Time Trends in the Diagnosis of Colorectal Cancer With Obstruction, Perforation, and Emergency Admission After the Introduction of Population-Based Organized Screening

Key Points Question What is the association of organized, population-based colorectal cancer screening with the rate of obstructions, perforations, and emergency admissions prior to a colorectal cancer diagnosis? Findings This cohort study of 1861 Canadian adults with a diagnosis of colorectal cancer found no change in the rate of obstructions and perforations, but found a significant decrease in the rate of emergency hospital admissions after the implementation of organized colorectal cancer screening. Individuals who were up to date for colorectal screening were less likely to receive a diagnosis of an obstruction or perforation or have an emergency admission. Meaning Targeted colorectal cancer screening strategies are needed that focus on individuals at higher risk of an emergency presentation.


Introduction
Colorectal cancer (CRC) is the second most common cancer in North America, accounting for 13% of all cancer diagnoses in Canada. 1 Although the incidence of CRC has decreased during the last 20 years, one-half of individuals with CRC receive the diagnosis when the cancer is in a late stage. 1,2 In addition, 15% to 30% of patients with a CRC diagnosis present as an emergency. 3-5 Emergencies include intestinal obstructions, perforations, or emergency admissions to the hospital (OPE) prior to diagnosis. 6 Short-term 7,8 and long-term 9 survival for individuals with a diagnosis of CRC after an OPE is worse compared with nonemergency diagnoses even when adjusted for age and comorbidity. 10 Moreover, the health care burden of emergency CRC presentations is substantial, as these patients spend greater than 50% more days in the hospital than those with nonemergency diagnoses and overall treatment costs are higher. 11 Therefore, the rate of OPE among individuals who present with a new CRC diagnosis is a useful quality indicator because it represents a missed opportunity to diagnose CRC early. [12][13][14][15][16] One strategy that may decrease OPE rates is screening. 17  In this study, we used administrative health data to examine the association of CRC screening with OPE among individuals with a diagnosis of CRC. We examined time trends in the rate of OPE, emergency department (ED) visits, and diagnoses of stage IV CRC after the start of organized CRC screening, as well as factors associated with OPE including up-to-date CRC screening status for patients with CRC at initial presentation in a population-based setting.

Data Sources
Manitoba Health, the publicly funded provincial health insurance agency, provides comprehensive universal health coverage for hospitalizations, procedures, and physician visits for provincial residents (approximately 1.35 million in 2018). Manitoba Health maintains several electronic databases to monitor health care use and reimburse clinicians for services delivered. Since 1984, provincial residents have been assigned a personal health identification number that can be used to link provincial health information databases, allowing health care use and outcomes to be tracked longitudinally.
We used 3 Manitoba Health administrative databases: the Manitoba Population Registry, the Medical Claims database, and the Hospital Abstracts database. The Manitoba Population Registry contains demographic, vital status, and migration information and was used to assess the duration of provincial health coverage. The Medical Claims database is generated by claims filed by health care professionals for reimbursement of service. Medical Claims data were used to identify outpatient contacts with the health care system and type of contact (PCC or specialist), colonoscopy, FS, and nonprogram FOBT use. The Hospital Abstracts database includes all hospital admissions for Manitoba residents and was used to identify emergency and nonemergency hospital admissions regardless of admission route. Medical Claims and Hospital Abstracts data were also used to examine comorbidities and identify individuals with a diagnosis of ulcerative colitis or Crohn disease using a previously validated algorithm. 24 The accuracy and completeness of Manitoba Health's administrative data has been previously established. [25][26][27] The Winnipeg Regional Health Authority administers the delivery of health care in Winnipeg (two-thirds of the provincial population). Winnipeg ED visits that did not lead to hospitalization were identified using the Winnipeg Regional Health Authority's Admissions, Discharge and Transfer and E-Triage data and Emergency Department Information System databases. ColonCheck's populationbased registry was used to identify individuals who completed a screening program FOBT. Statistics included because ColonCheck recommends that screening begin at 50 years of age and the study required at least 2 years of screening information to determine screening history. Only individuals with a first diagnosis of CRC were included. Individuals with a prior diagnosis of ulcerative colitis or Crohn disease were excluded, because these individuals are at higher than average risk of CRC and are closely followed up by health care professionals.

Definition of Variables
Up-to-date screening was defined as a program or nonprogram FOBT in the previous 2 years, FS in the previous 5 years (recommended interval during most of the study years), or colonoscopy in the previous 10 years. Program FOBTs included FOBTs provided by ColonCheck. Nonprogram FOBTs included FOBTs provided outside of the screening program (ie, by PCC or health clinics). Flexible sigmoidoscopy and colonoscopies in the 3 months prior to CRC diagnosis were excluded, as they were likely to be for diagnostic purposes. 21 Program and nonprogram FOBTs that occurred up to 30 days before diagnosis were also excluded, as the primary outcome assessed was OPE in the month Area-level mean household income was categorized by quintile from lowest (income quintile 1) to highest (income quintile 5). Primary care clinician continuity of care (whether or not an individual received most of their ambulatory care from a single nonspecialist clinician) was measured by identifying individuals with at least 50% of visits to the same PCC among those with at least 3 visits in the 6 to 30 months prior to diagnosis. 32 Comorbidity was measured by the resource use band calculated using the Johns Hopkins Adjusted Clinical Group System software. 33 The resource use band includes the following 6 categories: 0 indicates nonuser; 1, healthy user; 2, low morbidity; 3, moderate morbidity; 4, high morbidity; and 5, very high morbidity. The resource use band is based on age, sex, physician claims, and hospital discharges in the year prior to diagnosis.

Outcomes
The primary outcomes were intestinal obstruction, intestinal perforation, or emergency hospital admission. Intestinal obstructions and perforations were identified using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) code K56.6 (other and unspecified intestinal obstruction) and ICD-10 code K63.1 (perforation of the intestine, nontraumatic). 13 An emergency hospital admission was identified using the emergency entry code (ie, patients who have a life-threatening condition requiring immediate assessment, treatment, and admission to the hospital). 34 This includes all modes of entry to the hospital and does not necessarily mean that the patient entered the hospital via the ED. Emergency department visits in the 30 days prior to a CRC diagnosis and diagnosis of stage IV CRC were added as outcomes in sensitivity analyses.

Statistical Analysis
Statistical analysis was performed from January 22, 2019, to February 26, 2020. Univariable and multivariable logistic regression analyses were used to evaluate factors associated with OPE. The following factors were included: screening history (up to date, colonoscopy, FS, program FOBT, and nonprogram FOBT), sex, age, income quintile, CRC stage at diagnosis, tumor location, continuity of care, comorbidity, era of diagnosis, PCC visits, specialist visits, and hospitalizations. 13 For CRC screening history, only up-to-date screening was included in the multivariable model.

JAMA Network Open | Public Health
Trends in Obstruction, Perforation, and Emergency Admissions After the Introduction of Colon Screening

Discussion
We found that the overall OPE rate decreased significantly from 2007, when provincial organized CRC screening was implemented, to 2015. This trend was primarily owing to a decrease in the rate of emergency hospital admissions. There was no change in the rate of obstructions and perforations or stage IV CRCs. A history of up-to-date CRC screening was independently associated with a decreased risk of CRC presentation with OPE.   studies and increased acceptability compared with guaiac-based FOBTs) 42,43 would lead to higher CRC screening uptake among those who need it the most in North America needs to be determined.
The effect of longer follow-up after the onset of population-based screening will also need to be evaluated in future studies.
Our study found that individuals with CRC who had lower mean household incomes were more likely to have OPE. These results are consistent with other studies that have found socioeconomic differences among patients with a diagnosis of CRC through emergency routes despite the provision of universal health care. 12,44,45 The association between income and the increased risk of OPE may be owing to individuals seeking health care only after developing advanced symptoms and lower screening rates. 21  We found that individuals with right-sided CRC were more likely to present with OPE. The rectum is more capacious and less likely to be associated with obstruction. In addition, the rectum is thicker walled and mostly extraperitoneal and less likely to be associated with perforations. 46 Rectal cancers are more likely to lead to rectal bleeding, which may lead to earlier diagnostic workup compared with subtle symptoms due to right-sided CRC. In addition, CRC screening is less effective for right-sided CRCs. 47-50 Higher rates of OPE among women and older individuals could also be due to a higher proportion of CRC located in the right colon in older individuals and women in general, although this association persisted in site-adjusted analyses. Other reasons are not obvious and merit additional qualitative studies of those with OPE.

Strengths and Limitations
The results of this study should be interpreted in the context of its strengths and limitations. We used data from previously validated population-based administrative health databases. 25,26,51,52 Regardless, there may have been misclassification of patients with OPE owing to coding errors. We did not conduct a medical record review to evaluate the accuracy of the ICD-10 codes used to identify OPE. However, we did use codes that were used in prior studies. 13,14 This was an observational study and the potential for residual confounding by unmeasured or unrecognized factors exists. Because nonprogram FOBT and ED data were not available for non-Winnipeg residents, the analysis was restricted to Winnipeg. Finally, we used area-level income as a proxy measure for individual-level income which may have attenuated the association between individual income and OPE owing to the misclassification of a few individuals' actual income. However, prior studies in Manitoba have shown substantial correlation between neighborhood-level income and a self-reported income. 28,29

Conclusions
Reducing emergency presentations is an important step in reducing CRC mortality. In Winnipeg, Manitoba, the rate of emergency hospital admissions prior to a CRC diagnosis has decreased. The rate of obstructions and perforations has not decreased. Disparities by income for OPE are present despite organized screening and universal health care. Nevertheless, regular health care contact and a history of CRC screening decreases the likelihood of OPE. These results make a strong argument for targeted screening strategies that focus on lower-income neighborhoods where individuals are at higher risk of OPE.

JAMA Network Open | Public Health
Trends in Obstruction, Perforation, and Emergency Admissions After the Introduction of Colon Screening