Utilizing Quality Improvement Methodology to Address Disparities in Colorectal Cancer Screening Between Faculty and Resident Physicians

Abstract

Health system data from a large academic medical center revealed a 14% lower rate of colorectal cancer (CRC) screening in resident patient panels compared to faculty patient panels. This resident-led quality improvement (QI) work identified causes for disparities in CRC screening and implemented an innovative panel management intervention to reduce CRC screening disparities. Analysis was conducted across two academic primary care clinics at a single institution. Residents engaged key stakeholders in the CRC screening process to identify causes for disparities and potential solutions in the CRC screening process. A novel interprofessional panel management protocol was implemented to guide residents on how to perform population health strategies to increase CRC screening and to streamline the navigation process. The effectiveness of each intervention in improving CRC screening was analyzed. After four months of protocol implementation, CRC screening for resident patient panels improved from 62 to 68% based on analysis done on a run chart. The difference in CRC screening between faculty and resident patient panels decreased from 14 to 10%, reducing the disparity by 29%. This interprofessional panel management protocol significantly increased the CRC screening rates among patients receiving primary care from resident physicians. This further highlights the importance of multipronged interventions to improve disparities in CRC screening and to improve overall screening rates.

INTRODUCTION

Colorectal cancer (CRC) is the 3rd most common type of cancer and the 2nd leading cause of cancer death in the USA.^1,2 Screening significantly reduces colorectal cancer incidence and mortality due to the removal of precancerous polyps and detection of early stage CRC.^3–6 The most used screening tests for colorectal cancer are fecal testing for occult blood and endoscopic screening with colonoscopy, the latter being considered the gold standard.^1,7 There is a grade A recommendation from the US Preventive Services Task Force for CRC screening in individuals aged 50 to 75.⁸ However, only about 60% of adults aged 50 to 75 in the USA are up-to-date with CRC screening and screening rates are significantly lower in underserved populations, with only 54% of eligible Medicaid-insured adults receiving screening.^4,5

While multiple screening modalities exist, the efficacy of stool-based versus office-based screening modalities for underserved patient populations is still being evaluated.^1,2,5 In the USA, most of the screening occurs in the primary care setting and is dependent on visit-based interventions.⁴ However, non-office-based population health strategies may be more helpful in increasing screening rates among vulnerable populations.⁹ The most studied methods for increasing CRC screening in vulnerable populations are patient navigation and stool testing outreach.¹⁰ Furthermore, screening outreach programs are considered a highly effective organizational population health strategy that leads to a sustained decrease in colorectal cancer risk at 10 years.^1,4,11 Combining mailed outreach invitations with office-based screening increases one-time screening in racially diverse and socioeconomically disadvantaged populations.^4,5,9,10 Even though non-office-based interventions are essential for increasing CRC screening rates in underserved populations, disparities in CRC screening remain prevalent in academic primary care clinics that tend to include many vulnerable patients.^10,12

In many academic settings, resident physicians practice under the supervision of attending physicians and provide longitudinal care in practices that share resources with faculty physicians.¹³ However, resident physicians are more likely to care for Medicaid-insured patients and patients from socioeconomically underserved populations as compared to faculty physicians.^10,14 Studies have demonstrated worse population health metrics, including chronic disease management and preventive cancer screening, in resident patient panels compared to attending patients.^10,15 Lower rates of colorectal cancer screening in patient populations receiving care from resident physicians are not explained by knowledge gaps alone. Studies have suggested factors such as knowledge of modalities for CRC screening, length of relationship between patient and primary physician, and structural issues within the health system and knowledge of health system navigation as factors contributing to lower screening rates.^16,17

Baseline data from our single academic medical center revealed a 14% lower rate of CRC screening in resident patient panels compared to attending physician panels. There are, however, no standard interventions for resident physicians to decrease the disparity in CRC screening rates between resident and attending patient panels. This resident-led quality improvement (QI) work identified causes for disparities in CRC screening and implemented an innovative panel management intervention within the resident primary care clinics to reduce CRC screening disparities.

SETTING AND PARTICIPANTS

The study population included all patients, aged 50–75 years old, from two hospital-based primary care practices within a single large academic medical center in an urban metropolitan area. Each practice comprises 12 faculty and 60 residents on multidisciplinary teams that provide primary care for a combined 22,645 patients. Residents serve as the primary care physician for 8487 (37%) of these patients with supervision from faculty preceptors. Each resident has a patient panel of 90–110 patients during their residency and then transitions the care of those patients to an incoming intern upon graduation. Residents have a 2-week ambulatory block every 2 months that includes seven half-day clinic sessions. There were 3395 (40%) resident patients and 8184 (58%) faculty patients that were eligible for CRC screening (Table 1). In the resident group, 2505 were Black (74%) and 23% had Medicaid as their primary insurance. In the faculty group, 3450 were Black (42%) and 10% had Medicaid as their primary insurance. Patients were eligible for CRC screening if aged 50–75 and were excluded if they were up to date with Healthcare Effectiveness Data and Information Set (HEDIS) CRC screening measures, defined as a colonoscopy every 10 years, a flexible sigmoidoscopy or CT colonography every 5 years, a stool DNA test every 3 years, or an annual fecal immunochemical test.⁹

Table 1.

Characteristics of Faculty and Resident Patient Panels

	Faculty	Resident
Panel size	14,158	8487
High risk (per EHR)	405 (3%)	326 (4%)
% Medicaid	10%	23%
CRC screen overdue	3256 (23%)	3225 (38%)
CRC screen eligible	8184 (58%)	3395 (40%)
% CRC eligible Black	42%	74%

PROGRAM DESCRIPTION

Quality data from the two primary care practices within a large academic medical center revealed a 14% lower CRC screening rate in resident panels compared to attending panels. The authors engaged key stakeholders which included faculty physicians, faculty nurse practitioners, resident physicians, patients, and gastroenterologists to identify both causes for disparities in CRC screening and potential solutions in the CRC screening process. Qualitative data was gathered through structured interviews to understand the current process for CRC screening in the primary care setting. Twenty-two structured interviews revealed scheduling algorithms, limited time during clinical encounters, and patient-provider knowledge deficits as the top contributors for low CRC screening rates. A root cause analysis for the CRC screening process was conducted by the authors and two faculty primary care nurse practitioners and identified the two most significant causes for reduced CRC screening rates to be related to clinic operations and clinician knowledge. Specifically, there was inefficiency in the process of colonoscopy scheduling leading to delayed or missed scheduling. Resident clinicians also felt unsure of when or how to engage patients in fecal immunochemical test (FIT) screening in place of colonoscopy. Further analysis highlighted that difficulty in scheduling a colonoscopy stemmed from staffing shortages and previously delayed procedures due to the COVID-19 pandemic. While the logistical barrier of increasing capacity for colonoscopy-based screening was outside the scope of the primary care practice, there existed an opportunity to address, simplify, and streamline the process of ordering a FIT test for eligible average-risk patients.

A novel panel management protocol consisting of an interprofessional team of providers, nurses, and medical assistants was implemented to guide residents on how to perform population health strategies to increase CRC screening rates for their eligible patients. Every resident was assigned an annual half-day during their ambulatory block to identify patients from their individual patient panel who were due for CRC screening using an electronic health record (EHR)–based dashboard. Patients eligible and overdue for CRC screening received an intervention determined based on the protocol (Fig. 1). Residents reached patients who did not have an upcoming appointment in 3 months using telephone calls or electronic messaging. If the patient had an incomplete order for a colonoscopy in the past, patients were mailed a FIT kit by the clinic’s ancillary staff. If a patient did not have a colonoscopy order, then a referral for a colonoscopy was made by the resident after communicating with the patient. A mailed FIT kit comprised of a one-page letter with simple instructions for completing the FIT as well as basic information on the importance of CRC screening. Additionally, all contacted patients were also offered a telemedicine visit to discuss CRC screening options. If patients opted for a telemedicine visit, an appointment was requested during outreach. The completion of colorectal cancer screening, regardless of modality, was evaluated for all patients for whom a telemedicine appointment was requested, regardless of whether the patient attended the visit. Instead of contacting patients with upcoming clinic visits in the next 3 months, residents sent electronic reminders to their providers to discuss and offer CRC screening. If patients did not have the electronic messaging system set up and could not be contacted due to lack of answer or non-working phone numbers, they were mailed a FIT kit to the listed current address in their record.

Figure 1.

Simplified flowsheet for CRC screening outreach.

PROGRAM EVALUATION

Data was obtained prospectively for the first 4 months after implementation of the protocol. Residents reached 343 patients with lapsed CRC screening, comprising 10% of the resident patient panel who were overdue or eligible for screening. Seven patients (2.1%) were excluded due to having recently completed CRC screening based on manual chart review. An EHR-based dashboard was used to collect quantitative metrics, including the percentage of outreached patients who underwent CRC screening, the CRC rates pre-and post-intervention, and the relative reduction in screening rate disparity between resident and attending patient panels. The effectiveness of each intervention leading to completion of CRC screening was analyzed using chi-square tests using Stata/IC 15.1.

Among outreached patients, a total of 83 patients (24.7%) were referred for colonoscopies, 126 patients (37.5%) were mailed a FIT kit, 94 patients (28.0%) were provided with a telemedicine visit to discuss CRC screening, and 33 patients (9.8%) with upcoming appointments had reminders sent to their providers. A total of 51 patients (15.2%) completed their screening. There was no statistical difference in screening rates between patients referred for a colonoscopy (19.3%) and those who received a FIT kit (15.1%), p = 0.42. Screening rates were notably higher (27%) in patients who had reminders sent for their upcoming appointments compared to patients offered colonoscopy or FIT (16.7%), but the finding was not statistically significant, p = 0.46. Patients provided with a telemedicine visit to discuss CRC screening had significantly lower screening rates (7.4%) compared to those offered colonoscopy or FIT (16.7%), p = 0.03. Of the 51 patients who completed their screening, 28 were Black (54.9%), and 10 had Medicaid as their primary insurance (19.6%). Of the 51 completed screenings, 12 (23.5%) had a positive FIT test or adenomas found on colonoscopy.

The CRC screening rates for resident and attending patient panels were tracked and analyzed for 12 months, 4 months pre-intervention, and 6 months post-intervention using a run-chat. The CRC screening completion rate for resident patient panels improved from the mean of 62% [61.3–63.1] to the mean of 68% [63.7–70.3] with a consistent upward trend of at least six data points. The mean CRC screening rate difference between faculty and resident patient panels decreased from 14 to 10%, reducing the disparity between resident and attending panels by a relative reduction of 29%.

DISCUSSION

Due to structural and historical factors within graduate medical training programs, Internal Medicine residents are more likely to serve as the primary care physician for more socioeconomically disadvantaged patients than faculty in the same practices. Even when accounting for these observed sociodemographic differences, resident-delivered primary care is associated with lower rates of preventive health screening as compared to faculty-delivered primary care. Limited time during clinical encounters and experience in practice also contribute to disparities in CRC screening between residents and faculty. Screening outreach programs can reduce disparities in cancer screening for vulnerable populations and increase CRC screening rates in resident primary care panels. This QI initiative demonstrates that an interprofessional panel management intervention outside the clinical visit can significantly increase the CRC screening rate among patients receiving primary care from resident physicians. Further, our results highlight the importance of a multipronged intervention to improve overall screening rates.

This innovative panel management intervention offers CRC screening modalities individualized to patient health risk factors, barriers to care access, and personal preferences. The protocol facilitates direct patient access to CRC screening through an interdisciplinary team that assists in colonoscopy scheduling, coordinating support services for transportation or colonoscopy prep, and access to stool-based screening tests. Reducing the steps needed for patients to complete CRC screening increases the likelihood of CRC screening completion, as evidenced by the results of this intervention. While these findings innovate upon previous CRC screening work, it reveals that there is significant room for improvement in care delivery for patients receiving care from resident physicians, as well as emphasizes the need to tailor population health interventions to address barriers to care in underserved patient populations. Although a majority of resident patients have Medicaid as their primary insurance, only a minority of patients who completed their CRC screening after outreach had Medicaid as their primary insurance.

This proactive panel management approach was an effective modality for increasing CRC screening rates and reducing care disparities between resident and faculty patient panels. This novel panel management protocol streamlined patient-provider interactions outside of a time-restricted outpatient clinic visit and made it easier for residents to conduct direct outreach to patients. Future directions for innovation include automating the process given the annual need for screening in patients opting for fecal-based testing, improving the effectiveness of this intervention as a sustainable tool that addresses disparities in resident-delivered primary care, and translating this approach to reduce disparities between resident and faculty patient panels within other age-appropriate cancer screening interventions.

This study on addressing disparities in CRC screening has several limitations. Although it examines two primary care practices staffed by faculty and resident physicians with similar clinic operations involving multidisciplinary staff with a pre-dominantly English-speaking patient population, the generalizability of our findings may be limited to smaller academic and community-based training programs as well as communities that serve pre-dominantly non-English-speaking patients. Furthermore, we were unable to measure relevant social determinants of health screening for patients, such as financial or housing insecurity, health literacy, and transportation insecurity, which can impact access to care. Another limitation included the significant waiting time for colonoscopies due to the COVID-19 pandemic. While we did not find a significant difference in completed screening between colonoscopy and FIT-based screening, it is possible that the long delay in time-to-colonoscopy scheduling may have impacted these results. Lastly, the long-term effectiveness is still uncertain since FIT as a CRC screening modality is dependent on repeat annual testing and previous studies have suggested 40–70% of patients fail to undergo repeat FIT testing. Careful consideration must be given to patients who screen positive on FIT or who fail to undergo repeat FIT testing since effective CRC screening relies on follow-up throughout the entirety of the screening process.⁴

Abbreviations

CRC

Colorectal cancer

QI

Quality improvement

FIT

Fecal immunochemical test

EHR

Electronic health record

Data Availability

The data that support the findings of this study are not publicly available due to restrictions related to participant confidentiality. Data from this study can be requested from the authors upon reasonable request and with the permission of the University of Pennsylvania Human Research Protections Program. 

Declarations:

Conflict of Interest:

The authors declare that they do not have a conflict of interest.