Abstract
Background
Patient navigation is a recommended practice of the Community Guide for Preventive Services; little is known about whether patient navigation improves colonoscopy completion for adults who have received an abnormal stool test result.
Objective
To determine whether patient navigation delivered to individuals with an abnormal stool test result increased follow-up colonoscopy completion (primary) at 1 year.
Design
Randomized controlled trial.
Setting
A federally qualified health center (n = 32 clinics) in Washington State.
Patients
Individuals aged 50-75 with an abnormal fecal test result in the prior month.
Interventions
A six-topic, phone-based patient navigation program delivered by bilingual (English and Spanish) clinical staff.
Measurements
Receipt of follow-up colonoscopy at 1 year (primary); time to colonoscopy receipt; and program effectiveness by patient characteristics, including patients’ probability of obtaining a colonoscopy without navigation, derived using health record data (secondary).
Results
Of 985 individuals enrolled (mean age, 61 [SD, 6.8] years; 170 [18%] had Spanish language preference listed in the medical record), 967 were included in the primary intention-to-treat analysis (479 in patient navigation; 488 in usual care). Receipt of follow-up colonoscopy was higher in the patient navigation group than in the usual care group (55.1% vs 42.1%; risk difference, 13.0 percentage points; 95% CI, 6.5 to 19.4 percentage points). The intervention effect was not moderated by patients’ probability of obtaining a colonoscopy without navigation.
Limitations
Our study was primarily conducted during the height of the COVID-19 pandemic, which created additional health system- and patient- barriers to colonoscopy.
Conclusions
These findings support the effectiveness of patient navigation for follow-up colonoscopy completion.
Primary Funding Source
National Cancer Institute
Trial registration
clinicaltrials.gov Identifier: NCT03925883
Keywords: Risk prediction, endoscopy, colorectal cancer screening, community health center, navigation, comparative effectiveness
INTRODUCTION
Stool testing, including fecal immunochemical testing (FIT), is a recommended form of colorectal cancer (CRC) screening. Improving CRC screening rates could reduce CRC mortality by more than 50% (1, 2), but only if patients with abnormal stool-testing results receive follow-up colonoscopies. The US Multi-Society Task Force on Colorectal Cancer established an 80% target for follow-up colonoscopy completion.(3) While some integrated healthcare systems have successfully achieved this goal,(4) other healthcare setting face significant obstacles in meeting this benchmark. Analyses of data from more than 20,000 commercially insured US adults in 38 health systems for the years 2016 – 2020 show a 6-month follow-up colonoscopy rate of 47.9% to 51.4%(5, 6); and 6-month rates in federally qualified health clinics (FQHCs) are as low as 18%.(7) Delays in follow-up colonoscopy increase the risk of cancer and later-stage cancer detection.(8, 9) In response, national efforts are underway to establish follow-up colonoscopy completion as a standardized performance measure for healthcare systems.(6)
Patient navigation programs can help patients overcome CRC screening barriers and the Community Guide for Preventive Services recommends patient navigation to improve CRC screening among historically disadvantaged racial and ethnic populations and people with lower incomes.(10) A meta-analysis of studies that evaluated the effectiveness of patient navigation on screening rates for CRC showed that screening rates were 64% higher with patient navigation overall (risk ratio [RR], 1.64; 95% confidence interval [CI] 1.42 to 1.92; I2 = 93.7%; 22 trials).(11) In analysis limited to the studies that evaluated colonoscopy/endoscopy outcomes, patient navigation programs doubled colonoscopy/endoscopy receipt versus usual care (RR, 2.08; 95% CI, 1.08 to 4.56; I2 = 94.6%; 6 trials). However, none of the trials evaluated colonoscopy follow-up after an abnormal test result. Understanding the effectiveness of navigation in the abnormal-FIT subgroup is critical, given that this group presents an elevated CRC risk and potential opportunity for a cost-efficient intervention.(12, 13)
Two small, randomized trials (ns < 300) evaluating colonoscopy completion after an initial abnormal screening test separately reported patient navigation-associated improvements of 10 percentage-points at 180 days (14) and 21 percentage-points at 1 year.(15) (16) Nevertheless, combined analysis of these two studies showed that patient navigation was not associated with improved completion of follow-up colonoscopies (RR, 1.21; 95% CI, 0.92-1.60; risk difference [RD], 14 percentage points; 95% CI, 0 to 29 percentage points).(17)
In these two trials, 58% to 80% of non-navigated patients still obtained a colonoscopy.(14, 15) Knowing whether a patient is likely to obtain a colonoscopy without navigation could help target costly navigation services (estimated at $275 per navigated patient(18)). Our team developed a risk-prediction model to calculate patients’ probability of obtaining a follow-up colonoscopy without navigation, and previously tested its use to deliver navigation in an integrated health system.(19, 20)
The Predicting and Addressing Colonoscopy non-Adherence in Community Settings (PRECISE) study assessed effects of a patient navigation program compared to usual care outreach on follow-up colonoscopy receipt among patients with an abnormal FIT result. The program used a predictive risk model to calculate individuals’ probability of obtaining a colonoscopy without navigation services.(19) Our primary hypothesis was that the intervention would improve rates of follow-up colonoscopy compared to usual care; a secondary hypothesis was that the effect would be greatest in those with the lowest probability scores.(21, 22)
MATERIALS AND METHODS
Study Design and Participants
PRECISE was an individual randomized study conducted at Sea Mar Community Health Centers (Sea Mar), an FQHC operating 32 clinics serving ~220,000 patients in western Washington, 40% of whom are classified as Hispanic/ Latino ethnicity in the electronic health record (EHR). The study protocol (protocol no.: 00000779; Supplement A) was reviewed and approved by the Kaiser Permanente Northwest Institutional Review Board and the Sea Mar Research Committee. The need for informed consent was waived due to minimal risks to patients. The trial was overseen by a data and safety monitor; CONSORT reporting guidelines were followed. A community advisory board, that included patients and clinic staff, provided input on the design, conduct, reporting, and dissemination of the research.
Sea Mar offers FITs (OC-Auto, Polymedco, Cortlandt Manor, NY) to eligible patients during clinic visits and operates a mailed FIT outreach program; FITs, processed in an on-site centralized laboratory, use a standard cutoff of >20μg of hemoglobin per gram of stool as abnormal. Colonoscopy services are delivered at more than 50 community gastroenterology practices. Sea Mar screened 33% and 29% of eligible adults ages 50-75 for CRC in 2020 and 2021, respectively. In each year, ~10,000 patients completed a FIT and ~700 had an abnormal FIT result.
We identified Sea Mar patients ages 50-75 who had received an abnormal FIT test result in the previous month using EHR data. Patients were excluded if they had life-limiting comorbidities (e.g., on hospice, diagnosis of metastatic cancer), received a recent colonoscopy (e.g., within the past 5 years), had clinical conditions that made them ineligible for colonoscopy (e.g., dementia, COPD requiring oxygen), or had moved out of the area. Patient identification took place between July 2019 and April 2022, with a pause from March 2, 2020, to August 24, 2020, because of an EHR conversion and COVID-19-related care suspensions.
Randomization
The project statistician used Sealed Envelope (London, UK)(23) to randomize patients 1:1 to navigation or usual care in blocks of 4 to 8 individuals, stratified on county of residence. Patients were allocated to intervention or usual care by an analyst at the health center (n = 892 patients) or research center (n = 93 patients). Intervention patients were entered into a REDCap study database, a secure web application, that tracked navigation activities.(24, 25)
Predicted Probability of Colonoscopy Adherence
We calculated individual colonoscopy adherence probabilities for randomized patients using a model we have described in previous publications.(19, 20) Model variables are provided (Table 1 in the Supplement). The model had a bootstrap-corrected c-statistic of 0.65.(19) (19)Consistent with prior research, we used tertiles as cutpoints to define groups having low, moderate, and high probability of obtaining a colonoscopy without navigation.(20)
Table 1.
Characteristics of eligible patients with abnormal FIT result, ages 50-75
| Characteristic | Usual care (n = 488) |
Patient Navigation a (n = 479) |
|---|---|---|
| N (%) | N (%) | |
| Age | ||
| Mean (SD) | 60.6 (7.1) | 61.0 (6.6) |
| 50-54 | 118 (24.2) | 92 (19.2) |
| 55-59 | 112 (22.8) | 109 (22.8) |
| 60-64 | 106 (21.8) | 132 (27.6) |
| 65-69 | 80 (16.4) | 85 (17.7) |
| 70-75 | 72 (14.7) | 61 (12.7) |
| Gender | ||
| Female | 232 (47.5) | 210 (43.8) |
| Male | 256 (52.5) | 269 (56.2) |
| Race b | ||
| Asian | 16 (3.3) | 29 (6.1) |
| Black/African American | 15 (3.1) | 20 (4.2) |
| More than 1 race | 12 (2.4) | 8 (1.7) |
| Non-White, other | 69 (14.2) | 72 (15.0) |
| White | 376 (77.0) | 350 (73.1) |
| Ethnicity c | ||
| Non-Hispanic | 381 (78.1) | 389 (81.2) |
| Hispanic | 107 (21.9) | 90 (18.8) |
| Language Preference d | ||
| English | 360 (73.8) | 362 (75.6) |
| Spanish | 91 (18.6) | 79 (16.5) |
| Other | 36 (7.4) | 37 (7.7) |
| Insurance Status | ||
| Medicaid | 227 (46.5) | 226 (47.2) |
| Medicare | 84 (17.2) | 90 (18.8) |
| Self-Pay (Uninsured) | 66 (13.5) | 59 (12.3) |
| Commercial | 111 (22.8) | 104 (21.7) |
| Marital Status | ||
| Divorced/Separated | 76 (15.6) | 75 (15.6) |
| Married | 164 (33.6) | 171 (35.7) |
| Single | 219 (44.9) | 212 (44.3) |
| Widowed | 29 (5.9) | 21 (4.4) |
| Body Mass Index e | ||
| Mean (SD) | 31.1 (7.8) | 31.6 (8.0) |
| Normal | 100 (20.5) | 79 (16.5) |
| Obese | 241 (49.4) | 251 (52.4) |
| Overweight | 142 (29.1) | 141 (29.4) |
| Underweight | 5 (1.0) | 7 (1.5) |
| Gender/Mammography Screening | ||
| Female w/ Mammography (ever) | 195 (40.0) | 176 (36.7) |
| Female no Mammography (ever) | 37 (7.6) | 34 (7.1) |
| Male | 256 (52.4) | 269 (56.2) |
| Prior CRC Screening | ||
| No prior CRC Screening | 315 (64.5) | 294 (61.4) |
| Prior CRC Screening | 173 (35.5) | 185 (38.6) |
| Number of encounters (0-6+) | ||
| Mean (SD) | 3.1 (2.5) | 3.0 (2.4) |
| No visits | 135 (27.7) | 129 (26.9) |
| At least 1 visit | 353 (72.3) | 350 (73.1) |
| Prior missed appointments | ||
| Mean (SD) | 0.8 (0.9) | 0.8 (0.9) |
| No (Zero-no show) | 252 (51.6) | 260 (54.3) |
| Yes (At least one no-show) | 236 (48.4) | 219 (45.7) |
| Gagne Comorbidity Score | ||
| Mean (SD) | 0.3 (0.8) | 0.3 (0.9) |
| County | ||
| Clark | 100 (20.5) | 95 (19.8) |
| Grays Harbor | 76 (15.6) | 78 (16.3) |
| King | 92 (18.9) | 86 (18.0) |
| Pierce | 50 (10.2) | 49 (10.2) |
| Skagit | 43 (8.8) | 43 (9.0) |
| Snohomish | 36 (7.4) | 34 (7.1) |
| Thurston | 51 (10.4) | 51 (10.6) |
| Whatcom | 40 (8.2) | 43 (9.0) |
| Baseline Predicted Probability of Completing a Colonoscopy within 1 yr f | ||
| Low (0 to .51) | 151 (30.9) | 167 (34.9) |
| Moderate (.52 to .62) | 169 (34.7) | 158 (33.0) |
| High (.63 to 1.0) | 168 (34.4) | 153 (31.9) |
Excludes patients who were randomized in error (10 in patient navigation, 8 in usual care)
Non-White, other includes American Indian or Alaskan Native (n = 1), Hispanic/Latino (n = 75), Mexican/Mexican American (n = 3), Other Race (n = 58), Patient Declined to Answer (n = 1), and Unknown to Patient (n = 3)
Patients with no documented electronic health record evidence of Hispanic ethnicity were classified as non-Hispanic
Values were missing for 2 individuals (1 allocated to patient navigation; 1 allocated to usual care)
Values were missing for 1 individual allocated to usual care
Values were missing for 1 individual allocated to patient navigation
Study Procedures
Patient navigation intervention
We adapted the navigation intervention from the New Hampshire Colorectal Cancer Screening Program, covering six topic areas: introduction and barrier assessment, barrier resolution, bowel preparation instruction, bowel preparation reminder, colonoscopy procedure check-in, and understanding of colonoscopy result and re-testing interval.(26) The original program was delivered by phone by a registered nurse, and included free colonoscopy services. Our program eliminated the clinical credential requirement (our navigators generally had served as a medical assistant or care coordinator with patient-facing experience in a health care setting), allowed text messaging in place of phone calls for topic areas 4 and 5, and relied on available community resources for colonoscopy (Figure 1 in the Supplement).
Figure 1.
CONSORT study flow diagram
Flow diagram for the study showing study enrollment
The study employed one centralized, full-time navigator to serve the enrolled patients. Throughout the duration of the study, nine different staff members served in this role. All navigators were fluent in English and Spanish. Navigators mailed introductory letters and delivered live calls and text messages to patients addressing the six topic areas at scheduled times: introduction and barrier assessment occurred within 1 week of patient identification, barrier resolution followed on an unrestricted timeline, bowel preparation instruction and reminders were delivered up to 1 week before the scheduled colonoscopy, colonoscopy check-in occurred on the day of or 1 day following the scheduled procedure, and a final call occurred 2-4 weeks following the completed procedure. The navigator made up to six attempts between 8:00 a.m. and 6:30 p.m. Monday to Friday (unless the phone number was incorrect or disconnected) to reach patients for each scheduled call. Patients not initially reached were called up to 12 more times and sent a follow-up letter. When text messages were used instead of phone calls, they were sent during the same hours. The study database tracked patient navigation outcomes including patient-reported barriers to colonoscopy and services provided. Navigators used the EHR and Excel (Microsoft Corp., Redmond, WA) to track patient contact attempts, and placed chart notes in the EHR notifying the care team if a patient was not reached. Navigators requested records from gastroenterology facilities for completed colonoscopies not in the chart.
Initial training for patient navigators was the same as in the New Hampshire program: an 8-week schedule of didactic training, role playing, and supervised navigation practice. (27) Following the COVID-19 pandemic, navigator training was condensed and conducted using pre-recorded videos and interactive webinars.
Community Resources
To help patients obtain colonoscopies, patient navigators worked with the research project team to assess and engage with local gastroenterology practices and community organizations providing reduced-cost specialty care for uninsured or underinsured individuals. Patients allocated to navigation were offered free rideshare services for attending pre-procedure or colonoscopy appointments provided by a community grant from Lyft (San Francisco, CA).
Usual care condition
Patients not randomized to navigation received usual care. A centralized referral coordinator contacted patients with a colonoscopy referral to schedule an appointment; patients not initially reached were recontacted 30 and 45 days following the referral date. Up to 2 attempts were made to reach patients at each time point. For patients not reached, a letter was sent to the patient. Patients randomized to the navigation intervention were removed from the referral coordinator list and did not receive usual care.
Primary and Secondary Outcomes
The primary outcome was receipt of colonoscopy (yes/no) documented in the EHR within 1 year following an abnormal FIT result. Secondary outcomes were time to colonoscopy completion (days from date of abnormal FIT result) and bowel preparation adequacy (% adequate), among those who completed a colonoscopy. These data were extracted from Sea Mar’s EHR and confirmed through chart abstraction. Double adjudication was performed for colonoscopy receipt and time to colonoscopy; requests were sent to gastroenterology facilities in cases where colonoscopy was documented in the chart, but the procedure and/or pathology reports were missing. All discrepancies were resolved through consensus. Participants were followed until colonoscopy or for 365 days, whichever came first.
To assess intervention delivery, we report the proportion of intervention participants who received at least one outreach phone call, and the proportion who did not receive the intervention because they were not reached, declined, or were found to be ineligible. All analyses excluded 18 patients who were randomized in error (10 in intervention, 8 in usual care; these patients were determined to be ineligible for colonoscopy following randomization, based on conditions that existed prior to randomization).(28) Patients who were determined to be ineligible for colonoscopy based on clinical conditions diagnosed following randomization were retained in all analysis.
Blinding
Chart abstractors were blinded to randomization assignment. It was not possible to blind navigators or patients to randomization assignment, but as the intervention was delivered as part of standard care, patients were unaware of their study participation. Navigators were unaware of patients’ calculated probability of obtaining a colonoscopy.
Statistical analysis
We used modified intent-to-treat principles for all primary analyses. Participants were analyzed as part of the group to which they were randomized, excluding patients randomized in error. For the primary outcome of colonoscopy completion at 1 year, we performed a multilevel logistic regression with intervention assignment as an independent variable and county as a covariate (randomization variable). We calculated predicted probabilities using the LSMEANS function, which estimates the marginal probabilities over a balanced population. The intra-class correlation (ICC) was negligible (<.0001), reflecting minimal resemblance among individuals in the same clinic; thus, we report estimates from the fixed-effect logistic regression model. All analyses were conducted using SAS 9.4 (SAS Institute Inc. Cary NC). We defined statistical significance at a two-tailed P of 0.05.
For time to colonoscopy completion, we report mean days from abnormal FIT result and 95% confidence intervals, based on restricted mean survival time analysis. The proportional hazard assumption was assessed by 1) plotting the survival function against survival time, using standard and log-transformed data, then 2) adding an interaction term (study condition x survival time) to a model containing study condition and survival time. Kaplan–Meier plots of time to colonoscopy completion were generated. Among individuals who completed a colonoscopy, bowel prep adequacy was reported as the proportion adequate by study condition. Models were adjusted for county.
Heterogeneity of treatment effects
We performed predefined subgroup analyses by sex (male, female), age (50-54, 55-59, 60-64, 65-69, 70-75), language (English, Spanish), county (Clark, Grays Harbor, King, Pierce, Skagit, Snohomish, Thurston, Whatcom), insurance status (Uninsured, Medicaid, Medicare, Commercial), and baseline predicted probability of colonoscopy completion, with predictive-model cutpoints set at tertiles (low: 0 < .52; moderate: .52 < .63; high: ≥.63). We tested for treatment effect heterogeneity by adding the moderator and the product of the moderator and study condition to the primary outcome model; a significant product term, using the Wald chi-square test, would provide support for heterogeneity.
Power calculations
Power calculations assumed 80% power, an ICC of .03, follow-up colonoscopy completion of 44% in usual care, and a two-tailed alpha level of .05.(29, 30) A sample size of 1,200 randomly allocated (1:1) participants in 32 clinics could detect a 12.9 percentage-point difference in follow-up colonoscopy uptake accounting for ICC or a 9.2 percentage-point difference not accounting for ICC.(31, 32) COVID-19 care suspensions led to reductions in the number of study-eligible patients with abnormal FITs; we did not re-run power calculations for our reduced sample size of 985.
Role of funder
The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
RESULTS
We identified 1200 individuals with abnormal FIT results and excluded 215 prior to randomization, leaving 985. We then excluded 18 patients (10 in intervention and 8 in usual care) who were randomized in error; leaving 967 in the primary analysis (intervention n=479; usual care n=488). Baseline characteristics were similar among individuals in the two conditions (Table 1).
Intervention receipt
Among the 479 intervention patients in the primary analysis, 79 (16.5%) declined navigation and 56 (11.7%) were never reached (Figure 1). Thirty-three patients (6.9%) were clinically ineligible for CRC screening after randomization, and additional 16 (3.2%) had moved from the area or left the health system, 5 (1.0%) were deceased and 7 (1.6%) were ineligible for other reasons. The remaining 242 (49.5%) intervention patients received the allocated intervention.
Primary outcome
In modified intention-to-treat analysis, the adjusted percentage who received a colonoscopy in 12 months was 55.1% (95% CI, 50.3 to 59.7) in the patient navigation arm and 42.1% (95% CI, 37.5 to 46.8) in the usual-care arm (risk difference, 13.0 percentage points; 95% CI, 6.5 to 19.4 percentage points; adjusted odds ratio [OR], 1.69; 95% CI, 1.30 to 2.19; P = .0001; Table 2).
Table 2.
Effectiveness of PRECISE Patient Navigation program
| Analytic Sample | N | Outcomes | % Adjusted Difference (95% CI) a | Adjusted OR (95% CI) a | P value | |
|---|---|---|---|---|---|---|
| Patient Navigation Adjusted % (95% CI) a |
Usual Care Adjusted % (95% CI) a |
|||||
| Primary Outcome: Receipt of Colonoscopy within 12 months of abnormal FIT | ||||||
| Modified intention-to-treat b | 967 | 55.1 (50.3, 59.7) | 42.1 (37.5, 46.8) | 13.0 (6.5, 19.4) | 1.69 (1.30, 2.19) | <.001 |
| Secondary Outcome: Adequate Bowel Preparation (12-months) | ||||||
| Modified intention-to-treat c | 452 | 92.4 (87.9, 95.3) | 92.1 (86.9, 95.4) | 0.2 (−4.9, 5.3) | 1.03 (0.51, 2.09) | -- |
adjusted for county
Excludes patients who were randomized in error (10 in patient navigation, 8 in usual care); 479 allocated to intervention; 488 allocated to usual care
Within 12 months; excludes 19 patients with missing data
Treatment-effect heterogeneity
Treatment-effect heterogeneity was not significant for any variable we tested: age, sex, ethnicity, preferred language, insurance status, county of residence, nor predicted probability of follow-up colonoscopy receipt (Supplemental Table 3). For those with the lowest probability, navigation boosted follow-up colonoscopy receipt by 2.7 percentage point (46.1% vs. 43.4%), versus 15.5 percentage point (58.2% vs. 42.7%) in the moderate probability, and 14.8 percentage point (63.4% vs. 48.8%) in the high probability groups (test for treatment effect heterogeneity, P = .16).
Secondary outcomes and sensitivity analyses
Among the 242 individuals who received the allocated navigation intervention, 227 (94%) completed a follow-up colonoscopy. Among individuals who received a colonoscopy, bowel preparation adequacy was similar in both groups (92.5% vs. 91.6%); mean time to colonoscopy was 229 days in the intervention group (95% CI, 217-241 days), and 256 days in the usual care group (95% CI, 244-268; Figure 2). Among 471 individuals who completed a colonoscopy within 12 months across study conditions, 225 (47.8%) had a low-risk adenoma, 64 (13.6%) had an advanced adenoma, and 7 (1.5%) were found to have cancer.
Figure 2.
Cumulative incidence of colonoscopy across groups defined by probability of obtaining a colonoscopy without navigation
The graphs show cumulative incidence of colonoscopy, which are estimated using the Kaplan-Meier estimator. Between-group (intervention vs. usual care) differences in follow-up colonoscopy completion were overall: 13.0 percentage points (55.1% vs 42.1%); probability band <52%: 2.7 percentage points (46.1% vs. 43.4%); probability band 52% <63%: 15.5 percentage points (58.2% vs. 42.7%); probability band 63%+: 14.8 percentage points (63.4% vs. 48.8%; test for treatment effect heterogeneity, P = .16). Time to colonoscopy (days) calculated using restricted mean survival estimates with 95% confidence intervals by study conditions (patient navigation versus usual care), are reported overall and by probability band plots. Overall estimates were: 229 days (95% CI, 217-241 days) for patient navigation; 256 days (95% CI, 244-268 days) for usual care. Probability band <52%: 258 days (95% CI, 239-278 days) for patient navigation; 264 days (95% CI, 243-284 days) for usual care; probability band 52% <63%: 221 days (95% CI, 199-242 days) for patient navigation; 260 days (95% CI, 239-280 days) for usual care; probability band 63%+: 207 days (95% CI, 185-228 days) for patient navigation; 247 days (95% CI, 226-267 days) for usual care. One person is missing because of incomplete information to calculate the probability score.
DISCUSSION
Patient navigation resulted in a significant 13 percentage-point (95% CI, 6.5 to 19.4 percentage points) boost in 1-year follow-up colonoscopy completion over usual care in this randomized trial in a community health center setting. Patients allocated to navigation who completed a colonoscopy did so within a mean 229 days, 27 days shorter than usual care patients. Contrary to expectation, program effectiveness did not differ by patients’ probability of obtaining a colonoscopy without navigation, derived using health record data (test for treatment effect heterogeneity, P = .16).
Our team previously applied a follow-up colonoscopy probability score in an evaluation of a patient navigation program delivered in an integrated care setting.(20) That study, which included 415 patients with the lowest probability scores (70% or lower likelihood of obtaining a colonoscopy without navigation), reported an 11 percentage-point improvement in 6-month colonoscopy adherence for those allocated to navigation versus usual care.(20) The study targeted navigation to 495 of 1,962 patients with an abnormal FIT result, demonstrating efficiencies that could be gained. The model showed promise for identifying subgroups with the greatest need and who could benefit from navigation. Neither this previous study nor the current study were powered to show a statistically significant treatment effect heterogeneity. More research is needed to assess whether risk-stratification methods can effectively be used to direct navigation services.
Notably, 20.3% of patients could not be contacted or were lost to follow-up by the navigator, and an additional 29.7% did not receive navigation for various reasons, such as declining participation, being ineligible for colonoscopy, death, moving and/or transferring care. This underscores the complex challenges to providing navigation in this setting. Among the 242 patients who received the allocated intervention, 94.2% completed the colonoscopy procedure; this compared to 14.4% among the 247 who did not receive the intervention (P < .0001). This demonstrates the value of patient navigation among colonoscopy-eligible patients who can be contacted and agree to receive navigation services.
Several strategies could be useful for further enhancing the reach and effectiveness of patient navigation programs. A national quality measure for colonoscopy follow-up, as supported by several national organizations, could spur meaningful improvements.(6) Additionally, a prioritized scheduling approach that expedites appointments for individuals with high-risk indications – such as the occurrence of symptoms, abnormal test results, or risk factors -- may facilitate timely follow-up for patients with concerning screening outcomes, extending COVID-era recommendations.(33) This approach is met with high patient acceptability.(34) Novel care models that co-locate gastroenterology providers (clinician or nurse) in community health center practices to perform pre-procedure visits and facilitate scheduling could be explored in future research.
Securing sustainable funding for patient navigation will be important for its long-term viability and impact.(35) In 2023, the White House mandated reimbursement for navigation services for patients with a cancer diagnosis, but did not address navigation for screening or follow-up after an abnormal test result.(36) Some national organizations have advocated for broader coverage.(37) Ensuring adequate and reliable funding sources will be essential for these programs to reach their full potential. This may require demonstrating the potential for cost savings achieved through down-staged CRC detection and a reduction in missed/canceled gastroenterology appointments.(38)
Our findings suggest that patient navigation is highly effective for patients who are colonoscopy-eligible. Nevertheless, the absolute improvement we observed (55% vs. 42%; 13 percentage point difference, P <.001) was substantially smaller than that reported by the original New Hampshire program (96% vs. 69%; 27 percentage point difference, P<.001).(26) Several factors likely explain this discrepancy. The original evaluation relied on a small, non-randomized sample (n=206) of patients ages 50-64 who had an appointment for a colonoscopy. The program provided no-cost colonoscopy services using funds from the Centers for Disease Control and Prevention. The present study used a randomized design, intention-to-treat analysis, and included patients ages 50-75 who had an abnormal FIT result (irrespective of having a colonoscopy appointment). As a result, our findings are more robust and generalizable to a range of populations. Moreover, our improvement was consistent with prior small, randomized studies on this topic, reporting improvements of 10 and 21 percentage points with navigation, at 180 days and 1-year, respectively.(14, 15) Further, our observed 94.2% colonoscopy completion proportion among patients who received patient navigation approximates the findings from the original evaluation.
Strengths and limitations
Our study’s sample size was two times larger than any previous study of patient navigation for follow-up colonoscopy and included a diverse patient population. Navigation was delivered by clinic staff, primarily during standard clinic hours, increasing applicability of findings to real-world clinic settings. Among our study limitation is the fact that our program’s implementation occurred during the COVID-19 pandemic, which likely impacted its effectiveness. Care suspensions, policy changes (making, pre-procedure COVID-19 testing, etc.), combined with high navigator turnover and condensed, remote training likely depressed colonoscopy participation.(39, 40) Moreover, patient fears and hesitancies during this time likely created additional barriers to obtaining timely colonoscopy. Despite our program’s success, the long time to colonoscopy (more than 220 days overall) suggests that additional effort may be needed to expedite colonoscopy receipt to avoid worsening clinical outcomes. Finally, our program was implemented in a single, large FQHC, and further adaptations may be needed to successfully scale-up the program across small practices.
Conclusion
Patient navigation improved rates of colonoscopy follow up after abnormal FIT among patients at a large FQHC.
Supplementary Material
Funding.
Research reported in this publication was supported by the National Cancer Institute through award number NCI R01 CA218923. All authors received funding from this grant through their respective institutions. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute. The National Cancer Institute had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Conflicts of Interest and Financial Disclosures
From 2021 – 2023, Dr. Coronado served as PI on a contract funded by Guardant Health to the Kaiser Permanente Center for Health Research that assessed adherence to a commercially available blood test for colorectal cancer. All other authors declare no conflicts of interest.
Footnotes
Access to Data and Data Analysis. MCL and MS had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Research Reproducibility Statement.
Protocol: posted as data supplement at annals website available at www.authorswebsite.org
Statistical Code: Available to interested readers by contacting Dr. Coronado at gdcoronado@arizona.edu
Data: not available
Data Sharing Statement.
We do not have authorization to distribute or share externally the patient-level data used in this manuscript.
References
- 1.Zauber AG. The impact of screening on colorectal cancer mortality and incidence: has it really made a difference? Dig Dis Sci. 2015;60(3):681–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Levin TR, Corley DA, Jensen CD, Schottinger JE, Quinn VP, Zauber AG, et al. Effects of Organized Colorectal Cancer Screening on Cancer Incidence and Mortality in a Large Community-Based Population. Gastroenterology. 2018;155(5):1383–91.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Robertson DJ, Lee JK, Boland CR, Dominitz JA, Giardiello FM, Johnson DA, et al. Recommendations on fecal immunochemical testing to screen for colorectal neoplasia: a consensus statement by the US Multi-Society Task Force on colorectal cancer. Gastrointest Endosc. 2017;152(5):1217–37. [DOI] [PubMed] [Google Scholar]
- 4.Selby K, Senore C, Wong M, May FP, Gupta S, Liang PS. Interventions to ensure follow-up of positive fecal immunochemical tests: An international survey of screening programs. J Med Screen. 2021;28(1):51–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mohl JT, Ciemins EL, Miller-Wilson LA, Gillen A, Luo R, Colangelo F. Rates of Follow-up Colonoscopy After a Positive Stool-Based Screening Test Result for Colorectal Cancer Among Health Care Organizations in the US, 2017–2020. JAMA Netw Open. 2023;6(1):e2251384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ciemins EL, Mohl JT, Moreno CA, Colangelo F, Smith RA, Barton M. Development of a Follow-Up Measure to Ensure Complete Screening for Colorectal Cancer. JAMA Netw Open. 2024;7(3):e242693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bharti B, May FFP, Nodora J, Martinez ME, Moyano K, Davis SL, et al. Diagnostic colonoscopy completion after abnormal fecal immunochemical testing and quality of tests used at 8 Federally Qualified Health Centers in Southern California: Opportunities for improving screening outcomes. Cancer. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Corley DA, Jensen CD, Quinn VP, Doubeni CA, Zauber AG, Lee JK, et al. Association Between Time to Colonoscopy After a Positive Fecal Test Result and Risk of Colorectal Cancer and Cancer Stage at Diagnosis. JAMA. 2017;317(16):1631–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lee YC, Fann JC, Chiang TH, Chuang SL, Chen SL, Chiu HM, et al. Time to Colonoscopy and Risk of Colorectal Cancer in Patients With Positive Results From Fecal Immunochemical Tests. Clin Gastroenterol Hepatol. 2019;17(7):1332–40.e3. [DOI] [PubMed] [Google Scholar]
- 10.US Preventive Services Task Force 2022;Pages https://www.thecommunityguide.org/news/cpstf-recommends-patient-navigation-services-increase-cancer-screening-advance-health-equity.html.
- 11.Nelson HD, Cantor A, Wagner J, Jungbauer R, Fu R, Kondo K, et al. Effectiveness of Patient Navigation to Increase Cancer Screening in Populations Adversely Affected by Health Disparities: a Meta-analysis. Journal of General Internal Medicine. 2020;35(10):3026–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Knudsen AB, Rutter CM, Peterse EFP, Lietz AP, Seguin CL, Meester RGS, et al. Colorectal Cancer Screening: An Updated Modeling Study for the US Preventive Services Task Force. Jama. 2021;325(19):1998–2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mohan BP, Khan SR, Daugherty E, Chandan S, Ponnada S, Facciorusso A, et al. Pooled rates of adenoma detection by colonoscopy in asymptomatic average-risk individuals with positive fecal immunochemical test: a systematic review and meta-analysis. Gastrointest Endosc. 2022;96(2):208–22.e14. [DOI] [PubMed] [Google Scholar]
- 14.Green BB, Anderson ML, Wang C-Y, Vernon SW, Chubak J, Meenan RT, Fuller S. Results of Nurse Navigator Follow-up After Positive Colorectal Cancer Screening Test: A Randomized Trial. The Journal of the American Board of Family Medicine. 2014;27(6):789–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Raich PC, Whitley EM, Thorland W, Valverde P, Fairclough D. Patient navigation improves cancer diagnostic resolution: an individually randomized clinical trial in an underserved population. Cancer Epidemiol Biomarkers Prev. 2012;21(10):1629–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Selby K, Baumgartner C, Levin TR, Doubeni CA, Zauber AG, Schottinger J, et al. Interventions to Improve Follow-up of Positive Results on Fecal Blood Tests: A Systematic Review. Annals of Internal Medicine. 2017;167(8):565–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Dougherty MK, Brenner AT, Crockett SD, Gupta S, Wheeler SB, Coker-Schwimmer M, et al. Evaluation of Interventions Intended to Increase Colorectal Cancer Screening Rates in the United States: A Systematic Review and Meta-analysis. JAMA Intern Med. 2018;178(12):1645–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bernardo BM, Zhang X, Beverly Hery CM, Meadows RJ, Paskett ED. The efficacy and cost-effectiveness of patient navigation programs across the cancer continuum: A systematic review. Cancer. 2019;125(16):2747–61. [DOI] [PubMed] [Google Scholar]
- 19.Petrik AF, Johnson ES, Slaughter M, Leo MC, Thompson J, Mummadi RR, et al. The recalibration and redevelopment of a model to calculate patients’ probability of completing a colonoscopy following an abnormal fecal test. J Med Screen. 2023:9691413231195568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Coronado GD, Rawlings AM, Petrik AF, Slaughter M, Johnson ES, Hannon PA, et al. Precision patient navigation to improve rates of follow-up colonoscopy, an individual randomized effectiveness trial. Cancer Epidemiol Biomarkers Prev. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kent DM, Steyerberg E, van Klaveren D. Personalized evidence based medicine: predictive approaches to heterogeneous treatment effects. Bmj. 2018;363:k4245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kent DM, Nelson J, Dahabreh IJ, Rothwell PM, Altman DG, Hayward RA. Risk and treatment effect heterogeneity: re-analysis of individual participant data from 32 large clinical trials. Int J Epidemiol. 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ltd. SE 2019;Pages https://www.sealedenvelope.com/simple-randomiser/v1/lists on June 3 2019.
- 24.Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O’Neal L, et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform. 2019;95:103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rice K, Gressard L, DeGroff A, Gersten J, Robie J, Leadbetter S, et al. Increasing colonoscopy screening in disparate populations: Results from an evaluation of patient navigation in the New Hampshire Colorectal Cancer Screening Program. Cancer. 2017;123:3356–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Centers for Disease Control and Prevention NCfCDPaHP. New Hampshire Colorectal Cancer Screening Program. Patient Navigation Model for Increasing Colonoscopy Quality and Completion: A Replication Manual. 2016. [Google Scholar]
- 28.Tong G LC, Coronado GD, Li F. Randomized in Error in Pragmatic Clinical Trials Contemporary Clinical Trials (in press). 2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Coronado GD, Johnson ES, Leo MC, Schneider JL, Smith D, Mummadi R, et al. Patient randomized trial of a targeted navigation program to improve rates of follow-up colonoscopy in community health centers. Contemp Clin Trials. 2019;89:105920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Coronado GD, Kihn-Stang A, Slaughter MT, Petrik AF, Thompson JH, Rivelli JS, et al. Follow-up colonoscopy after an abnormal stool-based colorectal cancer screening result: analysis of steps in the colonoscopy completion process. BMC Gastroenterol. 2021;21(1):356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Demidenko E Sample size determination for logistic regression revisited. Statistics in Medicine; 2007:3385–97. [DOI] [PubMed] [Google Scholar]
- 32.Hsieh F, Block D, Larsen M. A simple method of sample size calculation for linear and regression. Statistics in Medicine; 1998:1623–34. [DOI] [PubMed] [Google Scholar]
- 33.American Society for Gastrointestinal Endoscopy 2021;Pageshttps://www.asge.org/docs/default-source/default-document-library/guidance-for-gi-endoscopy-and-practice-operations-during-the-covid-19-pandemic_updated_final-march-2021.pdf on December 16 2024.
- 34.Telford JJ, Rosenfeld G, Thakkar S, Bansback N. Patients’ Experiences and Priorities for Accessing Gastroenterology Care. J Can Assoc Gastroenterol. 2021;4(1):3–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Dwyer AJ, Weltzien ES, Harty NM, LeGrice KE, Pray SLH, Risendal BC. What makes for successful patient navigation implementation in cancer prevention and screening programs using an evaluation and sustainability framework. Cancer. 2022;128 Suppl 13:2636–48. [DOI] [PubMed] [Google Scholar]
- 36.Policy. TWHOoSaT. Cancer Navigation Improves Equity & Outcomes – Paying for It Matters, Too. 2023. [Google Scholar]
- 37.Paskett ED, Battaglia T, Calhoun EA, Chappell MC, Dwyer A, Fleisher LG, et al. Isn’t there enough evidence on the benefits of patient navigation? CA Cancer J Clin. 2023;73(6):562–4. [DOI] [PubMed] [Google Scholar]
- 38.Rice K, Sharma K, Li C, Butterly L, Gersten J, DeGroff A. Cost-effectiveness of a patient navigation intervention to increase colonoscopy screening among low-income adults in New Hampshire. Cancer. 2019;125(4):601–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Patel S, Issaka RB, Chen E, Somsouk M. Colorectal Cancer Screening and COVID-19. Am J Gastroenterol. 2021;116(2):433–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Shaukat A, Church T. Colorectal cancer screening in the USA in the wake of COVID-19. The Lancet. Gastroenterology & Hepatology. 2020;5(8):726–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
We do not have authorization to distribute or share externally the patient-level data used in this manuscript.