The longitudinal impact of patient navigation on equity in colorectal cancer screening in a large primary care network

Authors

  • Sanja Percac-Lima MD, PhD,

    Corresponding author
    1. Massachusetts General Hospital Chelsea HealthCare Center, Chelsea, Massachusetts
    2. General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts
    • Corresponding author: Sanja Percac-Lima MD, PhD, Massachusetts General Hospital Chelsea HealthCare Center, 151 Everett Avenue, Chelsea, MA 02150; Fax: (617) 889-8579; spercaclima@partners.org

    Search for more papers by this author
  • Lenny López MD, MPH, MDiv,

    1. General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts
    2. Mongan Institute for Health Policy and Disparities Solutions Center, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author
  • Jeffrey M. Ashburner MPH,

    1. General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author
  • Alexander R. Green MD, MPH,

    1. General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts
    2. Mongan Institute for Health Policy and Disparities Solutions Center, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author
  • Steven J. Atlas MD, MPH

    1. General Medicine Division, Massachusetts General Hospital, Boston, Massachusetts
    Search for more papers by this author

  • We thank Sarah A. Oo, the Director of the Community Health Team at the Massachusetts General Hospital Chelsea HealthCare Center, and the patient navigators for their support and work on the program: without them, this study would not have been possible.

Abstract

BACKGROUND

The long-term effects of interventions to improve colorectal (CRC) screening in vulnerable populations are uncertain. The authors evaluated the impact of patient navigation (PN) on the equity of CRC prevention over a 5-year period.

METHODS

A culturally tailored CRC screening PN program was implemented in 1 community health center (CHC) in 2007. In a primary care network, CRC screening rates from 2006 to 2010 among eligible patients from the CHC with PN were compared with the rates from other practices without PN. Multivariable logistic regression models for repeated measures were used to assess differences over time.

RESULTS

Differences in CRC screening rates diminished among patients at the CHC with PN and at other practices between 2006 (49.2% vs 62.5%, respectively; P < .001) and 2010 (69.2% vs 73.6%, respectively; P < .001). The adjusted rate of increase over time was higher at the CHC versus other practices (5% vs 3.4% per year; P < .001). Among Latino patients at the CHC compared with other practices, lower CRC screening rates in 2006 (47.5% vs 52.1%, respectively; P = .02) were higher by 2010 (73.5% vs 67.3%, respectively; P < .001). Similar CRC screening rates among non-English speakers at the CHC and at other practices in 2006 (44.3% vs 44.7%, respectively; P = .79) were higher at the CHC by 2010 (70.6% vs 58.6%, respectively; P < .001). Adjusted screening rates increased more over time for Latino and non-English speakers at the CHC compared with other practices (both P < .001).

CONCLUSIONS

A PN program increased CRC screening rates in a CHC and improved equity in vulnerable patients. Long-term funding of PN programs has the potential to reduce cancer screening disparities. Cancer 2014;120:2025–2031. © 2014 American Cancer Society.

INTRODUCTION

The incidence of colorectal cancer (CRC) is declining, but CRC remains the third most commonly diagnosed cancer and the second leading cause of cancer death in the United States.[1-3] Although CRC-related morbidity and mortality can be reduced through early detection and treatment, almost two-thirds of patients are diagnosed with advanced-stage disease.[1] Compared with other preventable cancers, CRC screening rates are lower,[4] particularly for racial/ethnic minorities, non-English speakers, and low-income individuals.[5, 6] When diagnosed with CRC, these vulnerable patient groups present with cancer at later stages and have higher mortality rates than white, English-speaking, and higher income patients.[7-10]

These observed differences are because of patient, physician, and health system barriers.[11, 12] Factors particularly relevant for low-income patients and those with limited English proficiency may include: insufficient time to effectively discuss CRC screening during clinic visits, complicated scheduling processes, financial difficulties, health insurance, problems with transportation, and/or lack of interpreters at procedure sites.[13, 14] A variety of strategies to improve CRC screening, particularly in vulnerable populations, have been evaluated, including patient navigation (PN).[15-26] Implemented in 1990 by Freeman and colleagues,[27, 28] PN uses culturally and linguistically matched outreach workers to explore patients' barriers to cancer care and to help them navigate the complexities of the health care system to obtain necessary care.[29, 30] Several studies have demonstrated that PN can significantly improve CRC screening rates.[31-37] However, there is little information about the sustainability and effectiveness of these programs over time. Also, few studies have directly measured the impact of PN programs in CRC on disparities and health equity.[38]

In 2005, analyses examining CRC screening rates among patients who received care in primary care practices associated with Massachusetts General Hospital (MGH) identified lower rates in Latino patients versus non-Latino white patients (41% vs 61%). Among MGH primary care practices, the Chelsea HealthCare Center (MGH Chelsea) had the lowest screening rates (35% among Latinos and 58% overall). To better understand these disparities, we conducted interviews with MGH Chelsea patients[13]; and, based on those findings, in January 2007, we implemented a culturally tailored, multifaceted CRC screening PN program at MGH Chelsea. Over a 9-month randomized controlled trial, patients in the PN group had significantly higher CRC screening rates compared with a usual care group (27% vs 12% for any screening [P < .001]; 21% vs 10% for colonoscopy completion [P < .001]).[39]

Based on these results, the CRC PN screening program was continued for all eligible patients at MGH Chelsea. The objective of the current study was to evaluate the impact of this CRC screening PN program on improving equity in CRC screening at MGH over a 5-year period.

MATERIALS AND METHODS

Setting and Participants

The MGH Practice-Based Research Network (PBRN) consists of 16 affiliated primary care practices that share an administrative and information technology infrastructure. One of 3 community health centers within this network, MGH Chelsea serves a low-income, predominantly Latino and immigrant population. Since the completion of the randomized control trial, all eligible patients who are overdue for CRC screening at MGH Chelsea are contacted by a patient navigator. We evaluated the PN CRC screening program over a 5-year period starting in 2006, 1 year before the initiation of the program, and continuing through the end of 2010. During this study period, there were no other PN or other quality-improvement programs that targeted CRC screening within the MGH PBRN. All study activities were approved by the Partners Internal Review Board. This included a waiver of consent, because PN is considered usual care in our institution.

Intervention

Details of navigator training and the intervention have been described previously.[39] The same college-educated, bilingual, Spanish/English-speaking Latina woman continued to work as the main CRC screening navigator during the entire study period. In 2010, a 0.5 full-time equivalent, Serbo-Croatian/English-speaking, college-educated male navigator was added. In addition, trained interpreters covering 6 other languages were involved in the PN program throughout the study period. The navigators were supervised by the study principal investigator and the community health director. The average cost of the program was $75,000 per year.

At the completion of the randomized trial, the navigator was provided an updated list of all CRC screening-eligible patients at MGH Chelsea who were overdue for testing. This list was updated every January. Overdue patients were sent introductory letters in their native language that explained the project and educational materials related to CRC screening. A week later, a navigator would call or meet the patient in the health center. Up to 7 telephone call attempts were made at different times of the day over 3 months. Patients who could not be reached were deferred for the remainder of the year but, if still overdue, would be present on the subsequent year's list for additional contact attempts. During the initial contact, the navigator educated the patient about CRC screening and explored the patient's barriers to screening. Further interventions were tailored to each individual patient with the intention of overcoming personal, cultural, and systemic barriers to successfully complete screening. Specific interventions included reviewing all available CRC screening methods, helping schedule appointments, reminding patients about appointments, reviewing and translating the required bowel preparation, organizing transportation, and accompanying patients without available family members. The navigator updated the patient's medical record with the CRC screening results. Navigators encouraged patients to undergo colonoscopy, which is the preferred CRC screening method at MGH. However, if patients preferred another CRC screening method, then the navigator also assisted them in getting screened through alternate methods.[39]

Outcomes and Analyses

Patient characteristics and CRC screening data were obtained from an electronic central data repository at Partners Healthcare.[40] A completed colonoscopy within the prior 10 years or a sigmoidoscopy, colonography, or barium enema within the prior 5 years was considered evidence of completed preventive CRC screening. Fecal occult blood testing (FOBT) and fecal immunochemical testing (FIT) were not included because of the poor reliability of FOBT data and because FIT is not used in our network. The primary outcome was the CRC screening rate assessed during each of the 5 calendar-year study periods.

To assess the impact of the PN program on CRC screening rates over time, we performed 4 sets of analyses. First, we compared CRC screening rates in patients who received care at MGH Chelsea with the rates in patients who received care at other MGH PBRN practices from 2006 (the year before the start of the PN program) through 2010. Second, because the PN program was designed to address low CRC screening rates in Latino and non-English–speaking patients, we examined outcomes over time in Latino patients who received care at MGH Chelsea with the rates of Latinos who received care at other practices. Similarly, we compared non-English–speaking patients at MGH Chelsea with non-English–speaking patients at other practices. Third, to assess the impact of the PN program on equity of care within vulnerable patient subgroups at MGH Chelsea, CRC screening rates over time were compared among Latino and non-Latino patients and among non-English–speaking and English-speaking patients. Finally, we examined the annual change in screening rates over time for all patients, for Latino patients, and for non-English–speaking patients and compared differences between those who received care at MGH Chelsea and those who received care at other practices.

Descriptive statistics of the study population were compared across groups using 2-tailed chi-square tests for categorical variables and using 1-way analyses of variance for continuous variables. Unadjusted CRC screening rates in individual years were compared using chi-square tests among all patients from MGH Chelsea versus patients from other practices and in separate analyses limited to Latino patients and non-English–speaking patients. To assess whether the rate of change over time in CRC screening differed between patients at MGH Chelsea and patients at other non-PN practices, adjusted slopes were calculated using an interaction term between practice site and calendar year with logistic regression models for repeated measures that adjusted for age, sex, race/ethnicity, language spoken, insurance, and number of clinic visits (Proc Genmod procedure in SAS; SAS Inc., Cary, NC). Similar models also were performed to compare rates of screening over time in Latino patients and non-English–speaking patients who received care at MGH Chelsea with those who received care at other practices.

RESULTS

There were 3115 patients eligible for CRC screening at MGH Chelsea and 43,905 patients in the other practices (Table 1). Eligible patients from MGH Chelsea were less likely to speak English (59.6% vs 95.8%), to be white (49.6% vs 86.3%), to have commercial health insurance (36.5% vs 62%), and had more clinic visits over the prior 3 years (10.3 visits vs 7.9 visits) compared with patients from other practices. There were no significant differences in age or sex between patients from MGH Chelsea and patients from other practices.

Table 1. Characteristics of Patients Overdue for Colorectal Cancer Screening
 Percentage of Patients 
Patient CharacteristicsChelsea, n = 3115Other Practices,n = 43,905P
  1. Abbreviations: SD, standard deviation.

  2. a

    Free care (including colonoscopy coverage) is provided by some hospitals in Massachusetts to low-income patients who do not have insurance and are not eligible for Medicaid.

Age: Mean ± SD, y61.4 ± 6.761.6 ± 6.5.09
Sex: Women57.155.4.06
Language: English59.695.8<.001
Race/ethnicity  <.001
Asian1.83.3 
Black5.24.8 
Latino39.52.3 
Other/unknown43.3 
White49.686.3 
Insurance status  <.001
Commercial36.562 
Medicaid17.57 
Medicare3327.6 
Free carea10.92.2 
Self-pay2.11.3 
No. of clinic visits over 3 y: Mean ± SD10.3 ± 7.07.9 ± 6.2<.001

In 2006, before the CRC screening PN program was established, 49.2% of all eligible patients at MGH Chelsea were up to date for CRC screening compared with 62.5% of patients who received care at other practices. In 2010, the CRC screening rate at MGH Chelsea was 69.2% compared with 73.6% (P < .001) at practices without PN programs (Fig. 1A). Latino patients at MGH Chelsea had lower CRC screening rates in 2006 compared with Latino patients at other practices (47.5% vs 52.1%; P = .02), but they had higher screening rates in 2010 (73.5% vs 67.3%; P < .001) (Fig. 1B). Non-English–speaking patients at MGH Chelsea had similar CRC screening rates in 2006 compared with non-English–speaking patients at other practices (44.3% vs 44.7%, respectively; P = .79), but they had higher screening rates in 2010 (70.6% vs 58.6%; P < .001) (Fig. 1C).

Figure 1.

Colorectal cancer screening rates are illustrated over time (A) among all patients at the Massachusetts General Hospital Chelsea HealthCare Center (MGH Chelsea) compared with all patients at other practices without patient navigation, (B) among Latino patients at MGH Chelsea compared with Latino patients at other practices without patient navigation, and (C) among non-English–speaking patients at MGH Chelsea compared with non-English–speaking patients at other practices without patient navigation.

Figure 2A illustrates the comparison in CRC screening in Latino and non-Latino patients who received care at MGH Chelsea over 5 years. In 2006, before implementation of the PN program, the screening rate in MGH Chelsea Latino patients was 47.5% compared with 50.4% in non-Latino patients (P = .09). In 2010, the rate in MGH Chelsea Latinos was significantly higher than the rate in non-Latinos (73.5% vs 66%; P < .001). In 2006, the CRC screening rate among non-English–speaking patients at MGH Chelsea was significantly lower than the screening rate among English-speaking patients at MGH Chelsea (44.3% vs 52.6%; P < .001). At the end of 2010, the screening rate in non-English–speaking patients was slightly higher than the rate among English speakers (70.6% vs 68%; P = .09) (Fig. 2B).

Figure 2.

Colorectal cancer screening rates are illustrated over time among (A) Latino and non-Latino patients at the Massachusetts General Hospital Chelsea HealthCare Center (MGH Chelsea) and (B) English-speaking and non-English–speaking patients at MGH Chelsea.

Among all patients, the adjusted rate of increase in CRC screening was higher in patients at MGH Chelsea compared with patients at other practices (5.2% vs 3.3% per year; difference, 1.9% per year; P < .001). In addition, the adjusted rate of increase was higher for Latino patients at MGH Chelsea compared with Latino patients at other practices (7.1% vs 3.4% per year; difference, 3.7% per year; P < .001) and non-English–speaking patients (7% vs 3.4% per year; difference, 3.6% per year; P < .001).

DISCUSSION

To our knowledge, this is the first study to evaluate the impact of a culturally tailored CRC screening PN program on equity in CRC screening in an academic primary care network over a 5-year period. At a heavily immigrant, non-English–speaking, and lower socioeconomic community health center with a CRC PN program, the rate of increase in CRC screenings was significantly higher than in other network practices without PN programs. By the end of the study period, the disparity in screening between the practice with a PN program and the other network practices was nearly eliminated. In addition, the impact of the PN program was strongest and was sustained for Latino and non-English–speaking patients over the 5-year period.

Our study demonstrated that using PN at a single practice in a primary care network can correct some disparities in care, but it may also create new ones. Non-English–speaking and Latino patients in practices without PN had significantly lower screening rates by the end of the study period, and the gap actually increased over time. Approximately 50% of Latino patients who receive care within the network are seen at MGH Chelsea, whereas the other half are scattered within other health centers and private practices.

After just 1 year of PN at MGH Chelsea, CRC screening rates in Latino patients were higher than in non-Latino patients at the health center. After 4 years of the program, the rates of screening among Latinos at the community health center with PN were as high as the rates among patients who received care in private practices within the primary care network. Although all patients who were overdue for CRC screening at MGH Chelsea were offered PN, our primary navigator was a bilingual Latina woman, which made the program more culturally tailored to the Latino population. This is comparable to most current PN programs that are targeted to specific, vulnerable populations.[33, 35] Our study demonstrates that, if these PN programs continue over time, then the disparities in care could be significantly decreased.

Our study has several important implications. First, at a relatively low cost compared with other intensive interventions, PN programs have the potential to effectively eliminate disparities in CRC screening.[37, 41] Developing and testing strategies to eliminate disparities is an active area of research. There are few consistently proven disparity-reduction strategies, but methods that bypass the physician to offer preventive services directly to patients independent of physician involvement have demonstrated promise.[42] The use of PN for CRC screening fits this paradigm, and our current study provides evidence of its sustained and long-term beneficial impact.

Second, although there are many patient, physician, and health care systems barriers for CRC screening, the PN model is sufficiently versatile and personal to allow individual patient-centered tailoring—something not easily done with other standardized methods like letters or simple telephone call reminders. Patient centeredness is 1 of the 6 pillars of high-quality and efficient care as defined by the Institute of Medicine report Crossing the Quality Chasm.[43] There are several dimensions of patient-centered care that are integral to PN. PN allows for respect of patients' values, preferences, and expressed needs; coordination and integration of care; and improved patient communication and education.[43]

Third, the most recent National Healthcare Disparities Report demonstrates many areas with continued racial/ethnic disparities that might be amenable to the PN model.[44] Our study demonstrated the long-term beneficial impact of PN in CRC screening; thus, a broader application of the PN model may be an essential component in achieving equity in care.

Finally, accountable care organizations (ACOs) are an integral part of the restructuring of health care delivery with the implementation of the Affordable Care Act. An ACO is a provider-led organization whose mission is to manage the full continuum of care and to be accountable for the overall costs and quality of care for a defined population.[45] The ACO goal is integrated and efficient care by fostering local organizational accountability for quality and costs through performance measurement. Equity is an essential part of this mission. Currently, most CRC screening PN programs are located in health centers that serve predominately vulnerable patients.[24, 33] To be able to reach out to all vulnerable patients within large primary care networks, information technology systems need to be developed that would identify high-risk patients who may benefit from PN for CRC screening.[46] PN programs could have an important role in ACOs. Currently, PN programs are supported primarily by grant funding[22, 34-37]; however, in this new model, ACOs would invest in PN programs and, over time, reduce costs and improve the quality and equity of care.

Several important limitations warrant consideration. PN was offered in only 1 health center within an academic primary care network; therefore, the results may not be generalizable to other clinical settings. Some patients may have obtained CRC screening outside our system. Our data do not include FOBT and FIT testing. However, had we included these tests, the rates in the navigated patients might have been higher, because navigators were trained to offer FOBT to all patients who refused to have a colonoscopy. Patients at the health center who received PN differed from those at other practices, and factors that were not accounted for in the multivariate models we used to adjust for differences may have underestimated the impact of the PN program.

In conclusion, a culturally tailored PN program designed to identify and overcome barriers to CRC screening not only improved CRC screening rates but also decreased disparities in CRC prevention in vulnerable populations. Further work is needed to design PN programs that can reach out to all vulnerable patients within a primary care network. The long-term support of PN programs can improve equity in cancer screening.

FUNDING SUPPORT

The patient navigator program was supported by a Clinical Innovation Award through Massachusetts General Hospital (CDC SIP 16), the Massachusetts Cancer Prevention Community Research Network, and the Trefler Foundation. Drs. Percac-Lima and Atlas were supported in part by a grant from the Agency for Healthcare Research and Quality (AHRQ R18-HS018161).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary