The incidence and mortality rates of colorectal cancer (CRC) have increased markedly in Japan. Age-adjusted death rates of CRC have doubled during the past few decades, from 8.6 and 7.5 per 100,000 males and females respectively, in 1950 to 22.8 to 13.5 per 100,000 in 2003 [1]. 5-year survival rates for CRC vary dramatically according to the stage of detection from 25% when there is distant spread of the disease, to 95% when the disease is localized [2]. Thus, the prevention and early detection of CRC is of great public health importance.

Screening using the fecal occult blood test (FOBT) has been shown to reduce the incidence and mortality of CRC in randomized clinical trials [3]. In addition, results from case-controlled studies have suggested that screening using immunochemical FOBT could reduce mortality from CRC by 60% or more for Japanese populations aged 40 year and older who screen annually [4]. Patients with an abnormal FOBT (FOBT+) in these trials were routinely followed up with a complete diagnostic evaluation (CDE), and the reduction in mortality demonstrated in FOBT screening trials can be attributed to the use of follow-up CDE. According to the Japanese Ministry of Health and Welfare's CRC screening guidelines, CDE includes colonoscopy or the combination of a double air contrast barium enema X-ray and flexible sigmoidoscopy, or in cases where an endoscopy cannot be completed, a double-contrast barium enema examination is also temporarily acceptable [5].

Under the auspices of the Health and Medical Service Law for the Aged, a national CRC screening program was initiated in Japan in 1992 as part of a public health policy. Asymptomatic populations over the age of 40 are recommended to participate in the screening program, which uses a 2-day immunochemical FOBT. Despite strong consensus from public health academics and cancer epidemiologists in supporting CRC screening, fewer than 60% of screening FOBT+ patients received follow-up evaluation, and this poor follow-up rate remained unchanged between 1992 and 2003 [6]. Limited follow-up rates pose an important obstacle to achieving overall CRC screening effectiveness. Moreover, non-adherence to CDE has been implicated as contributing to adverse outcomes in retrospective analysis of advanced CRC [7]. It remains a matter of urgency to improve clinical follow-up compliance. Therefore, to assist in the development of effective interventions, studies which provide knowledge regarding the factors that facilitate or impede clinical follow up positive FOBT results are needed.

Inadequate follow-up of FOBT+ patients have also been reported in the United States [8–13]. Recently, Nadel and colleagues have showed that 31.6% of the FOBT+ patients did not have any follow-up exams. Previous studies have focused on physicians' potential barriers and facilitating factors implicated in the provision of follow-up exams [10–16], including physicians' background (e.g., board certification, time in practice, specialty, etc.), cognitive and psychosocial representations (e.g., perceived CDE effectiveness, belief that CDE is standard practice, intention to evaluate a FOBT+ with CDE, concerns about CDE related cost, etc.), practice environment (e.g., number of physicians, etc.) and patient characteristics. Regarding patient characteristics, being a male [12], having high social class [16], with the greater number of positive Hemoccult windows, having family history of CRC, patients' CRC worry [10], and consultation with a gastroenterologist [8] have been reported to improve physicians' chances to order an adequate follow-up evaluation.

Non-adherence is perceived to be a global problem, in that both providers and patients appear to share responsibility for the problem [11, 14, 16, 17]. Very little, if any, few published studies focus on factors impacting on FOBT+ patients' decisions to undergo follow-up examinations. Previous research has suggested that noncompliant FOBT+ patients reported time restraints and an absence of current symptoms as major reasons for failure to undergo clinical follow-up after a positive FOBT [18]. However, it must be noted that the study sample employed was small and that the research methodology used descriptive methods only. Recently, Tashiro reported that in comparison with patients who complied with follow-up recommendations, noncompliant patients who were 70 years and older demonstrated worse mental health status and physical functioning [19]. However this study included a methodological limitation in that it measured the FOBT+ patients' quality of life only. In order to target patient education and intervention strategies to improve quality of care for patients with abnormal FOBT, and ultimately to decrease non-compliance to clinical follow-up recommendations, it is important to understand the characteristics motivating and inhibiting patients compliance with clinical follow-up recommendations.

Conceptual framework (Fig.1)

In order to contribute to intervention design, this study examined the factors that may impact intention to comply with clinical follow-up recommendations. We incorporated elements of the Health Belief Model (HBM) [20, 21] and the Theory of Planned Behavior [22] to direct the present research. According to the Theory of Planned Behavior, one of the strongest immediate determinants of behavior is a person's intention to perform it. Empirically, intention to undergo screening remains one of the strongest and most consistent factors associated with actual cancer preventive behavior, including breast [23, 24], prostate [25], cervical [24], and colon [26, 27] cancer screening. The main outcome of variable of this study was intention.

The present study used four original constructs from HBM as identifying factors that may facilitate or restrain follow-up intention. According to the original HBM, a person's decision regarding undertaking a preventive action related to a disease is influenced primarily by the following four beliefs: (a) perceived susceptibility (one's opinion of the chances of getting a condition), (b) perceived severity (one's opinion of how serious a condition and its sequelae are), (c) perceived barriers (one's opinion of the tangible and psychological costs of the advised action), and (d) perceived benefits (one's opinion of the efficacy of the advised action to reduce risk of seriousness of impact). Furthermore, knowledge regarding cancer [28–31] and social support from family and/or the health care professionals [32–37] that have been shown to be related or associated with screening behavior and intention were also included in our framework.

To the best of our knowledge, the present study is the first investigation into the factors relating to adherence to recommendations for clinical follow-up after an abnormal screening FOBT result. This theory-driven research presents useful findings that have not been previously investigated. Among the 288 individuals analyzed, 74.7% indicated that they would definitely adhere to recommendations for follow-up, we identified that higher perceived barriers were significantly associated with lower intention, and that high-perceived benefits and greater knowledge of CRC risk were associated with increased intention. In the following discussion we discussed in accordance with Table 2.

Several studies published to date have widely used HBM or part of HBM to design theoretical framework to facilitate understanding of CRC screening interests [31, 33, 36, 37, 46, 51] and behaviors [28, 30, 42–44]. During the past decade, HBM has also been used to investigate Japanese stomach cancer screening behavior [62] and preventive health behaviors [32]. Despite variations in study design and measurement of cancer screening attitudes and behaviors, considerable support for the HBM has been documented. In this study, perceived barriers and perceived benefits were found to significantly associate with intention, our findings suggest that the HBM provide a useful framework for understanding patients' attitudes and beliefs regarding follow-up for abnormal CRC screening results.

Barriers to performing recommended behaviors are one of the key components in the HBM and have been shown to be strong predictors of taking action [21, 63]. We assessed the four aspects of barriers, and our findings will help clinician and researchers to develop effective interventions to increase patients' compliance with recommendations for further diagnostic follow-up.

The misconception held by respondents that follow-up consultations were "unnecessary unless symptomatic" and "unnecessary for my age" reflect lack of adequate understanding regarding the purpose of screening to detect CRC or polyps at an asymptomatic stage [42]. Respondents may lack understanding of the meaning of positive FOBT results [35, 64], misinterpret recommendation [35], or lack of distinction regarding asymptomatic screening versus diagnostic testing [65]. In addition, we suggested that age accounted for nonparticipation in the following two ways: (1) that the young may perceive that only older people need follow-up examination, and (2) the older may believe that they are too old and too close to death to look for illness [66]. Based on previous studies, lack of understanding of the necessity for follow-up tests were also found among participants of breast [55] and cervical cancer screening studies [35, 52, 53], and noncompliance was attributed to lack of adequate communication regarding screening results [35] and necessity for follow-up testing [64]. Misperception may lead some patients to belief findings are normal and do not require further attention [17]. Although follow-up guidelines for abnormal FOBT have greater consensus and less variation, Japanese cancer screening programs currently use a generic letter providing information regarding screening results and notification for follow-up. Since a well-informed screening participant may be more likely to follow up promptly after an abnormal screening result [57], improving communication of the meaning of positive screening results, the purpose and necessity of follow-up examination, and clarifying misperceptions regarding follow-up may be important components for further interventions.

Perceived unpleasantness regarding preparation for CDE and discomfort of CDE procedures have also been identified as barriers to compliance with endoscopic screening [28, 30, 31, 36, 37, 41, 42, 44, 45, 47, 48, 50, 67]. Despite lack of experience, the majority of subjects thought that procedure of CDE was uncomfortable. Results from the current sample are partially inconsistent with the literature, in that we observe significant association of perceived embarrassment and intention. It may be that since these patients already understood that CDE is invasive and unpleasant, they had already anticipated a certain degree of pain or discomfort but believed they could cope. In fact, patient who had undergone endoscopy examination found that the procedure more comfortable than expected [68, 69]. Nevertheless, combined with effects to improve technical skills making CDE comfortable, health providers should assist patients in accepting the CDE [47].

If notified of abnormal FOBT results, a very high percentage of respondents perceived that they were "afraid of finding cancer", and after controlling co-variables, the perception of "uncomfortable talking about CRC" showed a negative association with intention. The psychological impact of abnormal cancer screening results, such as fear of diagnostic test, anxiety and worry about cancer were indicated in the literature regarding follow-up of cervical [53, 70, 71], breast [56, 57, 72] and prostate cancer screening [54]. Abnormal cancer screening findings and recalls for further investigation have the potential to generate psychological distress, and may be one reason of patients fail to undergo follow-up testing. Our findings indicate that the recommendation letter and the communication regarding follow-up should be created cautiously; as such communication will elicit and address patients' concern. Patients with special concerns or questions should be encouraged to contact the screening provider.

As well as findings in the FOBT and/or endoscopy screening for CRC [31, 36, 41, 47, 65, 73], respondents identified time constraints such as "being busy" or "having other priorities" to undergo follow-up check. Compared with FOBT screening, follow-up employing CDE is much more demanding and invasive. Motivation and time is required for proper bowel preparation, and undertaking the test usually necessitates time off work for the participant. In Japan there are local variations in the capacity of conducting endoscopic diagnostic examinations for positive FOBT [74], difficulties in obtaining consultation and long waiting periods for follow-up medical appointments have interdependent effects on the risk for noncompliance [75], thus shortening the interval from screening to diagnostic examination, making follow-up testing more convenient and accessible may help to motivate those persons to overcome competing demands [52, 55, 73]. A few case studies have suggested that diagnostic examination in CRC screening were improved by reducing inconvenience barriers [75], the intervention efficiency requires further evaluation more thoroughly using studies with a randomized controlled design. Our study also identified that the necessity to pay out-of-pocket patients' costs constitutes a barrier to intention. Fees for screening are partially met by the local government, and fees for follow-up tests are partially subsidized by patients' health insurance. For example, individual payment for FOBT screening is only about 800 yen (approximately US$ 7) and individual payment for colonoscopy is approximately 7,500 yen (approximately US$ 65), with other costs (e.g., colonoscopic polypectomies) running to approximately 24,000 yen (approximately US$ 209). Previous research suggested that being uninsured or underinsured was associated with delayed or incomplete follow-up of cervical cancer screening[17] and out-of pocket costs have influenced patients' decision making [76], thus factors in willingness to pay in encouraging follow-up for abnormal FOBT screening is needed in the future.

In this study, patients with high-perception of benefits were more likely to state that they intended to adhere to follow-up recommendations than patients having low-perception of benefits. These findings support a substantial body of research on the association between benefits and intentions in participating in cancer screening [36, 41]. In relation to a review of previous studies that used constructs from the HBM, Janz and Becker indicated that perceived benefits was a more important factor for sick-role behaviors than for preventive health behavior [63]. Follow-up examination appears to function more as an early detection or sick-role behavior, where anticipated benefits emerge as the stronger influencing factor [43]. Our findings can readily be interpreted within this context. Although CRC screening has convincing evidence regarding effectiveness, it appear that messages regarding CRC screening exaggerate its potential benefits and fail to mention the potential uncertainties (e.g., false-positive results of FOBT) and limitations of CRC screening (e.g., false-negative results of FOBT and CDE), as well as the possible complications associated with CDE (e.g., occasional risk of perforation, bleeding, and cardio-respiratory events from intravenous sedation). To better understand the follow-up procedure and to obtain a truly informed decision about CRC screening, we believe it is necessary to promote more balanced information rather than a one sided message, which states screening can only be universally beneficial [77, 78].

Overall, the percentage of correct answers for knowledge of CRC risk factors was high. We suggest that Japanese people are highly aware of CRC as a result of 10-year national CRC screening program. Screening participants also appear to be relatively health-conscious and thus, reported a greater knowledge of CRC. Similar to reports from previous studies [29, 31], few subjects are aware that increasing age is a risk factor for CRC. This lack of awareness may be explained by the suggestion that individuals do not conceptualize unmodifiable characteristics (e.g., gender and age) as risk factors [29]. Knowledge may be viewed as a distal predictor of intention, and in accordance with other research results [31]; our findings demonstrate that knowledge is positively associated with intention. As knowledge of risk factors for cancer has increased, the intention to undertake screening has increased [29]; educating public regarding the risk factors of CRC and increasing public awareness of CRC is required. A limitation of the current study was a restricted assess to knowledge of risk factors, thus further studies should be conducted to provide insight into the other domains of knowledge regarding CRC.

Perceived severity has a tendency to associate with intention, but failure to achieve statistical significance. Our finding is not consistent with results from previous studies that have reported that perceived susceptibility has an association with behavioral intention [33, 36, 44, 50, 67]. It is possible that perceived susceptibility might be indirectly connected to follow up intention mediated by the others variables [36, 79]. Social support was not a significant predictor of intention in this study, and it may due to the relative crudeness of our measurement. Furthermore the sample in our investigation was homogenous, limiting the ability to detect significant associations between patients' characteristics and intention.

This study should be perceived as the first step concerning the investigation of patients' factors potentially associated with inadequate follow-up of abnormal FOBT screening. Our findings should be interpreted in light of some important limitations. Firstly, since this is a cross-sectional study, a strong inference about causality cannot be drawn. Secondly, intention was utilized as the outcome variable in the present study. While a considerable body of research supports the predictive validity of intentions for a variety of behaviors, prospective studies should be conducted to determine how well the variables identified here, including intention, and other variables serve to predict actual compliance with follow-up tests. Thirdly, since this sample was recruited from a single community the generalization of findings is limited. Fourthly, since this study examined patients' beliefs regarding CRC and follow-up under a potentially abnormal FOBT, combined with the limited sample size of patients who reported previous bowel examination was not possible, further research is required to embrace those patients. Finally, the schedule for this study did not allow time for a large pilot study sample to establish reliability and validity of perceived barriers scales, and a recommendation for further research would be for further analysis to be conducted correlating social-psychosocial variables and intention in one model simultaneously.

Despite a number of limitations, this theory-driven research identified three factors that were significantly associated with intentions to undergo follow-up exams: perceived barriers, perceived benefits, and knowledge of CRC risk factors. Ours findings suggest that future interventions should focus on reducing modifiable barriers by clarifying misperceptions about follow-up, promoting the acceptance of CDE, addressing psychological distress, and making follow-up testing more convenient and accessible. Moreover, educating the public regarding the risk factors of CRC and increasing understanding of the benefits of follow-up is also important. We believe that further research is required to apply the findings from this study to develop appropriate information and effective methods of communication and to identify the best strategies for interventions.