for external evaluation. The model remained valid when applied to the local patient populations and performed as well as locally constructed models (concordance: .73-.92), demonstrating transferability. The performance of the locally constructed models reduced substantially when applied to each other's healthcare system (concordance: .62-.90). We believe that our modeling approach, in which a model is learned from a national cohort and is externally validated, produces a transferable model, allowing patients at smaller healthcare systems to benefit from precision medicine.

Predictors of Long-Term Survival among High-Grade Serous Ovarian Cancer Patients

BACKGROUND

Relatively little is known about factors associated with long-term survival (LTS) following a diagnosis of ovarian cancer.

METHODS

We conducted a retrospective study of high-grade serous ovarian cancer (HGSOC) to explore predictors of LTS (defined as ≥7 years of survival) using electronic medical record data from a network of integrated health care systems. Multivariable logistic regression with forward selection was used to compare characteristics of women who survived ≥7 years after diagnosis (n = 148) to those who died within 7 years of diagnosis (n = 494).

RESULTS

Our final model included study site, age, stage at diagnosis, CA-125, comorbidity score, receipt of chemotherapy, BMI, and four separate comorbid conditions: weight loss, depression, hypothyroidism, and liver disease. Of these, only younger age, lower stage, and depression were statistically significantly associated with LTS.

CONCLUSIONS

We did not identify any new characteristics associated with HGSOC survival.

IMPACT

Prognosis of ovarian cancer generally remains poor. Large, pooled studies of ovarian cancer are needed to identify characteristics that may improve survival.

Collaborating on Data, Science, and Infrastructure: The 20-Year Journey of the Cancer Research Network

The Cancer Research Network (CRN) is a consortium of 12 research groups, each affiliated with a nonprofit integrated health care delivery system, that was first funded in 1998. The overall goal of the CRN is to support and facilitate collaborative cancer research within its component delivery systems. This paper describes the CRN's 20-year experience and evolution. The network combined its members' scientific capabilities and data resources to create an infrastructure that has ultimately supported over 275 projects. Insights about the strengths and limitations of electronic health data for research, approaches to optimizing multidisciplinary collaboration, and the role of a health services research infrastructure to complement traditional clinical trials and large observational datasets are described, along with recommendations for other research consortia.

Cardiovascular medication use and risks of colon cancer recurrences and additional cancer events: a cohort study

Background

Cardiovascular medications may be associated with cancer development, but little is known about their association with cancer recurrence. Medications such as statins and antihypertensives may be commonly used among colon cancer survivors, who are, on average, diagnosed in their mid-60s. We described the associations between statins and antihypertensive medications and colon cancer recurrence in a large, population-based study.

Methods

We conducted a cohort study among adults with stage I-IIIA colon cancer diagnosed in 1995–2014 in two Kaiser Permanente regions, Colorado and Washington. Statin and antihypertensive use were obtained from electronic pharmacy dispensing data. People were classified as medication users on the date of their first dispensing after cohort entry, which started 90 days after completing cancer treatment, continuing through the earliest of death, health plan disenrollment, or chart abstraction. We collected outcome information from medical record abstraction and tumor registries on colon cancer recurrences and second primary cancers. Using Cox proportional hazards multivariable models, we estimated hazard ratios (HRs) with 95% confidence intervals (CIs) for colon cancer recurrences and any cancer event (recurrences and new primaries at any anatomic site) comparing medication users to non-users.

Results

Among 2039 people, 937 (46%) used statins and 1425 (70%) used antihypertensives at any point during a median of 4.9 years of follow-up; 460 people had any additional cancer event, including 152 with a colon cancer recurrence. Statin use was not associated with colon cancer recurrence (HR = 1.09, 95%CI = 0.65–1.85) or any cancer event (HR = 1.12, 95%CI = 0.85–1.47), nor was antihypertensive use associated with recurrence (HR = 0.73, 95%CI = 0.44–1.21) or any cancer event (HR = 0.93, 95%CI = 0.70–1.24).

Conclusions

Our results suggest no association between cardiovascular medication use and the risk of recurrence or any additional cancer, and may provide reassurance to colon cancer survivors.

A Queensland research support network in emergency healthcare

PURPOSE

The purpose of this paper is to describe the structure and impact of a Queensland Research Support Network (RSN) in emergency medicine (EM).

DESIGN/METHODOLOGY/APPROACH

This paper presents a descriptive summary of EM networks, network evaluations and the structure and development of the Emergency Medicine Foundation's (EMF) RSN in Queensland, including an observational pre- and post-study of research metrics.

FINDINGS

In two years, the RSN supported 33 Queensland emergency departments (EDs), of which 14 developed research strategies. There was an increase in research active clinicians, from 23 in 2015 to 181 in 2017. Collaborator engagement increased from 9 in 2015 to 276 in 2017 as did the number of research presentations, from 6 in 2015 to 61 in 2017. EMF experienced a growth in new researchers, with new investigators submitting approximately 60 per cent of grant applications in 2016 and 2017. EMF also received new applications from a further three HHS (taking EMF-funded research activity from 8 to 11 HHS).

RESEARCH LIMITATIONS/IMPLICATIONS

This paper describes changes in KPIs and research metrics, which the authors attribute to the establishment of the RSN. However, it is possible that attribution bias plays a role in the KPI improvements.

SOCIAL IMPLICATIONS

This network has actively boosted and expanded EM research capacity and capability in Queensland. It provides services, in the form of on-the-ground managers, to develop novice clinician-researchers, new projects and engage entire EDs. This model may be replicated nationwide but requires funding commitment.

ORIGINALITY/VALUE

The RSN improves front-line clinician research capacity and capability and increases research activity and collaborations with clear community outcomes. Collaborations were extended to community, primary health networks, non-government organisations, national and international researchers and academic institutions. Evaluating and measuring a network's benefits are difficult, but it is likely that evaluations will help networks obtain funding.

Medical Care Costs for Recurrent versus De Novo Stage IV Cancer by Age at Diagnosis

OBJECTIVE

To address the knowledge gap regarding medical care costs for advanced cancer patients, we compared costs for recurrent versus de novo stage IV breast, colorectal, and lung cancer patients.

DATA SOURCES/STUDY SETTING

Virtual Data Warehouse (VDW) information from three Kaiser Permanente regions: Colorado, Northwest, and Washington.

STUDY DESIGN

We identified patients aged ≥21 with de novo or recurrent breast (nde novo  = 352; nrecurrent  = 765), colorectal (nde novo  = 1,072; nrecurrent  = 542), and lung (nde novo  = 4,041; nrecurrent  = 340) cancers diagnosed 2000-2012. We estimated average total monthly and annual costs in the 12 months preceding, month of, and 12 months following the index de novo/recurrence date, stratified by age at diagnosis (<65, ≥65). Generalized linear repeated-measures models controlled for demographics and comorbidity.

PRINCIPAL FINDINGS

In the pre-index period, monthly costs were higher for recurrent than for de novo breast (<65: +$2,431; ≥65: +$1,360), colorectal (<65: +$3,219; ≥65: +$2,247), and lung cancer (<65: +$3,086; ≥65: +$2,260) patients. Conversely, during the index and post-index periods, costs were higher for de novo patients. Average total annual pre-index costs were five- to ninefold higher for recurrent versus de novo patients <65.

CONCLUSIONS

Cost differences by type of advanced cancer and by age suggest heterogeneous patterns of care that merit further investigation.

Risk of colon cancer recurrence in relation to diabetes

PURPOSE

To describe the association between diabetes and colon cancer recurrence.

METHODS

We conducted a cohort study at two integrated health care delivery systems in the United States. Using tumor registry data, we identified patients aged ≥ 18 years when diagnosed with stage I-IIIA adenocarcinomas of the colon during 1995-2014. Pre-existing diabetes was ascertained via diagnosis codes. Medical records were reviewed for eligibility and to abstract recurrence and covariate information. Recurrence was ascertained beginning 90 days after the end of colon cancer treatment (i.e., cohort entry). Recurrence of any cancer or a new primary cancer at any site was a secondary outcome. We used multivariable Cox proportional hazards models to estimate hazard ratios (HR) and 95% confidence intervals (CIs) for the associations between diabetes at cohort entry and study outcomes.

RESULTS

Among the 1,923 eligible patients, 393 (16.7%) had diabetes at cohort entry. Diabetes was not associated with recurrence (HR 0.87; 95% CI 0.56-1.33) or with any subsequent cancer (HR 1.09; 95% CI 0.85-1.40). When the definition of recurrence included second primary colorectal cancer, risk was non-significantly higher in patients with diabetes than without diabetes.

CONCLUSIONS

The risk of colon cancer recurrence appears to be similar in patients with and without diabetes at diagnosis.

IMPACT

Future studies should evaluate the association between diabetes and colorectal cancer outcomes, especially second primary colon cancers, in larger populations.

Monitoring Lung Cancer Screening Use and Outcomes at Four Cancer Research Network Sites

RATIONALE

Lung cancer screening registries can monitor screening outcomes and improve quality of care.

OBJECTIVES

To describe nascent lung cancer screening programs and share efficient data collection approaches for mandatory registry reporting in four integrated health care systems of the National Cancer Institute-funded Cancer Research Network.

METHODS

We documented the distinctive characteristics of lung cancer screening programs, and we provide examples of strategies to facilitate data collection and describe early challenges and possible solutions. In addition, we report preliminary data on use and outcomes of screening with low-dose computed tomography at each of the participating sites.

RESULTS

Programs varied in approaches to confirming patient eligibility, ordering screening low-dose computed tomographic scans, and coordinating follow-up care. Most data elements were collected from structured fields in electronic health records, but sites also made use of standardized order templates, local procedure codes, identifiable hashtags in radiology reports, and natural language processing algorithms. Common challenges included incomplete documentation of tobacco smoking history, difficulty distinguishing between scans performed for screening versus diagnosis or surveillance, and variable adherence with use of standardized templates. Adherence with eligibility criteria as well as the accuracy and completeness of data collection appeared to depend at least partly on availability of personnel and other resources to support the successful implementation of screening.

CONCLUSIONS

To maximize the effectiveness of lung cancer screening, minimize the burden of data collection, and facilitate research and quality improvement, clinical workflow and information technology should be purposefully designed to ensure that patients meet eligibility criteria and receive appropriate follow-up testing.

Data Linkage to Improve Geriatric Oncology Research: A Feasibility Study

Older adults (aged 65 years and older) diagnosed with cancer account for most cancer-related morbidity and mortality in the United States but are often underrepresented on clinical trials. Recent attention from a variety of professional, research, regulatory, and patient advocacy groups has centered on data linkage and data sharing as a means to capture patient information and outcomes outside of clinical trials to accelerate progress in the fight against cancer. The development of a more robust observational research data infrastructure would help to address gaps in the evidence base regarding optimal approaches to treating cancer among the growing and complex population of older adults. To demonstrate the feasibility of building such a resource, we linked information from a sample of older adults with cancer in North Carolina using three distinct, but complementary, data sources: (a) the Carolina Senior Registry, (b) the North Carolina Central Cancer Registry, and (c) North Carolina fee-for-service Medicare claims data. A description of the linkage process, metrics, and characteristics of the final cohort is reported. This study highlights the potential for data linkage to improve the characterization of health status among older adults with cancer and the possibility to conduct passive follow-up for outcomes of interest over time. Extensions of these linkage efforts in partnership with other institutions will enhance our ability to generate evidence that can inform the management of older adults with cancer.