The contribution of a negative colorectal screening test result to symptom appraisal and help‐seeking behaviour among patients subsequently diagnosed with an interval colorectal cancer

Abstract Background Colorectal cancer (CRC) screening programmes using a guaiac faecal occult blood test (gFOBt) reduce CRC mortality. Interval cancers are diagnosed between screening rounds: reassurance from a negative gFOBt has the potential to influence the pathway to diagnosis of an interval colorectal cancer. Methods Twenty‐six semi‐structured face‐to‐face interviews were carried out in Scotland and England, with individuals diagnosed with an interval colorectal cancer following a negative gFOBt result. Results Participants reported they were reassured by a negative gFOBt, interpreting their result as an “all clear”. Therefore, most did not suspect cancer as a possible cause of symptoms and many did not recall their screening result during symptom appraisal. Among those who did consider cancer, and did think about their screening test result, reassurance from a negative gFOBt led some to “downplay” the seriousness of their symptoms with some interviewees explicitly stating that their negative test result contributed to a delayed decision to seek help. Conclusion Screening participants need to be informed of the limitations of screening and the ongoing risk of developing colorectal cancer even when in receipt of a negative result: the importance of minimizing delay in seeking medical advice for colorectal symptoms should be emphasized.


| INTRODUCTION
In the UK and other high-income countries, colorectal cancer (CRC) is a leading cause of cancer-related mortality. 1,2 Population-based CRC screening programmes have been introduced in a number of countries and have been shown to reduce CRC mortality. 3 Since 2006, the UK has introduced CRC screening programmes based on the guaiac faecal occult blood test (gFOBt), with colonoscopy offered to those who receive a positive result. UK programmes are currently switching to use of faecal immunochemical testing (FIT).
Following a successful trial, 4 the UK is also currently implementing national flexible sigmoidoscopy screening programmes that will run alongside existing biennial screening.
Despite CRC screening programmes being in place, uptake of CRC screening in the UK is approximately 55% and the majority of CRC cases will present symptomatically. [5][6][7][8][9] Furthermore, gFOBt screening is associated with a high proportion of interval cancers: between 30% and 50% of all CRCs detected in the screened population in the Scottish and English pilot programmes. 10,11 Interval cancers include cancers that have developed between screening rounds and "missed" cancers following a false-negative screening result; they have poorer survival when compared to screendetected CRC. [12][13][14] A number of studies report that people who experience potential cancer symptoms rarely initially interpret them as such, often normalizing symptoms or attributing them to something else. 2 The Model of Pathways to Treatment is a conceptual framework developed to describe the complexity of pathways leading to a cancer diagnosis and defines the processes within 4 key intervals (appraisal, help-seeking, diagnostic and pre-treatment) comprising key events from the detection of a bodily change through to the start of treatment. 15,16 The patient interval encompasses appraisal and help-seeking. 17 Evidence suggests that one of the main factors contributing to a long patient interval is non-recognition of symptom seriousness resulting in increased time to presentation and diagnostic delay. 18,19 Previous reassurance from a healthcare provider for a similar symptom or receipt of a previous "all clear" diagnosis has been associated with delays in help-seeking for potential cancer symptoms. 20, 21 We and others have reported that participation in the NHS bowel screening programmes has the potential to lead to a delay in help-seeking for some patients following the onset of symptoms, through over-reassurance from a negative gFOBt result. 22,23 Unintended consequences, including over-reassurance, from a negative or "normal" result have been demonstrated in breast cancer and other screening programmes. [24][25][26] The aim of this study was to explore, through individual interviews with screening participants who were diagnosed with an interval colorectal cancer, if receiving a negative screening result in a CRC screening programme contributed to their subsequent response to symptoms potentially indicative of cancer, and their decision to seek medical advice.

| Design
Semi-structured individual face-to-face interviews were carried out with patients diagnosed with an interval colorectal cancer following a negative gFOBt result through colorectal cancer screening (bowel screening programmes). Interviews were carried out within 6 months of patients receiving their cancer diagnosis.

| Setting and recruitment
All patients newly diagnosed with a primary colorectal cancer and who had received a negative gFOBt result (within the preceding two years) from the Scottish bowel screening programme or the Midlands and North West screening programme hub were eligible to take part in the study. Eligible screening age for the two respective programmes was 50-74 years (Scotland) and 60-74 years

| Exclusion criteria
Suitability based on clinical grounds (physical or mental health) was at the discretion of the secondary care team and primarily the patient's colorectal surgeon in charge of their care.

| Data collection
Interviews were carried out by a female researcher, KB (MA, MSc, PhD), with previous training and experience in conducting qualitative research. Interviews took place between August 2013 and June 2014. Interviews lasted between 45 and 60 minutes, and signed consent was obtained before the interview commenced. All interviews were audio-recorded with the participant's permission, professionally transcribed verbatim and anonymized. Field notes were made both during and after each interview. Audio files were deleted once the written transcripts had been received and verified by the researcher.
Interviews took place in the patient's home, and some patients were accompanied by their spouse or friend during the interview.

| Topic guide
Interviews explored two key domains: (i) the participant's "pathway to diagnosis" including: initial symptom appraisal, symptom progression, first presentation to any healthcare provider and any perceived delays, and (ii) the contribution, if any, of their negative gFOBt result to their symptom appraisal or decision to seek help.

| Ethical review
The project was granted ethical approval from the South East Scotland Research Ethics Committee 01 (11/SS/0006).

| Analysis
All transcripts were read by both the project researcher (KB) and the principal investigator (CC). Emerging themes were discussed and a coding frame agreed with additional codes included when appropriate as data collection progressed. Final themes were agreed through an iterative process involving the core and wider research group.
Thematic analysis of the data was undertaken using NVivo software (QSR International, V.10) and was ongoing throughout the study to allow emerging themes to be fed back into the data collection. An inductive reasoning approach was adopted where themes (or categories) were identified through careful examination and comparison of the data which involved 6 stages: familiarization of the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes and producing a report. 27

| Recruitment
A total of 27 participants were recruited, with 17 interviews carried out in Tayside, Scotland, and 10 in the Midlands and North West, England. Recruitment was ongoing throughout the study; 50% of patients who received an invitation to take part in the study agreed to be interviewed. Twenty-one participants presented to their GP with symptoms prior to their diagnosis, two attended their GP for a routine diabetic health check, one participant was diagnosed following an emergency admission to hospital and two participants were diagnosed following a surveillance colonoscopy. One participant, wrongly identified, had received a positive gFOBt result in the screening programme, which led to their diagnosis: this interview was not included in the reported findings. A summary of study recruitment and participant demographics is provided in Table 1.

| Reported symptoms
Participants described a range of symptoms and symptom characteristics, which differed in frequency, duration and severity prior to receiving their colorectal cancer diagnosis. Reported symptoms included: noticeable blood on toilet paper or in the toilet bowl or stool following a bowel movement, constipation and/or diarrhoea, change in bowel habit, weight-loss, indigestion, nausea and sickness, fatigue and pain.
Participants who were diagnosed following a routine diabetic health check reported that they did not experience any symptoms preceding their cancer diagnosis.

| Main themes
We found the relationship between a negative gFOB result and a patient's diagnostic journey to be multifaceted, with varying opportunities or scenarios in which screening participation and the receipt of a negative result had the potential to influence the appraisal interval following the onset of symptoms ( Figure 1). We identified three key stages where a negative gFOBt result had the potential to contribute to the appraisal and/or help-seeking interval outlined in the Model of Pathways to Treatment, 15

| Trust and Reassurance
Screening participation was associated with low levels of apprehension, with the receipt of a negative test result leading to positive emotions such as pleasure and relief. For some participants, who felt well and perceived themselves to be healthy, a negative result was simply confirmation of their health status. Participants portrayed a high degree of confidence and trust in their negative screening result describing it as "an all clear" that they "were ok" or that they had "passed." Participants did not question the reliability of the test result when they received their result letter although some participants, following more in-depth reflection, acknowledged the information or disclaimer, provided with the screening materials promoting symptom vigilance and help-seeking between screening episodes.

| Time since last screen
There was some evidence to suggest that patients who had recently participated in colorectal screening, prior to the onset of symptoms, for example in the preceding few months, were more likely to reflect on their negative screening result and recall the information provided with the result letter. Some participants were more explicit in their accounts that the reassurance provided by their negative test result contributed to them "downplaying" the potential severity or urgency of their symptoms.
However, there was no strong indication that receipt of a negative result was the primary (or sole) driver for delaying consulting with a GP. It was described more as a contributing factor to the appraisal and help-seeking interval in which a number of other factors and competing demands, such as caring for a spouse or going on holiday, influenced the timing of their decision to seek help. It's important to note that although a small number of patients were influenced by their negative test result, most interviewees believed that they presented promptly to their GP following the onset of symptoms (generally within a number of days or weeks). Early recognition of symptoms was usually followed by a short period of self-monitoring where participants adopted a "watch and wait" approach to see if symptoms resolved by themselves. Generally, when symptoms persisted or progressed in severity, or following a period of self-treatment using overthe-counted medications, for example for suspected haemorrhoids or constipation, participants chose to consult with a GP or HCP.

I just thought I'll wait a few weeks and see if it's… see if it
improves and I was using proprietary stuff, you know. Erm, that I bought out the chemist.

(Male, Tayside #15)
The presence of pre-existing conditions or comorbidities and their treatment (prescribed or over-the-counter medications) was associated with an extended appraisal and diagnostic interval. This was evident among participants independent to any reflection on previous screening participation during their symptom appraisal. The impact of comorbidity on the pathway to diagnosis was two-fold: (i) some participants reported that they attributed their new symptoms to a pre-existing diagnosis and therefore did not seek immediate advice or view the symptoms as serious, and (ii) once a decision had been made to seek help, referral to secondary care was sometimes delayed due to an initial focus, both by the patient and by the GP, on alternate treatments or investigations for a pre-existing condition.

| Knowledge/acceptance of the limitations of gFOBt screening
While some participants recalled their negative gFOBt result during the appraisal interval, others only considered their participation in the colorectal screening programme after they had received their cancer diagnosis. These participants tended to respond in one of two ways when reflecting on their negative gFOBt result, which was dependent on their knowledge and acceptance of the limitations associated with gFOBt screening. For some, the process of being given a perceived "all clear" followed by a colorectal cancer diagnosis led to some degree of uncertainty regarding the accuracy of their test result and a loss of confidence in the screening programme (subtheme). In contrast, others reported that they understood that the gFOB test was not 100% effective and was not a guarantee that they did not have or would not develop cancer in the future, recognizing the importance of monitoring symptoms and recalling the disclaimer that accompanies the screening result letter.  ity of participants did not consider their previous gFOBt result (nor consider a potential colorectal cancer diagnosis) following the onset of symptoms, and therefore, reassurance from a negative gFOBt did not contribute to a delay in the patient interval. However, for a smaller group of patients who did consider colorectal cancer as a possible cause of their symptoms (or were experiencing symptoms intermittently or at the time they completed their screening test), a negative gFOBt result did influence the patient interval and contributed to a delay in seeking help. A negative gFOBt result led participants to "downplay" the seriousness of their symptoms by reducing their cancer suspicion and reinforcing "benign" causal beliefs having been reassured by a recent "all clear" test result. Preexisting conditions, particularly when the characteristics of any new symptoms were similar to those previously experienced, led to some delay across the symptom appraisal, help-seeking and diagnostic interval. 15

| Comparison with existing literature
The processes within the appraisal interval as part of the Model of Pathways to Treatment, 15 namely "patient appraisal and selfmanagement," draw on a number of psychological theories including Leventhal's Common Sense Model of Illness self-regulation and the Andersen Model of Total Patient Delay. 16,28 Consistent with our findings, Leventhal's model explains that initial symptom appraisal, or recognition of a bodily change, when symptoms do not exceed a certain level of interference, leads to symptoms being normalized or dismissed. However, when symptoms persisted, or increased in severity (eg exceeded a threshold of interference), patients adopted a different approach where they self-monitored their symptoms or sought self-treatments prior to making the decision to seek medical advice. Furthermore, as reported in the literature, we found that many participants did not attribute their symptoms to a potential cancer diagnosis despite some experiencing a colorectal cancer "alarm" symptom, for example blood in stool. 2 Lack of recognition of the seriousness of a symptom has been reported extensively in the literature as one of the main contributing factors to a delayed patient interval. 18,19 In both the present study and the study carried out by Solbjor and colleagues examining interval breast cancers following mammography screening, a negative screening result was shown to impact on the patient interval primarily through reassurance which led some participants to "downplay" the seriousness of their symptoms and to assume a benign cause. 26 Hall and colleagues found that a negative screening result, along with a number of other factors including absence of blood or pain, or "feeling well," was viewed as reassuring. 23 In the same way, reassurance from a previous "all clear" investigation has been associated with a delay in help-seeking even when symptoms persisted or occurred a number of years later with some patients believing that they would not be taken seriously if they returned with similar symptoms. 20,21 The added complexity of the presence of comorbidities (or competing diagnoses) and the associated delay on the diagnostic pathway potentially resulting in late-stage disease is an important and recognized challenge for healthcare providers, 29,30 particularly in the light of the recent evidence published by Torring and colleagues who report any delay to the primary care interval to be associated with more advanced colorectal cancer. 31 Interval colorectal cancer has been shown to have an adverse effect on trust in FOBt screening, with the same study reporting poorer quality of life among interval cancer patients when compared to those with screen-detected disease. 32 Similarly, this has been demonstrated in mammography screening where a diagnosis of an interval breast cancer influenced patient's trust, although not to the point of creating distrust, where women instead saw themselves as exceptions in an otherwise beneficial screening programme. 33 We also found that despite some interviewees suggesting a loss of confidence in the screening programme following their diagnosis, participants generally remained positive towards the overall benefits of screening and expressed a willingness for continued participation.

| Strengths and limitations
Participants were recruited from hospitals in both Scotland and England, providing narratives from cancer patients who participated in different area-based bowel screening programmes. Carrying out semi-structured interviews within six months of diagnosis permitted an in-depth exploration of the participant's pathway to diagnosis; nevertheless, accounts are retrospective and therefore subject to recall and framing bias. We did not have access to any clinical or demographic information for those who declined to be interviewed, nor the number that the hospitals did not consider eligible to contact: their experiences and perspectives may differ.

| Implications for practice and policy
Although reassurance from a negative gFOBt result is appropriate for most screening participants, there is a risk of over-reassurance for some. Methods of increasing understanding of the limitations of gFOBt screening (and indeed of FIT screening) and the ongoing risk of developing cancer despite receiving a negative test result are needed; the significance of any associated delay in seeking medical advice in terms of clinical outcomes also merits further exploration. Similarly, finding effective ways to engage with screening participants that complement existing initiatives in primary care with regard to symptom vigilance and prompt help-seeking behaviour remains an important challenge; there is a need to develop nuanced health messages to inform screening participants of the importance of help-seeking for vague as well as alarm symptoms following a negative test result, particularly in the presence of comorbidities.
The use of FIT as a diagnostic tool in primary should allow more streamlined diagnostic pathways for all patients. 34