Background
Use of complementary and alternative medicine (CAM) in the UK has increased over the last decade (Giveon et al., Reference Giveon, Liberman, Klang and Kahan2003) and, although mainly accessed privately by patients, 10% of consultations are accessed through NHS primary care (Ong et al., Reference Ong, Bodeker and Burford2005). Forty per cent of patients would use CAM if recommended by a medical practitioner; therefore, doctors’ views about CAM are important to patients (Lewis et al., Reference Lewis, Paterson, Beckerman and Sandilands2001; Lewith et al., Reference Lewith, Hyland and Gray2001). However, General Practitioners (GPs) describe their knowledge of CAM as poor or very poor (Wharton and Lewith, Reference Wharton and Lewith1986) and only a quarter report feeling confident talking about CAM with patients (Perry and Dowrick, Reference Perry and Dowrick2000). Many doctors report that their ability to refer patients to CAM is diminished by this lack of information and understanding (Thomas et al., 1995; White et al., Reference White, Mitchell and Ernst1996; Lewith et al., Reference Lewith, Hyland and Gray2001; Schmidt et al., Reference Schmidt and Barton2002).
Concerns about CAM that are commonly aired by doctors include patients stopping their conventional treatment, treating CAM as a substitute for conventional prescribed medicines, delayed or missed diagnosis, adverse effects of CAM, unqualified or overcharging practitioners, and lack of research evidence (Astin et al., Reference Astin, Marie, Pelletier, Hansen and Haskell1998; Zollman and Vickers, Reference Zollman and Vickers1999; Cohen et al., Reference Cohen, Penman, Pirotta and da Costa2005). Despite this, up to 40% of doctors view some CAM as having a sound theoretical basis (Anderson and Anderson, Reference Anderson and Anderson1987), and 70% believe in its safety (Schmidt et al., Reference Schmidt and Barton2002). Indeed, there is some evidence that GPs endorse CAM despite not necessarily understanding or believing in it (Perry and Dowrick, Reference Perry and Dowrick2000; Cohen et al., Reference Cohen, Penman, Pirotta and da Costa2005). Estimates of GP referral rates vary depending upon the time period assessed, the therapy and the timing of the studies (Table 1), but studies show that between 59% and 72% of GPs have referred to CAM in the preceding year (Wharton and Lewith, Reference Wharton and Lewith1986; Anderson and Anderson, Reference Anderson and Anderson1987).
Studies quantifying GP referrals to CAM.
CAM- complementary medicine
CP- Complementary medical practitioner
GP- General practitioner
No qualitative studies have specifically assessed the reason or motivation for these referrals. Limited and non-systematic data from open sections in a small number of surveys have suggested some reasons, for example patient request or a lack of response to conventional treatment (Borkan et al., Reference Borkan, Neher, Anson and Smoker1994; Boucher and Lenz, Reference Boucher and Lenz1998; Kaczorowski et al., Reference Kaczorowski, Patterson, Arthur, Smith and Mills2002; van Haselen et al., Reference van Haselen, Reiber, Nickel, Jakob and Fisher2004). However, these quantitative studies were focussed on the overall pattern of referral and not on understanding the processes that shape referral behaviour. We have addressed this research gap by exploring the key factors that shape GPs’ referring behaviour. This initial exploratory qualitative study formed part of a service evaluation within a UK Primary Care NHS contract and only relevant data from the study is presented.
Methods
Qualitative face-to-face semistructured interviews were used. An interview topic guide, based on the existing literature, was used to ensure the primary question was addressed systematically in each interview. Framework analysis was used to identify key themes (Ritchie and Spencer, Reference Ritchie and Spencer1994). In order to be eligible, UK GPs needed to have referred patients to the private CAM clinic within their NHS contract during the previous two years, so their experiences were ‘current’. There were no other criteria for inclusion. Invitation letters from the CAM clinic were sent to 33 GPs who had referred their NHS patients for CAM within the existing NHS contract during the previous two years. An interview was arranged with those GPs who expressed an interest and were suitable to take part.
The complementary practitioners (CPs) at the CAM clinic study site are all medically qualified. The NHS contract has been in existence for 12 years and covers referrals for irritable bowel syndrome, migraine, eczema, chronic fatigue syndrome, childhood behavioural disorders and non-specific allergy syndrome. Patients are treated for up to six sessions, which can be extended for a further six. Therapies offered are homeopathy, acupuncture and herbal, environmental and nutritional medicine. There is no specific upper limit for referral by a single GP, although a provisional limit of 500 referrals per year is set by the PCT.
Interviews took place in the GPs’ surgeries (EH). Informed written consent was obtained prior to the interview. Open-ended questions addressed GP perspectives about their referral practices and emergent issues relevant to the research question were added to the topic guide and explored at subsequent interviews. See Appendix 1 for the initial guide, which also shows the two additional questions included in the final guide. Interviews were audiotaped for transcription and the use of pseudonyms protected participants’ anonymity. This study was classed as a service evaluation by the relevant NHS authorities, so ethical approval was not required. The study met all of the standards that would be required by the ethics process, including independent peer review, data protection, informed written consent and researcher safety.
Data analysis
Analysis of the data used the framework approach, a structured methodology following five distinct but interconnecting stages (Ritchie and Spencer, Reference Ritchie and Spencer1994) and all stages were conducted by EH and SB. These steps include familiarization with the data, identification of a thematic framework, indexing of the interviews, charting and lastly mapping and interpretation. Overall, this permitted the analysts to identify key themes in a rigorous and systematic fashion, ensuring that the corpus of data was dealt with in its entirety. The initial framework was coded by EH, checked by SB and any disagreements resolved by consensus. Illustrative quotations are used in the findings to elucidate the key themes.
Results
Participants
Of the 33 GPs referring in the previous two years, 10 agreed to be interviewed; the remaining 23 declined to participate and no reason for refusal to participate was given. Interviews, lasting about 30 min, took place between October 2006 and January 2007. GP referral rates varied, with one GP referring three to four times a week, seven of the 10 GPs referring 3–4 times a year, and two once in the preceding two years. GP practices were mainly in small towns or villages in the UK but also in a deprived city area. The distance from the GP practices to the CAM practice varied from 22 to 73 miles. Eight interviewees were male and two were female.
The key factors that shape GPs’ referring behaviour
The key factors that shaped GPs’ referring behaviour are demonstrated in Figure 1. GPs decisions to make a referral to CAM were mediated by: (a) their personal and professional experiences of CAM, and the evidence and their knowledge of CAM; and (b) their understanding of their patients’ attitudes to CAM and their patients’ experience of conventional medicine. GPs would only initiate referral if they perceived matching between these two.
Thematic model of GP referrals. Note: the arrows do not imply causality
1. GPs experience, evidence and knowledge of CAM
GPs recognized that their experiences, evidence and knowledge of CAM were of central importance to the referral process (see Figure 2).
Model of GPs’ experience, evidence and knowledge of CAM. Note: the arrows do not imply causality
a. The experience and knowledge of the GP
GPs’ own experiences and/or knowledge of CAM, along with an awareness of the referral process, influenced their referral decisions, but a lack of knowledge did not preclude referral. The two GPs in the sample who practiced CAM reported that their experiences as CAM practitioners acted as a powerful catalyst for referral.
I’m a member of the Faculty of Homeopathy, … so I’m very aware of what conditions do well with homeopathy.
(Andrew: 9–12)
I think it is just that I am lucky, because I know about complementary medicine. I practise one (CAM) as well…So, I have read a lot about the others, so I know what they can do and they can’t., I know that complementary medicine might be a better option
(Andrew: 21–26)
Positive reports from patients, friends and colleagues shaped one GPs decision to visit a CAM practitioner. His positive personal experience provided a particularly strong motivation for then referring his own patients.
So I thought well golly that’s three of them, plus the school nurse’s husband is four, plus this other friend, five, six, refractory [disease], I must go and see him Um I came back from seeing him, and before I had tried the [medicine] myself I saw a patient here [with the complaint I had] I said to [patient] try a month of [medicine] and I bumped into him in the street and he came rushing up to me and he said I’m cured. And that was fifteen, eighteen months of [disease] and I did actually suggest to this other chap that he went and saw a CAM practitioner.
(Peter: 76–122)
Positive past patient referrals were reported as shaping not just the decision to refer but the actual timing of the referral.
I have got so many cases improved by complementary medicine, so I probably refer people earlier than other GPs.
(Adrian: 54–5)
In contrast to ‘early’ referral, those who considered themselves to be less knowledgeable about CAM tended only to refer once conventional routes had been exhausted. They viewed CAM as a last resort.
Help, what should I do? Tried everything I can think of! Maybe we have tried one or more conventional consultants, and they haven’t come up with anything… useful.
(Mary: 23–5)
Interviewees who reported the greatest epistemic uncertainty also tended to refer less than their more knowledgeable counterparts.
I wouldn’t say I am particularly instructed in what conditions are suitable more than others for a complementary approach.
(Susan: 37–40)
It was clear that the awareness and understanding of evidence for CAM therapies varied across the group and this shaped their referring behaviour in a number of ways. Some GPs indicated more information would help them improve their referrals.
I don’t think there is enough possibly circulated about what types of treatment can help…
(Susan: 104–6)
GPs with limited knowledge of CAM contrasted this with a relatively more enlightened and knowledgeable patient population.
I think people do look up things on websites; a lot of people are very well informed these days and sort of researched about themselves, um I am not that knowledgeable myself in complementary therapies.
(Susan: 34–40)
When GPs reported a lack of awareness of the scientific evidence base for CAM, experience and ‘anecdotal’ evidence formed the foundation for referral.
It is really it is a combination of anecdotal evidence and experience. You know I am not really well versed on research data for complementary medicine, you know but I accept it as a sort of part of the repertoire of what we have to offer.
(Nick: 57–61)
GP accounts indicated some discomfiture in justifying their referring behaviour because of the dearth of scientific evidence and their lack of knowledge about the available evidence.
I can’t remember seeing any evidence, um, of efficacy for whatever [CAM treatment]. I am always a little bit suspicious of evidence anyway, because evidence always involves lots of other people who are not my patients… All I can say is that I can’t give you chapter and verse, but I have had patients that I have sent to the centre who have come back better than when I sent them. And that is really what we are all about.
(John: 66–72)
Scientific evidence constituted a core focus for many of the interviewees. One GP indicated that while he had been equivocal about trying a new complementary medicine for a patient, his uncertainty was later quashed by a scientific study, which provided strong support for the approach.
The first time I saw him [i.e. the GP was a CAM patient] he suggested [some complementary medicine,] its interesting that he was ahead of the game, there’d been…an article in Gut since then from Cologne showing in a double blind trial …that [this medicine]was as effective as [conventional medicine]. I was impressed.
(Peter: 78–86)
Some adopted a pragmatic stance by suggesting that the gaps in knowledge for complementary medicine are no different from the gaps in knowledge for conventional medicine.
Well you could argue that [the evidence isn’t sufficient] for quite a lot of conventional treatments as well.
(Tony: 226–227)
Furthermore, the prima facie safety of CAM versus the known iatrogenic consequences of conventional medicine provided good reason to refer for the half of the GPs.
We cause an awful lot of harm in conventional medicine which complementary medicine rarely does, and we tend to overlook that
(John: 235–7)
b. Valuing CAM and trusting the complementary practitioner (CP)
Good communication with the CP to whom they refer to was also important to the GPs.
I mean the [CAM practitioner] is very good, if he sees them [privately] he does just drop us a line.
(Peter, 164–5)
In this study, the CAM practitioners were all medically qualified, and although interviewees did not explicitly discuss this as influencing their decision to refer it was implied but not explicitly stated that a medical qualification gave the GPs confidence and trust in the CPs’ judgments. For many, trust was borne out of positive experience from past referrals.
I had a very good relationship with [CAM practitioner], … he sees most of my patients…, through my knowledge of him and I have, I know exactly what we are allowed to refer…
(Andrew: 12–13)
GPs also valued the CPs’ ability to offer patients time, both diagnostically and therapeutically. This was seen as augmenting their ability to identify and treat complex patient problems.
I refer them on because I know that they … have more time and more experience and expertise than I have, so homoeopathically speaking, it makes sense.
(Andrew: 184–8)
Some GPs valued the expertise CPs could offer, such as their diagnostic ability, especially when conventional medicine had failed. Others, however, were a little more circumspect and only referred when there was a clear diagnosis.
You’ve got to exclude the nasty, the nasties first, the cancer and so on, but once those are excluded he can delve into his, into his areas of interest
(Peter: 27–9)
2. GPs perception of their patients’ attitudes, requests and preferences for CAM
GPs recognized that their patients’ experiences of and attitudes to both conventional medicine and CAM were of central importance to the referral process (see Figure 3). GPs offered five reasons for patient openness and requests for CAM:
Model of patient-related factors. Note: the arrows do not imply causality
(i) Patients perceive CAM as safe, possibly incorrectly.
It is safer they say… but I mean that is a very contentious issue isn’t it. Safety is not always the case with CAMs.
(Adrian: 73–7)
(ii) Patients like CAM because it does not involve drugs, viewing CAM as a positive drug-free alternative, which would drive a request for a referral.
Lots of people who don’t really want to take medicine all the time, don’t want to take medication to control something chronic.
(Mary: 31–3)
(iii) Some patients view health and illness holistically or are particularly receptive and/or have faith in it.
(iv) GPs’ patients, friends or family have previous positive experiences of CAM.
Sometimes there is recommendation, they have a friend who has been or has done well with complementary therapy.
(James: 73–4)
(v) Increased awareness from the media and knowledge from the internet
I think that patients are much more aware of complementary therapy now, because of the media, and people like Prince Charles.
(James: 77–8)
Matching between doctors’ attitudes and treatment preferences on the one hand and patients’ views on these issues on the other was described as a vital prerequisite for referral. Without a mutual commitment to ‘trying CAM’, referrals would be resisted by patients.
There have been quite a few patients for instance who I have offered referral and they haven’t wanted it
(Tony: 40–1)
Discussion
Summary of main findings
This is the first exploratory qualitative analysis of GPs reasons for NHS referral to CAM so the results cannot be viewed as definitive. We have systematically identified that CAM referral is complex and occurs when there is matching of both the GP factors: their attitude to CAM, and the patient’s diagnosis and clinical state and essential patient-related factors identified in this study: patient’s openness and motivation and/or requests for CAM treatment. As illustrated in Figures 1–3, the key original finding was that GPs’ perceive the final decision about CAM referral to be largely shaped by patient preference. In particular, GPs described patient enthusiasm for CAM as a pre-requisite to them investigating CAM referral. A mutual commitment was perceived as crucial and this contrasts with the evidence on referring behaviour for conventional medicine. Referrals to conventional medicine specialists are also not solely based on clinical need and GPs perception of the patient’s pressure to refer does affect their referral rates (see eg, Armstrong et al., Reference Armstrong, Fry and Armstrong1991; Newton et al., Reference Newton, Hayes and Hutchinson1991; Cockburn and Pit, Reference Cockburn and Pit1997; Britten, Reference Britten2004; Little et al., Reference Little, Dorward, Warner, Stephens, Senior and Moore2004). However, GPs still may refer patients to a conventional specialist based on clinical need without complete agreement from their patient (Reynolds et al., Reference Reynolds, Chitnis and Roland1991; Forrest, Reference Forrest2003) and take the time to persuade patients of the merits of conventional treatments, by citing the evidential basis for their recommendations and offering the benefits that will ensue should treatment recommendations be followed
We also identified that GPs consider CAM referral when there are no other therapeutic options available for their patients. Various factors increase the likelihood of referral: the positive attitude of the individual GP to, and experience of, CAM; the patient’s attitudes, openness and request for CAM treatment and confidence in the CP. Scientific evidence was a core focus for all interviews and knowledge of CAM was a factor in promoting or diminishing referral rates for some GPs but was not an absolute requirement for referral. Our data suggest that GPs consider anecdotal outcomes from CAM treatment as more influential on their decision to refer than the need for a systematic evidence base. The finding that clinical practice is not driven solely by research evidence has been identified previously. Despite the rapid escalation of evidence-based medicine in the 1990s (Sackett et al., Reference Sackett, Richardson, Rosenberg and Haynes2000), clinical actions are not solely based on available evidence. For example, clinicians may not follow best practice guidelines (Freemantle et al., Reference Freemantle, Harvey and Wolf2000; Raine et al., Reference Raine, Sanderson, Hutchings, Carter, Larkin and Black2004). The direct translation of research evidence into a clinical setting is complex. Interpretation of data is dependent on clinical judgement, individual patient need and the practitioners’ mindset and experiences (Garfield and Garfield, Reference Garfield and Garfield2000; Malterud, Reference Malterud2001). From the tentative evidence in this paper, the role of evidence in referral to CAM appears to be similarly complex as referrals to conventional medicine, but a key contrasting element remains and that is the greater weight given to patient enthusiasm in CAM when compared to conventional medical treatments.
Limitations of this study
The number of participants was limited because we were unable to obtain further interviews within this NHS contract and data saturation may not have been reached. Also, as no information could be obtained to explain the reasons for those GPs who refused to take part in the study, it is not possible to identify any differences between responders and non-responders and this limits the generalizability of these findings.
The CPs to whom these GPs referred were all medically qualified; referrals and attitudes towards referrals to non-medical CAM therapists may differ. These data are based on referral to a single CAM practice and our findings will need to be further explored in other clinical environments. However, the focus of this study was on the referring GP and not on the practice to which they referred.
Comparison with existing literature
Prior CAM experience influences referral decisions. GPs are more likely to refer if they themselves have used CAM (Boucher and Lenz, Reference Boucher and Lenz1998; Lewith et al., Reference Lewith, Hyland and Gray2001), had positive feedback from patients they have referred in the past (Wharton and Lewith, Reference Wharton and Lewith1986; Emanuel, Reference Emanuel1999), and consequently on anecdotal evidence from friends and colleagues (Wharton and Lewith, Reference Wharton and Lewith1986). Both non-medical factors (patient’s attitude to CAM and conventional medicine) and request for CAM as well as medical factors (adverse drug reactions) influence this decision (Paterson, Reference Paterson1997; Giveon et al., Reference Giveon, Liberman, Klang and Kahan2003; Sharples et al., Reference Sharples, van Haselen and Fisher2003). Some previously identified triggers for referral were also systematically identified in this study, for example treatment failure and the inappropriate use of conventional medicine (Borkan et al., Reference Borkan, Neher, Anson and Smoker1994; Boucher et al., 1998; van Haselen et al., Reference van Haselen, Reiber, Nickel, Jakob and Fisher2004) or a poor response to conventional treatment (Lewith et al., Reference Lewith, Hyland and Gray2001; Poynton et al., Reference Poynton, Dowell, Dew and Egan2006). Being a GP practitioner of CAM increased referrals (the GP is knowledgeable about which patients are most likely to benefit), but it was not essential (Wharton and Lewith, Reference Wharton and Lewith1986; Anderson and Anderson, Reference Anderson and Anderson1987; Thomas et al., 1995; White et al., Reference White, Mitchell and Ernst1996; Perry and Dowrick, Reference Perry and Dowrick2000; Lewith et al., Reference Lewith, Hyland and Gray2001; Schmidt et al., Reference Schmidt and Barton2002). A lack of emphasis on safety was at odds with previous literature (Cohen et al., Reference Cohen, Penman, Pirotta and da Costa2005), but possibly understandable in the context of this study where referral is to medical colleagues. This suggests that if the GP or patient knows little about CAM then referrals may not occur or not be made in a timely manner.
Conclusion: implications for future research and clinical practice
These initial exploratory findings suggest that referral to CP appears to differ from referral to conventional medicine and this may have important consequences. We speculate that these differences when GPs refer to a conventional and CAM practitioner occur on three levels:
(a) Patient clinical need. In conventional referrals, there is usually an evidence-based clinical need for specialist advice about treatment or diagnosis. CAM referrals are not usually evidence-based, but there is a possibility of clinical benefit for the patient in the management of their chronic illness.
(b) The Doctor’s skills. Referrals by GPs for either conventional or complementary expertise seems to follow the same patterns: (i) GPs may not have the skills to manage their patient’s clinical need so they refer their patient to a ‘specialist’; (ii) GPs with an interest in developing their specialist skills will increase their knowledge and management of handling complex patients through referral and specialist communication (Cummins et al., Reference Cummins, Jarman and White1981; Reynolds et al., Reference Reynolds, Chitnis and Roland1991; O’Donnell, Reference O’Donnell2000).
(c) The process of referral. There is a key difference between referrals to conventional and complementary specialists. In conventional medicine, the referral may be doctor driven based on clinical need, but can also be driven by the patient based on their demands and expectations (eg Little et al., Reference Little, Dorward, Warner, Stephens, Senior and Moore2004). In a CAM referral, there has to be a consensual agreement between the doctor and the patient for a referral to occur.
The GPs in this study did refer to CAM especially once conventional treatments have failed and when patients were enthusiastic. This implies that referrals may (a) not be best practice (due to lack of CAM knowledge and that patient request may be driving the referral); (b) be excluding those patients who do not have a prior knowledge or understanding of CAM; and (c) not be made in a timely fashion as referrals usually occur after exhausting conventional management routes. In addition, GPs conflict over the lack of evidence or lack of understanding of CAM may also preclude effective and timely referring behaviour.
A follow-on study is required to further explore these data in different NHS environments. Future work is needed to elucidate a fuller understanding of why some GPs find it valuable to refer patients to CAM within the NHS and others never do so. There is also a need to understand the referral processes to non-medical CPs. This exploratory study has identified that a substantial minority of GPs value the option for referral to CAM within the NHS. Increasing knowledge and evidence about CAM will assist GPs and their patients in this decision process.
Acknowledgements
Funding sources had no involvement in this review. Sarah Brien is funded by a Personal Award Scheme Researcher Development Award, for Complementary and Alternative Medicine, from the National Institute of Health Research (Department of Health) (ref no. PDA04/CAMs2/02). Emily Howells received no source of funding for this paper. Geraldine Leydon is funded by the University of Southampton. George Lewith’s post is funded by the Rufford Maurice Laing Foundation. The authors would like to thank those GPs who agreed to participate in this study. Funding sources had no involvement in this review. George Lewith is one of the CPs referred to under the PCT contract in this study.
Appendix 1: Initial interview topic guide
Opening question
How often would you say you refer patients to the [CAM clinic]?
When was the last time you did?
These patients might be good examples for you to think about when answering my questions.
Specific questions
1. Reasons for GP to refer patients for CAM therapy
If at times you refer without being asked by a patient, what prompts you to do this?
Can you tell me any more about your decision making?
Are they generally purely medical decisions?
Has personal/anecdotal evidence been your main guide?
Does the type of illness affect your decisions?
Does how the patient make you feel have any impact?
Additional questions added to the final guide:
What is it about a patient/situation that makes you think of CAM?
Why do you refer some patients with_and not others?
Is this pattern different for CAM referrals and other conventional medicine referrals when you decide to refer a patient?
Is your referral threshold higher or lower for CAM
Time before referral greater or less for CAM
2. Reasons patients ask for CAM referral
What reasons do patients give for requesting CAM referral?
Do patients ask for medical/Non-medical reasons?
What is important for them: Personal experience or ‘ evidence’?
Do any of them not want conventional treatment?
Are many patients aware of the opportunity?
Is this different for CAM than for other conventional medicine referrals?
Is your referral threshold higher or lower for CAM?
Time before referral greater or less for CAM?
How much do patient requests and the NHS Contract stipulations match up?
Can you send everyone you want or is the service too limited? Many people disappointed?
OH, WE HAVEN’T TALKED ABOUT…
End of Interview
