Development and validation of an algorithm for classifying colonoscopy indication

Fecal Immunochemical Test Screening and Risk of Colorectal Cancer Death

Importance

The fecal immunochemical test (FIT) is widely used for colorectal cancer (CRC) screening, but evidence of its effectiveness is limited.

Objective

To evaluate whether FIT screening is associated with a lower risk of dying from CRC overall, according to cancer location, and within demographic groups.

Design, Setting, and Participants

This nested case-control study in a cohort of screening-eligible people was conducted in 2 large, integrated health systems of racially, ethnically, and socioeconomically diverse members with long-term programs of mailed FIT screening outreach. Eligible participants included people aged 52 to 85 years who died from colorectal adenocarcinoma between 2011 and 2017 (cases); cases were matched in a 1:8 ratio based on age, sex, health-plan membership duration, and geographic area to randomly selected persons who were alive and CRC-free on case's diagnosis date (controls). Data analysis was conducted from January 2002 to December 2017.

Exposures

Completing 1 or more FIT screenings in the 5-year period prior to the CRC diagnosis date among cases or the corresponding date among controls; in secondary analyses, 2- to 10-year intervals were evaluated.

Main Outcomes and Measures

The primary study outcome was CRC death overall and by tumor location. Secondary analyses were performed to assess CRC death by race and ethnicity.

Results

From a cohort of 2 127 128 people, a total of 10 711 participants (3529 aged 60-69 years [32.9%]; 5587 male [52.1%] and 5124 female [47.8%]; 1254 non-Hispanic Asian [11.7%]; 973 non-Hispanic Black [9.1%]; 1929 Hispanic or Latino [18.0%]; 6345 non-Hispanic White [59.2%]) was identified, including 1103 cases and 9608 controls. Among controls during the 10-year period prior to the reference date, 6101 (63.5%) completed 1 or more FITs with a cumulative 12.6% positivity rate (768 controls), of whom 610 (79.4%) had a colonoscopy within 1 year. During the 5-year period, 494 cases (44.8%) and 5345 controls (55.6%) completed 1 or more FITs. In regression analysis, completing 1 or more FIT screening was associated with a 33% lower risk of death from CRC (adjusted odds ratio [aOR], 0.67; 95% CI, 0.59-0.76) and 42% lower risk in the left colon and rectum (aOR, 0.58; 95% CI, 0.48-0.71). There was no association with right colon cancers (aOR, 0.83; 95% CI, 0.69-1.01) but the difference in the estimates between the right colon and left colon or rectum was statistically significant (P = .01). FIT screening was associated with lower CRC mortality risk among non-Hispanic Asian (aOR, 0.37; 95% CI, 0.23-0.59), non-Hispanic Black (aOR, 0.58; 95% CI, 0.39-0.85) and non-Hispanic White individuals (aOR, 0.70; 95% CI, 0.57-0.86) (P for homogeneity = .04 for homogeneity).

Conclusions and Relevance

In this nested case-control study, completing FIT was associated with a lower risk of overall death from CRC, particularly in the left colon, and the associations were observed across racial and ethnic groups. These findings support the use of FIT in population-based screening strategies.

Predicting Risk of Colorectal Cancer After Adenoma Removal in a Large Community-Based Setting

INTRODUCTION

Colonoscopy surveillance guidelines categorize individuals as high or low risk for future colorectal cancer (CRC) based primarily on their prior polyp characteristics, but this approach is imprecise, and consideration of other risk factors may improve postpolypectomy risk stratification.

METHODS

Among patients who underwent a baseline colonoscopy with removal of a conventional adenoma in 2004-2016, we compared the performance for postpolypectomy CRC risk prediction (through 2020) of a comprehensive model featuring patient age, diabetes diagnosis, and baseline colonoscopy indication and prior polyp findings (i.e., adenoma with advanced histology, polyp size ≥10 mm, and sessile serrated adenoma or traditional serrated adenoma) with a polyp model featuring only polyp findings. Models were developed using Cox regression. Performance was assessed using area under the receiver operating characteristic curve (AUC) and calibration by the Hosmer-Lemeshow goodness-of-fit test.

RESULTS

Among 95,001 patients randomly divided 70:30 into model development (n = 66,500) and internal validation cohorts (n = 28,501), 495 CRC were subsequently diagnosed; 354 in the development cohort and 141 in the validation cohort. Models demonstrated adequate calibration, and the comprehensive model demonstrated superior predictive performance to the polyp model in the development cohort (AUC 0.71, 95% confidence interval [CI] 0.68-0.74 vs AUC 0.61, 95% CI 0.58-0.64, respectively) and validation cohort (AUC 0.70, 95% CI 0.65-0.75 vs AUC 0.62, 95% CI 0.57-0.67, respectively).

DISCUSSION

A comprehensive CRC risk prediction model featuring patient age, diabetes diagnosis, and baseline colonoscopy indication and polyp findings was more accurate at predicting postpolypectomy CRC diagnosis than a model based on polyp findings alone.

Surveillance Colonoscopy Findings in Older Adults With a History of Colorectal Adenomas

Importance

Postpolypectomy surveillance is a common colonoscopy indication in older adults; however, guidelines provide little direction on when to stop surveillance in this population.

Objective

To estimate surveillance colonoscopy yields in older adults.

Design, Setting, and Participants

This population-based cross-sectional study included individuals 70 to 85 years of age who received surveillance colonoscopy at a large, community-based US health care system between January 1, 2017, and December 31, 2019; had an adenoma detected 12 or more months previously; and had at least 1 year of health plan enrollment before surveillance. Individuals were excluded due to prior colorectal cancer (CRC), hereditary CRC syndrome, inflammatory bowel disease, or prior colectomy or if the surveillance colonoscopy had an inadequate bowel preparation or was incomplete. Data were analyzed from September 1, 2022, to February 22, 2024.

Exposures

Age (70-74, 75-79, or 80-85 years) at surveillance colonoscopy and prior adenoma finding (ie, advanced adenoma vs nonadvanced adenoma).

Main Outcomes and Measures

The main outcomes were yields of CRC, advanced adenoma, and advanced neoplasia overall (all ages) by age group and by both age group and prior adenoma finding. Multivariable logistic regression was used to identify factors associated with advanced neoplasia detection at surveillance.

Results

Of 9740 surveillance colonoscopies among 9601 patients, 5895 (60.5%) were in men, and 5738 (58.9%), 3225 (33.1%), and 777 (8.0%) were performed in those aged 70-74, 75-79, and 80-85 years, respectively. Overall, CRC yields were found in 28 procedures (0.3%), advanced adenoma in 1141 (11.7%), and advanced neoplasia in 1169 (12.0%); yields did not differ significantly across age groups. Overall, CRC yields were higher for colonoscopies among patients with a prior advanced adenoma vs nonadvanced adenoma (12 of 2305 [0.5%] vs 16 of 7435 [0.2%]; P = .02), and the same was observed for advanced neoplasia (380 of 2305 [16.5%] vs 789 of 7435 [10.6%]; P < .001). Factors associated with advanced neoplasia at surveillance were prior advanced adenoma (adjusted odds ratio [AOR], 1.65; 95% CI, 1.44-1.88), body mass index of 30 or greater vs less than 25 (AOR, 1.21; 95% CI, 1.03-1.44), and having ever smoked tobacco (AOR, 1.14; 95% CI, 1.01-1.30). Asian or Pacific Islander race was inversely associated with advanced neoplasia (AOR, 0.81; 95% CI, 0.67-0.99).

Conclusions and Relevance

In this cross-sectional study of surveillance colonoscopy yield in older adults, CRC detection was rare regardless of prior adenoma finding, whereas the advanced neoplasia yield was 12.0% overall. Yields were higher among those with a prior advanced adenoma than among those with prior nonadvanced adenoma and did not increase significantly with age. These findings can help inform whether to continue surveillance colonoscopy in older adults.

Impact of a scalable training program on the quality of colonoscopy performance and risk of postcolonoscopy colorectal cancer

BACKGROUND AND AIMS

Endoscopist adenoma detection rates (ADRs) vary widely and are associated with patients' risk of postcolonoscopy colorectal cancers (PCCRCs). However, few scalable physician-directed interventions demonstrably both improve ADR and reduce PCCRC risk.

METHODS

Among patients undergoing colonoscopy, we evaluated the influence of a scalable online training on individual-level ADRs and PCCRC risk. The intervention was a 30-minute, interactive, online training, developed using behavior change theory, to address factors that potentially impede detection of adenomas. Analyses included interrupted time series analyses for pretraining versus posttraining individual-physician ADR changes (adjusted for temporal trends) and Cox regression for associations between ADR changes and patients' PCCRC risk.

RESULTS

Across 21 endoscopy centers and all 86 eligible endoscopists, ADRs increased immediately by an absolute 3.13% (95% confidence interval [CI], 1.31-4.94) in the 3-month quarter after training compared with .58% per quarter (95% CI, .40-.77) and 0.33% per quarter (95% CI, .16-.49) in the 3-year pretraining and posttraining periods, respectively. Posttraining ADR increases were higher among endoscopists with pretraining ADRs below the median. Among 146,786 posttraining colonoscopies (all indications), each 1% absolute increase in screening ADR posttraining was associated with a 4% decrease in their patients' PCCRC risk (hazard ratio, .96; 95% CI, .93-.99). An ADR increase of ≥10% versus <1% was associated with a 55% reduced risk of PCCRC (hazard ratio, .45; 95% CI, .24-.82).

CONCLUSIONS

A scalable, online behavior change training intervention focused on modifiable factors was associated with significant and sustained improvements in ADR, particularly among endoscopists with lower ADRs. These ADR changes were associated with substantial reductions in their patients' risk of PCCRC.

Association of Physician Adenoma Detection Rates With Postcolonoscopy Colorectal Cancer

Importance

Although colonoscopy is frequently performed in the United States, there is limited evidence to support threshold values for physician adenoma detection rate as a quality metric.

Objective

To evaluate the association between physician adenoma detection rate values and risks of postcolonoscopy colorectal cancer and related deaths.

Design, Setting, and Participants

Retrospective cohort study in 3 large integrated health care systems (Kaiser Permanente Northern California, Kaiser Permanente Southern California, and Kaiser Permanente Washington) with 43 endoscopy centers, 383 eligible physicians, and 735 396 patients aged 50 to 75 years who received a colonoscopy that did not detect cancer (negative colonoscopy) between January 2011 and June 2017, with patient follow-up through December 2017.

Exposures

The adenoma detection rate of each patient's physician based on screening examinations in the calendar year prior to the patient's negative colonoscopy. Adenoma detection rate was defined as a continuous variable in statistical analyses and was also dichotomized as at or above vs below the median for descriptive analyses.

Main Outcomes and Measures

The primary outcome (postcolonoscopy colorectal cancer) was tumor registry-verified colorectal adenocarcinoma diagnosed at least 6 months after any negative colonoscopy (all indications). The secondary outcomes included death from postcolonoscopy colorectal cancer.

Results

Among 735 396 patients who had 852 624 negative colonoscopies, 440 352 (51.6%) were performed on female patients, median patient age was 61.4 years (IQR, 55.5-67.2 years), median follow-up per patient was 3.25 years (IQR, 1.56-5.01 years), and there were 619 postcolonoscopy colorectal cancers and 36 related deaths during more than 2.4 million person-years of follow-up. The patients of physicians with higher adenoma detection rates had significantly lower risks for postcolonoscopy colorectal cancer (hazard ratio [HR], 0.97 per 1% absolute adenoma detection rate increase [95% CI, 0.96-0.98]) and death from postcolonoscopy colorectal cancer (HR, 0.95 per 1% absolute adenoma detection rate increase [95% CI, 0.92-0.99]) across a broad range of adenoma detection rate values, with no interaction by sex (P value for interaction = .18). Compared with adenoma detection rates below the median of 28.3%, detection rates at or above the median were significantly associated with a lower risk of postcolonoscopy colorectal cancer (1.79 vs 3.10 cases per 10 000 person-years; absolute difference in 7-year risk, -12.2 per 10 000 negative colonoscopies [95% CI, -10.3 to -13.4]; HR, 0.61 [95% CI, 0.52-0.73]) and related deaths (0.05 vs 0.22 cases per 10 000 person-years; absolute difference in 7-year risk, -1.2 per 10 000 negative colonoscopies [95%, CI, -0.80 to -1.69]; HR, 0.26 [95% CI, 0.11-0.65]).

Conclusions and Relevance

Within 3 large community-based settings, colonoscopies by physicians with higher adenoma detection rates were significantly associated with lower risks of postcolonoscopy colorectal cancer across a broad range of adenoma detection rate values. These findings may help inform recommended targets for colonoscopy quality measures.

Impact of the Affordable Care Act on Colorectal Cancer Incidence and Mortality

INTRODUCTION

The Patient Protection and Affordable Care Act eliminated cost sharing for preventive services, including colorectal cancer screening for individuals aged 50-75 years with private health insurance. This study examines the impact of the Affordable Care Act's removal of cost sharing for colorectal cancer screening on colorectal cancer incidence and mortality.

METHODS

Trends in colorectal cancer incidence and colorectal cancer‒related mortality were modeled among 2,113,283 Kaiser Permanente Northern California members aged ≥50 years between 2003 and 2016 using an interrupted time-series design. As a sensitivity analysis, a controlled analysis utilized a comparison group of members covered with pre‒Affordable Care Act zero cost sharing for colorectal cancer screening. Analyses were performed in 2019 and 2020.

RESULTS

The colorectal cancer incidence dropped by 17% around the time the Affordable Care Act was enacted (change in level incidence rate ratio; 95% CI=0.77, 0.90, 2-sided p-value <0.0001), followed by a 3% further decrease per year (95% CI=0.93, 1.00, p=0.05). A similar pattern was observed for colorectal cancer‒related mortality. The controlled results indicated that the elimination of cost sharing for screening due to the Affordable Care Act was associated with greater improvements in colorectal cancer outcomes among members previously covered by health plans with out-of-pocket costs for screening than among those with health plans with zero cost sharing for screening before the Affordable Care Act.

CONCLUSIONS

The elimination of cost sharing for colorectal cancer screening due to the Affordable Care Act was associated with a decrease in age-, race/ethnicity-, and sex-adjusted colorectal cancer incidence and colorectal cancer‒related mortality, implying that policies that remove barriers to screening, particularly financial burden from cost sharing, can result in improved colorectal cancer outcomes.

Risk stratification for colorectal cancer in individuals with subtypes of serrated polyps

OBJECTIVE

The longitudinal risk of colorectal cancer (CRC) associated with subtypes of serrated polyps (SPs) remains incompletely understood.

DESIGN

This community-based, case-control study included 317 178 Kaiser Permanente Northern California members who underwent their first colonoscopy during 2006-2016. Nested within this population, we identified 695 cases of CRC and 3475 CRC-free controls (matched 5:1 to cases for age, sex and year of colonoscopy). Two expert pathologists reviewed the tissue slides of all SPs identified on the first colonoscopy and reclassified them to sessile serrated lesions (SSLs), hyperplastic polyps (HPs) and traditional serrated adenomas. SPs with borderline characteristics of SSLs but insufficient to make a definitive diagnosis were categorised as unspecified SPs. The association with development of CRC was assessed using multivariable logistic regression.

RESULTS

Compared with individuals with no polyp, the adjusted ORs (aORs) for SSL alone or with synchronous adenoma were 2.9 (95% CI: 1.8 to 4.8) and 4.4 (95% CI: 2.7 to 7.2), respectively. The aORs for SSL with dysplasia, large proximal SSL,and small proximal SSL were 10.3 (95% CI: 2.1 to 50.3), 12.8 (95% CI: 3.5 to 46.9) and 1.9 (95% CI: 0.8 to 4.7), respectively. Proximal unspecified SP also conferred an increased risk (aOR: 5.8, 95% CI: 2.2 to 15.2). Women with SSL were associated with higher risk (aOR: 4.4; 95% CI: 2.3 to 8.2) than men (aOR: 1.7; 95% CI: 0.8 to 3.8).

CONCLUSION

Increased risk of CRC was observed in individuals with SSLs, particularly large proximal ones or with dysplasia, supporting close endoscopic surveillance. Proximal unspecified SPs were also associated with increased risk of CRC and should be managed as SSLs.

Early Screening of African Americans (45-50 Years Old) in a Fecal Immunochemical Test-Based Colorectal Cancer Screening Program

BACKGROUND & AIMS

Some guidelines recommend starting colorectal cancer (CRC) screening before age 50 years for African Americans, but there are few data on screening uptake and yield in this population.

METHODS

We performed a prospective study of fecal immunochemical test (FIT) screening among African American members of the Kaiser Permanente Northern California health plan. We compared data from African American members screened when they were 45-50 years old (early screening group) in 2018 with data from previously unscreened African American, white, Hispanic, and Asian/Pacific Islander health plan members who were 51-56 years old. Screening outreach was performed with mailed FIT kits. Logistic regression models, adjusted for sex, were used to evaluate differences among groups in screening uptake, colonoscopy follow-up of abnormal test results, and test yield.

RESULTS

Among 10,232 African Americans in the early screening group who were mailed a FIT, screening was completed by 33.1%. Among the 4% with positive test results, 85.3% completed a follow-up colonoscopy: 57.8% had any adenoma, 33.6% had an advanced adenoma (adenoma with advanced histology or polyp ≥10 mm), and 2.6% were diagnosed with CRC. African Americans in the early screening group were modestly more likely to have completed screening than previously unscreened African Americans, whites, and Hispanics 51-56 years old. The groups did not differ significantly in positive results from the FIT (range, 3.8%-4.6%) and more than 74% received a follow-up colonoscopy after a positive test result. The test yields for any adenoma (range, 56.7%-70.7%), advanced adenoma (range, 20.0%-33.6%), and CRC (range, 0%-7.1%) were similar.

CONCLUSIONS

Proportions of African Americans who participated in early (aged 45-50 years) FIT screening and test yield were comparable to those of previously unscreened African Americans, whites, Hispanics, and Asian/Pacific Islanders who were 51-56 years old.

Evaluating Screening Participation, Follow-up, and Outcomes for Breast, Cervical, and Colorectal Cancer in the PROSPR Consortium

BACKGROUND

Cancer screening is a complex process encompassing risk assessment, the initial screening examination, diagnostic evaluation, and treatment of cancer precursors or early cancers. Metrics that enable comparisons across different screening targets are needed. We present population-based screening metrics for breast, cervical, and colorectal cancers for nine sites participating in the Population-based Research Optimizing Screening through Personalized Regimens consortium.

METHODS

We describe how selected metrics map to a trans-organ conceptual model of the screening process. For each cancer type, we calculated calendar year 2013 metrics for the screen-eligible target population (breast: ages 40-74 years; cervical: ages 21-64 years; colorectal: ages 50-75 years). Metrics for screening participation, timely diagnostic evaluation, and diagnosed cancers in the screened and total populations are presented for the total eligible population and stratified by age group and cancer type.

RESULTS

The overall screening-eligible populations in 2013 were 305 568 participants for breast, 3 160 128 for cervical, and 2 363 922 for colorectal cancer screening. Being up-to-date for testing was common for all three cancer types: breast (63.5%), cervical (84.6%), and colorectal (77.5%). The percentage of abnormal screens ranged from 10.7% for breast, 4.4% for cervical, and 4.5% for colorectal cancer screening. Abnormal breast screens were followed up diagnostically in almost all (96.8%) cases, and cervical and colorectal were similar (76.2% and 76.3%, respectively). Cancer rates per 1000 screens were 5.66, 0.17, and 1.46 for breast, cervical, and colorectal cancer, respectively.

CONCLUSIONS

Comprehensive assessment of metrics by the Population-based Research Optimizing Screening through Personalized Regimens consortium enabled systematic identification of screening process steps in need of improvement. We encourage widespread use of common metrics to allow interventions to be tested across cancer types and health-care settings.

Increased Risk of Colorectal Cancer in Individuals With a History of Serrated Polyps

BACKGROUND AND AIMS

Serrated polyp (SPs) are precursors to 20% to 30% of cases of colorectal tumors, but patients' long-term risk after removal of SPs is poorly understood. We investigated the risk of colorectal cancer (CRC) in individuals with a history of SPs.

METHODS

We performed a retrospective cohort study of Kaiser Permanente Northern California members who underwent colonoscopy from 2006 through 2016. Study participants were categorized based on the size and location of SPs. We used Cox proportional hazards modeling to estimate the hazard ratio (HR) and 95% confidence interval (CI) for the association of CRC diagnosed more than 1 year after colonoscopy, with polyp type vs no polyp after adjustment for year of colonoscopy, age, sex, race/ethnicity, and smoking history.

RESULTS

The study included 233,393 individuals, of whom 445 developed incident CRC. At 10 years, the cumulative incidence rates of CRC for individuals with no polyp, proximal small SPs, proximal large SPs, and distal SPs were 4.7 (95% CI, 4.0-5.6), 14.8 (95% CI, 9.0-24.3), 30.2 (95% CI, 13.2-68.4), and 5.9 (95% CI, 3.6-9.5) per 1000 persons, respectively. In patients with SPs, risk of CRC was not increased until 3 years or more after the first colonoscopy (HR for small proximal SPs 2.6; 95% CI, 1.7-3.9 and HR for large proximal SPs 8.0; 95% CI, 3.6-16.1). The presence of synchronous adenomas increased the risk for CRC (HR for proximal SPs with synchronous adenomas 4.0; 95% CI, 3.0-5.5 and HR for distal SPs with synchronous adenomas 2.4; 95% CI, 1.7-3.4).

CONCLUSIONS

In a retrospective analysis of a large cohort of individuals examined by colonoscopy, we found that risk of incident CRC increased in individuals with proximal SPs (large SPs in particular) 3 years or more after the colonoscopy. These findings support guidelines that recommend surveillance colonoscopy for individuals with SPs.

A Transparent and Adaptable Method to Extract Colonoscopy and Pathology Data Using Natural Language Processing

Key variables recorded as text in colonoscopy and pathology reports have been extracted using natural language processing (NLP) tools that were not easily adaptable to new settings. We aimed to develop a reliable NLP tool with broad adaptability. During 1996-2016, Kaiser Permanente Northern California performed 401,566 colonoscopies with linked pathology. We randomly sampled 1000 linked reports into a Training Set and developed an NLP tool using SAS® PERL regular expressions. The NLP tool captured five colonoscopy and pathology variables: type, size, and location of polyps; extent of procedure; and quality of bowel preparation. We used a Validation Set (N = 3000) to confirm the variables' classifications using manual chart review as the reference. Performance of the NLP tool was assessed using the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and Cohen's κ. Cohen's κ ranged from 93 to 99%. The sensitivity and specificity ranged from 95 to 100% across all categories. For categories with prevalence exceeding 10%, the PPV ranged from 97% to 100% except for adequate quality of preparation (prevalence 92%), for which the PPV was 65%. For categories with prevalence below 10%, the PPVs ranged from 62% to 100%. NPVs ranged from 94% to 100% except for the "complete" extent of procedure, for which the NPV was 73%. Using information from a large community-based population, we developed a transparent and adaptable NLP tool for extracting five colonoscopy and pathology variables. The tool can be readily tested in other healthcare settings.

Long-term Risk of Colorectal Cancer and Related Death After Adenoma Removal in a Large, Community-based Population

BACKGROUND & AIMS

The long-term risks of colorectal cancer (CRC) and CRC-related death following adenoma removal are uncertain. Data are needed to inform evidence-based surveillance guidelines, which vary in follow-up recommendations for some polyp types. Using data from a large, community-based integrated health care setting, we examined the risks of CRC and related death by baseline colonoscopy adenoma findings.

METHODS

Participants at 21 medical centers underwent baseline colonoscopies from 2004 through 2010; findings were categorized as no-adenoma, low-risk adenoma, or high-risk adenoma. Participants were followed until the earliest of CRC diagnosis, death, health plan disenrollment, or December 31, 2017. Risks of CRC and related deaths among the high- and low-risk adenoma groups were compared with the no-adenoma group using Cox regression adjusting for confounders.

RESULTS

Among 186,046 patients, 64,422 met eligibility criteria (54.3% female; mean age, 61.6 ± 7.1 years; median follow-up time, 8.1 years from the baseline colonoscopy). Compared with the no-adenoma group (45,881 patients), the high-risk adenoma group (7563 patients) had a higher risk of CRC (hazard ratio [HR] 2.61; 95% confidence interval [CI] 1.87-3.63) and related death (HR 3.94; 95% CI 1.90-6.56), whereas the low-risk adenoma group (10,978 patients) did not have a significant increase in risk of CRC (HR 1.29; 95% CI 0.89-1.88) or related death (HR 0.65; 95% CI 0.19-2.18).

CONCLUSIONS

With up to 14 years of follow-up, high-risk adenomas were associated with an increased risk of CRC and related death, supporting early colonoscopy surveillance. Low-risk adenomas were not associated with a significantly increased risk of CRC or related deaths. These results can inform current surveillance guidelines for high- and low-risk adenomas.

Colonoscopy Indication Algorithm Performance Across Diverse Health Care Systems in the PROSPR Consortium

BACKGROUND

Despite the importance of characterizing colonoscopy indication for quality monitoring and cancer screening program evaluation, there is no standard approach to documenting colonoscopy indication in medical records.

METHODS

We applied two algorithms in three health care systems to assign colonoscopy indication to persons 50-89 years old who received a colonoscopy during 2010-2013. Both algorithms used standard procedure, diagnostic, and laboratory codes. One algorithm, the KPNC algorithm, used a hierarchical approach to classify exam indication into: diagnostic, surveillance, or screening; whereas the other, the SEARCH algorithm, used a logistic regression-based algorithm to provide the probability that colonoscopy was performed for screening. Gold standard assessment of indication was from medical records abstraction.

RESULTS

There were 1,796 colonoscopy exams included in analyses; age and racial/ethnic distributions of participants differed across health care systems. The KPNC algorithm's sensitivities and specificities for screening indication ranged from 0.78-0.82 and 0.78-0.91, respectively; sensitivities and specificities for diagnostic indication ranged from 0.78-0.89 and 0.74-0.82, respectively. The KPNC algorithm had poor sensitivities (ranging from 0.11-0.67) and high specificities for surveillance exams. The Area Under the Curve (AUC) of the SEARCH algorithm for screening indication ranged from 0.76-0.84 across health care systems. For screening indication, the KPNC algorithm obtained higher specificities than the SEARCH algorithm at the same sensitivity.

CONCLUSION

Despite standardized implementation of these indication algorithms across three health care systems, the capture of colonoscopy indication data was imperfect. Thus, we recommend that standard, systematic documentation of colonoscopy indication should be added to medical records to ensure efficient and accurate data capture.

Changes in screening colonoscopy following Medicare reimbursement and cost-sharing changes

OBJECTIVES

To compare existing algorithms for classifying screening vs diagnostic colonoscopies and to quantify the increase in screening colonoscopy rates when Medicare began reimbursement in 2001 and when the Affordable Care Act (ACA) eliminated cost-sharing.

DATA SOURCES

Twenty percent random sample of fee-for-service (FFS) Medicare claims, 2000-2012.

STUDY DESIGN

Using recent administrative codes as tarnished gold standards, we examined the sensitivity and specificity of five published algorithms for classifying colonoscopies and calculated annual screening colonoscopy rates. We estimated the change in rates after Medicare began reimbursement and used difference-in-differences analysis to estimate the effects of eliminating cost-sharing by comparing states with and without a mandate to cover screening colonoscopy prior to the ACA.

FINDINGS

Model-based algorithms have higher sensitivity (0.53-0.99) than expert-based algorithms (0.35-0.39), but lower specificity (0.43-0.65 vs 0.79-0.88). All algorithms detected increases in screening from both Medicare's reimbursement change (range: 24-93/10 000) and the 2011 cost-sharing change (range: 1.1-34/10 000). Difference-in-difference estimates of the ACA's effect varied from 51 to 155 tests per 10 000 depending on the algorithm.

CONCLUSIONS

Screening colonoscopy rates increased after eliminating cost-sharing in 2011, but the increase's size varied depending on the algorithm used to classify the indication. Improvements are needed in Medicare coding for screening.

Long-term Risk of Colorectal Cancer and Related Deaths After a Colonoscopy With Normal Findings

Importance

Guidelines recommend a 10-year rescreening interval after a colonoscopy with normal findings (negative colonoscopy results), but evidence supporting this recommendation is limited.

Objective

To examine the long-term risks of colorectal cancer and colorectal cancer deaths after a negative colonoscopy result, in comparison with individuals unscreened, in a large, community-based setting.

Design, Setting, and Participants

A retrospective cohort study was conducted in an integrated health care delivery organization serving more than 4 million members across Northern California. A total of 1 251 318 average-risk screening-eligible patients (age 50-75 years) between January 1, 1998, and December 31, 2015, were included. The study was concluded on December 31, 2016.

Exposures

Screening was examined as a time-varying exposure; all participants contributed person-time unscreened until they were either screened or censored. If the screening received was a negative colonoscopy result, the participants contributed person-time in the negative colonoscopy results group until they were censored.

Main Outcomes and Measures

Using Cox proportional hazards regression models, the hazard ratios (HRs) for colorectal cancer and related deaths were calculated according to time since negative colonoscopy result (or since cohort entry for those unscreened). Hazard ratios were adjusted for age, sex, race/ethnicity, Charlson comorbidity score, and body mass index.

Results

Of the 1 251 318 patients, 613 692 were men (49.0%); mean age was 55.6 (7.0) years. Compared with the unscreened participants, those with a negative colonoscopy result had a reduced risk of colorectal cancer and related deaths throughout the more than 12-year follow-up period, and although reductions in risk were attenuated with increasing years of follow-up, there was a 46% lower risk of colorectal cancer (hazard ratio, 0.54; 95% CI, 0.31-0.94) and 88% lower risk of related deaths (hazard ratio, 0.12; 95% CI, 0.02-0.82) at the current guideline-recommended 10-year rescreening interval.

Conclusions and Relevance

A negative colonoscopy result in average-risk patients was associated with a lower risk of colorectal cancer and related deaths for more than 12 years after examination, compared with unscreened patients. Our study findings may be able to inform guidelines for rescreening after a negative colonoscopy result and future studies to evaluate the costs and benefits of earlier vs later rescreening intervals.

Accurate Identification of Colonoscopy Quality and Polyp Findings Using Natural Language Processing

OBJECTIVES

The aim of this study was to test the ability of a commercially available natural language processing (NLP) tool to accurately extract examination quality-related and large polyp information from colonoscopy reports with varying report formats.

BACKGROUND

Colonoscopy quality reporting often requires manual data abstraction. NLP is another option for extracting information; however, limited data exist on its ability to accurately extract examination quality and polyp findings from unstructured text in colonoscopy reports with different reporting formats.

STUDY DESIGN

NLP strategies were developed using 500 colonoscopy reports from Kaiser Permanente Northern California and then tested using 300 separate colonoscopy reports that underwent manual chart review. Using findings from manual review as the reference standard, we evaluated the NLP tool's sensitivity, specificity, positive predictive value (PPV), and accuracy for identifying colonoscopy examination indication, cecal intubation, bowel preparation adequacy, and polyps ≥10 mm.

RESULTS

The NLP tool was highly accurate in identifying examination quality-related variables from colonoscopy reports. Compared with manual review, sensitivity for screening indication was 100% (95% confidence interval: 95.3%-100%), PPV was 90.6% (82.3%-95.8%), and accuracy was 98.2% (97.0%-99.4%). For cecal intubation, sensitivity was 99.6% (98.0%-100%), PPV was 100% (98.5%-100%), and accuracy was 99.8% (99.5%-100%). For bowel preparation adequacy, sensitivity was 100% (98.5%-100%), PPV was 100% (98.5%-100%), and accuracy was 100% (100%-100%). For polyp(s) ≥10 mm, sensitivity was 90.5% (69.6%-98.8%), PPV was 100% (82.4%-100%), and accuracy was 95.2% (88.8%-100%).

CONCLUSION

NLP yielded a high degree of accuracy for identifying examination quality-related and large polyp information from diverse types of colonoscopy reports.

Index colonoscopy-related risk factors for postcolonoscopy colorectal cancers

BACKGROUND AND AIMS

Postcolonoscopy colorectal cancers (PCCRCs) are defined as those detected ≤10 years after an index colonoscopy negative for cancer, but modifiable risk factors are not well established in large, community-based populations.

METHODS

We evaluated risk factors from the index colonoscopy for PCCRCs diagnosed 1 to 10 years after an index colonoscopy using a case-control design. Odds ratios (OR) and 95% confidence intervals (CI) were adjusted for potential confounders.

RESULTS

A proximal polyp ≥10 mm (OR, 8.18; 95% CI, 4.59-14.60), distal polyp ≥10 mm (OR, 3.30; 95% CI, 1.65-6.58), adenoma with (OR, 3.23; 95% CI, 1.83-5.68) and without advanced histology (OR, 1.87; 95% CI, 1.37-2.55), and an incomplete colonoscopy (OR, 5.52; 95% CI, 2.98-10.21) were associated with PCCRC. Risk factors for early versus late cancers (12-36 months vs >36 months to 10 years after examination) included incomplete polyp excision in the colonic segment of the subsequent cancer (OR, 4.76; 95% CI, 2.35-9.65); failure to examine the segment (OR, 2.42; 95% CI, 1.27-4.60); and a polyp ≥10 mm in the segment (OR, 2.38; 95% CI, 1.53-3.70). A total of 559 of 1206 patients with PCCRC (46.4%) had 1 or more risk factors that were significant for PCCRC (incomplete examination, large polyp, or any adenoma).

CONCLUSIONS

In a large community-based study with comprehensive capture of PCCRCs, almost half of PCCRCs had potentially modifiable factors related to polyp surveillance or removal and examination completeness. These represent potential high-yield targets to further increase the effectiveness of colorectal cancer screening.

Effects of Organized Colorectal Cancer Screening on Cancer Incidence and Mortality in a Large Community-Based Population

BACKGROUND & AIMS

Little information is available on the effectiveness of organized colorectal cancer (CRC) screening on screening uptake, incidence, and mortality in community-based populations.

METHODS

We contrasted screening rates, age-adjusted annual CRC incidence, and incidence-based mortality rates before (baseline year 2000) and after (through 2015) implementation of organized screening outreach, from 2007 through 2008 (primarily annual fecal immunochemical testing and colonoscopy), in a large community-based population. Among screening-eligible individuals 51-75 years old, we calculated annual up-to-date status for cancer screening (by fecal test, sigmoidoscopy, or colonoscopy), CRC incidence, cancer stage distributions, and incidence-based mortality.

RESULTS

Initiation of organized CRC screening significantly increased the up-to-date status of screening, from 38.9% in 2000 to 82.7% in 2015 (P < .01). Higher rates of screening were associated with a 25.5% reduction in annual CRC incidence between 2000 and 2015, from 95.8 to 71.4 cases/100,000 (P < .01), and a 52.4% reduction in cancer mortality, from 30.9 to 14.7 deaths/100,000 (P < .01). Increased screening was initially associated with increased CRC incidence, due largely to greater detection of early-stage cancers, followed by decreases in cancer incidence. Advanced-stage CRC incidence rates decreased 36.2%, from 45.9 to 29.3 cases/100,000 (P < .01), and early-stage CRC incidence rates decreased 14.5%, from 48.2 to 41.2 cases/100,000 (P < .04).

CONCLUSIONS

Implementing an organized CRC screening program in a large community-based population rapidly increased screening participation to the ≥80% target set by national organizations. Screening rates were sustainable and associated with substantial decreases in CRC incidence and mortality within short time intervals, consistent with early detection and cancer prevention.

Validation of 5 key colonoscopy-related data elements from Ontario health administrative databases compared to the clinical record: a cross-sectional study

BACKGROUND

Colonoscopy is used widely, but its quality is highly variable, which may adversely affect patients. Research and quality-improvement initiatives in a variety of jurisdictions have sought to address this issue, often supported by the use of health administrative data. As these data are generally not collected for these purposes, it is critical to measure their validity before use. The aim of this study was to validate health administrative data definitions for 5 key colonoscopy elements through comparison to the clinical record.

METHODS

In a cross-sectional study, we randomly sampled 1968 colonoscopy and noncolonoscopy procedures performed at 23 hospitals and 5 nonhospital endoscopy clinics between April 2008 and March 2009 in Ontario. We compared definitions for 5 key colonoscopy elements (colonoscopy case, colonoscopy setting, colonoscopy completeness, anesthesiologist assistance and polypectomy) derived from the health administrative data to the clinical record. We calculated weighted and unweighted sensitivity, specificity and positive predictive value, adjusted for clustering of patients within physicians, for each definition relative to the reference standard.

RESULTS

We abstracted 1845 records; in 1282 cases (69.5%), colonoscopy was intended or performed. The weighted sensitivity and specificity of colonoscopy case, nonhospital colonoscopy setting and anesthesiologist assistance exceeded 95%. The weighted sensitivity for colonoscopy completeness and polypectomy exceeded 95%, but specificity was less than 90%.

INTERPRETATION

Ontario health administrative data definitions for 5 key colonoscopy data elements performed well, with sensitivity and specificity values acceptable for use in research and quality-improvement initiatives. In jurisdictions where health administrative data are similarly used for research or quality improvement, similar studies could be considered.

Diagnosis and predictors of sessile serrated adenoma after educational training in a large, community-based, integrated healthcare setting

BACKGROUND AND AIMS

Sessile serrated adenomas (SSAs) are precursors of 15% to 30% of colorectal cancers but are frequently underdiagnosed. We sought to measure the SSA detection rate (SDR) and predictors of SSA detection after educational training for community gastroenterologists and pathologists.

METHODS

Colonoscopy and pathology data (2010-2014) from 3 medical centers at Kaiser Permanente Northern California were accessed electronically. Gastroenterologists and pathologists attended a training session on SSA diagnosis in 2012. Mean SDRs and patient-level predictors of SSA detection post-training (2013-2014) were investigated.

RESULTS

Mean SDRs increased from .6% in 2010-2012 to 3.7% in 2013-2014. The increase in the detection of proximal SSAs was accompanied by a decrease in the detection of proximal hyperplastic polyps (HPs). Among 34,161 colonoscopies performed in 2013 to 2014, SDRs for screening, fecal immunochemical test positivity, surveillance, and diagnostic indication were 4.2%, 4.5%, 4.9%, and 3.0%, respectively. SSA detection was lower among Asians (adjusted odds ratio [aOR], .46; 95% confidence interval [CI], .31-.69) and Hispanics (aOR, .59; 95% CI, .36-.95) compared with non-Hispanic whites and higher among patients with synchronous conventional adenoma (aOR, 1.46; 95% CI, 1.15-1.86), HP (aOR, 1.74; 95% CI, 1.30-2.34), and current smokers (aOR, 1.78; 95% CI, 1.17-2.72). SDRs varied widely among experienced gastroenterologists, even after training (1.1%-8.1%). There was a moderately strong correlation between adenoma detection rate (ADR) and SDR for any SSA (r = .64, P = .0003) and for right-sided SSAs (r = .71, P < .0001).

CONCLUSIONS

Educational training significantly increased the detection of SSA, but a wide variation in SDR remained across gastroenterologists. SSA detection was inversely associated with Asian and Hispanic race/ethnicity and positively associated with the presence of conventional adenoma, HP, and current smoking. There was a moderately strong correlation between ADR and SDR.

Association Between Time to Colonoscopy After a Positive Fecal Test Result and Risk of Colorectal Cancer and Cancer Stage at Diagnosis

IMPORTANCE

The fecal immunochemical test (FIT) is commonly used for colorectal cancer screening and positive test results require follow-up colonoscopy. However, follow-up intervals vary, which may result in neoplastic progression.

OBJECTIVE

To evaluate time to colonoscopy after a positive FIT result and its association with risk of colorectal cancer and advanced-stage disease at diagnosis.

DESIGN, SETTING, AND PARTICIPANTS

Retrospective cohort study (January 1, 2010-December 31, 2014) within Kaiser Permanente Northern and Southern California. Participants were 70 124 patients aged 50 through 70 years eligible for colorectal cancer screening with a positive FIT result who had a follow-up colonoscopy.

EXPOSURES

Time (days) to colonoscopy after a positive FIT result.

MAIN OUTCOMES AND MEASURES

Risk of any colorectal cancer and advanced-stage disease (defined as stage III and IV cancer). Odds ratios (ORs) and 95% CIs were adjusted for patient demographics and baseline risk factors.

RESULTS

Of the 70 124 patients with positive FIT results (median age, 61 years [IQR, 55-67 years]; men, 52.7%), there were 2191 cases of any colorectal cancer and 601 cases of advanced-stage disease diagnosed. Compared with colonoscopy follow-up within 8 to 30 days (n = 27 176), there were no significant differences between follow-up at 2 months (n = 24 644), 3 months (n = 8666), 4 to 6 months (n = 5251), or 7 to 9 months (n = 1335) for risk of any colorectal cancer (cases per 1000 patients: 8-30 days, 30; 2 months, 28; 3 months, 31; 4-6 months, 31; and 7-9 months, 43) or advanced-stage disease (cases per 1000 patients: 8-30 days, 8; 2 months, 7; 3 months, 7; 4-6 months, 9; and 7-9 months, 13). Risks were significantly higher for examinations at 10 to 12 months (n = 748) for any colorectal cancer (OR, 1.48 [95% CI, 1.05-2.08]; 49 cases per 1000 patients) and advanced-stage disease (OR, 1.97 [95% CI, 1.14-3.42]; 19 cases per 1000 patients) and more than 12 months (n = 747) for any colorectal cancer (OR, 2.25 [95% CI, 1.89-2.68]; 76 cases per 1000 patients) and advanced-stage disease (OR, 3.22 [95% CI, 2.44-4.25]; 31 cases per 1000 patients).

CONCLUSIONS AND RELEVANCE

Among patients with a positive fecal immunochemical test result, compared with follow-up colonoscopy at 8 to 30 days, follow-up after 10 months was associated with a higher risk of colorectal cancer and more advanced-stage disease at the time of diagnosis. Further research is needed to assess whether this relationship is causal.

Endoscopist fatigue estimates and colonoscopic adenoma detection in a large community-based setting

BACKGROUND AND AIMS

Endoscopist fatigue may impact colonoscopy quality, but prior studies conflict, and minimal data exist from community-based practices where most colonoscopies are performed.

METHODS

Within a large, community-based integrated healthcare system, we evaluated the associations among 4 measures of endoscopist fatigue and colonoscopic adenoma detection from 2010 to 2013. Fatigue measures included afternoon versus morning colonoscopy and the number of GI procedures performed before a given colonoscopy, including consideration of prior procedure complexity. Analyses were adjusted for potential confounders using multivariate logistic regression.

RESULTS

We identified 126 gastroenterologists who performed 259,064 total GI procedures (median, 6 per day; range, 1-24), including 76,445 screening and surveillance colonoscopies. Compared with morning examinations, colonoscopies in the afternoon were not associated with lower adenoma detection for screening examinations, surveillance examinations, or their combination (OR for combination, .99; 95% CI, .96-1.03). The number of procedures performed before a given colonoscopy, with or without consideration of prior procedure complexity, was also not inversely associated with adenoma detection (OR for adenoma detection for colonoscopies in the fourth quartile of fatigue based on the number of prior procedures performed vs colonoscopies performed as the first procedure of the day, .99; 95% CI, .94-1.04).

CONCLUSIONS

In a large community-based setting, adenoma detection for screening and surveillance colonoscopies were not associated with either time of day or the number of prior procedures performed by the endoscopist, within the range of procedure volumes evaluated. The lack of association persisted after accounting for prior procedure complexity.

How Efficacious Are Patient Education Interventions to Improve Bowel Preparation for Colonoscopy? A Systematic Review

BACKGROUND

Bowel preparation is inadequate in a large proportion of colonoscopies, leading to multiple clinical and economic harms. While most patients receive some form of education before colonoscopy, there is no consensus on the best approach.

AIMS

This systematic review aimed to evaluate the efficacy of patient education interventions to improve bowel preparation.

METHODS

We searched the Cochrane Database, CINAHL, EMBASE, Ovid, and Web of Science. Inclusion criteria were: (1) a patient education intervention; (2) a primary aim of improving bowel preparation; (3) a validated bowel preparation scale; (4) a prospective design; (5) a concurrent control group; and, (6) adult participants. Study validity was assessed using a modified Downs and Black scale.

RESULTS

1,080 abstracts were screened. Seven full text studies met inclusion criteria, including 2,660 patients. These studies evaluated multiple delivery platforms, including paper-based interventions (three studies), videos (two studies), re-education telephone calls the day before colonoscopy (one study), and in-person education by physicians (one study). Bowel preparation significantly improved with the intervention in all but one study. All but one study were done in a single center. Validity scores ranged from 13 to 24 (maximum 27). Four of five abstracts and research letters that met inclusion criteria also showed improvements in bowel preparation. Statistical and clinical heterogeneity precluded meta-analysis.

CONCLUSION

Compared to usual care, patient education interventions appear efficacious in improving the quality of bowel preparation. However, because of the small scale of the studies and individualized nature of the interventions, results of these studies may not be generalizable to other settings. Healthcare practices should consider systematically evaluating their current bowel preparation education methods before undertaking new interventions.

Association Between Primary Care Visits and Colorectal Cancer Screening Outcomes in the Era of Population Health Outreach

BACKGROUND

Population outreach strategies are increasingly used to improve colorectal cancer (CRC) screening. The influence of primary care on cancer screening in this context is unknown.

OBJECTIVE

To assess associations between primary care provider (PCP) visits and receipt of CRC screening and colonoscopy after a positive fecal immunochemical (FIT) or fecal occult blood test (FOBT).

DESIGN

Population-based cohort study.

PARTICIPANTS

A total of 968,072 patients ages 50-74 years who were not up to date with CRC screening in 2011 in four integrated healthcare systems (three with screening outreach programs using FIT kits) in the Population-Based Research Optimizing Screening through Personalized Regimens (PROSPR) consortium.

MEASURES

Demographic, clinical, PCP visit, and CRC screening data were obtained from electronic health records and administrative databases. We examined associations between PCP visits in 2011 and receipt of FIT/FOBT, screening colonoscopy, or flexible sigmoidoscopy (CRC screening) in 2012 and follow-up colonoscopy within 3 months of a positive FIT/FOBT in 2012. We used multivariable logistic regression and propensity score models to adjust for confounding.

RESULTS

Fifty-eight percent of eligible patients completed a CRC screening test in 2012, most by FIT. Those with a greater number of PCP visits had higher rates of CRC screening at all sites. Patients with ≥1 PCP visit had nearly twice the adjusted-odds of CRC screening (OR = 1.88, 95 % CI: 1.86-1.89). Overall, 79.6 % of patients with a positive FIT/FOBT completed colonoscopy within 3 months. Patients with ≥1 PCP visit had 30 % higher adjusted odds of completing colonoscopy after positive FIT/FOBT (OR = 1.30; 95 % CI: 1.22-1.40).

CONCLUSIONS

Patients with a greater number of PCP visits had higher rates of both incident CRC screening and colonoscopy after positive FIT/FOBT, even in health systems with active population health outreach programs. In this era of virtual care and population outreach, primary care visits remain an important mechanism for engaging patients in cancer screening.

Metformin use and risk of colorectal adenoma after polypectomy in patients with type 2 diabetes mellitus

BACKGROUND

Existing literature suggests that metformin, the most commonly used biguanide, may lower colorectal cancer risk. Because most colorectal cancers originate in precancerous adenomas, we examined whether metformin use lowered colorectal adenoma risk after polypectomy in patients with type-2 diabetes.

METHODS

Retrospective cohort study of 40- to 89-year-old Kaiser Permanente Northern California patients who had type 2 diabetes, and ≥1 adenoma detected at baseline colonoscopy during 2000 to 2009 and a repeat colonoscopy 1 to 10 years from baseline adenoma diagnosis through 2012. Cox models evaluated the association between metformin use during follow-up and subsequent adenoma diagnoses, controlling for age, race/ethnicity, sex, body mass index, and repeat examination indication.

RESULTS

Study included 2,412 patients followed for a median of 4.5 years; cumulatively, 1,117 (46%) patients had ≥1 adenoma at repeat colonoscopy. Compared with patients not receiving diabetes medications (n = 1,578), metformin-only use (n = 457) was associated with lower adenoma recurrence risk [adjusted HR, 0.76; 95% confidence interval (CI), 0.65-0.89], and the association was stronger with increasing total metformin dose [quartile (Q) 1: HR, 0.90; 95% CI, 0.72-1.12; Q2: HR, 0.89; 95% CI, 0.70-1.12; Q3: HR, 0.80; 95% CI, 0.63-1.01; Q4: HR, 0.50; 95% CI, 0.42-0.60, Ptrend < 0.001]. Findings were unchanged in sensitivity analyses, including evaluating only outcomes during the 3- to 10-year period from baseline.

CONCLUSION

Our study suggests a potential benefit of metformin use in lowering the risk of subsequent adenomas after polypectomy in patients with type 2 diabetes.

IMPACT

Metformin may lower colorectal cancer risk by reducing the formation of precancerous lesions, reinforcing the potential additional benefits of its use.