Do Men Know Which Lower Bowel Symptoms Warrant Medical Attention? A Web-Based Video Vignette Survey of Men in Western Australia

Abstract

The current study aims to explor how men would advise other men about seeking help for lower bowel symptoms and also to determine the factors that may influence help seeking. A purposive sample of Western Australian men aged 18 years and older was recruited for the study. Participants completed 8 of the 28 randomly assigned video vignettes (video clips) displaying men (older or younger) with various combinations of one or more lower bowel symptoms. Participants were asked if the person in the vignette should seek health advice. Subsequently, the participants answered a set of questions based on the Health Belief Model. A total of 408 participants (response rate = 51%) answered 3,264 vignettes. Participants younger than 50 years, participants who were not tertiary educated and those who had lower incomes, or those living in regional or remote areas were less likely to advise help seeking from general practitioner (GP). Participants who visited their general practitioner less frequently were also less likely to advisehelp seeking. There was a trend to consider unintentional weight loss and diarrhea as minor symptoms not necessitating medical attention compared with rectal bleeding. The findings suggest for a need to improve public awareness among men about the need to seek timely medical advice for lower bowel symptoms in primary care. The importance of early presentation of persistent lower bowel symptoms must be specifically targeted at men younger than 50 years, those with lower incomes, or residing in regional or remote areas.

Keywords

men’s health lower bowel symptoms rectal bleeding vignettes help seeking health behaviors health belief model

Introduction

Colorectal cancer (CRC) is the third most common cancer in men with more than half the cases occurring in developed countries (Ferlay et al., 2013). Australia and New Zealand together account for the highest incidence of CRC in the world with age-standardized incidence rate of 44.8 per 100,000 men (Ferlay et al., 2013). CRC is often symptomatic and has a high potential for curative treatment when detected at an early stage (Lindholm, Brevinge, & Haglind, 2008). In contrast, late stage presentation is associated with a poor 5-year survival (McArdle & Hole, 2004). Although factors related to both patient delay and the health care system delay have been reported to be associated with diagnostic delays in CRC, patient delay accounts for the longest time delay (Langenbach, Johannes Schmidt, Neumann, & Hubert Zirngibl, 2003). The majority of CRC patients experience lower bowel symptoms such as rectal bleeding, change in bowel habit, abdominal pain, and weight loss (D. Smith et al., 2006); however, the rate of seeking medical advice for such symptoms is low (Chaplin, Curless, Thomson, & Barton, 2000; Hamilton & Sharp, 2004; Thompson, Pond, Ellis, Beach, & Thompson, 2000). These symptoms also occur in benign bowel diseases, are also common in the general population and are often self-limiting (Adelstein, Macaskill, Chan, Katelaris, & Irwig, 2011; Courtney et al., 2012b).

In their study on Australian population (n = 1,332), Cockburn, Paul, Tzelepis, McElduff, and Byles (2003) reported that although nearly 70% of adults were aware of rectal bleeding as a sign of CRC, less than half knew about the other potential symptoms of cancer. In the same study nearly 1 in 4 individuals were unaware of at least one symptom of CRC. The authors also identified that the awareness of these symptoms alone did not ensure timely help seeking among the participants. Similar views were echoed by Hashim et al. (2011) who found no association between the knowledge of cancer symptoms and delay in presentation of rectal bleeding and also highlighted the gap between knowledge and help-seeking behavior in Malaysian population. A recent Danish study by Pedersen, Hansen, and Vedsted (2013) reported that CRC patients with rectal bleeding experienced longer delays in seeking medical advice than patients without rectal bleeding when adjusting for other symptoms such as change in bowel habit or abdominal pain. The authors also suggested that the attribution of symptoms to benign causes was not the only reason for patient delay, and emphasized on the need to identify the barriers to help seeking. In another study on Australian adults (n = 1,592), Courtney et al. (2012a) reported that up to 18% of people experiencing rectal bleeding had never consulted a doctor and nearly the same proportion delayed for more than 1 month before seeking medical advice. Up to 20% of people who experienced a change in bowel habit had never sought medical advice and 37% delayed for more than 1 month before seeking advice.

No major improvement in the consultation rates for lower bowel symptoms have been reported over the past two decades (Cockburn et al., 2003; Crosland & Jones, 1995; Thompson et al., 2000). The issue is rather more important in an Australian context where the estimated rates for CRC are reportedly the highest in the world (Ferlay et al., 2013). Adelstein, Macaskill, Turner, and Irwig (2012) observed that patients, who considered their symptoms as less serious, had a threefold higher prevalence of cancer than those who considered their symptoms as serious. Qualitative studies focusing on Australian men have reported that the rationalization of the lower bowel symptoms or self-medication was a common practice among men and prevented them from seeking medical advice. The fear of diagnosis and the cost of consulting a general practitioner (GP) were other barriers to help seeking (Oberoi, Jiwa, McManus, & Hodder, 2015b). Although majority of GPs in Australia (nearly 83%) bulk bill their patients (i.e., the GPs accept Medicare rebate as full fee for their consultation), the patients may have to bear significant out-of-pocket costs (nearly AUD30 for a GP consultation and more than AUD200 for a specialist consultation) if they consult non–bulk billing doctor (Department of Health and Ageing, 2013). Furthermore, it is speculated that the proposed copayment fee to be paid by Australians for each GP visit as well as for diagnostic investigations and the proposed increased copayments for prescription medications in future is further likely to discourage people from seeking timely medical advice (Kirby, 2014). Past studies have also reported that support from female family members played a vital role in encouraging men to seek help (Oberoi, Jiwa, McManus, & Hodder, 2015a, 2015b).

To facilitate timely help seeking among men, it is important to understand the factors that affect their decision to seek help. The current understanding of the medical advice–seeking behavior for lower bowel symptoms is largely based on retrospective studies highlighting retrospective recall of symptoms and time intervals from symptom onset to help seeking (Basaranoglu et al., 2008; Chaplin et al., 2000; Cockburn et al., 2003; Macleod, Mitchell, Burgess, Macdonald, & Ramirez, 2009). In their literature review, Oberoi, Jiwa, McManus, and Hodder (2014) discussed various behavioral factors associated with delay in seeking medical advice for lower bowel symptoms. Such information may be subject to memory-recall bias as it may be difficult to remember the exact timing of vague and nonspecific symptoms and the health actions taken in response to the symptoms (Cockburn et al., 2003; Korsgaard, Pedersen, Sørensen, & Laurberg, 2006).

The current study aimed at exploring the factors that may affect men’s help-seeking decisions with regard to the symptoms of CRC through the use of standardized vignettes. Vignettes (hypothetical patient scenarios) may be useful in eliciting information about the questions that may be personally intimidating, or about the information that may be either underreported or distorted in a socially desirable direction (Barter & Renold, 2000; Brener, Billy, & Grady, 2003; Wallander, 2009). Another advantage is that the vignettes can reflect real-life scenarios, while controlling for external or potentially confounding variables. Vignettes are also reported to control for the contextual variations between the vignette characters (Kinicki, Hom, Trost, & Wade, 1995; Ramanathan, Baratiny, Stocks, Searles, & Redford, 2011), thereby allowing all the participants to respond to similar stimuli. The advantages of using the vignettes in social research have been summarized by Hughes and Huby (2002).

Method

Theoretical Framework

The health belief model (HBM) has been used as the theoretical framework in this study. The HBM purports that people’s decision to commit to a recommended health action (to seek medical advice) is associated with their perception of the severity of symptoms, susceptibility to the disease, the perceived benefits of the action, the absence of barriers to the action and confidence in the GP or family physician (Strecher & Rosenstock, 1997). The model also proposes that demographic variables (i.e., age, sex, or ethnicity), sociopsychological variables (i.e., personality or social class) and structural variables (i.e., knowledge about the disease or personal experience with a condition) may indirectly influence health-related behaviors by affecting an individual’s perception of susceptibility, severity, benefits, and barriers (Pinto, Lively, Siganga, Holiday-Goodman, & Kamm, 2006; Rosenstock, Strecher, & Becker, 1994). As the HBM was originally developed to explain preventive behaviors it is a suitable model for the current research (Shaw, Brittain, Tansey, & Williams, 2008).

Study Population

In this vignette-based study, participants were recruited from the general population in the state of Western Australia through a research recruitment agency and data were collected between August and October 2012. The potential participants were approached by the recruitment agency and those willing to participate were included in the study. Men aged 18 years or older who were able to communicate in English, had access to the Internet, and were willing to complete a self-administered online questionnaire were eligible. Although, the average age at diagnosis of CRC in Australia is around 69.3 years, approximately 6% of patients diagnosed with CRC in Australia are below 50 years of age (Bowel Cancer Australia, 2013; Wu et al., 2012). The sample was selected in a purposive manner so as to closely resemble the Western Australian population with regard to age, income, educational and employment status, relationship status, and country of birth (consistent with Census Western Australia, 2011). The participants were able to give informed consent before participating in the survey.

We focused exclusively on male participants, because men are not only observed to have a higher biological risk and prevalence of CRC compared with women, but are somewhat more likely to delay seeking medical advice for primary symptoms of CRC (Bray, Ren, Masuyer, & Ferlay, 2013; McArdle, McMillan, & Hole, 2003; McCashland, Brand, Lyden, & de Garmo, 2001; Oberoi et al., 2014; Wichmann, Müller, Hornung, Lau-Werner, & Schildberg, 2001).

Based on the number of explanatory variables (n = 20), and to achieve a power of 80% to detect variables exhibiting a moderate effect size, with α = .05, a minimum sample size of 156 participants was required (Green, 1991). The explanatory variables were clinical variables presented in the vignettes (Appendix A), the demographic variables (Appendix C, Section 1a) and the five domains of the HBM (Appendix C, Sections 1b and 2). In consistency with the HBM, which was used as a theoretical framework, demographic variables such as age, income, educational and employment status, relationship status, country of birth and/or duration of stay in Australia, inhabitance (metropolitan or regional/remote areas), and private health insurance were used as covariates.

Materials and Procedure

The Curtin University Human Research Ethics Committee (Approval No.: PH-02-12) approved the study. Informed consent was obtained from the participants.

Survey

Twenty-eight video vignettes (hypothetical scenarios) were developed. The content of the vignettes was informed by the NICE guidelines (National Institute of Health and Care Excellence, 2005) for symptoms indicating high risk for CRC and the clinical and population-based studies (Courtney et al., 2012a; Hashim et al., 2010; Lynch et al., 2008; McMinn, Steel, & Bowman, 2011; Philipp, Wood, Heaton, & Hughes, 1993; Stapley, Peters, Sharp, & Hamilton, 2006). CRC is assumed to be primarily a disease of the elderly and the incidence is higher in individuals aged more than 50 years. The average age at diagnosis of CRC in Australia in 2009 was 69.3 years. However, approximately 6% of patients diagnosed with CRC in Australia are younger than 50 years (Bowel Cancer Australia, 2013; Wu et al., 2012). Therefore, age was used as one of the clinical variables for the vignettes and vignettes were created using both older and younger men. The five binary clinical variables describing each vignette are shown in Appendix A. Each scenario was presented as a short story and styled in the manner shown in the examples in Appendix B. The vignettes were developed using professional actors and the consent to publish the videos was obtained.

The video vignettes were preferred in this study over text-vignettes as video vignettes remove sources of bias not addressed in other survey methods. Various clinical symptoms and psychosocial variables can be presented more realistically in the video vignettes, which may ignite the participant’s own emotions, memories, attitudes, and biases that may influence their responses (Hawley, Adams, & Galbraith, 2010).

Each participant answered a set of demographic questions before proceeding to the vignettes. Based on the scenario played in the vignettes, the participants were asked whether the man in the vignette should seek professional health advice, which health professional should he seek advice from, and how urgently should he seek advice.

As a next step, the participants were asked a set of questions related to clinical scenario presented in the video-vignettes and based on five themes of the HBM: severity of the symptoms (two items), susceptibility to the symptoms (one item), benefits of seeking advice (two items), barriers to seek advice (one item), and confidence (one item). The respondents were asked to assess their level of agreement with each of the statements relating to these themes on a 5-point Likert-type scale (strongly agree to strongly disagree). These questions were informed by literature review of questionnaires previously used in studies with a direct application HBM-based questionnaire on CRC and lower bowel symptoms (Jacobs, 2002; Macrae, Hill, St John, Ambikapathy, & Garner, 1984; Ozsoy, Ardahan, & Özmen, 2007; Rawl et al., 2001). The questionnaire was tested for reliability and validity. The survey was administered to the participants via web-based software (https://www.qualtrics.com/). Eight randomly selected vignettes were presented to each participant. However, considering that some men in the vignettes would need to seek advice for their symptoms from a GP for diagnosis and referral, the survey was also administered to an expert panel comprising five GPs. The expert panel was invited to identify the cases that would need to seek help from GP and the participants’ responses were matched against the recommendations of the expert panel. The complete questionnaire is reported in Appendix C.

Reliability

The test–retest reliability of the HBM questionnaire ranged between .88 and .98 indicating a highly reliable questionnaire. Cronbach’s alphas for the Severity questions and Benefits questions were .93 and .91, respectively. This shows very high consistency for the pairs of questions used in each of these domains.

Construct Validity

Factor analysis of the HBM responses (after converting the Likert-type scale responses to the numbers 1-5) indicated that a single factor was responsible for 74.1% of the variance in the data. This factor involved all the five domains used in the questionnaire, with four of them (Severity, Susceptible, Confidence, and Benefits) having approximately equal weight, and the weight of the Barrier domain being approximately half (and in a negative direction compared with the others). The factor loadings for the data set were found to be 0.941 (Severity), 0.936 (Susceptible), 0.934 (Confidence), 0.918 (Benefits), and −0.480 (Barriers). This analysis confirms that all domains were relevant and important in explaining the variance in the data set.

Statistical Analysis

A general estimating equation (GEE) was used to identify the relative importance of the HBM, the symptom profile portrayed in the vignette and the demographic profile of the participant in regard to the participants’ decision for vignette referral to seek medical attention. The GEE regression model was used because it took into account the correlation in the data because of the multiple responses made by each participant (eight vignettes per person). Three different GEE models were developed with three dependent variables, respectively: the odds of recommending a help seeking, the odds of recommending help seeking from a GP, and the odds of the participants’ responses agreeing with expert panel comprising five GPs. Observations for the last outcome included those in which the expert panel decided that the men portrayed in the vignettes should seek medical advice for their symptoms. Furthermore, a single HBM score was calculated for each of the participants’ vignette responses based on the factor loadings (a weighted sum of the five responses using the factor loadings as weights). The purpose of the factor analysis was to obtain a single summary score obtained from the linear combination of the HBM questions that could be used in the subsequent regression analyses. To avoid the assumption of a linear relationship between the score and the odds of each outcome, the score for each person was categorized into one of four groups based on the quartiles of the score distribution (scores: 10, 15, 17). The grouped score variable was then introduced into the models as a categorical variable, with the highest quartile group (the most likely to seek advice) used as reference.

Statistical analyses were performed using the SPSS Version 20 and SAS Version 9.2 statistical software programs. A p value < .05 indicates a statistically significant association in all tests.

Results

In the final survey, 408/800 participants recruited to the study met the inclusion criteria and completed the study (response rate = 54.8%). Each participant answered Section 1 of the questionnaire (only once) followed by Section 2 (eight times) based on the eight video vignettes. A total of 3,259 (408 × 8 = 3,264) complete responses to the video vignettes were received (5 responses were incomplete or missing). The demographic profile of the study participants is reported in Table 1. Compared with the Western Australian census data, the survey sample was reasonably representative in terms of age, marital status, country of birth, highest level of education, and occupation.

Table 1.

Demographic Profile of the Study Participants (N = 408).

Variable	Frequency (%)
Age (years)
18-29	23 (5.6)
30-39	100 (24.5)
40-49	120 (29.4)
50-59	91 (22.3)
60-69	57 (14.0)
≥70	17 (4.2)
Place of birth
Australia	263 (64.5)
Other countries	145 (35.5)
Time spent in Australia (years)
≤5	24 (5.9)
6-10	17 (4.2)
>10	367 (89.9)
Education
Year 10 or less	109 (26.7)
Year 12	99 (24.2)
Technical/trade	85 (20.8)
Tertiary/university	115 (28.1)
Employment status
Unemployed	12 (2.9)
Student	12 (2.9)
Employed part-time	46 (11.3)
Employed full-time	273 (66.9)
Retired/pensioner/on social security	57 (14.0)
Other	8 (2.0)
Annual income (AUD)
<20,000	33 (8.1)
20,000-50,000	228 (55.9)
51000-100,000	93 (22.8)
>100,000	40 (9.8)
Prefer not to answer	14 (3.4)
Relationship status
Married	236 (57.8)
De facto	69 (16.7)
Separated/divorced/widowed	42 (10.3)
Never married	61 (15.0)
Private health insurance
No	229 (56.14)
Yes	179 (43.8)
Frequency of general practitioner visit
Never	33 (8.1)
Once a year or less	176 (43.1)
Several times a year	185 (45.3)
1-4 times a month	14 (3.4)

Results of the GEE Models

The proportion of the participants seeking advice from GP for vignettes featuring older men with symptom duration of 3 weeks ranged between 42% and 78% across all vignettes while for vignettes featuring younger men with similar symptom duration ranges between 26% and 84% across all vignettes with different symptom combinations (Appendix D). Additionally, a disagreement was observed between the participants and the expert panel (comprising five GPs) with respect to referring the vignettes to seek help. Nearly 50% of the participants did not agree with the expert panel with respect to seeking medical advice for rectal bleeding lasting for 3 weeks.

Outcome 1: Vignettes Recommended for Help Seeking

Of the 3,259 completed vignette assessments, 2,076 (63.7%) were recommended to seek advice. The odds of recommending the vignettes for help seeking were lower if the participants resided in regional or remote areas as opposed to metropolitan areas (odds ratio [OR] = 0.44, 95% confidence interval [CI] = 0.23-0.83; p < .002) or if their annual income was ≤AUD50,000 (OR = 0.30, 95% CI = 0.18-0.5; p < .0001). The odds of seeking help were also directly related to the participants’ HBM score and their own frequency of GP visits. Those who never visited the GP themselves were 72% less likely to advise men in the vignettes to seek help compared with those who visited their GPs several times in a year (OR = 0.28, 95% CI = 0.15-0.54; p < .0001). Those with HBM score ≥15 were nearly 1,000 times more likely to advise help seeking than those with HBM scores <10 (OR = 0.001, 95% CI = 0.0003-0.002; p < .0001). With regard to the clinical factors, the odds of help seeking increased with the increasing number and duration of the symptoms. Table 2 reports the results of fitting a GEE model to these data and gives the ORs for seeking advice along with 95% CIs and p values.

Table 2.

Relative Odds of Men From Vignette Being Referred to Seek Help.

Variables	n/N (%)	OR	95% CI	p value
Symptom duration
3 days	896/1,600 (56.0)	1 (ref)
3 weeks	1,180/1,659 (71.1)	2.59	[1.97, 3.39]	<.0001
Symptoms
Rectal bleeding	243/466 (52.2)	1 (ref)
Weight loss	118/467 (25.3)	0.15	[0.09, 0.25]	<.0001
Rectal bleeding and diarrhea	379/469 (80.8)	2.12	[1.31, 3.44]	.0022
Rectal bleeding, diarrhea and weight loss	395/471 (83.9)	2.78	[1.71, 4.52]	<.0001

Note. OR = odds ratio; CI = confidence interval. The “n/N” column shows the number of vignette responses seeking advice out of the total number of vignette responses in that category of the independent variable. ORs, CIs, and p values were obtained from a general estimating equation model, which takes into account correlations between responses from the same participant. The analysis is based on 3,259 records (all data). Odds ratios greater than 1 indicated a higher likelihood of seeking advice, whereas odds ratios less than 1 indicated a lower likelihood of doing so.

Outcome 2: Vignettes Recommended for Help Seeking From GP

Of the 2,076 vignettes in which the participants recommended seeking health advice, only 1,517 (73%) were recommended to seek advice from a GP as opposed to other health professionals such as a pharmacist or a nurse practitioner. The odds of recommending help seeking from a GP were higher in participants aged ≥ 50 years than those aged < 50 years (OR = 1.65, 95% CI = 1.07-2.54; p = .023). The odds were also lower for participants who studied up to Year 12 compared with those who had received tertiary education (OR = 0.31, 95% CI = 0.20-0.5; p < .0001). Participants who earned between AUD20,000 and AUD50,000 annually were much less likely to recommend than those who earned >AUD50,000 (OR = 0.13, 95% CI = 0.07-0.22; p < .0001).

Men who believed that visiting a GP was time consuming were 61% less likely to recommend help seeking from GP (OR = 0.39; 95% CI = 0.26-0.59; p < .0001). Additionally, those who had HBM scores <10 were half as likely as those with scores ≥15 to recommend help-seeking advice from GP. With regard to the clinical factors presented in the vignettes, the odds of consulting a GP increased with the increase in the number and the duration of the symptoms. For lone symptoms, the odds were higher for rectal bleeding than for weight loss and diarrhea. Table 3 reports the results of the final GEE model for this outcome.

Table 3.

Relative Odds of Men From Vignette Being Referred to Seek Help From a General Practitioner.

Variables	n/N (%)	OR	95% CI	p value
Symptom duration
3 days	604/896 (67.4)	1 (ref)
3 weeks	913/1,180 (77.4)	3.09	[2.35, 4.05]	<.0001
Symptoms
Rectal bleeding	218/243 (89.7)	1 (ref)
Diarrhea	144/310 (46.5)	0.13	[0.08, 0.21]	<.0001
Weight loss	107/118 (90.7)	0.46	[0.22, 0.95]	.0353
Rectal bleeding and diarrhea	264/379 (69.7)	0.65	[0.44, 0.95]	.0262
Rectal bleeding and weight loss	276/303 (91.1)	1.65	[1.06, 2.58]	.0267
Diarrhea and weight loss	186/328 (56.7)	0.27	[0.18, 0.40]	<.0001
Rectal bleeding, diarrhea and weight loss	322/395 (81.5)	1.60	[1.07, 2.41]	.0235

Note. OR = odds ratio; CI = confidence interval. The analysis is based on 2,076 records (records where no referral was thought to be necessary were excluded from this analysis). While odds ratios greater than 1 indicated a higher likelihood of seeking advice from a general practitioner, odds ratios less than 1 indicated a lower likelihood of doing so.

Outcome 3: Vignette Agreement With the Expert Panel

Vignettes appropriately identified as warranting medical attention were defined as participants’ referral of a vignette for seeking advice that matched the opinion of the expert panel in terms of whether a GP should be consulted. Only vignettes that were considered by the expert panel to require referral to a GP were included in this analysis.

Participants aged ≥50 years were more nearly twice as likely to agree with the expert panel in terms of advising the men in the vignettes to seek advice from the GP compared with those aged <50 years (OR = 1.88, 95% CI = 1.23-2.87; p < .003). The participants who lived in regional or remote areas as opposed to metropolitan areas (OR = 0.43, 95% CI = 0.26-0.72; p < .001) and those who were studied up to Year 10 as opposed to those who had received tertiary education (OR = 0.43, 95% CI = 0.25-0.75; p < .002). Participants with higher incomes were also more likely to agree with the expert panel. Participants who visited their GPs at least once a month were nearly 16 times more likely to agree than those who never visited their GP (OR = 15.9, 95% CI = 2.71-93.81; p < .002) and those who believed that visiting a GP was time consuming were 60% less likely to agree (OR = 0.4, 95% CI = 0.26-0.60; p < .0001).

Participants who had lower HBM scores, visited their GPs less frequently, or believed that visiting a GP was time consuming had lower odds of agreeing with the experts. With regard to the clinical symptoms, the odds of agreement were higher for rectal bleeding as opposed to other symptoms and were also higher when all the three symptoms occurred concurrently or lasted for 3 weeks. The results of the final GEE model for this outcome are reported in Table 4.

Table 4.

Relative Odds of the Participants Agreeing With Expert Panel on the Need for Help Seeking From a General Practitioner.

Variables	n/N (%)	OR	95% CI	p value
Symptom duration
3 days	72/218 (33.0)	1 (ref)
3 weeks	808/1,425 (56.7)	8.22	[4.62, 14.62]	<.0001
Symptoms
Rectal bleeding	119/236 (50.4)	1 (ref)
Diarrhea	119/344 (34.6)	0.26	[0.16, 0.41]	<.0001
Weight loss	31/118 (26.7)	0.14	[0.07, 0.29]	<.0001
Rectal bleeding and weight loss	159/239 (66.5)	1.92	[1.23, 3.00]	.0039
Diarrhea and weight loss	122/230 (53.0)	0.50	[0.33, 0.76]	.0011
Rectal bleeding, diarrhea and weight loss	256/356 (71.9)	4.37	[2.66, 7.19]	<.0001

Note. OR = odds ratio; CI = confidence interval. The analysis is based on 1,643 records (videos where the experts considered that referral to a general practitioner were not necessary are excluded from this analysis). Odds ratios greater than 1 indicated a higher likelihood of seeking advice from a general practitioner.

Discussion and Conclusion

Discussion

In the current study, the HBM score (a score based on five domains of the HBM, i.e., severity of the symptoms, susceptibility to major bowel pathology, benefits of and barriers to seeking medical advice, and the participants’ confidence in their GP to help with their symptoms)—was strongly associated with the odds of the participants recommending help seeking. In general, a high overall HBM score was associated with higher odds to seek advice thus highlighting the role of health beliefs in regard to seeking advice for lower bowel symptoms. These findings were consistent with previous studies that used the HBM as a theoretical framework (Jacobs, 2002; Ozsoy et al., 2007; Ueland, Hornung, & Greenwald, 2006). The role of health belief factors such as the severity of symptoms and confidence in the recommended health action in response to the symptoms have also been recently supported by the findings of Esteva et al. (2013).

The odds of vignette referral were also strongly associated with participants’ own pattern of GP visits and were much lower for participants who visited their GPs less frequently. Finding time to visit a GP and the monetary costs associated with consultation were other barriers to refer the vignettes for seeking help. Furthermore, the odds of referral of vignettes to GP were lower for participants who believed that visiting a GP was time consuming or expensive. The participants with lower household incomes and those with lower educational qualifications (educated up to or less than high school vs. tertiary educated) had lower odds of seeking medical advice. These variables have been independently shown to play a favorable role in men’s help-seeking behavior in previous studies, which suggests that health care cost could be an important factor in determining the use of health care services in patients requiring medical attention (McCaffery, Wardle, & Waller, 2003).

Among other demographic factors, the participants who resided in regional or remote areas were less likely to refer compared with those in metro areas. Factors such as availability and access to health services and the service cost are important in determining health care utilization (Corboy, McDonald, & McLaren, 2011). These factors are relevant particularly for Australian men living in regional and remote areas characterized by limited access to health services due to geographic isolation, poor transport links, and shortage of health care providers (Australian Institute of Health and Welfare, 2005). Previous studies report that increased awareness of symptoms and the risk of disease through mass media or through lay support networks is known to trigger help seeking, whereas lack of awareness of symptoms or of treatment act as barriers to help seeking (Smith, Pope, & Botha, 2005).

In this study, a disagreement was observed between the participants and the expert panel (comprising five GPs) with respect to referring the vignettes to seek help. Nearly 50% of the participants did not agree with the expert panel with respect to seeking medical advice for rectal bleeding lasting for 3 weeks. This finding is consistent with previous studies reporting low rate of help seeking for rectal bleeding (Chaplin et al., 2000; Cockburn et al., 2003; Courtney et al., 2012a; Esteva et al., 2013; Kocher & Saunders, 1999). Nearly one third of the participants disagreed with the expert panel on seeking medical advice for symptoms of unintentional weight loss or change in bowel habit occurring concurrently with rectal bleeding and lasting for 3 weeks. This may be a cause for concern as persistent rectal bleeding (particularly in the absence of peri-anal symptoms) may be indicative of significant bowel pathology and warrants medical attention (Ellis & Thompson, 2005). Nearly a quarter of the participants disagreed with the expert panel for vignettes comprising all the three symptoms—rectal bleeding, diarrhea, and weight loss—lasting for 3 weeks. The age of the men in the vignettes had no impact on the level of agreement as the respondents had similar advice for both younger and older men. This is again a cause for concern as these symptoms are known to increase the risk of bowel cancer particularly in the elderly (Adelstein et al., 2011; Bekkink et al., 2009; Bhargava, Aldameh, Stewart, & Hill, 2007). A number of participants sought health advice for symptoms and symptom combinations presented in the vignettes (rectal bleeding, diarrhea, and unintentional weight loss) from health professionals other than a GP such as pharmacists and nurse practitioners. This suggests that other health professionals especially community pharmacists could play a vital role in educating men about the risks of CRC based on their symptoms and subsequently advise them to seek medical advice from their GP (Sriram, McManus, Emmerton, Parsons, & Jiwa, 2014).

Strengths and Limitations

The use of video vignettes, instead of typical opinion surveys is a major strength of this study. The behaviors that have a social stigma attached to them or are personally threatening, may be either underreported or distorted in a socially desirable direction (Brener et al., 2003). The vignettes are thus, appropriate for studying the actual determinants of human judgments with more valid and reliable measures of the participants’ responses compared with typical opinion surveys (Wallander, 2009) as the questions may be personally less intimidating and therefore less susceptible to social desirability bias (Alexander & Becker, 1978; Finch, 1987). The use of video vignettes as a method of data collection is well established (Hawley et al., 2010). The various clinical symptoms and other variables, such as patients’ emotions, their body language and appearance cues, and social and psychological information, embedded in the vignettes can be presented more effectively in the video format than in the text vignettes, which may ignite the respondent’s own emotions, memories, attitudes, and biases that may influence their responses (Hawley et al., 2010).

Past studies (Chaplin et al., 2000; Cockburn et al., 2003; Crosland & Jones, 1995; Langenbach et al., 2003; Talley & Jones, 1998) have explored the help-seeking behavior in people who had experienced lower bowel symptoms, such as rectal bleeding and change in bowel habit, in the past. In one of the recent studies, Courtney et al. (2012b) highlighted the factors associated with early medical advice seeking behavior for potential symptoms of CRC. However, the time interval since exposure to the symptoms is known to affect the recall of exposures (Cockburn et al., 2003). The current study is designed to understand the help-seeking behavior of people in the general population regardless of symptoms, thereby minimizing the risk of recall bias and gaining better understanding of their help-seeking behavior.

As the sample was nonrandom, the findings may have limited generalizability. A nonrandom sample and the survey’s application in other settings will further validate the survey and will also increase confidence in the generalizability of the results.

Furthermore, the HSB questionnaire used in the survey did not include a subscale to measure “cues to action,” another domain of the HBM. The development and testing of the items for “cues to action” and additional items for each construct of the HBM scale may further increase the amount of common variance and further explain men’s help-seeking behavior as would a revision to the wording and scoring of the existing items. It may also help in future studies to include additional symptoms such as constipation or abdominal pain in the vignettes that were excluded in this study.

In this study, participants’ knowledge of the symptoms relating to CRC was not measured directly. Hence, it would be difficult to suggest if men’s underestimation of the symptoms severity that warranted medical attention was related to their suboptimal help-seeking behavior and if the programs leading to increase in men’s knowledge about the symptoms would lead to help seeking. It may be helpful in future research to study men’s knowledge about the potential causes for the symptoms presented in the vignettes and its correlation with their help-seeking behavior. It would also be interesting to conduct a similar survey among women in order to explore gender differences.

Another limitation is that the current study did not control for men’s recent receipt of a CRC investigation or diagnosis. Another limitation of the study is that only men who had access to Internet and a computer (desktop, laptop, or tablet) were able to answer the survey. It remains to be seen if the population of men with no Internet access would respond to the vignettes in a similar or different manner. Lastly, in regards to CRC risk stratification the authors in this study did not develop individual models for different age groups. Future studies may undertake models for individuals in lower risk (age groups ≤ 50 and > 50 years).

Conclusion

This study provides insight into the influence of the demographic and clinical characteristics and the personal health beliefs on the intention to seek timely advice from a health professional about lower bowel symptoms. The findings report that a considerable proportion of men would not advise other men even with significant lower bowel symptoms to seek help from GP. The HBM was an applicable model to study men’s help-seeking behavior and the HBM score, a measurement of men’s health beliefs regarding the symptoms, was directly related with the odds of seeking help.

Practical Implications

The study explores the factors affecting men’s decision to seek timely medical advice for symptoms that may mimic CRC orcolorectal pathology. The knowledge of these factors may be considered in developing interventions aimed at encouraging timely medical advice seeking for lower bowel symptoms and facilitating early diagnosis among men. It may be possible to encourage men to seek timely medical advice if public health messages on the importance of early presentation of persistent lower bowel symptoms are targeted at men with lower incomes or residing in regional or remote areas. As men younger than 50 years were less likely to seek health advice, such men should be targeted more specifically in public health campaigns.

Footnotes

Appendix A

Appendix B

Appendix C

Appendix D

Acknowledgements

We would like to thank Prof. Jeff Hughes, Curtin University for his suggestions in the design of the study and for proof reading the article.

Authors’ Note

This article was presented at the Cancer and Primary Care Research international Network Conference, 2014, Winnipeg, Canada and the abstract from the conference was published in the supplementary issue of the European Journal of Cancer Care. The funding body had no role in the data collection or report writing.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded through an internal grant from Curtin University.

References

Adelstein

B.-A.

Macaskill

Chan

Katelaris

Irwig

(2011). Most bowel cancer symptoms do not indicate colorectal cancer and polyps: A systematic review. BMC Gastroenterology, 11(1), 65.

Adelstein

B.-A.

Macaskill

Turner

R. M.

Irwig

(2012). Patients who take their symptoms less seriously are more likely to have colorectal cancer. BMC Gastroenterology, 12(1), 130.

Alexander

C. S.

Becker

H. J.

(1978). The use of vignettes in survey research. Public Opinion Quarterly, 42, 93-104.

Australian Institute of Health and Welfare. (2005, May 15). Rural, regional and remote health: Indicators of Health. Canberra: Author. Retrieved from http://www.aihw.gov.au/publication-detail/?id=6442467718

Barter

Renold

(2000). “I wanna tell you a story”: Exploring the application of vignettes in qualitative research with children and young people. International Journal of Social Research Methodology, 3, 307-323.

Basaranoglu

Celebi

Ataseven

Rahman

Deveci

S. E.

Acık

(2008). Prevalence and consultation behavior of self-reported rectal bleeding by face-to-face interview in an Asian community. Digestion, 77, 10-15.

Bekkink

M. O.

McCowan

Falk

Teljeur

Van de Laar

Fahey

(2009). Diagnostic accuracy systematic review of rectal bleeding in combination with other symptoms, signs and tests in relation to colorectal cancer. British Journal of Cancer, 102, 48-58.

Bhargava

Aldameh

Stewart

Hill

A. G.

(2007). Prioritisation of patients with rectal bleeding for urgent outpatient colonoscopy—A pilot study [100 Years Ago in the NZMJ]. New Zealand Medical Journal, 120(1258), 21-26.

Bowel Cancer Australia. (2013). Understanding bowel cancer: Don’t sit on your symptoms. Retrieved from http://www.bowelcanceraustralia.org/bca/index.php?option=com_content&view=article&id=206&Itemid=360

10.

Bray

Ren

J.-S.

Masuyer

Ferlay

(2013). Global estimates of cancer prevalence for 27 sites in the adult population in 2008. International Journal of Cancer, 132, 1133-1145. doi:10.1002/ijc.27711

11.

Brener

N. D.

Billy

J. O.

Grady

W. R.

(2003). Assessment of factors affecting the validity of self-reported health-risk behavior among adolescents: Evidence from the scientific literature. Journal of Adolescent Health, 33, 436-457.

12.

Chaplin

Curless

Thomson

Barton

(2000). Prevalence of lower gastrointestinal symptoms and associated consultation behaviour in a British elderly population determined by face-to-face interview. British Journal of General Practice, 50, 798-802.

13.

Cockburn

Paul

Tzelepis

McElduff

Byles

(2003). Delay in seeking advice for symptoms that potentially indicate bowel cancer. American Journal of Health Behavior, 27, 401-407.

14.

Corboy

McDonald

McLaren

(2011). Barriers to accessing psychosocial support services among men with cancer living in rural Australia: Perceptions of men and health professionals. International Journal of Men’s Health, 10, 163-183.

15.

Courtney

R. J.

Paul

C. L

Sanson-Fisher

R. W.

Macrae

Attia

McEvoy

(2012a). Current state of medical-advice-seeking behaviour for symptoms of colorectal cancer: Determinants of failure and delay in medical consultation. Colorectal Disease, 14, e222-e229.

16.

Courtney

R. J.

Paul

C. L.

Sanson-Fisher

R. W.

Macrae

Attia

McEvoy

(2012b). Factors associated with consultation behaviour for primary symptoms potentially indicating colorectal cancer: A cross-sectional study on response to symptoms. BMC Gastroenterology, 12(1), 100.

17.

Crosland

Jones

(1995). Rectal bleeding: Prevalence and consultation behaviour. BMJ, 311, 486-488.

18.

Department of Health and Ageing. (2013). Quarterly Medicare statistics (March quarter). Retrieved from http://health.gov.au/internet/main/publishing.nsf/Content/1A9DB6D72BD5879ACA257BF0001AFE28/$File/MBSStatistics20143SepQtr20141023.pdf

19.

Ellis

B. G.

Thompson

M. R.

(2005). Factors identifying higher risk rectal bleeding in general practice. British Journal of General Practice, 55, 949-955.

20.

Esteva

Leiva

Ramos

Pita-Fernández

González-Luján

Casamitjana

. . . Magallón

(2013). Factors related with symptom duration until diagnosis and treatment of symptomatic colorectal cancer. BMC Cancer, 13(1), 87.

21.

Ferlay

Soerjomataram

Ervik

Dikshit

Eser

Mathers

. . . Bray

(2013). GLOBOCAN 2012 v1. 0: Cancer incidence and mortality worldwide (IARC CancerBase No. 11). Lyon, France: International Agency for Research on Cancer.

22.

Finch

(1987). The vignette technique in survey research. Sociology, 21, 105-114.

23.

Green

S. B.

(1991). How many subjects does it take to do a regression analysis. Multivariate Behavioral Research, 26, 499-510.

24.

Hamilton

Sharp

(2004). Diagnosis of colorectal cancer in primary care: The evidence base for guidelines. Family Practice, 21, 99-106.

25.

Hashim

S. M.

Fah

T. S.

Omar

Radzniwan

Rashid

Shah

S. A.

Sagap

(2011). Knowledge of colorectal cancer among patients presenting with rectal bleeding and its association with delay in seeking medical advice. Asian Pacific Journal of Cancer Prevention, 12, 2007-2011.

26.

Hashim

S. M.

Omar

Fah

T. S.

Rashid

R. A.

Daud

T. I.

Shah

S. A.

Sagap

(2010). Factors influencing late consultation among patients with rectal bleeding in University Kebangsaan Malaysia Medical Centre. Asian Pacific Journal of Cancer Prevention, 11, 1335-1339.

27.

Hawley

Adams

Galbraith

(2010, January). Video vignette study of attitudes and behaviour of health professionals in consultations (Road Safety Research Report No. 91). London, England: Department of Transport.

28.

Hughes

Huby

(2002). The application of vignettes in social and nursing research. Journal of Advanced Nursing, 37, 382-386.

29.

Jacobs

L. A.

(2002). Health beliefs of first-degree relatives of individuals with colorectal cancer and participation in health maintenance visits: A population-based survey. Cancer Nursing, 25, 251-265.

30.

Kinicki

A. J.

Hom

P. W.

Trost

M. R.

Wade

K. J.

(1995). Effects of category prototypes on performance-rating accuracy. Journal of Applied Psychology, 80, 354-370.

31.

Kirby

(2014). Health and science suffer major cuts in Australia’s budget. Lancet, 383, 1874-1876.

32.

Kocher

Saunders

(1999). Complacency or ignorance about rectal bleeding? Colorectal Disease, 1, 332-333.

33.

Korsgaard

Pedersen

Sørensen

H. T.

Laurberg

(2006). Reported symptoms, diagnostic delay and stage of colorectal cancer: A population-based study in Denmark. Colorectal Disease, 8, 688-695.

34.

Langenbach

M. R.

Johannes Schmidt

Neumann

Hubert Zirngibl

(2003). Delay in treatment of colorectal cancer: Multifactorial problem. World Journal of Surgery, 27, 304-308.

35.

Lindholm

Brevinge

Haglind

(2008). Survival benefit in a randomized clinical trial of faecal occult blood screening for colorectal cancer. British Journal of Surgery, 95, 1029-1036.

36.

Lynch

B. M.

Youlden

Fritschi

Newman

Pakenham

K. I.

Leggett

. . . Aitken

J. F.

(2008). Self-reported information on the diagnosis of colorectal cancer was reliable but not necessarily valid. Journal of Clinical Epidemiology, 61, 498-504.

37.

Macleod

Mitchell

Burgess

Macdonald

Ramirez

(2009). Risk factors for delayed presentation and referral of symptomatic cancer: Evidence for common cancers. British Journal of Cancer, 101, S92-S101.

38.

Macrae

F. A.

Hill

D. J.

St John

D. J. B.

Ambikapathy

Garner

J. F.

(1984). Predicting colon cancer screening behavior from health beliefs. Preventive Medicine, 13, 115-126.

39.

McArdle

Hole

(2004). Emergency presentation of colorectal cancer is associated with poor 5-year survival. British Journal of Surgery, 91, 605-609.

40.

McArdle

C. S.

McMillan

D. C.

Hole

D. J.

(2003). Male gender adversely affects survival following surgery for colorectal cancer. British Journal of Surgery, 90, 711-715.

41.

McCaffery

Wardle

Waller

(2003). Knowledge, attitudes, and behavioral intentions in relation to the early detection of colorectal cancer in the United Kingdom. Preventive Medicine, 36, 525-535.

42.

McCashland

Brand

Lyden

de Garmo

(2001). Gender differences in colorectal polyps and tumors. American Journal of Gastroenterology, 96, 882-886. doi:http://dx.doi.org/10.1016/S0002-9270(00)02431-X

43.

McMinn

Steel

Bowman

(2011). Investigation and management of unintentional weight loss in older adults. BMJ, 342, d1732.

44.

National Institute of Health and Care Excellence. (2005). Referral guidelines for suspected cancer (Clinical Guideline CG27). London, England: Author. Retrieved from http://www.nice.org.uk/nicemedia/pdf/cg027niceguideline.pdf

45.

Oberoi

D. V.

Jiwa

McManus

Hodder

(2014). Colorectal cancer: Applying a gender lens. Quality in Primary Care, 22, 71-79.

46.

Oberoi

D. V.

Jiwa

McManus

Hodder

(2015a). Barriers to help seeking in men diagnosed with benign colorectal diseases. American Journal of Health Behavior, 39, 22-33. doi:10.5993/AJHB.39.1.3

47.

Oberoi

D. V.

Jiwa

McManus

Hodder

(2015b). Men’s help seeking behaviour with regards to lower bowel symptoms. American Journal of Health Behavior, 39, 212-221. doi:10.5993/AJHB.39.2.8

48.

Ozsoy

S. A.

Ardahan

Özmen

(2007). Reliability and validity of the colorectal cancer screening belief scale in Turkey. Cancer Nursing, 30, 139-145.

49.

Pedersen

A. F.

Hansen

R. P.

Vedsted

(2013). Patient delay in colorectal cancer patients: Associations with rectal bleeding and thoughts about cancer. PloS One, 8(7), e69700.

50.

Philipp

Wood

Heaton

K. W.

Hughes

A. O.

(1993). Perceptions and reactions of the public to diarrhoea. Journal of the Royal Society for the Promotion of Health, 113, 128-131.

51.

Pinto

S. L.

Lively

B. T.

Siganga

Holiday-Goodman

Kamm

(2006). Using the health belief model to test factors affecting patient retention in diabetes-related pharmaceutical care services. Research in Social & Administrative Pharmacy, 2, 38-58.

52.

Ramanathan

S. A.

Baratiny

Stocks

N. P.

Searles

A. M.

Redford

R. J.

(2011). General practitioner referral patterns for women with gynaecological symptoms: A randomised incomplete block study design. Medical Journal of Australia, 195, 602-606.

53.

Rawl

Champion

Menon

Loehrer

P. J.

Sr. Vance

G. H.

Skinner

C. S.

(2001). Validation of scales to measure benefits of and barriers to colorectal cancer screening. Journal of Psychosocial Oncology, 19(3-4), 47-63.

54.

Rosenstock

I. M.

Strecher

V. J.

Becker

M. H.

(1994). The health belief model and HIV risk behavior change. In DiClemento

R. J.

Petersen

J. L.

(Eds.), Preventing AIDS: Theories and methods of behavioral interventions (pp. 5-24). New York, NY: Springer.

55.

Shaw

Brittain

Tansey

Williams

(2008). How people decide to seek health care: A qualitative study. International Journal of Nursing Studies, 45, 1516-1524.

56.

Smith

Ballal

Hodder

Soin

Selvachandran

S. N.

Cade

(2006). Symptomatic presentation of early colorectal cancer. Annals of the Royal College of Surgeons of England, 88, 185-190. doi:10.1308/003588406X94904

57.

Smith

L. K.

Pope

Botha

J. L.

(2005). Patients’ help-seeking experiences and delay in cancer presentation: A qualitative synthesis. Lancet, 366, 825-831.

58.

Sriram

McManus

Emmerton

Parsons

Jiwa

(2014). Development and validation of a clinical decision-making aid for screening bowel symptoms in community pharmacies. Journal of Evaluation in Clinical Practice, 20, 260-266. doi:10.1111/jep.12120

59.

Stapley

Peters

T. J.

Sharp

Hamilton

(2006). The mortality of colorectal cancer in relation to the initial symptom at presentation to primary care and to the duration of symptoms: A cohort study using medical records. British Journal of Cancer, 95, 1321-1325.

60.

Strecher

V. J.

Rosenstock

I. M.

(1997). The health belief model. In Baum

Newman

Weinman

West

McManus

(Eds.), Cambridge handbook of psychology, health and medicine (pp. 113-117). Cambridge, England: Cambridge University Press.

61.

Talley

N. J.

Jones

(1998). Self-reported rectal bleeding in a United States community: Prevalence, risk factors, and health care seeking. American Journal of Gastroenterology, 93, 2179-2183.

62.

Thompson

Pond

Ellis

Beach

Thompson

(2000). Rectal bleeding in general and hospital practice; “the tip of the iceberg”. Colorectal Disease, 2, 288-293.

63.

Ueland

A. S.

Hornung

P. A.

Greenwald

(2006). Colorectal cancer prevention and screening: A health belief model-based research study to increase disease awareness. Gastroenterology Nursing, 29, 357-363.

64.

Wallander

(2009). 25 Years of factorial surveys in sociology: A review. Social Science Research, 38, 505-520.

65.

Wichmann

M. W.

Müller

Hornung

H. M.

Lau-Werner

Schildberg

F. W.

(2001). Gender differences in long-term survival of patients with colorectal cancer. British Journal of Surgery, 88, 1092-1098.

66.

Chen

. . . Li

(2012). Association of estrogen receptor beta variants and serum levels of estradiol with risk of colorectal cancer: A case control study. BMC Cancer, 12, 276.