Effectiveness of a clinic-based colorectal cancer screening promotion program for underserved Hispanics
Abstract
BACKGROUND:
Hispanics in the United States are less likely than other groups to receive screening services for colorectal cancer.
METHODS:
The authors conducted a clinic-based individual randomized trial that enrolled Hispanic patients ages 50 to 79 years who had been seen in the Seattle-based community clinic in the past 5 years. A total of 501 patients met the eligibility criteria and were randomized to 1 of 3 conditions: 1) usual care; 2) mailed fecal occult blood test (FOBT) card and instructions on how to complete the test (mailed FOBT only); and 3) mailed FOBT card and instructions on how to complete the test, telephone reminders, and home visits (mailed FOBT and outreach). The authors assessed postintervention differences in rates of FOBT screening in intervention and usual care groups using computerized medical records reviewed from June 2007 to March 2008.
RESULTS:
Data analysis occurred between November 2008 and September 2009. Nine-month postintervention screening rates were 26% among patients who received the mailed packet only intervention (P < .001 compared with usual care) and 31% in the group that received the mailed packet and outreach intervention (P < .001 compared with usual care). This compared with 2% in the group that received usual care. Screening rates in the mailed FOBT only group and in the mailed FOBT and outreach group were not significantly different (P = .28).
CONCLUSIONS:
Culturally appropriate clinic-based interventions may increase colorectal cancer screening among underserved Hispanics. Cancer 2011. © 2010 American Cancer Society.
Compared with non-Hispanic whites, Hispanics in the United States (US) experience lower incidence of colorectal cancer (Hispanic males, 50.0 per 100,000 vs non-Hispanic whites males, 58.9; Hispanic females, 35.1 per 100,000 vs non-Hispanic females, 43.2 per 100,000) and lower mortality from the disease (Hispanic males, 16.1 per 100,000 vs non-Hispanic white males, 21.7 per 100,000; Hispanic females, 10.7 per 100,000 vs non-Hispanic white females, 15.1 per 100,000).1 Despite these relatively favorable statistics, colorectal cancer incidence among Hispanics is reported to increase with increasing time spent in the US.2 Further, Hispanics with colorectal cancer are more likely than non-Hispanic whites to have been diagnosed at advanced stages,3-5 when the chances of survival are reduced.
Colorectal cancer can be treated successfully if detected early, and the 5-year relative survival drops precipitously according to stage at diagnosis. Survival for localized, regional, and distant stages is 90%, 68%, and 10%, respectively.6 Preventive organizations concerned with colorectal cancer emphasize the importance of screening beginning at age 50 for those at average risk of the disease, as the vast majority, an estimated 70% to 80%, of all colorectal cancers occur among individuals with no personal or family history of the disease.8, 9 The US Preventive Services Task Force updated its colorectal cancer screening guidelines in 2008 to recommend interval screening with colonoscopy, sigmoidoscopy, or fecal occult blood test (FOBT) for all individuals aged 50 to 75 years.7 Annual FOBT screening has been shown in randomized controlled trials to reduce mortality from colorectal cancer by 15% to 33%.8-10
In the general population, screening for colorectal cancer is increasing;6, 11 however, screening among Hispanics lags behind that of non-Hispanic whites. Data from the 2006 Centers for Disease Control (CDC) Behavioral Risk Factor and Surveillance Survey (BRFSS) show that 15.1% of age-eligible Hispanics had received an FOBT in the past 2 years compared with 24.5% of non-Hispanic whites.12 Similarly, lifetime participation in colonoscopy or sigmoidoscopy screening is lower among Hispanics than non-Hispanic whites (46.6% vs 59.1%, respectively).12 Findings from other studies suggest that failure to receive colorectal cancer screening is more common among foreign-born Hispanics (compared with US-born Hispanics)4 and among those who are uninsured (compared with those who are insured).12-15
Some studies have examined the effectiveness of intervention programs to increase colorectal cancer screening among members of racial/ethnic groups. The CDC Guide to Community Preventive Services recommends clinic reminder systems for the prevention of colorectal cancer,16 and such systems are demonstrated to be effective in several individual randomized trials, showing increases in FOBT uptake ranging from 9% to 42%.17-21 Nevertheless, there remains a paucity of research designed to provide tailored, culturally appropriate interventions to increase colorectal cancer screening rates in Hispanics.
The purpose of this study was to evaluate a clinic-based intervention to improve participation in colorectal cancer screening. Specifically, we tested an individual randomized trial wherein patients of Hispanic ethnicity between the ages of 50 and 79 years were assigned to 1 of 3 conditions: 1) usual care; 2) mailed FOBT card and instructions on how to complete the test (mailed FOBT only); and 3) mailed FOBT card and instructions on how to complete the test, telephone reminders, and home visits (mailed FOBT and outreach).
MATERIALS AND METHODS
Setting
This study took place in South King County, Washington, where Hispanics comprise 7.2% of the total county population.22 The majority (66%) of Hispanics in King County are of Mexican origin.23 For the purposes of this study, we chose to work with a community clinic that specializes in care for Hispanic patients. Sea Mar Community Health Centers is a community-based organization that operates 11 community health centers in western Washington; we worked in collaboration with a single Sea Mar clinic that provides medical care in South Park, King County.
Study Design and Research Procedures
Using an individual randomized study design, we sought to implement and test a clinic-based intervention to increase participation in screening services for colorectal cancer among Hispanics. Working with the clinic, we designed a study in which we ascertained baseline rates of colorectal cancer screening (using clinic medical records only), conducted a clinic-based intervention, and examined postintervention rates of compliance with screening. We hypothesized that those in the “mailed FOBT and outreach” group would have the highest rates of participation in FOBT after the intervention. We chose to focus on FOBT screening for 2 primary reasons: 1) it is offered at low cost and is associated with few clinical risks; and 2) follow-up colonoscopies could be provided at low cost to patients with an abnormal FOBT result through a CDC-funded pilot program serving King County residents and operated through the Breast and Cervical Health Program (BCHP).24 After the clinic-based intervention, we conducted a telephone survey to assess differences across intervention groups in self-reported colorectal cancer screening awareness, participation, intention, and normative influences. The survey was designed to assess postintervention differences across randomization groups in intermediate outcomes related to colorectal cancer screening. The research protocol and all data-gathering instruments were approved by Human Subjects Review Board at the Fred Hutchinson Cancer Research Center and the Research and Management Teams of Sea Mar Community Health Centers.
Participant Selection
All Hispanic patients aged 50 to 79 years who had a clinic visit at the South Park clinic from January 1, 2002 through May 31, 2006 were drawn from the computerized clinic records system at Sea Mar. Patient names, identification numbers, age, gender, and ethnicity were entered into a computerized tracking file for the project. Baseline screening data were abstracted from patient medical records; the data included whether the patient had ever participated in colorectal cancer screening (FOBT, colonoscopy, sigmoidoscopy, and other) and the dates of screening. After exclusion of patients with a recent colorectal screening exam (a colonoscopy or sigmoidoscopy within the last 5 years or an FOBT before January 1, 2006), a computer-generated randomized list was used to assign patients to 1 of 3 conditions: 1) usual care; 2) mailed FOBT card and instructions on how to complete the test (mailed FOBT only); and 3) mailed FOBT card and instructions on how to complete the test, telephone reminders, and home visits (mailed FOBT and outreach).
The Intervention
The clinic intervention had 3 arms. Patients in the first arm (n = 165) received usual care, which consisted of no formal prompting of colorectal cancer screening, other than what is provided during a physician visit. Patients in the second arm (n = 168) were mailed a packet containing a letter, an FOBT card, a pamphlet with instructions on how to complete the FOBT, and a prestamped, addressed envelope for mailing the card to the clinic. The letter, written in English and Spanish, encouraged patients to complete the test and was signed by the clinic medical director. The pamphlet provided written instructions (in English and Spanish) and photos showing the 5 steps in completing an FOBT (prepare the card, collect stool, place stool on card, close the cover, repeat, and mail). The pamphlet was adapted from one originally developed by Tu et al, which was successfully used to prompt FOBT screening in a Chinese population.18
For the third arm, we combined the mailed FOBT card with a promotora-delivered educational intervention. Patients in this arm (n = 168) were mailed the same packet and, in addition, were given telephone reminders and a home visit. The promotora model is a peer education model whereby trained community members educate peers in a culturally appropriate manner. One Spanish-speaking male health promoter was hired and trained to deliver the intervention. For the home visits, a Spanish-speaking female medical assistant at Sea Mar accompanied the male health promoter to enhance acceptance of the household visits. Training of the health promoters was conducted by the project principal investigator who has much experience in training health promoters and health educators. The training (conducted in English and Spanish) addressed definitions of cancer, colorectal cancer, and strategies for prevention and early detection of colorectal cancer. During the training, the principal investigator demonstrated how to use rubber models of a colon that contained precancerous polyps and the FOBT instruction cards. The training also addressed documentation procedures to track the activities of the intervention.
For individuals randomized to the “mailed FOBT and outreach” condition, telephone reminders were conducted by the health promoter and began 2 weeks after the mailing. During a typical phone call, the health promoter described the importance of colorectal cancer screening, verified that the patient had received the mailed FOBT card, encouraged the patient to complete the test, and answered any questions about colorectal cancer. Phone calls generally lasted 10 minutes. During a typical home visit, the health promoters described colorectal cancer, explained the need for FOBT screening, demonstrated how to use the FOBT card, and discussed when and how to collect a sample as well as foods and medications to avoid in advance of taking a sample. The health promoters used a series of educational materials in their home visits. These included flip charts that displayed images of the colon and the steps in completing the test. Rubber models of a colon containing precancerous polyps and FOBT cards also were used. The patient was then instructed to mail the FOBT card, using an addressed stamped project envelope, to the clinic laboratory for processing. All project materials were developed for a low-literacy audience and aimed to overcome barriers to colorectal cancer screening documented in previous research, such as low awareness of colorectal cancer and misconceptions about the causes of the disease.25 Home visits generally lasted 50 minutes.
Final Assessment of Intervention
Our primary outcomes were postintervention differences between intervention and usual care groups in FOBT screening participation. The intervention was assessed in 2 ways. First, on an ongoing basis, completed FOBT cards that were mailed to the clinic were tracked by the health promoter, and laboratory records were reviewed to document test results. The date of test receipt, the result of the test, and any other relevant information about the testing procedure were recorded in the tracking system. Second, after the 9-month intervention, final screening rates were assessed by a second abstraction of clinic medical records and baseline and final rates were compared. Dates for which the final computerized data were collected were June 4, 2007 to March 4, 2008.
Final Survey
At the conclusion of the study, all study participants were contacted by phone and asked to complete a short interview. The survey was a 24-item schedule that addressed the following topics: awareness of FOBT, participation in FOBT screening, intention to receive an FOBT, awareness of colonoscopy, participation in colonoscopy screening, knowledge and attitudes about colorectal cancer screening, and normative influences toward screening (whether a doctor, family, or friends had recommended colorectal cancer screening, and whether the respondent had asked a doctor for such screening). Respondent demographic characteristics (number of years of education, ethnicity, country of birth, year of immigration to US, and language spoken at home) were also collected. The interview (available in English and Spanish) took about 10 minutes to complete. All participants provided verbal consent to participate.
Statistical Analysis
Pearson chi-square test was used for the comparison of follow-up FOBT screening frequencies in the 2 intervention arms relative to the usual care arm. Using final survey data, we also examined differences by intervention arm in our secondary outcomes: awareness and intention to receive an FOBT test, awareness of colonoscopy, colorectal cancer screening knowledge and attitudes, and normative influences (whether one's doctor, friends, or family had recommended screening). We present associated P-values and 95% (Wilson) confidence intervals (CI). All analyses were performed using Stata 10 (College Station, TX).26 Data analysis occurred between November 2008 and September 2009.
RESULTS
A total of 635 Hispanic patients ages 50 to 79 years who had visited the Seattle clinic between January 1, 2002 and May 31, 2006 were included in our study (Fig. 1). Participants were excluded if they were compliant with recommendations for colorectal cancer screening (n = 119) or if they were patients in the Sea Mar nursing home (n = 10). The evaluation was based on 165 patients in the usual care group, 168 patients in the “mailed FOBT only” group, and 168 patients in the “mailed FOBT and outreach” group. One patient assigned to the “mailed FOBT only” arm was later found to have been compliant (had an FOBT after January 1, 2006 and before May 31, 2006) at the onset of the study; this patient was kept in the analysis. More than half of the participants were aged 50 to 59 years, and the majority were female (Table 1). Less than one-third had ever completed an FOBT.
| Sociodemographic Factor | Usual Care (n=167) No. (%) | Mailed Packet (n=168) No. (%) | Mailed Packet + Outreach (n=168) No. (%) |
|---|---|---|---|
| Age, y | |||
| 50-59 | 96 (57.5) | 96 (57.1) | 95 (56.5) |
| 60-69 | 55 (32.9) | 56 (33.3) | 55 (32.7) |
| 70-79 | 16 (11.4) | 16 (9.5) | 18 (10.7) |
| Sex | |||
| Female | 88 (52.7) | 89 (52.7) | 89 (53.0) |
- a Among 503 patients with nonmissing medical records.
Reaching Participants
A total of 46 packets (20 in the “mailed FOBT only” group and 26 in the “mailed FOBT and outreach” group) were returned to the clinic because they were undeliverable. Of these, 8 in the “mailed FOBT only” group and 10 in the “mailed FOBT and outreach” group were sent to updated addresses obtained from medical records. Among the 168 patients in the “mailed FOBT and outreach” group who were eligible to receive a telephone reminder, 37 had disconnected phones or had wrong numbers (after attempts to find alternative numbers) and 16 had invalid addresses (as noted above). This left 115 eligible patients, of which 50 (43%) returned their cards (based on laboratory records) and 28 had moved out of the region. The remaining 37 (22%) patients were eligible for home visits. Of households eligible for a home visit, 7 had wrong addresses, 10 could not be reached after 7 attempts, 2 were ineligible for other reasons, and 5 refused. Home visits were completed with 13 patients. Ten patients (77%) from this group returned their cards after the home visits.
Effectiveness of the Intervention
Our primary outcome was receipt of an FOBT during the 9-month postintervention period. Four individuals (2%; 95% CI, 0.9-6) in the usual care group received an FOBT during the 9-month follow-up period (Fig. 2). Screening rates were 43 of 168 (26%; 20, 33) and 52 of 168 (31%; 25, 38) in the “mailed FOBT only” and “mailed FOBT and outreach” groups, respectively; both were statistically larger than the usual care group (P < .001 in both cases). The 5-percentage-point difference in screening between the 2 intervention groups was not statistically significant (P = .28).
Study design and follow-up.
Percentage of Latino adults ages 50 to 79 years who received a fecal occult blood test during intervention, by intervention arm.
Final Survey
Telephone interviews were attempted among all 501 patients who were randomized. Of these, 137 had disconnected phones, did not answer after 7 attempts, or were otherwise unreachable, leaving 357 reachable households. Of these, 177 households completed the interview for a response rate of 49.6%.
Demographics
More than half of respondents were aged 50 to 59 years, the majority were female, and three-quarters had less than a high school education (Table 2). The remainder had a high school diploma or GED (13%) or had completed some college (12%). Slightly more than one-quarter of respondents had private healthcare coverage. Less than 20% were covered by the Washington State Basic Health Care Plan. Nearly one-quarter of respondents reported either relying on a government plan (Medicare, Medicaid, or medical coupons) or having no healthcare insurance.
| Demographic Characteristic | Randomization Status | ||
|---|---|---|---|
| Usual Care (n=53) No. (%) | Mailed Packet Only (n=57) No. (%) | Mailed Packet + Outreach (n=67) No. (%) | |
| Age, y | |||
| 50-59 | 36 (67.9) | 30 (52.6) | 37 (55.2) |
| 60-69 | 14 (26.4) | 20 (35.1) | 23 (34.3) |
| 70-74 | 3 (5.7) | 7 (12.3) | 7 (10.5) |
| Gender | |||
| Female | 27 (50.9) | 36 (63.2) | 36 (53.7) |
| Education | |||
| 4th grade or less | 18 (35.3) | 23 (41.8) | 25 (38.5) |
| 5th-8th grade | 13 (25.5) | 14 (25.5) | 14 (22.6) |
| 9th-12th grade, no diploma | 7 (13.7) | 4 (7.3) | 10 (15.4) |
| High school graduate | 10 (19.6) | 4 (7.3) | 8 (12.3) |
| Some college | 3 (5.9) | 10 (18.2) | 8 (12.3) |
| Birthplace | |||
| Mexico | 35 (70.0) | 36 (66.7) | 42 (65.6) |
| US | 2 (4.0) | 2 (4.0) | 3 (4.7) |
| Othera | 13 (26.0) | 16 (30.0) | 19 (29.7) |
| Language spoken at home | |||
| Spanish | 45 (88.2) | 52 (94.6) | 58 (89.2) |
| Other | 6 (12.0) | 3 (5.4) | 7 (11.0) |
| Health insurance | |||
| Private health insurance | 20 (38.5) | 17 (29.8) | 16 (25.0) |
| Basic health plans | 9 (17.3) | 12 (21.1) | 12 (18.5) |
| Medicare, Medicaid, or coupons | 9 (17.3) | 18 (31.6) | 17 (26.6) |
| None | 14 (26.9) | 10 (17.5) | 19 (29.7) |
- a Other birthplace included El Salvador (n=11), Guatemala (n=10), Peru (n=8), Columbia (n=7), Honduras (n=4), Chile (n=2), Cuba (n=2), and other (n=4).
Colorectal Screening Awareness, Participation, and Intention
Awareness of FOBT varied across intervention arm, with nearly 60% of those in the usual care arm, 84% in the “mailed FOBT only” arm, and 79% in the “mailed FOBT and outreach” groups reporting having heard of an FOBT (Table 3). Slightly more than two-thirds of respondents in the usual care group had ever had an FOBT, and these percentages were similar in the “mailed FOBT only” (69%) and “mailed FOBT and outreach” (75%) groups. Substantial differences were noted in the proportions of respondents in the 3 groups who were considering having an FOBT in the next few months: 35% for the usual care group compared with 57% and 68% for the “mailed packet only” and the “mailed packet and outreach” groups, respectively. Having heard of a colonoscopy and attitudes about colorectal cancer and screening were similar across intervention arms.
| Randomization Status | ||||
|---|---|---|---|---|
| Percent Who Agreea | Usual Care No. (%) | Mailed Packet Only No. (%) | Mailed Packet + Outreach No. (%) | Pb |
| FOBT awareness, participation, and intention to receive test | ||||
| Heard of FOBT (n=177) | 32 (60.4) | 48 (84.2) | 53 (79.1) | <.01 |
| Ever had an FOBT (n=133) | 22 (68.8) | 33 (68.8) | 40 (75.5) | .70 |
| Considering having an FOBT in the next few monthsc (n=84) | 7 (35.0) | 17 (56.7) | 23 (67.7) | .07 |
| Colonoscopy awareness and participation | ||||
| Heard of colonoscopy/sigmoidoscopy (n=175) | 28 (53.9) | 36 (64.3) | 35 (52.2) | .40 |
| Ever had colonoscopy/sigmoidoscopy (n=99) | 16 (57.1) | 11 (30.6) | 12 (34.3) | .07 |
| Colorectal cancer knowledge (% correct) | ||||
| Colorectal cancer is caused by eating spicy foods (n=172) | 9 (18.0) | 19 (33.9) | 10 (15.2) | .08 |
| Colorectal cancer can be prevented if detected early (n=173) | 48 (94.1) | 53 (94.6) | 63 (95.5) | .83 |
| Colorectal cancer attitudes (% agree) | ||||
| I am afraid of getting colorectal cancer (n=173) | 41 (80.4) | 48 (85.7) | 54 (81.8) | .87 |
| Colorectal cancer testing is embarrassing (n=173) | 31 (60.8) | 38 (67.9) | 37 (56.1) | .72 |
| Normative influences | ||||
| Doctor recommended colorectal screening test (n=175) | 23 (44.2) | 20 (35.7) | 29 (43.3) | .63 |
| Family member recommended colorectal screening test (n=172) | 12 (23.1) | 13 (24.1) | 16 (24.2) | .98 |
| Friend recommended colorectal screening test (n=172) | 7 (13.7) | 11 (20.0) | 11 (16.7) | .69 |
| Asked a doctor for colorectal screening test (n=174) | 13 (25.0) | 12 (21.8) | 13 (19.4) | .75 |
- FOBT indicates fecal occult blood test.
- a Percentages based on nonmissing values.
- b Based on Pearson chi-square test.
- c Among those who have not had an FOBT (n=84).
Patient Concerns About FOBT
Despite providing a prestamped, addressed envelope, a small number of participants delivered their FOBT cards in person to the clinic. A total of 10 cards (10%) were improperly completed and new packets were mailed to study participants. Usually the errors resulted from patients providing one sample from a single stool, providing one sample from each of 3 consecutive stools (rather than 2 samples from each stool), or putting the sample on the reverse side of the card. Among patients who expressed reluctance to complete the FOBT either during the telephone reminders or home visits, 5 general concerns were raised: 1) that the FOBT was costly; 2) that there was not enough time to do the test; 3) that dietary and drug restrictions in preparation for the test were difficult to follow; 4) that infrequent bowel movements hampered proper completion of the test; and 5) that competing health and personal issues took precedence. No adverse events related to the intervention were reported.
DISCUSSION
A low-intensity clinic-based intervention program targeting low-income Hispanic patients using mailed FOBT cards and instructions on how to perform the home-based test successfully improved participation in colorectal cancer screening over usual clinical care. A high-intensity intervention involving home-based education in addition to the mailed FOBT card resulted in even higher rates of colorectal cancer screening uptake, although the difference between the low- and high-intensity intervention arms was not statistically significant. Because regular FOBT has been shown to significantly reduce mortality from colorectal cancer, interventions to improve rates of screening in underscreened populations have substantial public health importance.
Our study showed substantial improvements in FOBT use among patients in the 2 intervention arms compared with the usual care condition. However, patients in the “mailed FOBT only” group were nearly as likely as those in the “ mailed FOBT and outreach” group to have received an FOBT after the intervention. Health promoters have been successfully used among Hispanic populations to promote use of screening services for breast and cervical cancer27 and to improve dietary intake.32 Such lay educators are thought to enhance cultural relevance of a given intervention program by capitalizing on knowledge of cultural-based values and practices and accessing culturally relevant channels of communication. Nevertheless, for our patients, simply removing the structural barriers (providing free FOBT screening) significantly improved screening rates. Further research is needed to assess the added value of promotoras for interventions that target reductions in structural barriers to screening.
In a systematic review, the US Preventive Services Task Force identified client reminders as an effective strategy for improving participation in colorectal cancer screening programs, showing a median increase of 17.6% in FOBT participation and flexible sigmoidoscopy (based on 6 studies or study arms).28 The Task Force was unable to recommend one-on-one education to colorectal cancer screening as there were insufficient numbers of published studies evaluating their effectiveness.28 Although our study was not designed to assess the effectiveness of a one-on-one educational intervention per se, our findings showed that the approach combined with a mailed reminder significantly improved rates of screening over usual clinical care. In analyses comparing this with a lower-intensity telephone recall approach, however, no significant differences were observed.
In the past several years, various government agencies and nonprofit organizations have promoted early detection of colorectal cancer, such as the celebrity endorsement by Katie Couric,29 the CDC's Screen for Life Campaign,27 the American Cancer Society's public service announcement whose main character is “Polyp Man,”30 and former president Bill Clinton's pronouncement in 2000 of March as National Colorectal Cancer Awareness Month.31 This surge of publicity has raised awareness of the importance of screening in the general population, but there remain important gaps in reaching certain groups, specifically Hispanics. During this same time period, Medicare reimbursement for screening services for colorectal cancer were changed, first in 1998 to provide reimbursement for FOBT and flexible sigmoidoscopy and then in 2001 to provide reimbursement for colonoscopy. In Washington State, 2007 legislation led to the requirement that all insurance providers reimburse for colorectal cancer screening. Funding from the CDC and Washington State for the Breast, Cervical & Colon Health Program provided low-cost colorectal cancer screening to patients in King County. During the intervention period, the program provided low-cost FOBTs and colonoscopy to individuals with abnormal FOBT results. Since these programs began, there has been a modest rise in screening participation nationally.12 Despite these media campaigns and programs, a relatively small proportion of individuals (2%) who received usual clinical care in our study were screened during the 9-month intervention period.
Notably, we achieved a low response rate from a telephone survey that was conducted at the end of the study. We believe that this is primarily due to 2 factors. The first is that we conducted no baseline survey and relied on contact information in medical records to reach participants. Given that many participants had not been seen in the clinic for several years, many had outdated telephone numbers or had moved from the area. Further, the conversion of many household telephone services from land lines to cells phones likely further impeded our efforts to contact individuals. For similar reasons, we experienced difficulty in reaching households for the health promoter-delivered aspect of the high-intensity intervention. The small numbers of households that completed the health promoter aspect of the study may explain the nonsignificant difference in the proportions of individuals screened in the “mailed FOB only” and “mailed FOBT and outreach” intervention groups.
Data from the CDC from 2006 show that, for the state of Washington, the proportion of age-eligible Hispanics who completed an FOBT in the past 2 years was 23.8%; the national average for Hispanics that year was 24.2%. The screening rate of 2% in our usual care arm, though only spanning a period of 9 months, might suggest much room for improvement in colorectal cancer screening in this region. Caution should be exercised in comparing these figures, as our sample was limited to individuals who had not had a recent test. However, low rates of screening have been reported in communities in eastern Washington State.32
The proportion of tests that had positive results in this study (10.2%) was higher than that reported in other studies. FOBT screening trials and pilot studies have reported that from 1% to 9% of screened individuals show a positive FOBT.33 This may be due to the underserved nature of our study population or a lower proportion of individuals who adequately complied with dietary and drug restrictions in preparation for the test (eg, avoiding red meat, aspirin, and other foods that may result in a false%positive finding). Notably, a commonly reported barrier to completing the FOBT noted by our participants was the dietary restriction in preparation for the test (specifically difficult for Latino food preferences).
This study had some limitations. As this was a pilot investigation, we relied solely on clinic patient records to select participants and assess our main outcome, completion of colorectal cancer screening, during the intervention period. Although medical record information is widely considered the gold standard in assessments of cancer screening behaviors, our sole reliance on these data meant that our sample was not known to be reachable. As a result, several packets were returned to the clinic because of invalid addresses (n = 46) and our postintervention survey achieved a relatively low response rate (49.6%). Nevertheless, we relied on an “intention-to-treat” analysis for our primary outcome (the most conservative approach) and still achieved significant results. We expect that the proportion screened would have been much higher had we limited our analysis to those known to have received the intervention. Further, we did not administer a baseline survey to assess self-reported colorectal cancer screening behaviors as well as intentions, and knowledge and attitudes about colorectal cancer; thus our study could not assess changes in these “intermediate” outcomes. We do, however, present self-reported data on these factors by randomization status obtained from a telephone survey conducted after the intervention. We collected no information on the costs associated with administering the intervention across intervention arms nor did we collect data to assess the cost per individual screened (cost-effectiveness). Although we believe that this does not limit the conclusions that can be drawn, future studies may benefit from the collection and analysis of these data. Further, the short time frame for the intervention prevented us from capturing colorectal screening that may have been motivated by the intervention but occurred after the assessment period.
Our study involved notable strengths relative to previous investigations, which included our randomized design and the comparison of the effectiveness of multiple intervention conditions. Our study involved a collaboration with a busy community-based primary care clinic, rather than an academic-affiliated practice.34
Acknowledgements
The authors acknowledge the contributions of Ana Martinez, medical assistant at Sea Mar Community Health Centers.
CONFLICT OF INTEREST DISCLOSURES
This research was supported by a grant from the Fred Hutchinson Cancer Research Center. The authors report no conflicts of interest and have no financial disclosures.
