Improving Equity in Deep Learning Medical Applications with the Gerchberg-Saxton Algorithm.

Deep learning (DL) has gained prominence in healthcare for its ability to facilitate early diagnosis, treatment identification with associated prognosis, and varying patient outcome predictions. However, because of highly variable medical practices and unsystematic data collection approaches, DL can unfortunately exacerbate biases and distort estimates. For example, the presence of sampling bias poses a significant challenge to the efficacy and generalizability of any statistical model. Even with DL approaches, selection bias can lead to inconsistent, suboptimal, or inaccurate model results, especially for underrepresented populations. Therefore, without addressing bias, wider implementation of DL approaches can potentially cause unintended harm. In this paper, we studied a novel method for bias reduction that leverages the frequency domain transformation via the Gerchberg-Saxton and corresponding impact on the outcome from a racio-ethnic bias perspective.

T-DM1 after Pertuzumab plus Trastuzumab: Treatment Sequence-Induced Selection Bias in HER2-Positive Metastatic Breast Cancer.

Real-world studies have suggested decreased trastuzumab emtansine (T-DM1) effectiveness in patients with metastatic breast cancer (mBC) who received prior trastuzumab plus pertuzumab (H + P). However, these studies may have been biased toward pertuzumab-experienced patients with more aggressive disease. Using an electronic health record-derived database, patients diagnosed with mBC on/after 1 January 2011 who initiated T-DM1 in any treatment line (primary cohort) or who initiated second-line T-DM1 following first-line H ± P (secondary cohort) from 22 February 2013 to 31 December 2019 were included. The primary outcome was time from index date to next treatment or death (TTNT). In the primary cohort (n = 757), the percentage of patients with prior P increased from 37% to 73% across the study period, while population characteristics and treatment effectiveness measures were generally stable. Among P-experienced patients from the secondary cohort (n = 246), median time from mBC diagnosis to T-DM1 initiation increased from 10 to 14 months (2013-2019), and median TTNT increased from 4.4 to 10.2 months (2013-2018). Over time, prior H + P prevalence significantly increased with no observable impact on T-DM1 effectiveness. Drug approval timing should be considered when assessing treatment effectiveness within a sequence.

An overview of real-world data sources for oncology and considerations for research.

Generating evidence on the use, effectiveness, and safety of new cancer therapies is a priority for researchers, health care providers, payers, and regulators given the rapid pace of change in cancer diagnosis and treatments. The use of real-world data (RWD) is integral to understanding the utilization patterns and outcomes of these new treatments among patients with cancer who are treated in clinical practice and community settings. An initial step in the use of RWD is careful study design to assess the suitability of an RWD source. This pivotal process can be guided by using a conceptual model that encourages predesign conceptualization. The primary types of RWD included are electronic health records, administrative claims data, cancer registries, and specialty data providers and networks. Careful consideration of each data type is necessary because they are collected for a specific purpose, capturing a set of data elements within a certain population for that purpose, and they vary by population coverage and longitudinality. In this review, the authors provide a high-level assessment of the strengths and limitations of each data category to inform data source selection appropriate to the study question. Overall, the development and accessibility of RWD sources for cancer research are rapidly increasing, and the use of these data requires careful consideration of composition and utility to assess important questions in understanding the use and effectiveness of new therapies.

Adoption and effectiveness of de-escalated radiation and endocrine therapy strategies for older women with low-risk breast cancer.

PURPOSE: Recent clinical trials support de-escalation of adjuvant radiation therapy following lumpectomy in some older women with low-risk HR+ breast cancers planning to take endocrine therapy. The adoption of these findings into clinical practice, and the effectiveness of de-escalated therapy in real-world populations, remain under investigation. MATERIALS AND METHODS: We evaluated use of adjuvant radiation therapy and/or endocrine therapy among older women with T1-2 node-negative, HR+ breast cancer in the United States between 2007 and 2011. The study included patients from the Surveillance, Epidemiology and End Results-Medicare linked database and the North Carolina Cancer Information and Population Health Resource database. RESULTS: Radiation therapy was received by 65.5% of patients, with no decrease over time. Older women and those with T2 (compared to T1) tumors were less likely to receive radiation therapy. In propensity-adjusted analyses, both radiation therapy alone (HR 0.75, 95% CI 0.67-0.84) and radiation + endocrine therapy (HR 0.62, 95% CI 0.54-0.69) were associated with significantly lower recurrence risk compared to endocrine therapy alone. Non-adherence to endocrine therapy was common (37%) and similar across groups. With a median follow-up of 48 months (range 13-84), we were not able to detect an association of non-adherence with recurrence risk in endocrine therapy-containing treatment arms. CONCLUSION: Most older women with stage I HR+ breast cancers continue to receive radiation, at higher rates than patients with node-negative stage II tumors. These findings suggest that while multiple evidence-based treatment options exist in these patients, improvements are needed to ensure that radiation therapy is applied equitably and rationally.

Data linkage in pharmacoepidemiology: A call for rigorous evaluation and reporting.

PURPOSE: The purpose of this paper is to provide guidance on the evaluation of data linkage quality through the development of a checklist for reporting key elements of the linkage process. METHODS: Responding to a call for manuscripts from the International Society for Pharmacoepidemiology (ISPE), a working group including international representation from the academic, industry, and contract research, and regulatory sectors was formed to develop a checklist for evaluation of data linkage performance and reporting data linkage specifically for pharmacoepidemiologic research. This checklist expands on the reporting of studies conducted using observational routinely collected health data specific to pharmacoepidemiology (RECORD-PE) guidelines. RESULTS: A key aspect of data linkage evaluation for pharmacoepidemiology is to articulate how a linkage process was performed and its accuracy in terms of validation and verification of the resulting linked data. This study generates a checklist, which covers domains including data sources, linkage variables, linkage methods, linkage results, and linkage evaluation. For each domain, specific recommendations provide a clear and transparent assessment of the linkage process. CONCLUSIONS: Linking data sources can help to enrich analytic databases to more accurately define study populations, enable adjustment for confounding, and improve the capture of health outcomes. Clear and transparent reporting of data linkage processes will help to increase confidence in the evidence generated from these data by allowing researchers and end users to critically assess the potential for bias owing to the data linkage process.

Development and validation of a 5-year mortality prediction model using regularized regression and Medicare data.

PURPOSE: De-implementation of low-value services among patients with limited life expectancy is challenging. Robust mortality prediction models using routinely collected health care data can enhance health care stakeholders' ability to identify populations with limited life expectancy. We developed and validated a claims-based prediction model for 5-year mortality using regularized regression methods. METHODS: Medicare beneficiaries age 66 or older with an office visit and at least 12 months of pre-visit continuous Medicare A/B enrollment were identified in 2008. Five-year mortality was assessed through 2013. Secondary outcomes included 30-, 90-, and 180-day and 1-year mortality. Claims-based predictors, including comorbidities and indicators of disability, frailty, and functional impairment, were selected using regularized logistic regression, applying the least absolute shrinkage and selection operator (LASSO) in a random 80% training sample. Model performance was assessed and compared with the Gagne comorbidity score in the 20% validation sample. RESULTS: Overall, 183 204 (24%) individuals died. In addition to demographics, 161 indicators of comorbidity and function were included in the final model. In the validation sample, the c-statistic was 0.825 (0.823-0.828). Median-predicted probability of 5-year mortality was 14%; almost 4% of the cohort had a predicted probability greater than 80%. Compared with the Gagne score, the LASSO model led to improved 5-year mortality classification (net reclassification index = 9.9%; integrated discrimination index = 5.2%). CONCLUSIONS: Our claims-based model predicting 5-year mortality showed excellent discrimination and calibration, similar to the Gagne score model, but resulted in improved mortality classification. Regularized regression is a feasible approach for developing prediction tools that could enhance health care research and evaluation of care quality.

Examining determinants of geographic variation in colorectal cancer mortality in North Carolina: A spatial analysis approach.

PURPOSE: A recent study using national data from 2000 to 2009 identified colorectal cancer (CRC) mortality "hotspots" in 11 counties of North Carolina (NC). In this study, we used more recent, state-specific data to investigate the county-level determinants of geographic variation in NC through a geospatial analytic approach. METHOD: Using NC CRC mortality data from 2003 to 2013, we first conducted clustering analysis to confirm spatial dependence. Spatial economic models were then used to incorporate spatial structure to estimate the association between determinants and CRC mortality. We included county-level data on socio-demographic characteristics, access and quality of healthcare, behavioral risk factors (CRC screening, obesity, and cigarette smoking), and urbanicity. Due to correlation among screening, obesity and quality of healthcare, we combined these factors to form a cumulative risk group variable in the analysis. RESULTS: We confirmed the existence of spatial dependence and identified clusters of elevated CRC mortality rates in NC counties. Using a spatial lag model, we found significant interaction effect between CRC risk groups and socioeconomic deprivation. Higher CRC mortality rates were also associated with rural counties with large towns compared to urban counties. CONCLUSION: Our findings depicted a spatial diffusion process of CRC mortality rates across NC counties, demonstrated intertwined effects between SES deprivation and behavioral risks in shaping CRC mortality at area-level, and identified counties with high CRC mortality that were also deprived in multiple factors. These results suggest interventions to reduce geographic variation in CRC mortality should develop multifaceted strategies and work through shared resources in neighboring areas.

The relationship of travel distance with cystectomy access and outcomes.

PURPOSE: Regionalization of care and travel distance may result in unintended consequences for complex surgery such as cystectomy. Our objective was to evaluate effect of differential distance on cystectomy receipt among patients with muscle-invasive bladder cancer (MIBC) and investigate the association between travel distance and cystectomy outcomes such as readmission. METHODS: Using a linked data resource combining the NC Central Cancer Registry with claims data from Medicare, Medicaid, and private insurance plans, we included 736 patients with MIBC and 1,082 who underwent cystectomy. To evaluate access, differential distance was calculated as the difference between the nearest urologist and nearest cystectomy provider. To assess outcomes, logistic regression was used to evaluate rehospitalization and major complications, and Cox proportional hazards model for survival analysis. RESULTS: To evaluate access and outcomes, 736 patients with MIBC and 1,082 patients undergoing cystectomy were evaluated, respectively. Overall, 29% (211 of 736) with MIBC underwent cystectomy. Differential distance was not a predictor of cystectomy receipt (odds ratio = 1.0; 95% CI: 1.00, 1.01). Among patients undergoing cystectomy, travel distance from cystectomy provider was not a significant predictor of 30- or 31 to 90day readmissions (odds ratio = 1.0; 95% CI: 1.00, 1.00) although patients who lived further from their cystectomy provider were more likely to be readmitted to a nonindex hospital (P<0.001) when controlling for other factors. Although travel distance did not have a significant effect on overall survival, patients readmitted between 31 to 90days had worse overall survival (P<0.0001). CONCLUSIONS: The additional distance needed to reach a cystectomy provider did not predict receipt of surgery for MIBC. Furthermore, travel distance from cystectomy provider was not a significant predictor for subsequent readmission after cystectomy and did not affect overall survival.

Influence of provider factors and race on uptake of breast cancer gene expression profiling.

BACKGROUND: Gene expression profiling (GEP) has been rapidly adopted for early breast cancer and can aid in chemotherapy decision making. Study results regarding racial disparities in testing are conflicting, and may reflect different care settings. To the authors' knowledge, data regarding the influence of provider factors on testing are scarce. METHODS: The authors used a statewide, multipayer, insurance claims database linked to cancer registry records to examine the impact of race and provider characteristics on GEP uptake in a cohort of patients newly diagnosed with breast cancer between 2005 and 2012. Incidence proportion models were used to examine the adjusted likelihood of testing. Models were stratified by lymph node status (N0 vs N1). RESULTS: Among 11,958 eligible patients, 23% of black and 26% of non-Hispanic white patients received GEP. Among patients with N0 disease, black individuals were 16% less likely to receive testing after adjustment for clinical factors and the provider's specialty and volume of patients with breast cancer (95% confidence interval, 0.77-0.93). Adjustment for provider characteristics did not attenuate the effect of race on testing. Patients of middle-volume providers were more likely to be tested compared with those with either high-volume or low-volume providers, whereas patients seeing a medical oncologist were more likely to be tested compared with those whose only providers were from surgical specialties. CONCLUSIONS: Provider volume and specialty were found to be significant predictors of GEP use, but did not explain racial disparities in testing. Further research concerning the key contributors to lagging test use among black women is needed to optimize the equitable use of GEPs and support personalized treatment decision making for all patients. Cancer 2018;124:1743-51. © 2018 American Cancer Society.

Association of Delays in Surgery for Melanoma With Insurance Type.

Importance: Timely receipt of treatment for cancer is an important aspect of health care quality. It is unknown how delays of surgery for melanoma vary by insurance type. Objective: To analyze factors associated with delays between diagnosis and surgery for melanoma in patients with Medicare, Medicaid, or private insurance. Design, Setting, and Participants: Retrospective cohort study of patients who received a diagnosis of melanoma between 2004 and 2011 in North Carolina using data from the North Carolina Cancer Registry linked to administrative claims from Medicare, Medicaid, and private insurance. Inclusion criteria were incident patients with a diagnosis of melanoma stage 0 to III and with continuous insurance enrollment from at least 1 month prior to the month of diagnosis to 12 months after diagnosis of melanoma. Main Outcomes and Measures: Surgical delay, defined as definitive surgical excision occurring more than 6 weeks after melanoma diagnosis. Generalized linear models with log link, Poisson distributions, and robust standard errors were used to estimate adjusted risk ratios (RRs) to model risk of delay in definitive surgery. Results: A total of 7629 patients were included (4210 [55%] female; mean [SD] age, 64 [15] years), 48% (n = 3631) Medicare, 48% (n = 3667) privately insured, and 4% (n = 331) Medicaid patients. Privately insured patients were least likely to experience a delay in definitive surgery, followed by Medicare and Medicaid patients (519 [14%], 609 [17%], and 79 [24%], respectively; P < .001). After demographic adjustment, the risk of surgical delay was significantly increased in patients with Medicaid compared with private insurance (RR, 1.36; 95% CI, 1.09-1.70). Delays were more likely in nonwhite patients (RR, 1.38; 95% CI, 1.02-1.87). Surgical delays were less likely if the physician performing the surgery (RR, 0.82; 95% CI, 0.72-0.93) or the diagnosing clinician (RR, 0.81; 95% CI, 0.71-0.93) was a dermatologist as compared with a nondermatologist. Conclusion and Relevance: Surgical treatment delays were common but were less prevalent in patients diagnosed or surgically treated by a dermatologist. Medicaid patients experienced the most surgical delays. A reduction in delays in melanoma surgery could be achieved through better access to specialty care and cross-disciplinary coordination.

Reducing Racial Disparities in Breast Cancer Care: The Role of 'Big Data'.

Advances in a wide array of scientific technologies have brought data of unprecedented volume and complexity into the oncology research space. These novel big data resources are applied across a variety of contexts-from health services research using data from insurance claims, cancer registries, and electronic health records, to deeper and broader genomic characterizations of disease. Several forms of big data show promise for improving our understanding of racial disparities in breast cancer, and for powering more intelligent and far-reaching interventions to close the racial gap in breast cancer survival. In this article we introduce several major types of big data used in breast cancer disparities research, highlight important findings to date, and discuss how big data may transform breast cancer disparities research in ways that lead to meaningful, lifesaving changes in breast cancer screening and treatment. We also discuss key challenges that may hinder progress in using big data for cancer disparities research and quality improvement.

Comparative Toxicity and Effectiveness of Trastuzumab-Based Chemotherapy Regimens in Older Women With Early-Stage Breast Cancer.

Purpose The combination of chemotherapy and trastuzumab is the standard of care for adjuvant treatment of human epidermal growth factor receptor 2-positive breast cancer. Two regimens have been widely adopted in the United States: doxorubicin, cyclophosphamide, paclitaxel, and trastuzumab (ACTH) and docetaxel, carboplatin, and trastuzumab (TCH). No head-to-head comparison of these regimens has been conducted in a clinical trial, and existing trial data have limited generalizability to older patients. Methods We used SEER-Medicare data from 2005 to 2013 to compare outcomes of ACTH versus TCH among patients age older than 65 years. Propensity score matching was used to balance cohort characteristics between treatment arms. Outcomes included toxicity-related hospitalization, survival, and trastuzumab completion. Data from 1,077 patients receiving ACTH or TCH were analyzed, and the propensity-matched subsample included 416 women. Results There was a significant shift toward TCH over time, with 88% of patients receiving ACTH in 2005 compared with 15% by 2011. Among propensity score-matched patients, we found no difference between regimens in health care use overall or for chemotherapy-related adverse events (ACTH, 34% v TCH, 36.5%; P = .46). Patients receiving TCH were significantly more likely to complete trastuzumab (89% v 77%; P = .001). There was no difference in 5-year breast cancer-specific survival (ACTH, 92% v TCH, 96%; hazard ratio, 2.08; 95% CI, 0.90 to 4.82) or overall survival. Conclusion Among a matched sample of older patients, ACTH compared with TCH was not associated with a higher rate of serious adverse events or hospitalizations, but it was associated with less completion of adjuvant trastuzumab. We did not detect a difference in 5-year survival outcomes for ACTH compared with TCH. In the context of limited evidence in older patients, selection between these two regimens on the basis of concerns about differential toxicity or efficacy may not be appropriate.

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.

Endocrine Therapy Initiation and Medical Oncologist Utilization Among Women Diagnosed with Ductal Carcinoma in Situ.

BACKGROUND: Though randomized clinical trials have demonstrated a reduction in second breast events with endocrine therapy among women with ductal carcinoma in situ (DCIS), use of these therapies remains highly variable. The purpose of this study was to evaluate patient and treatment-related factors associated with endocrine therapy initiation and medical oncology specialty utilization after DCIS. MATERIALS AND METHODS: We identified women with a DCIS diagnosis during 2006-2010 in the North Carolina Central Cancer Registry with linked public and private insurance claims in the University of North Carolina Integrated Cancer Information Surveillance System data resource. Multivariable generalized linear regression models were used to estimate risk ratios (RR) and 95% confidence intervals (CI) for endocrine therapy initiation in the year following DCIS diagnosis. RESULTS: Among 2,090 women with a DCIS diagnosis, 37% initiated endocrine therapy. Initiation was less common among women aged 75+ at diagnosis (RR = 0.79; 95% CI: 0.64-0.97 vs. age 45-54) and women treated with breast-conserving surgery (BCS) who did not receive radiation (RR = 0.63; 95% CI: 0.50-0.78 vs. BCS plus radiation). Consultation with a medical oncologist was strongly associated with endocrine therapy initiation (RR = 1.40; 95% CI: 1.23-1.61). Women who saw a medical oncologist more often had private insurance, higher census tract-level income, hormone receptor positive disease, and treatment with BCS and radiation. CONCLUSION: Treatment strategies for DCIS remain controversial. Our data suggest that endocrine therapy is more often used in addition to standard therapies such as BCS plus radiation, rather than as an alternative to radiation. The Oncologist 2017;22:535-541 IMPLICATIONS FOR PRACTICE: Randomized trials have demonstrated a reduction in second breast cancer events with use of endocrine therapy for ductal carcinoma in situ (DCIS). However, notable variation exists in the uptake of these therapies among DCIS patients. In this study, factors associated with endocrine therapy initiation in the year following a DCIS diagnosis included consultation with a medical oncologist and treatment with breast-conserving surgery with radiation. Our findings help to explain the wide variation in endocrine therapy initiation and suggest the need for clear guidelines in the treatment of DCIS.

Association Between Choice of Radical Prostatectomy, External Beam Radiotherapy, Brachytherapy, or Active Surveillance and Patient-Reported Quality of Life Among Men With Localized Prostate Cancer.

Importance: Patients diagnosed with localized prostate cancer have to decide among treatment strategies that may differ in their likelihood of adverse effects. Objective: To compare quality of life (QOL) after radical prostatectomy, external beam radiotherapy, and brachytherapy vs active surveillance. Design, Setting, and Participants: Population-based prospective cohort of 1141 men (57% participation among eligible men) with newly diagnosed prostate cancer were enrolled from January 2011 through June 2013 in collaboration with the North Carolina Central Cancer Registry. Median time from diagnosis to enrollment was 5 weeks, and all men were enrolled with written informed consent prior to treatment. Final follow-up date for current analysis was September 9, 2015. Exposures: Treatment with radical prostatectomy, external beam radiotherapy, brachytherapy, or active surveillance. Main Outcomes and Measures: Quality of life using the validated instrument Prostate Cancer Symptom Indices was assessed at baseline (pretreatment) and 3, 12, and 24 months after treatment. The instrument contains 4 domains-sexual dysfunction, urinary obstruction and irritation, urinary incontinence, and bowel problems-each scored from 0 (no dysfunction) to 100 (maximum dysfunction). Propensity-weighted mean domain scores were compared between each treatment group vs active surveillance at each time point. Results: Of 1141 enrolled men, 314 pursued active surveillance (27.5%), 469 radical prostatectomy (41.1%), 249 external beam radiotherapy (21.8%), and 109 brachytherapy (9.6%). After propensity weighting, median age was 66 to 67 years across groups, and 77% to 80% of participants were white. Across groups, propensity-weighted mean baseline scores were 41.8 to 46.4 for sexual dysfunction, 20.8 to 22.8 for urinary obstruction and irritation, 9.7 to 10.5 for urinary incontinence, and 5.7 to 6.1 for bowel problems. Compared with active surveillance, mean sexual dysfunction scores worsened by 3 months for patients who received radical prostatectomy (36.2 [95% CI, 30.4-42.0]), external beam radiotherapy (13.9 [95% CI, 6.7-21.2]), and brachytherapy (17.1 [95% CI, 7.8-26.6]). Compared with active surveillance at 3 months, worsened urinary incontinence was associated with radical prostatectomy (33.6 [95% CI, 27.8-39.2]); acute worsening of urinary obstruction and irritation with external beam radiotherapy (11.7 [95% CI, 8.7-14.8]) and brachytherapy (20.5 [95% CI, 15.1-25.9]); and worsened bowel symptoms with external beam radiotherapy (4.9 [95% CI, 2.4-7.4]). By 24 months, mean scores between treatment groups vs active surveillance were not significantly different in most domains. Conclusions and Relevance: In this cohort of men with localized prostate cancer, each treatment strategy was associated with distinct patterns of adverse effects over 2 years. These findings can be used to promote treatment decisions that incorporate individual preferences.

Multilevel predictors of colorectal cancer testing modality among publicly and privately insured people turning 50.

Understanding multilevel predictors of colorectal cancer (CRC) screening test modality can help inform screening program design and implementation. We used North Carolina Medicare, Medicaid, and private, commercially available, health plan insurance claims data from 2003 to 2008 to ascertain CRC test modality among people who received CRC screening around their 50th birthday, when guidelines recommend that screening should commence for normal risk individuals. We ascertained receipt of colonoscopy, fecal occult blood test (FOBT) and fecal immunochemical test (FIT) from billing codes. Person-level and county-level contextual variables were included in multilevel random intercepts models to understand predictors of CRC test modality, stratified by insurance type. Of 12,570 publicly-insured persons turning 50 during the study period who received CRC testing, 57% received colonoscopy, whereas 43% received FOBT/FIT, with significant regional variation. In multivariable models, females with public insurance had lower odds of colonoscopy than males (odds ratio [OR] = 0.68; p < 0.05). Of 56,151 privately-insured persons turning 50 years old who received CRC testing, 42% received colonoscopy, whereas 58% received FOBT/FIT, with significant regional variation. In multivariable models, females with private insurance had lower odds of colonoscopy than males (OR = 0.43; p < 0.05). People living 10-15 miles away from endoscopy facilities also had lower odds of colonoscopy than those living within 5 miles (OR = 0.91; p < 0.05). Both colonoscopy and FOBT/FIT are widely used in North Carolina among insured persons newly age-eligible for screening. The high level of FOBT/FIT use among privately insured persons and women suggests that renewed emphasis on FOBT/FIT as a viable screening alternative to colonoscopy may be important.

Cost-Effectiveness Analysis of Four Simulated Colorectal Cancer Screening Interventions, North Carolina.

INTRODUCTION: Colorectal cancer (CRC) screening rates are suboptimal, particularly among the uninsured and the under-insured and among rural and African American populations. Little guidance is available for state-level decision makers to use to prioritize investment in evidence-based interventions to improve their population's health. The objective of this study was to demonstrate use of a simulation model that incorporates synthetic census data and claims-based statistical models to project screening behavior in North Carolina. METHODS: We used individual-based modeling to simulate and compare intervention costs and results under 4 evidence-based and stakeholder-informed intervention scenarios for a 10-year intervention window, from January 1, 2014, through December 31, 2023. We compared the proportion of people living in North Carolina who were aged 50 to 75 years at some point during the window (that is, age-eligible for screening) who were up to date with CRC screening recommendations across intervention scenarios, both overall and among groups with documented disparities in receipt of screening. RESULTS: We estimated that the costs of the 4 intervention scenarios considered would range from $1.6 million to $3.75 million. Our model showed that mailed reminders for Medicaid enrollees, mass media campaigns targeting African Americans, and colonoscopy vouchers for the uninsured reduced disparities in receipt of screening by 2023, but produced only small increases in overall screening rates (0.2-0.5 percentage-point increases in the percentage of age-eligible adults who were up to date with CRC screening recommendations). Increased screenings ranged from 41,709 additional life-years up to date with screening for the voucher intervention to 145,821 for the mass media intervention. Reminders mailed to Medicaid enrollees and the mass media campaign for African Americans were the most cost-effective interventions, with costs per additional life-year up to date with screening of $25 or less. The intervention expanding the number of endoscopy facilities cost more than the other 3 interventions and was less effective in increasing CRC screening. CONCLUSION: Cost-effective CRC screening interventions targeting observed disparities are available, but substantial investment (more than $3.75 million) and additional approaches beyond those considered here are required to realize greater increases population-wide.

Using Procedure Codes to Define Radiation Toxicity in Administrative Data: The Devil is in the Details.

BACKGROUND: Systematic coding systems are used to define clinically meaningful outcomes when leveraging administrative claims data for research. How and when these codes are applied within a research study can have implications for the study validity and their specificity can vary significantly depending on treatment received. SUBJECTS: Data are from the Surveillance, Epidemiology, and End Results-Medicare linked dataset. STUDY DESIGN: We use propensity score methods in a retrospective cohort of prostate cancer patients first examined in a recently published radiation oncology comparative effectiveness study. RESULTS: With the narrowly defined outcome definition, the toxicity event outcome rate ratio was 0.88 per 100 person-years (95% confidence interval, 0.71-1.08). With the broadly defined outcome, the rate ratio was comparable, with 0.89 per 100 person-years (95% confidence interval, 0.76-1.04), although individual event rates were doubled. Some evidence of surveillance bias was suggested by a higher rate of endoscopic procedures the first year of follow-up in patients who received proton therapy compared with those receiving intensity-modulated radiation treatment (11.15 vs. 8.90, respectively). CONCLUSIONS: This study demonstrates the risk of introducing bias through subjective application of procedure codes. Careful consideration is required when using procedure codes to define outcomes in administrative data.

Framework for Deploying a Virtualized Computing Environment for Collaborative and Secure Data Analytics.

INTRODUCTION: Large amounts of health data generated by a wide range of health care applications across a variety of systems have the potential to offer valuable insight into populations and health care systems, but robust and secure computing and analytic systems are required to leverage this information. FRAMEWORK: We discuss our experiences deploying a Secure Data Analysis Platform (SeDAP), and provide a framework to plan, build and deploy a virtual desktop infrastructure (VDI) to enable innovation, collaboration and operate within academic funding structures. It outlines 6 core components: Security, Ease of Access, Performance, Cost, Tools, and Training. CONCLUSION: A platform like SeDAP is not simply successful through technical excellence and performance. It's adoption is dependent on a collaborative environment where researchers and users plan and evaluate the requirements of all aspects.

Distance to a Plastic Surgeon and Type of Insurance Plan Are Independently Predictive of Postmastectomy Breast Reconstruction.

BACKGROUND: The psychosocial benefits of postmastectomy breast reconstruction are well established; however, health care barriers persist. The authors evaluated statewide patient population to further identify obstacles to reconstruction. METHODS: A linked data set combining the North Carolina Central Cancer Registry with administrative claims from Medicare, Medicaid, and private insurance plans identified women diagnosed with breast cancer from 2003 to 2006. For inclusion in the study, women must have had a mastectomy within 6 months of diagnosis and had continuous insurance enrollment at least 2 years postoperatively (n = 5381). Multivariable logistic regression was used to model odds of reconstruction. RESULTS: Approximately 20 percent underwent reconstruction (n = 1130). Distance to a plastic surgeon-10 to 20 miles (OR, 0.78) and greater than 20 miles (OR, 0.73; p < 0.05)-was significantly predictive of no reconstruction, independent of other well-known disparities, including age, race, rural location, and lower household income. Women with government-funded health care, such as Medicare (OR, 0.58) and Medicaid (OR, 0.24; p < 0.001), were also significantly less likely to undergo reconstruction. Consistent with previous study, advanced cancer stage and receipt of radiation therapy decreased the likelihood of reconstruction. Furthermore, when the authors compared immediate to delayed reconstruction, rural location, chemotherapy, and radiation therapy were significantly predictive of delay. CONCLUSIONS: This is the first population-based study to demonstrate distance to care and insurance plan as significant predictors of receipt of reconstruction. Additional research is needed to understand health care barriers and to determine whether distance to a plastic surgeon can be ameliorated by outreach programs. CLINICAL QUESTION/LEVEL OF EVIDENCE: Risk, III.