Abstract
Introduction: Screening for colorectal cancer is beneficial. Yet, screening remains suboptimal, and underserved populations are at greater risk for not being appropriately screened. Although many barriers to screening are understood, less is known about how the decision-making process on whether to receive colonoscopy or stool testing influences screening.
Methods: As part of a randomized controlled trial to test engaging underserved populations in preventive care through online, personalized, educational material, 2417 patients aged 50 to 74 years were randomly selected from the 70,998 patients with an office visit the year prior and mailed a survey to assess decision-making for colorectal cancer screening. Twenty practices in practice-based research networks from 5 diverse states participated. Survey data were supplemented with electronic health record data.
Results: Among respondents, 64% were or became up to date with screening within 3 months of their office visit. The main factor associated with being up to date was the length of the patient-clinician relationship (<6 months vs 5+ years: odds ratio [OR], 0.49; 95% CI, 0.30-0.80). Sharing the decision about screening options with the clinician was a predictor for being up to date compared with patients who made the decision for themselves (OR, 1.75; 95% CI, 1.27-2.44). Only 36% of patients reported being given a choice about screening options. Traditional factors like race, employment, insurance, and education were not associated with screening.
Conclusions: Having a long-term relationship with a primary care clinician and sharing decisions may be key drivers to ensure evidence-based preventive care for underserved populations.
- Colorectal Cancer
- Decision-Making
- Early Detection of Cancer
- Mass Screening
- Physician-Patient Relations
- Practice-Based Research
- Primary Health Care
- Surveys and Questionnaires
- Vulnerable Populations
Introduction
Colorectal cancer screening saves lives; yet, less than two-thirds of Americans who should be screened are up to date with screening.1 Minorities, individuals living in poverty, uninsured, and Medicaid beneficiaries disproportionately suffer greater morbidity and mortality from colorectal cancer.2 Although there is some debate as to whether biologic factors contribute to disparities in colorectal cancer outcomes for minorities, there is clear evidence that a lack of screening, delays in starting screening, inadequate follow-up of abnormal results, and treatment delays are all factors.3,4 Accordingly, a major focus of care has been to ensure that all adults age 50 to 74 years complete colon cancer screening.
Top barriers to screening include not knowing that screening is indicated, not being told by a clinician to be screened, financial concerns, misconceptions on benefits and harms of screening, fear about finding cancer, and test-specific barriers (eg, not wanting to handle stool or not wanting to do a bowel preparation).5 Interventions such as one-on-one education, clinician and patient alerts and reminders, reduced out-of-pocket expenses, and reduced structural barriers have all been shown to increase screening rates.6
Although the barriers to screening are well documented and there are interventions to promote screening, less is known about the relative importance of patient, clinician, health system, and communication factors associated with recommended screening, particularly for more underserved populations. Knowing the relative importance of these barriers may help to inform which interventions should be prioritized for implementation.
Methods
As part of a randomized controlled trial to evaluate engaging underserved patients in preventive care, we analyzed electronic health records (EHRs) and patient survey responses about clinical encounters. The full study protocol has been previously published, and the study and survey were approved by the Virginia Commonwealth University Institutional Review Board (HM15307).7
This study was conducted in 2 practice-based research networks (PBRNs): the Virginia Ambulatory Care Outcomes Research Network and the OCHIN community health information network. Twenty safety net practices and community health centers in 5 states with a wide geographic distribution participated. Among 70,998 patients seen for an office visit between November and July 2016, 4336 patients aged 40 to 75 were randomly mailed a survey, of which 2417 included questions about the colorectal cancer screening decision-making process.
Survey questions asked patients about their demographics, desired and actual locus of decision-making control,8 satisfaction with clinician communication,9 decisional conflict,10 length of patient-clinician relationship (ie, “How long have you been going to your doctor?”), and the single-item quality of life question. Locus of decision-making included 5 options: “I prefer to make the decision,” “I prefer to make the decision after considering my clinician's opinion,” “I prefer to share the decision with my clinician,” “I prefer for my clinician to make the decision after considering my opinion,” and “I prefer my clinician to make the decision.”8 Satisfaction with clinician communication was assessed using the 4 clinician communication questions from the Consumer Assessment of Health care Providers and Systems Clinician & Group Survey.9 Response options included never, sometimes, usually, and always, which are scored as 1, 2, 3, and 4, respectively. Decisional conflict was assessed using the 10-question simplified Decisional Conflict Scale. Questions were grouped into 4 subscales and scored per standard protocol, with an overall score under 25 being correlated with a greater likelihood of patients being able to make a decision, and a score over 37 being correlated with a greater likelihood that a patient will not be able to make a decision.10
EHR data included diagnoses, colorectal cancer screening tests and dates, race/ethnicity, and insurance type. Being up to date with screening was based on the 2016 US Preventive Services Task Force recommendation 3 months after the index office visit (eg, colonoscopy within 10 years and stool test within 1 year).1 Information about clinician characteristics (age, gender, race/ethnicity, specialty training, and concordance with patient) were collected from PBRN member records.
Associations between clinician, patient, and decision-making characteristics and being up to date with colorectal cancer screening were compared. Categorical variables were summarized with frequencies, and continuous variables were summarized with means and standard deviations. χ2 tests compared frequencies, except in cases of small sample sizes, for which Fisher's exact test was used. t test compared differences in means between patients who were and were not up to date. Analyses were completed using the SAS version 9.4 software (Cary, NC) in January 2019.
Results
Of 2417 patients mailed surveys with colorectal cancer screening questions, 1068 completed the survey (44%), and we were able to match EHR data for 959 patients eligible for screening. A total of 587 of 959 patients (61%) were up to date with screening at the time of their visit and 31 (3%) became up to date within 3 months (overall up to date rate of 64%). Table 1 shows patient characteristics for patients who were and were not up to date. Age (odds ratio [OR], 1.03; 95% confidence interval [CI], 1.01–1.05) and length of the patient-clinician relationship (<6 months vs 5+ years: OR, 0.49; 95% CI, 0.30–0.80; 6 to 12 months vs 5+ years: OR, 0.60; 95% CI, 0.36–0.99; 1 to 3 years vs 5+ years: OR, 0.67; 95% CI, 0.48–0.95) were associated with being up to date, and there was no evidence of associations between being up to date and other clinician and patient factors.
Association between Patient Characteristics and Colorectal Cancer Screening Up-to-Datedness.*
Table 2 shows the patient report of colorectal cancer screening discussions. Only 36% of patients reported being given a choice about the type of screening test. Patients generally preferred to make their screening decision independently (51%) or share the decision with their clinician (40%). A total of 71% of patients reported that the locus of control for their decision was consistent with how they wanted to make decisions. Although patients reported high satisfaction with clinician communication, 33% reported high decisional conflict (ie, score above 37), with greater uncertainty for the “informed” and “value clarity” subscales.
Association between Patient-Reported Decision-Making and Colorectal Cancer Screening Up-to-Datedness (n = 959)
Patients who shared the decision with their clinician about screening options were more likely to be screened than those who reported independently making the decision (OR, 1.75; 95% CI, 1.27–2.41). Even those who let their clinician make the decision for them were more up to date than those who made the decision themselves (OR, 1.94; 95% CI, 1.20–3.13).
Discussion
In this sample of practices that care for more underserved patients, the longer the patient-clinician relationship, the more likely a patient was to be screened for colorectal cancer. Sharing the decision about how to be screened or even letting the clinician make the decision further increased a patient's chance of being screened, suggesting that the trust established through these relationships may be a key driver in patients overcoming barriers to get recommended screening. Patient trust in their clinician and clinician knowledge of the patient have long been shown to improve not only patient satisfaction but also adherence with care.11 These findings are further consistent with multiple studies that have linked the continuity of both primary care and specialty care relationships with reduced disease-specific and all-cause mortality.12 Others have shown that the receipt of primary care specifically is associated with improved quality of care, better health care experiences, lower costs, and even increased life expectancy.13⇓–15 The clinician-patient relationship may be a key factor that leads to these benefits.
In this population, more traditional patient-level barriers to screening (race, employment, insurance, and education) and practice or clinician factors were not associated with being up to date.16 Clearly, on a national level, they remain significant barriers to colorectal cancer screening. However, in this sample of PBRN practices that are participating in research, are likely working on systems to improve evidence-based care, and where patients have a primary care clinician and have had an office visit, these barriers may be less significant. This idea is supported by the finding that the colorectal cancer screening rate was higher than national averages. Furthermore, now that first dollar coverage for colorectal cancer cost is mandated by the Affordable Care Act, cost may be less of a barrier; also, with multiple screening options, logistics may be less of a barrier. Another study in safety net practices found the top barrier for not receiving any screening was patients not knowing they were due for screening or their clinician not recommending screening.17 The longer clinician-patient relationships and greater clinician involvement in decisions observed in this study may be addressing these barriers.
A key strength of this study is that it uniquely linked the patient-reported process for care delivery (eg, decision-making) to the outcome of being up to date on colorectal cancer screening. Most prior studies have focused solely on patient-reported barriers to screening. In addition, this study examines multiple levels of influence on screening—the clinician, patient, and decision-making. A limitation is that we lacked power to compare subgroups of patients or factors influencing the receipt diagnostic follow-up. Of interest would be to compare those who were not up to date before the visit and became up to date with those who remained not up to date. Future work is needed to compare these groups of patients.
Overall, having a long-term relationship with a primary care clinician seems important to help ensure that more underserved patients get evidence-based preventive care. Promoting longitudinal patient-clinician relationships may be 1 strategy for reducing health inequities.
Notes
This article was externally peer reviewed.
Funding: Funding for this study is provided by the National Cancer Institute (R01CA166375-01A1), the National Center for Advancing Translational Sciences (UL1TR002649), and VCU Massey Cancer Center Biostatistical Shared Resource, supported, in part, with funding from NIH-NCI Cancer Center Support Grant P30 CA016059. The opinions expressed in this manuscript are those of the authors and do not necessarily reflect those of the funders.
Conflict of interest: Dr. Krist is a member of the United States Preventive Services Task Force (USPSTF).
Disclaimer: This article does not necessarily represent the views and policies of the USPSTF.
To see this article online, please go to: http://jabfm.org/content/33/5/779.full.
- Received for publication October 15, 2019.
- Revision received May 1, 2020.
- Accepted for publication May 9, 2020.