Factors Associated With Complementary and Alternative Medicine Use in Irritable Bowel Syndrome: A Literature Review

Aim: Irritable Bowel Syndrome (IBS) is a chronic functional bowel condition, which has substantial impact on quality of life and use of healthcare services. Patients often report using complementary and alternative medicine (CAM) for symptom management despite limited evidence to support its use. Psychological factors have been shown to be important in both influencing CAM use and as avenues of intervention to assist in managing IBS symptoms. Therefore, this review assessed prevalence of and psychological factors associated with CAM use by people with IBS. Method: Five electronic databases (including AMED, EMBASE and PsychINFO) were searched for studies that examined both the extent of and the reasons for CAM use. Five studies met the inclusion criteria. Results: Prevalence of CAM use ranged from 9% to 38%. CAM use was associated with psychosocial factors, including concerns about conventional medical care (i.e., the perceived harmful effects of medication, perception that conventional medicine had failed, and lack of satisfaction with conventional care) and anxiety. Conclusion: These findings identify psychological factors associated with CAM use which could be targeted through psychologically oriented management strategies for those affected with IBS.


Introduction
. Review search terms and strategy.

Selection Process and Data Analysis
One researcher (LU) conducted initial searches and selection of abstracts. Duplicates were removed from the search and all abstracts subsequently read. Where it was unclear if a study fitted the review criteria, the full text was obtained. There were two main inclusion criteria. Firstly, studies had to include measurement regarding the extent of CAM use in IBS. Secondly, included studies needed to have examined psychological factors that influence CAM use in those affected by IBS (such as beliefs about treatment for IBS). No restrictions were placed on the type of analysis or design studies used and only studies published in English were included. Studies that focused exclusively on conventional care seeking or organic bowel conditions were excluded from the review. Reference lists of obtained articles were also checked for relevant studies. Two researchers (KM & PF), then cross-validated the final selection of studies from the 13 full text articles down to the final five that were included. Agreement to include the studies amongst the three researchers was unanimous. The process of identification of studies is presented in Figure 2.
Included studies focused on participants with a functional bowel disorder (FBD, referring to IBS) or made sole reference to IBS. Recruitment and data collection varied from postal questionnaires (Smart et al., 1986), to recruitment from a general practitioner clinic (Donker et al., 1999) and outpatient clinic (Smart et al., 1986;Verhoef et al., 1990). Two studies used data collected from previous work by the respective authors, for example from a previous healthcare survey for those with FBD (van Tilburg et al., 2008), and previous population surveys (Koloski et al., 2003). The studies reported female predominance in samples, ranging from 60 to 75%.

Study Methodologies
All studies examined group differences. Two studies (van Tilburg et al., 2008;Verhoef et al., 1990) used one group of IBS/FBD outpatients and analysed participants in terms of those who had used or not used CAM. Smart et al. (1986) compared 96 IBS patients to 143 patients with other unspecified organic upper GI disorders and 222 Crohn's disease patients. Donker et al. (1999) and Koloski et al. (2003) compared an IBS group to healthy controls in addition to healthcare (including alternative healthcare) consulters and non-consulters.    Smart et al. (1986) assessed the frequency of CAM use in patients with a diagnosis of IBS according to the Manning criteria (Manning et al., 1978) and for whom a clinical examination revealed no bowel abnormalities.
Patients with organic GI disorders were recruited from the same outpatient clinic as those with IBS and Crohn's patients were contacted via post. All participants completed a questionnaire on alternative medicine consultations. Verhoef et al. (1990) however, examined patients who sought alternative treatment for the problem which had required a consultation with a GI specialist in the past two years. Differences in demographic profile and health status between CAM users and non-users were compared. Of the 395 GI patients recruited, 55 were classified as having a functional GI disorder with diagnosis made by four GI specialists. This study excluded patients who used CAM for health problems other than their diagnosed GI disorders. Participants completed a three item index based on scepticism towards conventional medicine and were asked about alternative medicine use during the previous two years. Donker et al. (1999) focused on the health status of 53 patients with IBS recruited from general practices participating in the Dutch National Survey of Morbidity and Intervention in General Practice and compared their use of healthcare services including CAM to a population sample of 10787. Participants were asked about healthcare use, health problems in the two weeks prior to being questioned and completed the General Health Questionnaire (GHQ) a screening tool for psychiatric illness (Goldberg, 1972) and the biographic problem list (BIOPRO) which measured social problems (Hosman, 1983). Health related behaviour (e.g., smoking, exercise) and the amount of healthcare sought was also measured which ranged from seeing a doctor (previous three months), a physical therapist (previous 12 months), a specialist (last two years) and an alternative therapist (previous five years).
Using semi-structured interviews and questionnaires, Koloski et al. (2003) considered usage of both conventional and alternative healthcare in 207 patients with functional GI diagnoses (IBS or functional dyspepsia). Participants were recruited from one of two previous surveys carried out by the authors and separated into consulters or nonconsulters for both conventional and alternative healthcare. Participants were asked about frequency, access and satisfaction with healthcare. The structured interview for bowel symptoms was administered to give a functional diagnosis based on the Rome I criteria (Drossman et al., 1994). This structured interview also accounted for aspects related to quality of life and extent of symptoms. In addition, participants were given the Composite International Diagnostic Interview (World Health Organization, 1997) designed to assess past and current psychological disturbance.
Using data from 1012 FBD patients recruited from an "outpatient" healthcare maintenance organisation in a previous study (Nyrop et al., 2007), van Tilburg et al. (2008) examined CAM use in IBS and FBD. Participants were assessed for symptom severity at a "baseline" visit to the clinic with the Irritable Bowel Syndrome Severity Scale (IBS-SS) (Francis, Morris, & Whorwell, 1997), quality of life using the Irritable Bowel Syndrome Quality of Life scale, IBS-QOL (Patrick et al., 1998), psychological distress using the Brief Symptom inventory (BSI) (Derogatis, 1993), perceived treatment effectiveness, and CAM use.

Extent of CAM Use
The reviewed studies indicated CAM use at between 9 and 38.4%. Smart et al. (1986) found significantly more of those with IBS had visited an alternative practitioner compared to Crohn's and organic GI patients. Current alternative medicine use was significantly greater in the IBS group and herbal treatments and homeopathy were used most frequently (Smart et al., 1986). Verhoef et al. (1990) reported that 50% of CAM-users had FBD (compared to 13% of non-CAM users) and 9% used CAM for the condition they presented to a gastroenterologist. Chiropractors (for conditions other than GI), herbalists, naturopaths and reflexologists were that most frequently CAM practitioners and 46% of participants had visited more than one type of CAM practitioner.  Donker et al. (1999) found the IBS patient group had paid significantly more visits to an alternative practitioner than the population group (32% compared with 15%). Koloski et al. (2003) revealed that 86.5% of the functional GI group had sought conventional healthcare at some point, and a reported 20.8% of participants had sought alternative healthcare, with only 9% using any CAM in the previous 12 months. The most frequently accessed treatment was naturopathy. Van Tilburg et al. (2008) found 35% of those with FBD and 38.4% of those with IBS had used CAM with ginger, massage therapy and yoga being the most frequently used treatments.

Reasons for CAM Use
Demographics and Functional Diagnosis - Koloski et al. (2003) found 88.4% of the 20.8% of CAM users were female. However, 79.8% of participants (64% who were female) did not use alternative healthcare. Neither Donker et al. (1999), Smart et al. (1986) or Verhoef et al. (1990) specifically examined the role of gender although Donker et al. (1999) reported that a majority of IBS patients were female. Van Tilburg et al. (2008) found being female and higher educational attainment predicted CAM use. Verhoef et al. (1990) reported that a functional diagnosis was an independent predictor of CAM use compared to those with organic GI disorders. Similarly, Smart et al. (1986) reported more patients with IBS than Crohn's used CAM. Donker et al. (1999) found significantly more IBS outpatients used CAM than the population group and those with IBS reported significantly more intense symptoms (e.g., abdominal pain and "secondary" symptoms including tiredness, backache and headaches). Koloski et al. (2003) found that physical symptoms of IBS significantly predicted conventional care seeking rather than CAM use. Van Tilburg et al. (2008) found stronger perceptions of symptom severity were associated with CAM use but not when controlling for demographics and other variables (e.g. the IBS-QOL and IBS-SS) in a logistic regression model. Smart et al. (1986) found that the IBS group were more likely to report using alternative treatments if they perceived conventional treatment had failed. Verhoef et al. (1990) observed that 54% of CAM-users with GI disorders (including IBS) were satisfied with conventional treatment compared with 85% of non-CAM participants and that GI patients who used CAM were significantly more sceptical (49%) of conventional medicine than those not using CAM (13%). Verhoef et al. (1990) also revealed associations between a functional diagnosis and scepticism towards conventional medicine, and that these variables both (independently) significantly predicted the use of CAM. In relation to communication between conventional practitioner and patient, CAM-users were less satisfied with responses from conventional practitioners than non-CAM users (77% vs. 91%). Koloski et al. (2003) found dissatisfaction did not significantly influence CAM use, although there was some difference between CAM-users and non-CAM users. Van Tilburg et al. (2008) found no association between CAM use and satisfaction with physician care during their primary visit and that CAM-users did not rate their conventional prescription medication as being less effective than non-CAM users.

Beliefs About CAM Treatments and Therapies -One study reported expectations of CAM efficacy as a rationale
for IBS patients to use CAM. Koloski et al. (2003) found a desire to treat the GI problem with a more natural approach, the potential for alternative treatments to work and personal recommendation were all factors (albeit not significant) that appear to influence CAM use. Donker et al. (1999) reported that 92% of CAM-users felt CAM had helped. Verhoef et al. (1990)  GHQ than the population group. Differences were observed in the two groups' BIOPRO scores where those with IBS reported greater concerns about the future, lower self-confidence, fewer social interactions, and relationship difficulties. The IBS group reported more occupational absence in the two months prior to the study. Additionally, having IBS resulted in significantly more visits to the family GP, a physical therapist and a GI specialist. The significant differences observed in healthcare seeking (including CAM use) between the two groups may be influenced by such psychosocial factors (Donker et al., 1999). Koloski et al. (2003) observed differences, albeit not significant, between both sets of healthcare consulters (CAM and conventional treatment) and non-consulters in psychological disturbance and perception of symptoms. Van Tilburg et al. (2008) found CAM-users reported significantly poorer IBS-QOL scores compared to those not using CAM. CAM-users also reported higher scores on the somatisation, anxiety and depression subscales of the BSI. Logistic regression analysis revealed that anxiety was the sole significant psychosocial predictor of CAM use.

Discussion
Prevalence of CAM was lower than CAM use reported by Kong et al. (2005) and Langmead et al. (2002). However, this might be explained by reviewed studies focusing on consultation with a CAM practitioner as opposed to selfdirected treatments (Kong et al., 2005). The reviewed studies included patient group and healthy control population comparisons, in addition to comparisons of those with IBS who used CAM and those not using CAM. Findings indicate that those with IBS who use CAM may report more severe symptoms and have concerns about conventional medicine including unhappiness and dissatisfaction with conventional care.
Female predominance was evident, which is concurrent with what is known about IBS (e.g. Andrews et al., 2005), however, in terms of CAM, there was limited agreement with findings from non-illness specific investigation into CAM use. CAM use in general has been reported to have female predominance (e.g. Astin, 1998), although only one reviewed study (van Tilburg et al., 2008) found that being female was an independent predictor of CAM use in those with a functional GI diagnosis. It should be noted however that the reviewed studies had a predominantly female representation which is consistent with previous findings that a greater proportion of females than males seek healthcare for IBS/functional GI symptoms (Andrews et al., 2005;Hungin et al., 2005). It has yet to be ascertained whether this disparity between males and females is due to biological or environmental distinctions related to gender or might be explained by differences in healthcare seeking (Corazziari, 2004).
Although there has been limited examination of many psychological aspects in general populations of CAM-users, points of contrast and similarity exist with what is currently known about possible psychological influences on CAM use. Illness related perceptions, for example, were not found to be strong influences on CAM use. This is in contrast to findings from Bishop et al. (2006) who found stronger perceived consequences of illness (from a general population) predicted use of CAM. This finding may be more indicative of a lack of measurement of such constructs. With illness perceptions being a key component for intervention in IBS (e.g. Oerlemans et al., 2011) this represents one important aspect that could be addressed by future research. CAM-users however, reported poorer quality of life and more severe symptoms, which concurs to an extent with Astin (1998) who established CAM-users from a general population reported poorer health than those not using CAM. There are further similarities in CAM-users affected by IBD who report amplified symptom perception and poorer quality of life (Hilsden et al., 1998;Langmead et al., 2002;Scott et al., 2003). One study (Koloski et al., 2003) found increased perception of IBS symptoms predicted conventional care seeking suggesting it worthwhile to investigate this further in relation to use of CAM. Treatment related beliefs however, appeared more influential on CAM use than illness perceptions. For example, findings showed a negative perception of conventional medical care influenced CAM use as did a desire to treat GI symptoms with a more natural approach (Bishop et al., 2006;Vincent & Furnham, 1996). A degree of dissatisfaction with conventional medicine also appears evident, as found by Scott et al. (2003) in IBD patients. However, future research should clarify whether this is due to issues with conventional treatment itself and the healthcare consultation, or both factors.
A major strength of this review is that it is the first to synthesise both prevalence estimates and evidence of why people affected with IBS use CAM. Several psychological factors that have a role in influencing CAM use have been highlighted. These findings may be beneficial in informing areas of potential intervention in conventional medical consultations. The review further highlights the paucity of research in this important area, thus highlighting the need for additional research that would aid understanding of influences on CAM use. This review has identified specific areas of investigation that could be addressed by future research. This may include addressing specific psychological components of illness and treatment beliefs and assessing potential influence of these factors on CAM use in those affected by IBS. Furthermore, exploring aspects of the health practitioner-client relationship to determine where dissatisfaction may arise could have potential benefits in targeting and implementing future improvements in healthcare.
Conversely this review has a number of limitations that need to be considered. Due to differences in sample size and measures used in the selected studies, meta-analysis was not deemed appropriate due to variation in both study design and reporting of findings. However, across the five reviewed studies, there were common themes why those with IBS turn to CAM although the methodological variation in the studies makes generalisation of findings problematic and potentially limited. Group comparisons in each study differed notably and it might be suggested that a protocol of studying CAM-users compared with those not using CAM may be advantageous in terms of explaining psychological influences on CAM use more precisely. There may also be benefit in examining factors that pull people to different forms of CAM (Bishop et al., 2006) as findings suggested that CAM is, at least partially, sometimes viewed as a single entity in terms of treatment. The reviewed studies were also conducted in different countries where differences in healthcare service provision may exist in addition to cultural differences.
Furthermore, there was variation in participant numbers in each of the studies. Two studies for example (Donker et al., 1999;Verhoef et al., 1990) had relatively small numbers of participants with IBS/FBD. There is some ambiguity concerning the scope of functional GI diagnosis (Verhoef et al., 1990) and if all participants had IBS or different functional diagnoses (such as functional dyspepsia) (Koloski et al., 2003).
It was notable that none of the reviewed studies considered emotional response to illness, something that is often a concern in those with IBS (e.g. Tanaka et al., 2011). CAM use was however predicted by higher reported anxiety in one study (van Tilburg et al., 2008) but it is unclear if anxiety was present pre or post-illness. The psychosocial factors (e.g. BIOPRO responses) reported by Donker et al. (1999) in the IBS group warrant further investigation as this particular study focused on a small group of those affected by IBS and a considerably larger population group. Further longitudinal investigation may reveal the extent to which these factors influence CAM use and if differences exist on such constructs between CAM-users and those not using CAM. Additionally, there were notable differences in the measurement of symptom experience. Smart et al. (1986) did not consider symptom duration important in predicting CAM use in IBS as more patients with IBS were currently using CAM than the Crohn's group (who have similar symptoms). Two studies (Donker et al., 1999;van Tilburg et al., 2008) considered the severity of GI symptoms and one considered ratings of quality of life (van Tilburg et al., 2008). Verhoef et al. (1990)  however, did not consider participants' reported symptoms. Further investigation into perceived severity of IBS symptoms is therefore warranted to determine if CAM-users report more severe IBS symptoms.
Furthermore in relation to concerns about conventional medical treatment, "failure" of (Smart et al., 1986) and "dissatisfaction" with conventional treatment (Koloski et al., 2003) may reflect sub-dimensions of the same construct.
Both refer to treatment, the consultation or both. In the studies reviewed, measurement of these factors varied considerably. Koloski et al. (2003) conducted a healthcare seeking interview while Smart et al. (1986) asked specifically about failure of conventional treatment. Verhoef et al. (1990) assessed scepticism towards conventional medicine. Van Tilburg et al. (2008) considered first healthcare visits and perceived effectiveness of prescription medication. Moreover, four studies (Donker et al., 1999;Koloski et al., 2003;Smart et al., 1986;Verhoef et al., 1990) focused solely on CAM consultations thus neglecting "off the shelf" products from the analysis. Clarification of the nature of dissatisfaction and a more expansive inclusion of CAM, which would include self-purchased treatments and consideration of different forms of CAM, are additional factors that could be addressed in future research.

Conclusion and Implications
CAM's use has been shown to be associated with psychological factors which could be targeted through psychologically based management strategies for those affected with IBS. Such interventions may be beneficial in addressing negative symptom or treatment perceptions and emotional distress that may accompany IBS symptoms (e.g. van Dulmen et al., 1996) in addition to focusing on providing information about IBS (e.g. Jarrett et al., 2009).
It is possible that CAM-users may initially benefit more from such intervention as evidence suggests those with IBS using CAM report an amplified or more intense illness experience than those not using CAM. This may extend to CAM-users reporting poorer quality of life despite using CAM although future longitudinal studies are required to support this. The array of psychological factors identified by this review also suggests that the application of a theoretical framework to future research may aid understanding and inform translation to practical interventions.
One such model, the common-sense model of illness representation (Leventhal, Brissette, & Leventhal, 2003) has incorporated both illness perceptions and treatment beliefs (e.g. Bishop et al., 2006). This model has had success in both exploration and translation of findings into practical change in illness perceptions resulting in benefits to health (McAndrew et al., 2008).