Fecal occult blood testing (FOBT) is a low-risk test that has been shown to reduce mortality from colorectal cancer,1,2 but it can be effective only when patients with an abnormal result receive follow-up with a complete evaluation of the colon (colonoscopy or barium enema plus sigmoidoscopy), which has clinically significant risks.3 Therefore, guidelines recommend targeting FOBT screening to patients whose comorbidity and preferences would permit complete evaluation of the colon to be pursued if the result was abnormal.4-6 However, prior studies7,8 have found that FOBT screening is often performed in patients with clinically significant comorbidities who would be at increased risk for serious adverse events (eg, colon perforation, major bleeding, or stroke) from follow-up colonoscopy. In addition, it is well known that many patients with positive screening FOBT results do not receive a complete evaluation of the colon.9-12 However, it remains unclear as to what extent low rates of follow-up are due to screening older patients whose poor health or preferences would not permit complete evaluation of the colon (and therefore they should not have been screened) vs failure to complete intended follow-up of healthy older patients.
Developing a better understanding of factors contributing to low rates of follow-up of abnormal screening FOBT results in older adults is critical for improving the effectiveness of this widely used test; FOBT is the most common colorectal cancer screening test used in the Veteran's Affairs (VA) health care system.13 For example, if poor follow-up occurs predominantly in patients in poor health, who should not have been screened in the first place, then initiatives mandating follow-up colonoscopy within 60 days of a positive result may cause harm if patients with serious comorbidity are automatically sent for colonoscopy.14 While the hope is that clinician judgment would prevail to appropriately target follow-up, previous studies of the impact of age, comorbidity, and other factors on the receipt of follow-up after an abnormal FOBT result have been limited by small sample sizes, single site data, data from intervention studies or clinician surveys that may not reflect real-world practices, and failure to describe the spectrum of abnormal screening results (eg, number of FOBT cards showing positive results [hereinafter, positive cards], incomplete tests).9-13,15-19 In addition, none of the previous VA studies have measured follow-up outside the VA,10,13,19,20 which is critical because many veterans receive care outside the VA.21
Therefore, to address these issues we took a novel approach of combining VA and Medicare claims and medical chart review to provide objective, real-world data characterizing the spectrum of screening FOBT results, and type of follow-up received within 1 year among patients 70 years or older screened at 4 geographically diverse VA facilities. For patients with a positive FOBT result, we determined the impact of age, comorbidity, and other factors on receipt of a complete colon evaluation within 1 year, and using VA medical records we report documented reasons for lack of follow-up as well as the number of patients who ultimately received a complete colon evaluation during the subsequent 5 years.
Data sources and patients
Outpatient claims from the VA National Patient Care Database were used to identify a cohort of 6853 patients 70 years or older whose first colorectal cancer screening test from January 1 to December 31, 2001 was FOBT at 1 of 4 VA facilities (Minneapolis, Minnesota; Durham, North Carolina; Portland, Oregon; and West Los Angeles, California). The FOBT was identified by Current Procedural Terminology (CPT) codes 82270, 82273, and 82274, Level II Healthcare Common Procedure Coding System (HCPCS) code G0107, and laboratory data.7 We used national VA and Medicare data to follow the cohort for 1 year after their FOBT for the performance of follow-up testing, including colonoscopy, sigmoidoscopy, barium enema, or repeated FOBT. To ensure complete comorbidity and follow-up data, we required all patients to have continuous enrollment in Medicare Parts A and B and fee-for-service coverage from January 1, 2000, through December 31, 2002 (Figure 1). In addition, patients had to be eligible for screening to be included in our cohort. Therefore, we used VA and Medicare inpatient and outpatient claims (dating back to October 1, 1992, for VA claims and January 1, 1999, for Medicare claims) to exclude 799 patients (12%) with a history of colorectal cancer, colitis, colorectal polyps, colectomy, or colostomy and 779 patients (11%) who had any history of a colonoscopy or had had a sigmoidoscopy or barium enema within 5 years and were therefore not due for screening. We also used claims from 6 months before their FOBT as well as medical chart review to exclude 378 patients (6%) who had signs or symptoms that would justify performance of FOBT for nonscreening purposes (Figure 1). This left a final cohort of 2410 patients who had a screening FOBT in 2001 at 1 of 4 VA facilities.
Data collection and measurement
Individual FOBT results were extracted from the chemistry section of the Veterans Health Information Systems and Technology Architecture (VISTA) laboratory package at the 4 VA facilities. If any FOBT cards were positive for occult blood, then the FOBT was classified as “positive.” If all 3 cards were negative for occult blood, then the FOBT was classified as “negative.” If fewer than 3 cards were submitted and the results were negative, then the FOBT was classified as “incomplete.” Less than 1% of patients had administrative codes for FOBT but no result in the electronic medical record. These FOBT results were classified as “missing,” and follow-up was not assessed. Three VA facilities (deidentified as sites A, C, and D) used Hemoccult FOBT without rehydration, whereas site B used Hemoccult SENSA (Beckman Coulter, Brea, California), a more sensitive test.
We assessed the receipt of follow-up during the 1 year after screening FOBT across the VA health care system and Medicare because many elderly veterans use more than 1 VA medical center and most are enrolled in Medicare.21 We identified follow-up testing in National VA Data Systems and linked Medicare payment data (inpatient and outpatient files) by using International Classification of Disease, Ninth Revision (ICD-9), CPT, and HCPCS codes for colonoscopy (ICD-9: 45.22, 45.23, 45.25, 45.41, 45.42, 45.43; CPT: 44388-44394, 45355, 45378-45385; HCPCS: G0105, G0121), sigmoidoscopy (ICD-9: 45.24, 48.22-48.24, 48.26, 48.35, 48.36; CPT: 45300, 45303, 45305, 45308, 45309, 45315, 45320, 45330-45334,45337-45339; HCPCS: G0104), barium enema (ICD-9: 87.64; CPT: 74270, 74280; HCPCS: G0106, G0120, G0122), and repeated FOBT.7,22 For patients with a positive FOBT result, we assessed whether a complete colon evaluation (defined as colonoscopy or barium enema plus sigmoidoscopy) was performed within 1 year.
In addition, 3 of us (C.M.C., M.A.C., and L.C.W.) reviewed VA electronic medical records independently for documented reasons why a complete colon evaluation was not performed and together grouped the reasons into categories. Any discrepancy in categorization was adjudicated by consensus among the reviewers. Medical record review was also used to determine if patients who did not receive a complete colon evaluation within 1 year went on to have a complete colon evaluation documented during the subsequent 5 years.
Age was determined on the date of the screening FOBT. Comorbidity was measured using the Deyo adaptation of the Charlson comorbidity index, a summary measure of 19 chronic disease diagnoses from administrative data.23 The risk of death from an increase of 1 in Charlson comorbidity score is approximately equal to that from an additional decade of age.24 Charlson-Deyo scores were calculated from VA and Medicare inpatient and outpatient claims during the 12 months before the date of the FOBT.
We also assessed other factors known to influence receipt of follow-up after positive FOBT results (Table 1) using VA and Medicare data and linkage to the 2000 US Census.25 The committee on human research at the University of California, San Francisco; the committee for research and development at the San Francisco VA Medical Center; and the institutional review board at the Minneapolis VA Medical Center approved the study.
For all estimates of the cumulative incidence of follow-up testing, we observed patients from the date of their screening FOBT in 2001 until their follow-up test or death or the end of the study period (1 year after the date of the FOBT). Patients were censored at the date of follow-up or 1 year, whichever came first, and deaths were treated as competing risk events.26 Date of death was obtained from the VA Vital Status File, which is similar to the National Death Index in terms of accuracy and completeness.27 We estimated unadjusted 1-year cumulative incidence of follow-up, stratified according to screening FOBT result. For patients who had a positive FOBT result, we used Cox regression models to assess the association between baseline characteristics and receipt of a complete colon evaluation. To describe the association of age and comorbidity combined, we determined the cumulative incidence of a complete colon evaluation for patients categorized into 3 subgroups: those aged 70 to 74 years with a Charlson score of 0, who are most likely to benefit from screening and follow-up (7-year mortality, 21%); those aged 80 years or older with a Charlson score of 1 or higher, who are least likely to benefit from screening and follow-up (7-year mortality, 62%; life expectancy, 5.6 years); and everyone else.7 We used SAS (version 9.1; SAS Institute, Cary, North Carolina) and Stata/SE (version 10.0; StataCorp LP, College Station, Texas) statistical software for all analyses.
Table 1 shows the characteristics of the 2410 patients in our cohort, stratified according to their FOBT result. Overall, 23% of patients were 80 years or older. Consistent with an elderly veteran population, 98% were men and 88% were white. A total of 212 patients (9%) had a positive FOBT result; 2091 (87%) had a negative FOBT result; and 83 (3%) had an incomplete FOBT result. A total of 106 patients (4%) died within 1 year of their screening FOBT.
Patients with positive FOBT results were more likely to have had their FOBT performed at site B. At site B, 16% of screened patients had a positive result compared with 5% to 6% of patients screened at the other VA facilities. A total of 109 patients with a positive FOBT result (51%) had 1 positive card, 53 (25%) had 2 positive cards, and in 50 (24%) all 3 cards were positive.
Incidence of follow-up testing
Of the 212 patients with positive FOBT results, only 90 (42%) received a complete colon evaluation within 1 year; 47 (22%) received an evaluation within 6 months. Eight-six patients received a colonoscopy, and 4 received a barium enema plus sigmoidoscopy. Fifty-eight (27%) received a complete colon evaluation within the VA, and 32 (15%) received this follow-up through Medicare. Among those who received a complete colon evaluation within 1 year, the mean (SD) time to follow-up was 192 (91) days within the VA and 119 (90) days for those who went outside the VA for follow-up through Medicare (P < .001). A total of 122 patients (58%) did not receive a complete colon evaluation. Overall, 4 (2%) received a barium enema alone, 11 (5%) received a sigmoidoscopy alone, 30 (14%) received a repeated FOBT, and 77 (36%) received no follow-up within 1 year. In addition, only 27 of 83 patients who had an incomplete FOBT result (33%) received any follow-up within 1 year. Repeated FOBT was the most common follow-up test (24%) after an incomplete result. To provide context of background testing rates, even among the 2091 patients with negative FOBT results, 29% received some type of colorectal cancer testing within 1 year (Figure 2).
Table 2 indicates there were no significant age-based, comorbidity, racial/ethnic, or socioeconomic characteristics associated with receiving a complete colon evaluation within 1 year after a positive FOBT result (P > .05). However, VA site, number of positive FOBT cards, and number of VA outpatient visits were associated with receipt of a complete colon evaluation. For example, 37 of 109 persons with 1 positive card (34%) received a complete colon evaluation compared with 53 of 103 persons with 2 or 3 positive cards (51%) (P = .01). In addition, most patients had opportunities for referral for a complete colon evaluation; the median number of VA outpatient visits within 1 year after a positive FOBT result was 3. A total of 110 of 122 patients who did not receive a complete colon evaluation (90%) had a VA outpatient visit within 1 year (range, 0-15 visits). Eleven patients (9%) who did not receive a complete colon evaluation died within 1 year of their positive FOBT result.
Evaluating the combined effect of age and comorbidity, the incidence of a complete colon evaluation within 1 year of a positive FOBT result was similar among the 46 patients 70 to 74 years old with a Charlson score of 0 compared with the 29 patients 80 years or older with a Charlson score of 1 or higher (48% vs 41%; P = .28). Only 18 patients (8%) received a complete colon evaluation within 60 days of a positive result (Figure 3).
Documented reasons for lack of complete colon evaluation after positive fobt results
Among the 122 patients with a positive FOBT result who did not receive a complete colon evaluation within 1 year,the most common findings in VA medical records were lack of any acknowledgment of the positive FOBT result in progress notes (43%) and patient refusal of colonoscopy (26%). Patients whom clinicians documented as refusing colonoscopy or having severe health problems or other miscellaneous problems that did not permit complete colon evaluation (Table 3) were classified as having been inappropriately screened (38%). Patients who failed to receive a complete colon evaluation for other reasons listed in Table 3 (62%) were classified as having failed to complete follow-up. Only 29 patients who did not receive a complete colon evaluation within 1 year (24%) went on to have a complete colon evaluation documented during the next 5 years.
We found that many veterans 70 years or older are not receiving any follow-up after incomplete or positive screening FOBT results. In addition, while 64% of elderly veterans with positive FOBT results received some type of follow-up within 1 year, only 42% received a complete colon evaluation. The cumulative incidence of complete colon evaluation after a positive result was low regardless of age or comorbidity. Instead, other factors, such as VA site, number of positive FOBT cards, and number of VA outpatient visits, were predictive of receiving a complete colon evaluation, although the incidence never exceeded 60%. Also, medical chart documentation indicated that patients who should not have been screened in the first place accounted for 38% of older patients who did not receive a complete colon evaluation within 1 year, and most patients (>75%) who did not receive a complete colon evaluation in the first year never did receive a complete follow-up, even during the next 5 years.
Low rates of follow-up after a positive screening FOBT result are found in many health care systems, with most studies showing that less than 60% of older patients receive a complete colon evaluation within 1 year.9-13,15-17,19,28 Despite this widespread problem, previous studies have not determined the extent to which lack of follow-up is due to screening patients whose age, comorbidity, or preferences would not permit a complete colon evaluation. Findings from previous studies have been mixed regarding the effect of age and comorbidity on receipt of a complete colon evaluation after positive FOBT results, they have been limited by narrow geographic area or included FOBT performed for nonscreening purposes, and prior VA studies have not included follow-up outside the VA.10-13,15-20,29 Novel aspects of our study are that we used a geographically diverse screened population and its spectrum of FOBT results (eg, number of positive cards, incomplete tests) and identified follow-up using objective real-world data, including claims data from outside the VA through Medicare, which accounted for 15% of complete colon evaluations.
While we found that age and comorbidity were not associated with receipt of a complete colon evaluation after positive FOBT results, 38% of patients without a complete follow-up had medical chart documentation that they refused follow-up or had long-standing health problems that did not permit follow-up. We also found that among patients without a complete follow-up within 1 year, less than 25% received a complete follow-up even over the subsequent 5 years. This suggests screening patients in whom a complete colon evaluation would never be pursued substantially contributes to lack of follow-up among older patients. This may be explained in 2 possible ways. First, a discussion about the need to follow up a positive result with colonoscopy may not happen at the time of screening. Prior studies have shown that clinicians often fail to discuss key information about colorectal cancer screening with patients, especially the risks of follow-up procedures.30,31 Second, previous studies have shown that FOBT screening is poorly targeted to healthy older patients for many reasons, including quality indicators that often promote screening regardless of comorbidity or preferences.7,32
Failure to complete follow-up in older patients for whom age, comorbidity, or preferences were not documented as contraindications to follow-up was also common (62% of patients). In fact, 43% of patients who did not receive a complete colon evaluation lacked acknowledgment of the positive FOBT result in progress notes, suggesting that clinicians may not have known about these results. In addition, clinicians were less likely to complete follow-up if only 1 FOBT card was positive or if a patient reported not following dietary instructions. However, guidelines recommend a complete colon evaluation if any FOBT card is positive.4,5 Similarly, dietary indiscretion is not a reason to avoid follow-up.33 Also, access to timely colonoscopy may be difficult at some VA facilities if the number of gastroenterologists is low or the rate of positive FOBT is especially high, such as site B, which used a more sensitive FOBT than the other facilities. Conversely, some facilities, such as sites A and C, had electronic data systems that informed clinicians of screening results and tracked endoscopy procedures, which increased follow-up.34,35
While increasing use of electronic reminder and notification systems for positive FOBT results has led to some increases in the percentage of patients undergoing a complete colon evaluation in both the VA and integrated health systems,19,34,35 these likely encourage follow-up in all patients indiscriminately, despite substantial differences in potential benefits and risks. Our findings that a substantial number of older patients should not have been screened in the first place suggest that interventions mandating follow-up may have the unintended consequence of worsening quality of care in these patients. In addition, the number of positive FOBT cards, VA facility, and number of visits should not be the major predictors of follow-up. Rather, quality improvement initiatives should encourage clinicians to weigh risks and benefits at each step in the screening process, in the context of a patient's age, comorbidity, and preferences.36
Our study has several limitations. Although we supplemented 2 national claims databases with medical chart review, we may have missed some tests performed outside the VA system, as Medicare does not capture tests paid for privately and Medicare claims are not reliably reported for patients enrolled in Medicare Managed Care so they were excluded from our study. Second, as our study predates quality improvement efforts initiated at the VA in 2005, current patterns may be different. However, a prior study17 found that follow-up has not been increasing over time (1991-2006), and even in 2007 less than a third of patients received a complete colon evaluation within 60 days of a positive FOBT result, suggesting that problems with follow-up persist.37 Third, while our study included over 2400 patients screened with FOBT, only 212 had a positive result, which may have limited our power to detect small differences between predictors of follow-up. Fourth, medical records sometimes lack details about why a patient refused or did not receive follow-up. However, the medical record is the official document of the follow-up decision that was actually made. This is an advantage over using clinician self-report, which may not represent real-world decision-making in a busy practice setting.38 Fifth, our cohort primarily comprised men who use the VA, so the generalizability of our findings to persons who do not use the VA is uncertain. Yet, understanding follow-up within the VA is important in its own right, because the VA is the largest health care system in the United States and a leader in improving health care quality.
In conclusion, low follow-up rates after a positive FOBT result are seen regardless of whether patients are 70 to 74 years old without comorbidity or 80 years or older with comorbidity. These findings argue against the approach of screening indiscriminately with FOBT and then targeting follow-up based on age and comorbidity since, like colorectal cancer screening, FOBT follow-up in older adults is not well targeted. Medical chart documentation indicates that failure to complete follow-up is due to problems with screening patients in whom follow-up is not appropriate as well as failing to complete follow-up in patients who should have received follow-up. Quality improvement initiatives should encourage individualized screening decisions in older patients and facilitate timely follow-up of positive FOBT results in patients whose comorbidity and preferences make follow-up appropriate.
Correspondence: Louise C. Walter, MD, Division of Geriatrics, VA Medical Center 181G, 4150 Clement St, San Francisco, CA 94121 (Louise.Walter@ucsf.edu).
Accepted for Publication: May 30, 2010.
Published Online: October 11, 2010. doi:10.1001/archinternmed.2010.372
Author Contributions: Dr Walter had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Carlson and Walter. Acquisition of data: Carlson, Casadei, Partin, and Walter. Analysis and interpretation of data: Carlson, Kirby, Kistler, and Walter. Drafting of the manuscript: Carlson, Kirby, and Walter. Critical revision of the manuscript for important intellectual content: Carlson, Casadei, Partin, and Kistler. Statistical analysis: Kirby and Kistler. Obtained funding: Walter. Administrative, technical, and material support: Casadei, Partin, and Walter.
Financial Disclosure: None reported.
Funding/Support: Dr Walter is supported by a VA Health Services Research and Development grant IIR-04-427 and by grant 1R01CA134425 from the National Cancer Institute and is a Robert Wood Johnson Physician Faculty Scholar. Dr Kistler is supported by a T-32 Training Grant AG000212-16. Dr Partin is supported by the Center for Chronic Disease Outcomes Research, VA Health Services Research and Development grant HFP 98-001.
Disclaimer: The funding sources had no role in the design, conduct, or analysis of this study or in the decision to submit the manuscript for publication.
O Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet
1467- 1471PubMedGoogle ScholarCrossref
et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet
1472- 1477PubMedGoogle ScholarCrossref
AI Risk of perforation after colonoscopy and sigmoidoscopy: a population-based study. J Natl Cancer Inst
230- 236PubMedGoogle ScholarCrossref
U.S. Preventive Services Task Force, Screening for colorectal cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med
627- 637PubMedGoogle ScholarCrossref
OW Cancer screening in the United States, 2009: a review of current American Cancer Society guidelines and issues in cancer screening. CA Cancer J Clin
27- 41PubMedGoogle ScholarCrossref
KE Cancer screening in elderly patients: a framework for individualized decision making. JAMA
2750- 2756PubMedGoogle ScholarCrossref
et al. Impact of age and comorbidity on colorectal cancer screening among older veterans. Ann Intern Med
465- 473PubMedGoogle ScholarCrossref
et al. Adverse events after outpatient colonoscopy in the Medicare population. Ann Intern Med
849- 857, W152PubMedGoogle ScholarCrossref
et al. A national survey of primary care physicians' methods for screening for fecal occult blood. Ann Intern Med
86- 94PubMedGoogle ScholarCrossref
DA Colorectal cancer screening, comorbidity, and follow-up in elderly patients. Am J Med Sci
159- 163PubMedGoogle ScholarCrossref
HG Diagnostic testing following fecal occult blood screening in the elderly. J Natl Cancer Inst
1641- 1646PubMedGoogle ScholarCrossref
CW Colorectal cancer screening and surveillance practices by primary care physicians: results of a national survey. Am J Gastroenterol
1551- 1556PubMedGoogle ScholarCrossref
K Barriers to full colon evaluation for a positive fecal occult blood test. Cancer Epidemiol Biomarkers Prev
1232- 1235PubMedGoogle ScholarCrossref
L Screening for colorectal neoplasia: physician's adherence to complete diagnostic evaluation. Am J Public Health
1620- 1622PubMedGoogle ScholarCrossref
et al. Physician-reported reasons for limited follow-up of patients with a positive fecal occult blood test screening result. Am J Gastroenterol
2078- 2081PubMedGoogle ScholarCrossref
et al. Reasons patients with a positive fecal occult blood test result do not undergo complete diagnostic evaluation. Ann Fam Med
11- 16PubMedGoogle ScholarCrossref
TD Challenges in the management of positive fecal occult blood tests. J Gen Intern Med
356- 360PubMedGoogle ScholarCrossref
et al. Using a multifaceted approach to improve the follow-up of positive fecal occult blood test results. Am J Gastroenterol
942- 952PubMedGoogle ScholarCrossref
et al. Measuring the quality of colorectal cancer screening: the importance of follow-up. Dis Colon Rectum
1002- 1010PubMedGoogle ScholarCrossref
et al. Veterans' access to and use of Medicare and Veterans Affairs health care. Med Care
214- 223PubMedGoogle ScholarCrossref
MA Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol
613- 619PubMedGoogle ScholarCrossref
J Validation of a combined comorbidity index. J Clin Epidemiol
1245- 1251PubMedGoogle ScholarCrossref
BE Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med
695- 706PubMedGoogle ScholarCrossref
DM Accuracy and completeness of mortality data in the Department of Veterans Affairs. Popul Health Metr
et al. Physician and patient factors associated with ordering a colon evaluation after a positive fecal occult blood test. J Gen Intern Med
357- 363PubMedGoogle ScholarCrossref
et al. Factors that influence the decision to do an adequate evaluation of a patient with a positive stool for occult blood. Am J Gastroenterol
196- 203PubMedGoogle ScholarCrossref
B Talking to patients about screening colonoscopy: where conversations fall short. J Fam Pract
E1- E9PubMedGoogle Scholar
G Physician-patient communication about colorectal cancer screening. J Gen Intern Med
1493- 1499PubMedGoogle ScholarCrossref
NJ Impact of comorbidity on colorectal cancer screening in the veterans healthcare system. Clin Gastroenterol Hepatol
991- 996PubMedGoogle ScholarCrossref
C Meta-analysis of dietary restriction during fecal occult blood testing. Eff Clin Pract
150- 156PubMedGoogle Scholar
JA Effectiveness of a provider reminder on fecal occult blood test follow-up. Dig Dis Sci
1991- 1996PubMedGoogle ScholarCrossref
B Interventions to improve follow-up of abnormal findings in cancer screening. Cancer
1188- 1200PubMedGoogle ScholarCrossref
KE Pitfalls of converting practice guidelines into quality measures: lessons learned from a VA performance measure. JAMA
2466- 2470PubMedGoogle ScholarCrossref
MR Timely follow-up of positive fecal occult blood tests strategies associated with improvement. Am J Prev Med
87- 93PubMedGoogle ScholarCrossref
WR Cancer screening by primary care physicians: a comparison of rates obtained from physician self-report, patient survey, and chart audit. Am J Public Health
795- 800PubMedGoogle ScholarCrossref