NATIONAL SURVEY OF ACCESS TO COMPLEMENTARY HEALTH CARE VIA GENERAL PRACTICE

Final Report to Department of Health

August 1995

Kate Thomas Margaret Fall Gareth Parry Jon Nicholl

Medical Care Research Unit SCHARR

Regent Court 30 Regent Street Sheffield S1 4DA

KJT /MCRU / 1995

CONTENTS Page List of tables and figures Abstract 1 Background 3 Study aims 4 Methods 4 Sample size and character 4

Pilot study 5

Data collection 5

Results 6 Response rate 6

Representativeness of the sample 12

Available data sets 19

Availability of complementary therapies via general practice 20

• Access to different complementary therapies 22

• Provision within the practice: estimates and characteristics 25

• NHS referrals for complementary therapies 31

GP behaviour in consultations in past week 33

• GP behaviour in relation to different therapies 33

• Differentials in behaviour by characteristics of GPs 36

• Estimates of GP activities in an average week 38

Discussion 43 Acknowledgements 48 References 49

Index of Tables and Figures Page Table 1 Response rate: by mailings

7

Table 2 Comparison of response rates: all national studies, involving postal questionnaires to GPs, published in 1992-1994 in the British Journal of General Practice

7

Table 3 Comparison of response rate: all local studies involving postal questionnaires to GPs, published in 1994 in the British Journal of General Practice

8

Table 4 Representativeness of data: analysis of partnership size

10

Table 5 Representativeness of data: analysis of fund-holding status

11

Table 6 Spearman Rank correlation co-efficients for relationship between response rates for each FHSA and for known characteristics of FHSAs

14

Table 7 Access to complementary therapies: estimates based on unweighted data compared with estimates based on data weighted to take variation in FHSA response rates into account

14

Table 8 Unweighted data: proportion of practices indicating access to complementary therapies by number of mailings received

15

Table 9 Representativeness of data: analysis by GP age group

15

Table 10

Representativeness of data: analysis by sex of GPs

16

Table 11 Unweighted data: proportion of GPs indicating use of complementary therapies by number of mailings received

17

Table 12 Weighted and unweighted data by partnership size of GP

17

Table 13 Proportion of partnerships providing access to complementary therapies via treatment within the practice or NHS referrals. Estimates of provision weighted by type of response

21

Table 14 Characteristics of practices offering access to complementary therapies via primary health care team, independent therapist or NHS referral

21

Table 15 Complementary therapy provided by type of provision

23

Table 16 Complementary therapy via general practice by FHSA of responding practice and response rate

24

Table 17 Provision within the practice by type of practitioner and therapy offered

26

Table 18 Mode of provision within practice by therapy

27

Table 19 Who pays for complementary therapies provided within general practice by therapy

28

Table 20 If NHS provision, source of funding for complementary therapies

in general practice by type of practitioner

30

Page

Table 21 NHS Referrals outside the practice for complementary therapies

by place of reference

32

Table 22 NHS referrals outside the practice for complementary therapies by source of funding

32

Table 23 Complementary therapies in consultations in one week: a) The number of GPs treating patients with complementary therapies, referring for such therapies, or recommending/endorsing treatments; b) the number of occasions each action was performed; and c) the average weekly interventions per GP reporting the action, d) the average weekly intervention per GP in England by therapy (weighted data)

34

Table 24 Complementary therapies in consultations in one week: a) The number of GPs giving a neutral response to a patient enquiry about complementary therapies or advising against their use; b) the number of occasions each action was performed; and c) the average weekly interventions per GP reporting the action, d) average weekly intervention per GP in England by therapy (weighted data)

35

Table 25 Complementary therapies in consultations in the last week: estimated proportion of GPs treating, referring or endorsing treatment by age group, sex and status of GP (weighted data)

37

Table 26 Complementary therapies in consultations in the past week: estimated proportion of GPs treating, referring or endorsing treatment by location of practice (weighted data)

37

Table 27 Estimated proportion of GPs treating, referring or endorsing complementary therapies by therapy (weighted data)

39

Table 28 Estimated proportion of GPs giving neutral or negative response to enquiries about complementary therapies by therapy (weighted data)

39

Table 29 Complementary therapies in consultations in the last week: GPs who gave a neutral response or advised against, by GPs giving treatment, referring, recommending or endorsing treatment (weighted data)

41

Table 30 National estimates of treatment, referral and recommendation/endorsement of complementary therapies in GP consultations in an average week

42

Figure 1 Groups of patients or conditions mentioned by GPs treated by complementary therapies within the practice

30

ABSTRACT

Study aims: To describe the scale and scope of access to complementary therapies

obtained via general practice with particular reference to acupuncture, chiropractic,

homoeopathy, hypnotherapy, medical herbalism and osteopathy.

Design: A postal questionnaire relating to provision of complementary therapies in

the practice as a whole, and to consultations in the past week, was sent to 1226

individual GPs in a random cluster sample of GP partnerships in England, taken from

24 FHSA lists. GPs received up to two reminders. A follow-up survey of all non-

responders was undertaken, requesting answers to three key questions.

Subjects: GPs from a random sample of 1226 (one in eight) GP partnerships in

England.

Main outcomes: Description of the scale and scope of access to complementary

therapists via general practice, including provision within the practice and NHS

referrals outside the practice. Estimates of the proportion of GPs treating, referring

and endorsing the use of complementary therapies in consultations in a one week

period.

Results: Seven hundred and sixty GPs returned the completed questionnaire, a

response rate of 62%. In addition to this, 204 (16.6%) non-responders replied giving

basic information. Responders appear to be representative of GP practices in

England with respect to known characteristics, and to the provision of complementary

therapists when compared with non-responders.

Analysis by practice showed that an estimated 39.5% (95% CI 35%-43%) of GP

partnerships in England now provide access to some form of complementary therapy

for their NHS patients. An estimated 21.4% (95% CI 19%-24%) are offering access

via the provision of treatment by a member of the primary health care team, 6.1%

(95% CI 2%-10%) employ an ‘independent’ complementary therapist, and an

estimated 24.6% of partnerships (95% CI 22%-27%) make NHS referrals for

complementary therapies. Of the therapies investigated, acupuncture and

homoeopathy are the most commonly provided, although the most frequently

employed independent practitioners were osteopaths.

Fund-holding practices are significantly more likely to offer complementary therapies

via a member of the primary health care team than non-fund-holding practices,

(P = <0.05), and single-handed GPs are significantly less likely to offer this service

(P = <0.01

In most cases, the complementary therapies provided within the practice were offered

by GPs (64%), and provision was split equally between regular clinics and normal

surgery time. 17.4% of the provision within the practice (including that offered by

independents) was paid for by the patient. 12% of fund-holding practices in the

sample (20/161) used savings or practice funds to purchase complementary therapies

for their patients within the practice.

Of the referrals, those to NHS homoeopathic hospitals were the most commonly cited,

followed by referrals to other NHS hospitals for acupuncture. The scale of this

provision cannot be ascertained accurately, but one referral per month was the

frequency most commonly cited by those GPs who made any such referrals. Of the

fund-holding practices an estimated 9% (14/161) reported using savings to fund such

referrals.

It is estimated that 45% of GPs recommend or endorse a complementary therapy in

their consultations in an average week, 21% refer a patient for complementary therapy

(private or NHS), and 10% treat a patient with one of the named complementary

therapies. On this basis, it is estimated that 14,900 (95% CI 12507/17302) treatments

with one of these complementary therapies are given by GPs in an average week,

750,000 in a year.

Conclusions: Access to complementary therapies in general practice is widespread

amongst practices, but appears to affect a relatively small number of patients.

Acupuncture and homoeopathy are the therapies most commonly offered within

practices by the primary health care team, and also the therapies for which NHS-

funded referrals are most commonly made. Manipulative therapies are more likely to

be offered by independent therapists working within practices, the majority of whom

appear to be NHS funded and offer their services free to NHS patients. This type of

provision is found in relatively few practices, but has the potential to affect a greater

number of patients.

BACKGROUND

The popularity of complementary medicine continues to be asserted by the

professional associations and umbrella organisation’s related to these therapies(1) and

this has been confirmed to some extent in pilot work recently undertaken here in the

MCRU(2). The BMA report on complementary therapies was very much more

favourable than its predecessor published in the 1980’s(3,4) and work undertaken by

the National Association of Health Authorities and Trusts (NAHAT) of the views of

NHS purchasers towards complementary therapies in 1992 also revealed largely

positive attitudes towards its provision on the NHS.

Within primary care, provision has been facilitated by changes in the GP contract (in

1990) and the subsequent introduction of GP fundholding. Non-fundholding GPs, for

example, are using the ancillary staff budget to employ complementary therapists,

whilst fundholders are able to use the staff element of their budgets and `practice

savings’ for this purpose (for which prior approval of the FHSA is not required) (5) In

addition, GPs may make private referrals or provide a complementary therapy, such

as homoeopathy, themselves.

A study of GP fundholders undertaken by NAHAT in 1992 has reported that

independent complementary therapists practised in 14 out of a sample of 101 fund-

holding practices, half of whom provided the service free of charge to NHS patients(6).

However, the low response rate achieved (43%) gives cause for concern that the

results may be biased in favour of those practices which have a positive attitude

towards complementary health care, and it is therefore unlikely that 14% of all fund-

holding practices offer such a service to their patients. The NAHAT survey did not

obtain information on activity within non-fund-holding practices.

A survey of a representative national sample of GP practices was therefore

undertaken to ascertain the extent to which access to complementary health care is

currently gained through general practice.

AIMS AND OBJECTIVES OF STUDY

This nationally representative survey has the following three aims;

1) To describe the patterns of access to complementary health care via general

practice (mode of delivery, type of practitioner) with particular reference to

acupuncture, chiropractic, homoeopathy, medical herbalism and osteopathy.

2) To quantify the scale of provision, including NHS referrals to practitioners

outside the practice, and the volume of treatments conducted within the

practice.

3) To describe the relationship between practice characteristics (including

location) and the distribution of access to complementary health care gained

via general practice.

METHODS Sample size and character

The study focuses on GP partnerships as the main unit of analysis and this was

reflected in the sampling strategy employed.

Random cluster sampling was used to select partnerships from all those in England

(fund-holding and non-fund-holding) within a geographically distributed sample of

FHSAs. As activity with respect to complementary therapies may be related to local

FHSA policy, a large sample of 24 FHSAs (one in four) were sampled, three chosen at

random from each of the eight new Health Regions. Within each FHSA we randomly

sampled one in two practices. In this way, a sample of approximately 1226 practices

was identified (one in eight practices in England).

One GP in each sampled partnership received a letter requesting their participation in

the study. Within each practice, the GP was chosen randomly from the list provided

by the FHSA, so as to achieve a distribution of senior partners and other partners

across the sample, thus giving a further sample of GPs for the analysis of data relating

to individual behaviour.

Pilot study

A small pilot study was undertaken locally to assess the feasibility of the data

collection method, and to obtain feedback on the design of the survey instrument.

Following the pilot, the format of the questionnaire was revised substantially to aid

ease and speed of completion and comprehension. Writing to a random GP within

each practice did not appear to adversely affect response, and this method was

employed in the main study.

Two further questions were added to the questionnaire relating to activity in the

responding GP’s own consultations in the previous week.

The final questionnaire design consisted of two distinct parts which were colour-

coded. The first two sides contained questions which related to the GP’s personal

experiences in their consultations in the past week and the three key “screening”

questions relating to provision of complementary therapies in the practice as a whole.

The second, more substantial part of the questionnaire was structured with a page for

each therapy covering details of provision. This section was only completed by GPs

reporting activity in their practice.

Data collection

The majority of FHSAs provided printed address labels for use in the study and these

were used in all correspondence with the GPs. The sampled GP from each of the

1226 partnerships received the questionnaire with a covering letter from the

researcher.

After a period of two weeks approximately half the GPs who had not yet responded

received a post-card reminding them about the study and requesting them to return a

completed questionnaire. All other non-responders received a reminder letter and a

second questionnaire. After a further two weeks the post-card group received a

second questionnaire. A further letter and third questionnaire were dispatched where

necessary.

A follow-up of all those GPs in the sample who did not respond to any of these

contacts was conducted nine weeks after the initial mailing. This entailed a brief letter

and a request to answer the three key “screening” questions from the questionnaire.

For all mailings, letters were sent out in franked envelopes bearing the University

crest. Printed, reply-paid envelopes were provided for the return of completed

questionnaires.

RESULTS

Response rate

After three mailings, 760 completed forms were returned, giving a response rate of

62%. Of those who did not return forms, 33 wrote declining to participate (2.7%) and

9 questionnaires were returned with an indication that the GP had retired, was on

long-term sick leave, or had left the practice.

A fourth mailing, containing the letter and 3 “screening” questions only, was sent to

the 423 non-responding GPs. Of these, 204 (48.2%) replied, answering the three

questions as requested. Including these responses, information was obtained on a

total of 964 partnerships, 78.6% of the original sample of 1226 partnerships (Table 1).

Table 1 Response rate: by mailings

Sent Returned Completed

Questionnaires returned completed

No. % No. Cumulative %

1st mail 1226 469 38.2 469 38.2

2nd mail 747 150 12.2 619 48.8

3rd mail 556 141 11.5 760 62.0

4th mail (letter only) 423 204 16.5 964 78.6

Table 2 Comparison of response rates: all national studies, involving postal questionnaires to GPs, published in 1992-1994 in the British Journal of General Practice.

Title Coverage Sample size No of GP’s in sample

Reported sick, retired, 1 gone away

Response rate %

Involvement of the primary health care team in coronary heart disease prevention(7)

England 2000 305 64.4

(n=1092)

Annual assessment of patients aged 75 years & over: GPs’ & practice nurses’ views and experiences(8)

England & Wales

10002 none reported 69.3

(n=693)

Telephone & postal surveys of GPs: methodological considerations(9)

England & Wales

1732 49 52.3

(n=881)

Attitudes towards practice nurses - survey of a sample of GPs in England & Wales(10)

England & Wales

48002 41.9

(n=2013)

Patient access to GPs by telephone: the doctor’s view(11)

England & Wales

19802 none reported 74.0

(n=1459)

Access to complementary medicine via general practice

England 1226 33 refused

9 retired gone away.

62.0 (questionnaire) (n = 760)

78.6 (questionnaire & letter) (n = 964)

1 No refusals were reported. 2 Response from any GP in practice accepted.

Table 3 Comparison of response rate: all local studies, involving postal

questionnaires to GPs, published in 1994 in the British Journal of General Practice.

Title Coverage Sample

Reported sick, retired, gone away

Response rate %

Provision of obstetric care by GPs in the SW region of England(12)

South West RHA

424 random

GP’s

none reported

78.5

Anti depressant prescribing: a comparison between GPs and psychiatrists(13)

Cardiff East

123 random

GP’s

none reported

60.0

Written lists in the consultation! Attitudes of GPs to lists and the patients who bring them.(14)

Leicester 58 GP trainers

none reported

84.0

Monitoring anticoagulant control in general practices(15)

Lothian & Fife H. B.

198 senior partners

none reported

Fife 89.1

Lothian 89.6

Fear of aggression at work among GPs who have suffered a previous episode of aggression(16)

West Midlands RHA

2694 random GPs

not reported 40.6

Exploratory study of GPs’ orientation to general practice and response to change(17)

Leicester 110 young principals

none reported

44.5

Fundholders’ referral patterns and perceptions of service quality in hospital provision of elective general surgery(18)

Trent RHA

115 senior partners

none reported

67.0

GPs are frequently pressed for time and do not tend to give research questionnaires

priority. For these reasons, surveys of GPs tend not to achieve high response rates.

Those surveys which are published are likely to have the best response rates. We

identified all national surveys of GPs published in the British Journal of General

Practice between January 1992 and December 1994. The response rate of 62%

obtained from three mailings in our survey compares well with that achieved by these

five national surveys published recently (Table 2). Local surveys are often done to

obtain higher response rates due to saliency and ease of access for follow-up (e.g.

telephone). However, published data suggest such surveys obtain a wide range of

response rates, and that local surveys with smaller sample sizes cannot guarantee

response rates (Table 3).

An analysis of the effect of the post-card reminder sent to half our sample who did not

respond initially, shows that final response rates achieved were identical in this group

and in the group not receiving the post-card (46.3% v. 45.4% respectively). However,

the post-card may have had an effect on the timing of the response; after the second

mailing, 32% of the 404 GPs who received the post-card had responded, compared to

22.4% of those who had not.

Table 4 Representativeness of data: analysis of partnership size1

Single handed Partnerships of:

% 2-3 %

4-6 %

7+ %

Total n = 100%

All partnerships England2 31.5 35.4 28.2 4.9 9111

Partnerships in 24 sampled FHSAs 27.5 34.1 32.5 5.9 2452

1 in 2 sample of partnerships from 24 FHSAs

27.7 33.6 32.9 5.8 1226

Partnerships responding to questionnaire

23.5 34.7 34.9 6.9 760

Partnerships responding to 4th letter with “screening” questions only

29.4 32.4 32.8 5.4 204

All partnerships responding to letter or questionnaire

24.8 34.2 34.4 6.5 964

1 Calculated from information provided by FHSAs, summer 1994.

2 GMS statistics 1st April 1994 England and Wales, National and Regional Tables Department of Health, NHS Exec. HQ., PD (STATS) C, 6E43, Quarry House, Leeds, LS2 7UE Table E&W 07.

Table 5 Representativeness of data: analysis of fund-holding status1

Fundholder2 Non Fundholder Total

n (%) n (%) n = 100%

Partnerships in England and Wales3 2040

(21.1) 7616 (78.9) 9656

Partnerships in 24 sampled FHSAs

517 (21.1) 1935 (78.9) 2452

1 in 2 sample of partnerships from 24 FHSAs

254 (20.7) 972 (79.3) 1226

Partnerships responding to questionnaire

163 (21.4) 597 (78.6) 760

Partnerships responding to 4th letter with “screening” questions only

48 (23.5) 156 (76.5) 204

Total partnerships responding by questionnaire or letter

211 (21.9) 753 (78.1) 964

1 Calculated from FHSA information obtained summer 1994.

2 As at April 1st 1994 (includes 4th wave)

3 Department of Health, NHS Exec. HQ., PD (STATS) C, 6E43, Quarry House, Leeds, LS2 7UE information on 1.12.94.

Representativeness of the samples

The representativeness of the sample of 1226 partnerships and GPs was assessed

by comparing it with all practices and GPs in England with respect to known

characteristics of practice size, fund-holding status, age and sex of GPs. The

achieved sample was also assessed using the same characteristics (Tables 4-7).

Table 4 shows that the sample of 1226 partnerships contained a smaller proportion of

single-handed GPs than are found nationally (27.5% v. 31.5%) and a correspondingly

higher proportion of larger partnerships. The achieved sample of 760 partnerships

also contained proportionally fewer single-handed GPs, although respondents to the

4th mailing containing only the three “screening” questions were more representative

of the population as a whole (Table 4).

Fund-holding partnerships were appropriately represented in the sample of 1226

partnerships, and the achieved sample was equally representative of the population

as a whole in this respect (Table 5).

Despite the good overall representativeness of the achieved sample as measured,

there was a large variation in the response rate achieved for the questionnaires

between the 24 FHSAs sampled, with a range of 47% to 81% around the mean of

62%. Many of this range of responses (16 out of 24 FHSAs) fall outside the response

expected within a 95% confidence interval for response if there was no `clustering’ of

response (59% to 65%), and we therefore examined available information on the

FHSAs to test whether the response rates obtained across the FHSAs confirmed any

of the following hypotheses:

a) Response rates could be negatively correlated with the proportion of single-

handed practices in the FHSA as these practices were known to be under-

represented in the sample as a whole.

b) Response rates might be negatively correlated with the proportion of GPs in

the FHSAs known to have been born outside the UK It has been suggested

that this group of GPs may be less likely to respond.

c) Response rates might be negatively correlated with Jarman Deprivation

Scores for each FHSA.

d) Response rates might be positively correlated with the proportion of

responders in each FHSA stating that in the past week they had treated a

patient with a complementary therapy, referred them for such treatment, or

recommended/endorsed such treatment, or stating that access to

complementary therapies was provided in their practice, i.e. the issue of

saliency and response.

Spearman Rank correlation co-efficients calculated for each of these relationships

showed only poor correlations in each case. With only 24 observations, none of these

correlations reached statistical significance (Table 6).

While no significant, systematic relationship was found between any of the

measurable characteristics and FHSA response rates, it remains possible that the

differences in response rates could still bias the results towards the characteristics of

the FHSAs with the higher response rates. We therefore weighted the data by FHSA,

according to the response rates achieved, and compared weighted and unweighted

estimates of access to complementary therapies (Table 7). The weighted estimates

remained similar to the unweighted estimates. The data were not, therefore, weighted

according to FHSA prior to the analyses presented in this report.

Finally, and perhaps most importantly, are the results obtained from the three

“screening” questions asked in the follow-up letter to non-responders, and returned by

204 GPS (48% of non-responders) These provide strong evidence that the 760

responders to the main questionnaire are broadly representative of the population as

a whole with respect to access to complementary therapies via the practice. The

higher level of GPs reporting any NHS referrals from their practice amongst the ‘non-

responders’ (Table 8) is probably attributable to the inclusion of an exemplification of

the category of NHS referral (e.g. “homoeopathic hospital”) in the screening questions

sent in the follow-up letter to non-responders. This suggests that the proportion of

responders stating that their practice make any referrals for Complementary therapies

might have been higher had we included a prompt for them.

The principal estimates calculated in this report relating to access to complementary

therapies via GP practices have taken the answers given by those non-responders

who returned the follow-up letter to be representative of all ‘non-responders’.

Estimates for the whole practice population have therefore been calculated by

combining estimates for responders to the main questionnaire with estimates for non-

responders, based on responses to the non-responders’ follow-up letter.

An alternative, more conservative, estimate is offered for some key results in which

all practices not replying to the non-responders follow-up letter are assumed to be

non-active.

Table 6 Spearman Rank correlation co-efficients for relationship between response rates for each FHSA and known characteristics of FHSAs

Characteristic of FHSA r N p-value

Proportion of practices single-handed

- .28 24 0.18

Ranked Jarman deprivation scores - .16 24 0.45

Proportion of GPs born outside UK - .32 24 0.13

Proportion of GPs recommending complementary therapies in past week

.21 24 0.31

Proportion of responders from practices offering access to comp. therapies

- .00 24 0.99

Table 7 Access to complementary therapies: Estimates based on unweighted data compared with estimates based on data weighted to take variation in FHSA response rates into account

Access/provision Unweighted data estimates

n = 760

Weighted data estimates

n = 760

Provision by primary health care team

178 173.0

Provision by independent therapist

48 47.3

Referral 160 162.6

Any of these 280 279.6

Table 8 Unweighted data: proportion of practices indicating access to complementary therapies by number of mailings received Provision by

PHCT %

Provision by indep.

therapist %

Any NHS referrals

%

Any

%

N

Response to 1st mailing

27.5

6.6

22.6

40.1

469

2nd mailing 18.7 7.3 21.9 37.3 150

3rd mailing 15.2 4.3 17.4 27.7 141

Non-responders follow-up letter

18.1 5.9 30.4 43.1 204

Representativeness of individual GP data

Questionnaires were returned by 760 GPs. These included questions (in the form of a

grid or matrix) relating to their behaviour regarding the specified complementary

therapies in consultations in the past week. This sample of GPs is representative of

all GPs with respect to age and sex (Tables 9 and 10).

Table 9 Representativeness of data: analysis of age of GPs in years1

Under 30 %

30-34 %

35-44 %

45-54 %

55-64 %

64+ %

Total n =

100%

England2 1.9 16.2 37.1 28.6 14.6 1.6 26387

As % of those respondents giving age3

1.8 12.8 37.6 30.9 16.1 0.8 615

1 Study data obtained from respondents.

2 GMS Statistics 1st April 1994 England and Wales, National and Regional Tables. Department of Health, NHS Exec. HQ., PD (STATS) C, 6E43, Quarry House, Leeds, LS2 7UE. Table E & W 04

3 81% of responding GPs indicated their age group

Table 10 Representativeness of data: analysis by sex of GPs1

Male %

Female %

Total n = 100%

England2 70.8 29.2 28587

Study respondents as %

of those stating sex3

70.7 29.3 605

1 Study data obtained from respondents. 2 GMS statistics 1st April 1994 England or Wales, National and Regional Tables Department of Health,

NHS Exec. HQ. PD (STATS) C, 6E43, Quarry House, Leeds, LS2 7UE Table E&W 07. 3 79.6% stated sex.

The accuracy of the numerical data provided by GPs regarding actions in the past

week is subject to a number of caveats. While a one-week investigation period should

minimise problems associated with recall, it is possible that some GPs have described

a ‘typical’ week rather than an actual week. This may have resulted in an inflation in

the levels of activity reported. Activity levels may also have been misrepresented by

the small number of GPs (30/730) who entered ticks in the cells of the question rather

than giving a number, these entries have been counted as single events. The majority

of GPs (78%) did not fully complete either matrix. However, this appears to have

been due to interpretation of the instructions rather than random omission (i.e. they

did not want to write 0 in the, majority, negative cells). In these cases, a blank cell in a

matrix where at least one cell was completed has been counted as an indication that

no such action took place.

Finally, there is the effect of the response rate to consider; the follow-up exercise with

non-responders did not include questions about individual GP behaviour. However,

the non-responders were found to be similar to responders with respect to practice

provision. The estimates calculated in the following tables are based on the

assumption that GP activity among non-responders is similar to the lower level

reported by those GPs responding to the third mailing of the questionnaire (Table 11).

Where appropropriate, lowest likely estimates are also given, based on the

assumption that all non-responders were non-active with respect to complementary

medicine in the week surveyed. However, these lower estimates are likely to

underrepresent activity as it is unlikely that all non-responding GPs undertook none of

these actions in the week surveyed.

Table 11 Unweighted data: proportion of GPs indicating use of complementarytherapies by number of mailings received.

Response to;

Treatments

n (%)

Referrals

n (%)

Endorsements

n (%)

N

(100%)

1st mailing 68 (14.5) 110 (23.5) 212 (45.2) 469

2nd mailing 18 (12.0) 31 (20.7) 67 (44.7) 150

3rd mailing 10 (7.1) 25 (17.7) 49 (34.8) 141

The achieved sample included responses from a high proportion of single-handed

GPs, compared to the population of GPs in England. This is due to the fact that the

initial sample was of practices rather than GPs. As partnership size may be related to

activity in complementary therapies the data on GP behaviour have been weighted

according to partnership status to ensure that they are representative of all GPs in this

respect (Table 12).

Table 12 Weighted and unweighted data by partnership size of GP

Size of partnership

Unweighted sample n %

Weighted sample n %

England 1

n %

Solo 165 (22.1%) 133 (10.8) 2,870 (11%)

2-3 256 (34.2%) 367 (29.9) 7,880 (30%)

4-6 270 (36.1%) 566 (46.2) 12,192 (46%)

7+ 57 (7.6%) 160 (13.1) 1,485 (13%)

All 7482 (100.0) 1226 (100.0) 26,387 (100%)

1 Unrestricted principals: Source GMS Statistics 1 April 1994 England and Wales Table ETW O6 2 12 NK

Available data sets

Coded questionnaires were entered on to the computer using EPI-Info5 data entry

programme and exported into SPSS for analysis. The data form three subsets as

follows;

1) The data relating to provision and access to complementary therapies in the

practice as a whole, obtained from 760 partnerships initially, plus an additional

204 in response to the letter. (For the purpose of calculating national

estimates of access to complementary therapies, the data relating to these

204 partnerships are included and treated as representative of all non-

responders. In this way an overall estimate is calculated combining

separately weighted estimates for responders and non-responders.)

2) Data giving details of provision (e.g. provision by whom and how it is funded)

is available for the 280 partnerships which reported offering either treatment

within the practice or NHS referrals for complementary therapies (302

instances of provision).

3) Data relating to individual GP behaviour in the past week, obtained from a

sample of 760 GPs, weighted according to partnership size and response

category (n=1226).

THE AVAILABILITY OF COMPLEMENTARY THERAPIES VIA GENERAL PRACTICE

This part of the analysis utilises the 760 questionnaires received which indicated

whether or not the practice, as a whole, provided access to complementary therapies

via provision by the primary health care team, provision by an ‘independent’

complementary therapist not offering the therapy as part of a wider job remit, or

access via NHS referrals for treatment involving complementary therapies. In

addition, where appropriate, the 204 responses received to the fourth mailing are

included in the analysis. These responses are treated as being representative of the

sub-group of non-responders, rather than pooled with the data from questionnaire

respondents.

Table 13 shows that a significant proportion of GP partnerships in England, 39.5%

(95% CI 35%-43%), now provide access to complementary therapies for their NHS

patients. An estimated 21.4% of practices in England (95% CI 19%-24%) are offering

access to one of these therapies through the provision of treatment by a member of

the primary health care team and 24.6% (95% CI 22%-27%) make NHS funded

referrals for complementary therapies. The presence of an ‘independent’

complementary therapist within the practice is relatively rare, an estimated 6.1% of

practices (95% CI 2%-10%). This estimate has the widest confidence interval, but it is

very stable across the two samples and the estimate is probably more reliable than

these intervals suggest

These estimates are based on the assumption that responders are representative of

the population of practices. If non-responders are assumed to be non-providers, the

following estimates can be made; provision of complementary therapies by a member

of the primary health care team 17.5%, NHS referrals by practice 18.1% and provision

via an ‘independent’ complementary therapist 4.9%. However, there is no a priori

reason to believe that none of the non-responders make any provision at all and these

estimates should therefore be understood as the lowest likely estimates or ‘bottom

line’ with respect to provision of complementary therapies.

An analysis of the characteristics of practices offering complementary therapies via

the primary health care team, an ‘independent’ therapist or NHS referrals are shown in

Table 14. ‘Practice location’ was constructed from the answers given by responding

GPs relating to the best description of their practice and the population it served.

‘Inner city’ includes all practices who mentioned this as the best description for all or

part of their practice population. The description ‘rural’ was constructed in a similar

way. ‘Else’ is composed mainly of practices described as having a ‘town’ or suburb

location. Information on fund-holding status includes fourth wave fundholders and

was obtained from the FHSAs, as was the information on the number of partners in

each practice. Table 14 shows that the estimates for the proportion of partnerships

making NHS referrals do not vary substantially with these practice characteristics and

all are below, or at the lower end of the range suggested by the 95% CI for the overall,

weighed estimate (22%-27%). Estimated provision via the primary health care team

varies more with these characteristics. Fund-holding practices are significantly more

likely to offer this type of provision, 27% compared with 21% (P = <.05), and single-

handed GPs are significantly less likely to offer such provision compared with larger

practices (14.3% v 24.8% and 25.4%, P = <.01). Practices serving mainly rural

populations were more likely to offer complementary medicine via the primary health

care team.

The number of practices reporting an ‘independent complementary medicine therapist

in the practice is small, and none of the differences observed in Table 14 reach

statistical significance. All of these estimates are within the 95% CI for the overall

estimate of provision (2%-10%) and the majority fall in the top half of this range.

Table 13 Proportion of practices providing access to complementary

therapies via treatment within the practice or NHS referrals. Weighted estimates of provision and 95% confidence intervals

Respondents

N

Weighted estimates of provision

95% CI for % sampling error1

Primary health care team provision

178 21.4 (19-24)

Independent complementary therapy practices working in practice

48 6.1 (2-10)

Any referral to NHS for treatment

160 24.6 (21-28)

‘Yes’ to any of these questions

283 39.5 (35-43)

1 Confidence intervals have been widened by rounding up and down to help adjust for additional variation not taken into account by treating the achieved sample as fixed.

Table 14 Characteristics of practices offering access to complementary

therapies via primary health care team, independent therapist or NHS referral

Practice characteristic

PHCT

n (%)

Independent1

n (%)

NHS referral2

n (%)

Any of these

n (%)

N

Fund-holding:

Yes 57 (27.0)* 14 (6.7) 53 (23.9) 95(45.5)** 209

No 152 (21.0) 46 (6.1) 152 (21.0) 276 (36.6) 755

964

Practice location:3

Inner City 21 (18.6) 8 (7.1) 25 (22.1) 39 (34.5) 113

Rural 47 (28.7) 14 (8.5) 33 (20.1) 67 (40.9) 164

Else 110 (22.9) 26 (5.4) 101 (21.0) 176 (36.7) 480

757

Partnership size:

1 GP 35 (14.3)** 20 (8.1) 61 (24.7) 86 (34.8) 247

2-3 GPs 81 (24.8) 19 (5.8) 80 (24.5) 137 (41.9) 327

4+ GPs 99 (25.4) 21 (5.4) 81 (20.8) 148 (37.9) 390

964

All (weighted) 21.4% 6.1% 24.6% 39.5%

1 NK 3 2 NK 17 3 NK 3 (excluding 4th mailing Chi square for difference in provision according to practice characteristic (* P = <0.05, ** P = <0.01)

Estimates for provision of any of the three types of provision suggest that

complementary therapy provision is more common in fund-holding practices

(45.5% compared with 36.6%, P = <.01), and appears to be less likely in single-

handed practices (34.8% compared with 39.7% for all other practices), although this

difference does not reach statistical significance.

Access to different complementary therapies

Data on the type of therapy offered is available for respondents to the full

questionnaire only, as the fourth mailing letter did not seek this information. Table 15

shows that access to the different types of complementary health care is not uniform.

Acupuncture and homoeopathy are clearly the most commonly provided forms of

complementary therapy provided by or via general practice. GPs have a long tradition

of offering homoeopathy as part of primary care, and this is reflected in the distribution

of therapies provided ‘in house’ by the primary health care team. More surprising

perhaps is the relative popularity of acupuncture amongst GPs and other memebrs of

the team. A much smaller proportion of ‘in house’ provison relates to the manipulative

therapies (chiropractic and osteopathy). This may be due to the training and

equipment requirements of certain therapies, rather than a reflection of their relative

popularity. Osteopathy is the most commonly provided therapy where an

‘independent’ therapist works in the practice and this form of provision may involve

relatively large numbers of patients.

The sample structure was designed to provide a random one in eight GP partnerships

in England. These partnerships were chosen as a one in two sample from 24 FHSAs.

There was no strong evidence of large variations in provision between the FHSAs

(p = 0.022), with only two FHSAs Humberside and West Sussex, indicating

statistically significant differences from the overall rate (Table 16).

Table 15 Complementary therapy provided by type of provision offered Therapy

Primary health care team

n (%)

Independent

n (%)

NHS referral

n (%)

All instances of provision

n (%)

Acupuncture 96 (43.0) 13 (16.9) 68 (30.0) 177 (33.6)

Chiropractic 5 (2.2) 5 (6.4) 15 (6.6) 25 (4.7)

Homoeopathy 51 (22.9) 6 (7.8) 95 (41.9) 152 (28.8)

Hypnotherapy 42 (18.8) 8 (10.4) 15 (6.6) 65 (12.3)

Medical Herbalism 4 (1.8) 3 (3.9) 4 (1.8) 11 (2.1)

Osteopathy 11 (4.9) 21 (27.3) 25 (11.0) 57 (10.8)

‘Other’ therapy2 14 (6.3) 21 (27.3) 5 (2.2) 40 (7.6)

Total

223 (100.0) 77 (100.0) 227 (100.0) 527 (100.0)

1 Practices may offer more than one therapy and/or have more than one type of provision 2 Other therapies mentioned more than once included aromatherapy (12 instances), reflexology (8), massage (4),

Alexander Technique (3) and ‘manipulation’ (2)

Table 16 Complementary therapy via general practice by FHSA1 of responding practice and response rate

FHSA

Any provision by practice

Responses to questionnaire

or to letter

n (%) (n = 100%)

Newcastle 7 (33.3) 21

S Tyneside 2 (20.0) 10

Humberside 14 (25.5) 55

Derbyshire 17 (26.6) 64

Barnsley 4 (22.2) 18

Rotherham 6 (37.5) 16

Suffolk 11 (33.3) 33

Cambridge 9 (31.0) 29

Oxfordshire 14 (42.4) 33

Essex 50 (43.1) 116

Redbridge 22 (46.8) 47

Kensington & Chelsea 19 (48.7) 39

W Sussex 24 (58.5) 41

Surrey 25 (40.3) 62

Croydon 5 (21.7) 25

Hampshire 37 (38.1) 97

Somerset 14 (46.7) 30

Avon 25 (35.2) 71

Coventry 12 (52.2) 23

Dudley 6 (26.1) 23

Shropshire 9 (34.6) 26

Cumbria 23 (50.0) 46

Oldham 4 (22.2) 18

Wirral

12 (52.2) 23

All 371 (39.5) 2 964

1 Overall X2 = 38.64 23df p = 0.0217

2 Weighted estimate using non-responders data to represent all non-responders

The sample of 24 FHSAs was not designed to produce accurate regional estimates

and the data have not, therefore, been aggregated regionally. Together, however, the

24 FHSAs form a good representative sample for England as a whole.

Provision within the practice: estimates and characteristics

Data were obtained giving details of the various types of complementary therapy

provision from 280 partnerships (302 instances of provision). There is no reason to

believe that there is any systematic bias in these data with respect to the information

provided. The data from these 280 partnerships are therefore treated as

representative of all partnerships currently offering complementary medicine in

England, an estimated 3,500 partnerships.

Complementary therapies were reported as being provided within the practice, either

by a member of the primary health care team or by an ‘independent’ therapist who

could be working on a sessional basis with NHS funding or on a private basis, making

a charge to patients attending. Although it is technically possible for any member of

the primary health care team to offer a therapy, in practice the majority of provision

reported (64%) was offered by one of the GPs. However, the manipulative therapies

were more likely to be provided by someone outside the primary health care team

(Table 17).

Much of the provision of these therapies within the practice was in regular (weekly,

fortnightly) clinics (41%), although almost half was provided as part of normal surgery,

with 26.7% being offered on a daily basis within surgery time (Table 18).

Overall, 17.4% of the instances of provision are paid for entirely or in part by the

patients. This appears to be more common if the provision is for manipulative

therapies, or “other” therapies of which aromatherapy and massage predominated. In

contrast, homoeopathy provision is almost entirely free to NHS patients (Table 19).

Table 17 Provision within the practice by type of practitioner and therapy offered

Therapy Provider

GP Practice Nurse

Other Independent therapist

All n %

Acupuncture 831 1 142 13 111 (36.8)

Chiropractic 4 - 1 5 10 (3.3)

Homoeopathy 49 1 1 6 57 (18.9)

Hypnotherapy 39 - 33 8 50 (16.6)

Med. Herbalism

4 - - 3 7 (2.3)

Osteopathy 11 - - 21 32 (10.6)

‘Other’ therapies4

6 2 6 21 35 (11.6)

All 196 (64.2%)

4 (1.3%)

25 (8.3%)

77 (18.5%)

3025 (100.0%)

1 In 2 cases this was provided with a practice nurse/physiotherapist 2 13 out of 14 were physiotherapists 3 2 out of 3 were community psychiatric nurses 4 Mostly aromatherapy and massage 5 Instances of provision

Table 18 Mode of provision within practice by therapy Therapy Provision

By regular clinic n (%)

In surgery daily n (%)

In surgery - ad hoc n (%)

By appointment only n (%)

Total (n = 100%)

Acupuncture 45 (42.1) 27 (25.2) 22 (20.6) 13 (12.1) 107

Chiropractic 4 (44.4) 3 (33.3) - 2 (22.2) 9

Homoeopathy 9 (17.0) 29 (54.7) 14 (26.4) 1 (1.9) 53

Hypnotherapy 19 (39.6) 4 (8.3) 14 (29.2) 11 (22.9) 48

Med. Herbalism 3 (50.0) 1 (16.7) 2 (33.3) - 6

Osteopathy 17 4 (54.8) 8 (25.8) (12.9) 2 (6.5) 31

‘Other’ therapies1 19 (61.3) 4 (12.9) 6 (19.4) 2 (6.5) 31

All 116 (40.8%)

76 (26.7%)

62 (21.8%)

31 (10.9%)

285 (100%)2

1 Mostly aromatherapy and massage 2 NK = 17

Table 19 Who pays for complementary therapies provided within general practice by therapy

Therapy Free on NHS

n (%)

Patient pays

n (%)

Mixture of both

n (%)

Other 1

n (%)

All

(n = 100%)

Acupuncture 87 (79.8) 11 (10.1) 10 (9.2) 1 (0.9) 109

Chiropractic 5 (55.5) 3 (33.3) - 1 (11.1) 9

Homoeopathy 47 (85.5) 3 (5.5) 5 (9.1) - 55

Hypnotherapy 42 (84.0) 5 (10.0) 3 (6.0) - 50

Med. Herbalism 5 (83.3) 1 (16.7) - - 6

Osteopathy 17 (54.8) 12 (38.7) 2 (6.5) - 31

‘Other’ therapies 14 (42.4) 16 (48.5) 2 (6.1) 1 (3.0) 33

All 1 217 (74.1%)

51 (17.4%)

22 (7.5%)

3 (1.0%)

293 2 (100.0%)

1 Donation, local business, patients’ association 2 NK = 9

Complementary therapies provided within the practice and paid for by the patient are

mostly provided by ‘independent’ therapists (42/51). GPs reported charging patients

for complementary therapies on nine occasions. However, independent therapists do

provide care free of charge to NHS patients where their post is funded by a Health

Authority or purchased with GP fund-holding moneys (Table 20). Independent

therapists are cited as providing osteopathy most frequently (21/78 instances of

provision), followed by acupuncture (14/78). “Other” therapies account for 21/78

instances of provision by ‘independent’ therapists, aromatherapy (5), reflexology (5),

massage (5), Alexander technique (3), relaxation (2) and spiritual healing (1).

Table 20 also shows that there were 22 occasions when a fund-holding practice used

practice funds to purchase complementary therapies for patients within their practice.

These 22 ‘occasions’ relate to 20 practices, or 12% (20/161) of all fund-holding

practices in the sample. On 21 occasions FHSA or DHA moneys were cited as the

source of funding (most frequently FHSA special development money). These 21

‘occasions’ relate to 19 practices, or 2.5% of all practices surveyed.

GPs were asked to indicate if the therapy provided was directed to a particular

condition or group of patients. The majority of GPs answered this negatively,

indicating that the therapies were provided for a range of conditions. Figure 1 lists the

conditions mentioned.

Table 20 If NHS provision, source of funding for complementary therapies1 in general practice by type of practitioner

Practitioner FHSA/DHA

n (%)

Practice-Fund-holding

moneys n (%)

Practice- non-Fund-

holding n (%)

No Costs

n (%)

NK n

GP/primary health care team

8 13 22 117 37

Independent therapist

13 9 0 3 7

All 21 22 22 120 44

% of known (n = 184)

(11.4) (12.0) (12.0) (65.2)

1 All, including ‘other’ therapies

Figure 1 Groups of patients or conditions mentioned by GPs treated by complementary therapies within the practice

Acupuncture: Smokers Back pain Joint pain Other pain Acute stress Migraine Chiropractic: Back pain Joint pain Homoeopathy: Depression Migraine Diabetes Pain Warts Hypnotherapy: Smokers Over 75’s Anxiety Psychological problems Acute stress Dental extraction Obesity Medical Herbalism: None given Osteopathy: Back pain Joint pain

4.43 NHS referrals for complementary therapies

One hundred and sixty respondents reported that their practice made any referral to

NHS funded provision for treatment with a complementary therapy. A total of 227

instances of such activity were described. The scale of this provision, in terms of the

number of patients affected, was not easy to ascertain from the questionnaire, but

those GPs offering an estimate of the number of such refrrals made by themselves in

one month indicated a range between one and five; one referral per month was the

most commonly cited frequency.

NHS hospitals, excluding homoeopathic hospitals, make up 40.1% of these instances

of referral, the majority of which are for acupuncture. Referral for homoeopathy was

the most common. Most of this activity related to NHS homoeopathic hospitals,

although a proportion (13%) were to ordinary NHS hospitals, and a similar proportion

entailed referral to care located in the private sector. Where the treatment was for

osteopathy, referral for treatment in the private sector was more common than referral

to an NHS location. (Table 21)

Funding for this type of referral appears to come largely from District Health

Authorities, although a significant proportion of this activity (27.5%) was funded

directly by the GP practices (Table 22). GP fund-holding practices reported 23

instances of referral for complementary therapies. This involved 14 practices, 8.7% of

all fund-holding practices in the survey.

Table 21 NHS Referrals outside the practice for complementary therapies by place of reference

Place of Referral

NHS Hospital

NHS homoeopathic hospital

Private Clinic or Consulting rooms

Other GP surgery

Other or not known

All

Therapy n n n n n n (%)

Acupuncture 52 3 5 3 5 68 (30.0)

Chiropractic 6 0 6 0 3 15 (6.6)

Homoeopathy 12 65 12 4 14 95 (41.9)

Hypnotherapy 8 0 3 2 2 15 (6.6)

Medical Herbalism

1 1 0 1 1 4 (1.9)

Osteopathy 9 0 13 1 2 25 (11.0)

Other therapies

3 0 1 0 1 5 (2.2)

Total 91

(40.1)

69

(30.4)

40

(17.6)

11

(4.8)

28

(12.3)

227

(100.0)

Table 22 NHS referrals outside the practice for complementary therapies by source of funding

Source of funding1

Therapy Fund-holding savings

Practice budget

FHSA DHA Other NHS All

Acupuncture 6 11 7 39 3 66

Chiropractic 3 2 2 5 1 13

Homoeopathy 9 15 21 41 1 87

Hypnotherapy 0 1 3 7 2 13

Medical Herbalism

0 1 0 1 1 3

Osteopathy 4 2 3 2 0 11

Other therapies 1 2 0 2 0 5

Total 23

(11.6)

34

(17.2)

36

(18.2)

97

(49.0)

8

(4.0)

198

(100.0)

1 NK source of funding = 29

GP BEHAVIOUR IN CONSULTATIONS IN THE PAST WEEK

Responding GPs completed a questionnaire matrix in which they gave the number of

occasions in consultations in the past week when they personally treated patients with

a complementary therapy, referred them for treatment (private or NHS) or

recommended/endorsed such treatment. Data on recommendations and

endorsements is perhaps more a measure of the GP’s opinion about complementary

therapy than an indication of the patient having received it. However, this category

serves to help distinguish between more formal ‘referrals’ and actions falling short of

this, i.e. endorsements of patients’ decisions to seek help from complementary

therapies, and thus makes the data on referrals more reliable. In the question matrix,

each of these three categories is applied to the listed therapies, and a category for

‘other’ (explicitly excluding psychotherapy and counselling).

A similar matrix was completed showing the number of occasions in the past week

when a consultation resulted in a ‘neutral’ response with respect to a particular

therapy or where the GP advised a patient against the use of a particular therapy on

that occasion. Neutral responses were exemplified on the questionnaire by the

phrase “It’s up to you”.

These data were weighted according to the partnership size of the GP and their

response category.

GP behaviour in relation to different therapies

Table 23 shows the weighted number of GPs treating patients with each of the listed

complementary therapies, referring a patient (NHS or privately) for such treatment or

recommending/endorsing such treatment. The distribution of these interventions

across the therapies suggests that osteopathy is the therapy most frequently

associated with treatment, referral or endorsement, followed by chiropractic,

Table 23: Complementary therapies in consultations in one week: a) the number of GPs treating patients with

complementary therapies, referring for such therapies, or recommending/endorsing treatments; b) the number of occasions each action was performed; and c) the average weekly interventions per GP reporting the action, d) the average weekly interventions per GP in England by therapy (weighted data)

N=1226 Treatments1 Referrals2 Recommend/ endorsements

Acupuncture GPs 62 57 165 Interventions3 310 68 188 Av.’active’ GP

Av. All GPs

5.0 0.25

1.2 0.05

1.1 0.15

Chiropractic GPs 8 94 249 Interventions3 13 119 338 Av.’active’ GP

Av. All GPs

1.6 0.01

1.3 0.09

1.4 0.27

Homoeopathy GPs 49 46 143 Interventions3 260 49 181 Av.’active’ GP

Av. All GPs

5.3 0.21

1.1 0.03

1.3 0.14

Hypnotherapy GPs 20 18 98 Interventions3 19 33 118 Av.’active’ GP

Av. All GPs

1.0 0.01

1.3 0.02

1.2 0.09

Medical Herbalism GPs 4 4 35 Interventions3 4 5 51 Av.’active’ GP

Av. All GPs

1.1 <0.00

1.3 <0.00

1.5 0.04

Osteopathy GPs 27 131 334 Interventions3 80 217 465 Av.’active’ GP

Av. All GPs

3.0 0.06

1.7 0.17

1.4 0.37

Reflexology GPs 0 8 30 Interventions3 0 12 32 Av.’active’ GP

Av. All GPs

0 0

1.5 <0.00

1.1 0.02

Aromatherapy GPs 1 8 54 Interventions3 1 9 58 Av.’active’ GP

Av. All GPs

1.0 <0.00

1.1 <0.00

1.10.04

1 Treated by responding GP 2 Referrals to NHS or private provision within and outside the practice 3 Probably equivalent to individual patients as all actions within a one week period

Table 24: Complementary therapies in consultations in a one week period a) number of GPs giving a neutral response to a patient enquiry

about complementary therapies or advising against their use; b) number of occasions each action was performed and c) the average interventons per GP reporting the action d) the average per GP in England, by therapy (weighted data)

N=1226 Neutral

response1 Advised against

Ratio of neutral responses to

advice against Acupuncture GPs 134 3 51:1 Interventions2 154 3 Av.’active’ GP

Av. All GPs

1.1 0.12

1.0 <0.00

Chiropractic GPs 146 28 8:1 Interventions2 186 28 Av.’active’ GP

Av. All GPs

1.3 0.05

1.0 0.02

Homoeopathy GPs 153 11 12:1 Interventions2 185 16 Av.’active’ GP

Av. All GPs

1.2 0.05

1.5 0.01

Hypnotherapy GPs 66 5 14:1 Interventions2 68 5 Av.’active’ GP

Av. All GPs 1.0

0.05 1.0

<0.00

Medical

GPs

54

20

3:1

Herbalism Interventions2 64 20 Av.’active’ GP

Av. All GPs

1.2 0.05

1.0 0.01

Osteopathy GPs 157 30 7:1 Interventions2 215 30 Av.’active’ GP

Av. All GPs

1.4 0.17

1.0 0.02

Reflexology GPs 27 9 3:1 Interventions2 29 9 Av.’active’ GP

Av. All GPs

1.1 0.02

1.0 <0.00

Aromatherapy GPs 39 4 12:1 Interventions2 48 4 Av.’active’ GP

Av. All GPs

1.2 0.03

1.0 <0.00

1 e.g. “Its up to you” 2 Probably equivalent to individual patients as all actions within a one week period

acupuncture and homoeopathy. GP treatments are more common for acupuncture

and homoeopathy, while referrals are more common for the manipulative therapies,

especially osteopathy.

Table 23 also gives the average number of interventions per GP reporting the action.

For example, GPs reporting the use of acupuncture undertook an average of 5.3

treatments in the week. For each therapy, GPs treating patients themselves reported

the highest number of interventions per action, whereas the greatest volume of activity

overall is associated with recommendations or endorsements.

As specific levels of neutral and negative responses reported for each therapy depend

on the number of occasions when the opportunity to offer either of these responses

arose, it may not be appropriate to make direct comparisons between therapies for

each of these responses. However, it is possible to gauge the relative balance of

neutral to negative response for each therapy. Thus Table 24 suggests that GPs may

view acupuncture differently from the other therapies, in that recommendations

against its use form a much smaller proportion of all non-positive response compared

to the other therapies. These ratios are, however, derived from small numbers and

these findings should be viewed as indicative rather than definitive.

Differentials in behaviour by characteristics of GPs

The data were analysed to test whether any of the known characteristics of the

responding GPs were associated with different behaviours relating to complementary

therapies in consultations in the past week. Female GPs appear to be more likely to

endorse or recommend than treat or refer patients, and age of GP seems to be

negatively related to endorsement. GPs who were identified as senior partners on the

FHSA lists were more likely to have treated patients with complementary therapies in

the past week than single-handed GPs or partners from group practices (Table 25).

The numbers in all these sub-groups are relatively small and none of these

differences reaches statistical significance. No significant differences were observed

Table 25: Complementary therapies in consultations in the last week: the

estimated number and percentage of GPs treating, referring or endorsing treatment by age group, sex and status of GP (weighted data)

Age of GP Treatments n (%)

Referrals n (%)

Recommend/ endorsements n (%)

All N

under 35 10 (6.6) 35 (22.7) 96 (61.9) 155

35-44 41 ( 11.7) 82 (23.4) 180 (51.6) 349

45-54 36 (12.7) 70 (24.4) 109 (38.2) 285

55+ 10 (5.6) 29 (16.6) 65 (37.6) 172

age not known 30 (11.4) 46 (17.3) 100 (37.8) 264 Female 28 (9.7) 60 (20.9) 172 (59.6) 289 Male 67 (10.2) 150 (22.7) 272 (41.0) 663 sex not known 31 (11.5) 50 (18.3) 105 (38.4) 274 Senior Partner 47 (13.3) 65 (18.5) 133 (37.9) 352 Single Handed 18 (10.7) 33 (18.8) 59 (33.3) 174 Other 62 (8.8) 163 (23.3) 358 (51.0) 700 All 127 (10.4) 261 (21.3) 550 (44.8) 1226

Table 26: Complementary therapies on consultations in the past week: the number and percentage of GPs treating, referring or endorsing treatment by location of practice (weighted data)

Location of practice

Treatments

n (%)

Referrals

n (%)

Recommend/ endorsements

n (%)

All

N

Inner city 17 (11.5) 27 (18.3) 36 (24.9) 146 Rural 35 (12.1) 62 (21.9) 156 (54.7) 285 Else 75 (9.5) 172 (21.8) 358 (45.3) 790

All 127 (10.4) 261 (21.3) 550 (44.8) 12211

1 NK = 5

between these groups with respect to negative or neutral responses. However, GPs

located in inner city areas were less significantly likely to recommend or endorse

treatments with complementary therapies (Table 26). This may be due to a higher

prevalence of patients in these areas who are unable to meet the costs of private

provision.

Estimates of individual GP activity

GP treatments and NHS referrals both have clear resource implications with respect

to GP time or NHS facilities (e.g. a homoeopathic hospital referral may entail

approved extra contractual referrals). A recommendation or endorsement is resource-

free, but may still be constrained by access issues (i.e. available practitioners in the

area of suitable calibre). Estimates of the percentage of GPs undertaking treatments,

making referrals or endorsing/recommending treatments are given in Table 27. These

are based on the assumption that non-responding GPs were most similar to those

GPs who responded after the third mailing. Lowest estimate are calculated assuming

that all non-responders were non-active in the past week are reported in the text.

An estimated 10.4% (lowest estimate 8%) of GPs treat patients with one of the listed

complementary therapies in a week, 21.3% (lowest estimate 13.5%) of GPs refer a

patient for complementary medicine treatment (NHS or private) and an estimated

44.8% (lowest estimate 27%) of GPs recommend or endorse treatment (Table 27). In

contrast, an estimated 31.2% (lowest estimate 20.6%) of GPs give a neutral response

in the past week, and 7.4% (lowest estimate 4.4%) give specific advice against the

use of one of the named complementary therapies. (Table 28)

Of the individual therapies, acupuncture was the most frequently reported treatment,

offered by an estimated 5% of GPs (lowest estimate 3.8%). Referrals for osteopathy

were made by an estimated 10.7% of GPs (lowest estimate 7.1% ). Treatments or

referrals for homoeopathy were made by an estimated 7.2%of GPs (lowest estimate

Table 27 Estimated proportion of GPs treating, referring or endorsing complementary therapies in one week by therapy (weighted data)

Treatments1 Referrals2 Treatments or Recommend/ Any of these N=1226 referrals endorsements

n % n % n % n % n % Acupuncture 62 (5.0) 57 (4.6) 118 9.6 165 (13.5) 266 21.7 Chiropractic 8 (0.7) 94 (7.6) 101 8.2 249 (20.3) 335 27.3 Homoeopathy 49 (4.0) 46 (3.7) 92 7.5 143 (11.6) 208 17.0 Hypnotherapy 20 (1.6) 18 (1.5) 38 3.1 98 (8.0) 130 10.6 Medical Herbalism

4 (0.3) 4 (0.3) 8 0.7 35 (2.8) 40 3.3

Osteopathy 27 (2.2) 131 (10.7) 149 12.2 334 (27.3) 451 36.8 Reflexology 0 (0.0) 8 (0.6) 8 0.7 30 (2.5) 38 3.1 Aromatherapy 1 (0.1) 8 (0.7) 8 0.7 54 (4.4) 61 5.0

Any of these therapies

127 (10.4) 261 (21.3) 337 26.4 550 (44.8) 729 57.1

1 Treated by responding GP 2 Referrals to NHS or private provision within and outside the practice

Table 28 Estimated proportion of GPs giving neutral or negative response to

enquiry about complementary therapies in one week by therapy (weighted data)

N = 1226 Neutral

n (%) Advice against

n (%)

Acupuncture 134 (10.9) 3 (0.2)

Chiropractic 146 (11.9) 28 (2.3)

Homoeopathy 153 (12.5) 11 (0.9)

Hypnotherapy 66 (5.4) 5 (0.4)

Medical Herbalism 54 (4.4) 20 (1.7)

Osteopathy 157 (12.9) 30 (2.4)

Reflexology 27 (2.2) 9 (0.7)

Aromatherapy 39 (3.2) 4 (0.4)

Any of these therapies 399 32.5 94 7.7

6.2%) (Table 27). All GPs offering aromatherapy or reflexology also offered at least

one of the six established therapies.

The extent to which GPs are able to report the outcome of discussing a

complementary therapy with a patient is strongly related to the number of occasions

on which patients raise this issue in consultations. GPs with a reputation for being

open to the possibility of the use of alternative therapies are more likely to be

approached by patients than those not known to hold such views. Thus a higher

proportion of those treating, referring or recommending in the past week also gave a

neutral or negative response in the same period compared with those not reporting

such actions. (Table 29).

Using the estimates based on the assumption that non-responders were most similar

to responders to the third mailing, estimates have been calculated for the level of

activity by GPs in England in an average week of GP consultations (Table 30). Thus it

is estimated that 14,900 treatments with acupuncture, chiropractic, osteopathy,

homoeopathy, hypnotherapy and medical herbalism (95% CI 12,500 to 17,300) are

performed per week by GPs, and 10,800 referrals are made, (95% CI 9,650 to

11,900). An estimated 29,600 recommendations or endorsements are made (95% CI

27,800 to 31,400). On the assumption that these estimates are representative of an

average week, this gives crude annual estimates of 750,000 treatments, and half a

million referrals.

Table 29: Neutral responses and advice against the use of complementary therapy by postive actions reported using weighted data

Neutral response

Advised against

N

n (%) n (%) Treatments:

Yes 49 (38.3%) 10 (7.6%) 127

No 350 (31.8%) 84 (7.7%) 1099

Referrals:

Yes 123 (47.1%) 44 (16.7%) 261

No 276 (28.6%) 50 (5.2%) 965

Recommend/ endorsements:

Yes 216 (39.3%) 42 (7.7%) 550

No 183 (27.0%) 51 (7.6%) 676

Total 399 (31.2%) 94 (7.4%) 1226

Table 30 National estimates of treatment, referral and recommendation/endorsement of established complementary therapies1 in GP consultations in an average week n2 Estimated Average Estimated number 95% CI for estimated number of interventions per of interventions4 interventions/week active GPs3 active GPTreatments 127 2733 5.45 14905 (12507, 17302)Referrals 261 5617 1.92 10789 (9655, 11924)Treatment or referral 337 7253 3.54 25694 (23412, 27976) Recommend / endorsements 550 11838 2.50 29590 (27796, 31383) Any of these 729 15690 3.52 55284 (52787, 57780) Acupuncture: Treatments 62 1334 5.00 6675 (5094, 8257) Referrals 57 1227 1.20 1472 (1108, 1837)Chiropractic: Treatments 8 172 1.61 277 (90, 464) Referrals 94 2023 1.26 2559 (2073, 3045)Homeopathy: Treatments 49 1055 5.31 5597 (4097, 7097) Referrals 46 990 1.06 1045 (755,1334)Hypnotherapy: Treatments 20 430 1.36 585 (336, 833) Referrals 18 387 1.25 483 (267, 700)Medical Treatments 4 86 1.19 102 (5, 200)Herbalism: Referrals 4 86 1.20 103 (5, 202)Osteopathy: Treatments 27 581 2.97 1727 (1098, 2357) Referrals 131 2819 1.66 4671 (3932, 5409) 1 Acupuncture, chiropractic, homoeopathy, hypnotherapy, medical herbalism and osteopathy only 2 N = 1226 sample size = 760 3 GP population of England = 26387 unrestricted principals: source GMS Statistics 1 April 1994 4 Probably equivalent to ‘patients’ as data relates to one week period

DISCUSSION

A recent national survey of GP attitudes conducted on behalf of the GMSC suggests

that, assuming adequate resources, between 18% and 29% of GPs currently favour

the expansion of GP services to cover such treatments as homoeopathy, chiropractic,

hypnotherapy, osteopathy, and acupuncture respectively. However, even given

increased resources, the provision of such services would clearly have to compete

with the expansion of GP services in more established areas such as physiotherapy

and chiropody, which may have higher priority for the majority of GPs.

Some information about the availability of complementary therapies in general

practice is available from a number of case-studies of experiments in the provision of

such treatments(19,20,21), but little is known about the scale and scope of provision

nationally. This study was designed to offer a ‘snapshot’ of the current situation using

a large random sample of GP partnerships in England (one in eight). There is always

a restriction of the type of information which it is possible to obtain via a postal

questionnaire, and the range of provision possible in terms of mode of delivery and

therapy offered made the questionnaire design quite complex. As a result of a local

pilot study, a questionnaire was developed which was relatively long (10 pages) but

easy to complete with easily identifiable sections for each possible therapy. Research

suggests that the sponsoring body for the research, the target population, the length

of the survey, the salience of the issue to responders and the number of times contact

is made are the key factors in influencing final response rates(22), however postal

surveys involving GPs tend to get relatively low response rates due to the heavy

workload of GPs and the fact that such requests tend to be allocated low priority by

GPs. As one GP commented in declining to participate in the study:

“I very much regret that I am at the present time totally overwhelmed and exhausted by my basic workload. I just do not have any spare capacity to undertake any work that is not directly necessary under my general medical services commitment. I can assure you that I am far from happy about this situation and wish you well with your research.”

The final response of 62% was not, therefore, unexpected. Responders were not

systematically biased according to available characteristics, but it was felt to be

important to investigate non-responders further in order to establish the

representativeness of responders with respect to the provision of complementary

therapies within the practice. The short letter, containing the three key “screening”

questions, was returned completed by almost half the non-responders. Given the time

lapse of 9 weeks some of these may not have remembered receiving three

questionnaires previously. However, it seems likely that the majority simply

responded to a request which could be completed in minutes and that, for this group

of respondents, the length of the questionnaire was a key factor in the response rates

achieved. The responses given to the screening questions provide strong evidence

that the 760 responding practices are representative of the population as a whole with

respect to the provision of complementary therapies.

With regard to the provision of complementary therapies via the general practice as a

whole, the data on 760 representative practices shows that such provision is

widespread but currently offered in a relatively small scale in terms of the patient

numbers involved. We estimate that 39.5% of all practices (95% CI 35%-43%) now

provide access to complementary therapies for their NHS patients.

The most frequently cited type of provision is NHS referral, reported by one in four

practices, mostly for homoeopathy or acupuncture at NHS hospitals. However, this

type of provision may influence the management of patients less than the availability

of treatment within the practice, which has the potential to affect a larger number of

individual patients.

Provision within the practice by a member of the primary health care team is relatively

common, affecting an estimated one in five practices. The employment of an

independent therapist is relatively rare, occurring in an estimated 6.1% of practices.

This is perhaps to be expected, as it is the least easily established form of provision,

but there may be considerable scope for growth in this area.

The study sample structure makes it hard to analyse these data by geographical area.

However, the national results are similar to those reported in a recent survey of 87 GP

practices in S.W. Thames(24). More local area studies are needed to establish any

regional variations in the pattern of provision. In contrast, the study findings do not

support the finding in the NAHAT survey that 14% of GP fund-holding practices

employ an independent practitioner (6). The level of such provision in this study was

similar for fund-holders and non-fund-holders (6.7% v 6.1% respectively). The level of

provision reported by the NAHAT survey is likely to be inflated by bias resulting from

their low response rates and the non-random selection of practices. Fund-holding

status was, however, associated with a significantly higher level of provision from the

primary health care team (27% v 21%) and a higher overall level of access (45.5%

compared with 36.6% respectively, Chi square P <0.01).

The practice data reflect the data on individual GP behaviour in so far as primary

health care team provision is dominated by GPs offering acupuncture and

homoeopathy but access via referrals is also most commonly associated with these

therapies in the practice-based data, with provision for referrals to manipulative

therapies being much less common. This reflects the change in emphasis from all

referrals (NHS and private) to those undertaken on the NHS only, and highlights a gap

between GP referring behaviour and NHS provision.

Reported funding for access to complementary therapies via primary care suggests

that patients are paying for a significant proportion of the treatments provided within

the practice, as well as for private referrals. This will inevitably lead to an uneven

distribution of provision and access between practices located in areas where patients

can afford to pay and those in areas where this is not an option.

Where the NHS is funding provision, the majority of provision within practices is

accomplished by absorbing costs into the practice. A significant proportion of fund-

holding practices surveyed, 22/161, reported using identified practice funds to cover

the costs of in-house provision. Just over 10% of all such provision was reported as

being funded directly by the FHSA or DHA, usually via special development moneys.

NHS referrals were largely reported as being funded by the DHA (51%) or FHSA

(17%). A smaller proportion, 11%, of referral schemes mentioned were reported as

being purchased with fund-holding savings (23 instances affecting 19 practices or

11.8% of all fundholding practices).

The study design provided a sample of 1226 GPs in England, covering single-handed

GPs, senior partners and other partners, from whom information was sought about the

previous week’s consultations. These data were weighted to take into account the

high proportion of single handed GPs in the sample. The most frequently reported

action was a recommendation or endorsement of complementary therapy treatment,

an estimated 45% of GPs. Such actions do not require any direct resources and may

be viewed as an indication of support for these therapies in a situation in which no

NHS provision is available via the GP. However, recommendations and

endorsements are likely to be influenced by two factors external to the GP; firstly, the

extent to which the GP is perceived to be sympathetic to enquiries about

complementary therapies and, secondly, the extent to which GPs believe there to be

provision in the private sector which is of a sufficient standard and which can be

afforded by patients. In this context, it is interesting to note that

recommendations/endorsements were less frequently reported by GPs working in

practices located largely in inner city areas. If recommendations/endorsements are

seen as indicators of unmet expressed demand for NHS complementary therapy

treatment, the additional referrals would be in the region of 25,000 per week.

However, this is likely to be an under-estimate of the potential demand as it is based

on a public perception of a lack of NHS provision(2) .

Recommendations and endorsements were explicitly distinguished from referrals in

order to strengthen the latter category. Even after allowing for this distinction, one in

five GPs (21%) reported making what they considered to be a referral for a

complementary therapy in the past week. Referrals (private and NHS) were most

frequently reported for manipulative therapies, especially osteopathy. A smaller

proportion of GPs (10.4%) reported having treated a patient with a complementary

therapy in the past week, and these treatments were most likely to be for acupuncture

or homoeopathy. This group of ‘active’ GPs made an average of 4 treatments each in

the week surveyed. Using these data, it is estimated that 2733 GPs in England are

active and perform a total of 14,900 complementary therapy treatments in an average

week (95% CI 12,500-17,300). Similarly it is estimated that 10,800 NHS referrals are

made (95% CI 9,650-11,900). However, to put these estimates into context,

assuming that the 14,900 treatments per week each represent a single patient

episode, this is equivalent to approximately 5 per 1,000 patients consulting their GP

(23) .

GPs were also asked about occasions on which they responded neutrally (e.g. “it’s up

to you”) or negatively to a patient enquiry about a complementary therapy in the past

week. GPs reported far more neutral than negative responses and GPs treating,

referring, recommending or endorsing were more likely than those not undertaking

any of these actions to report also having given a neutral or negative response. This

indicates a selective support of complementary therapies by GPs and reflects likely

levels of exposure to enquiries, i.e. some GPs will have a reputation for being

sympathetic. The fact that an estimated 31% GPs give a neutral response in an

average week suggests that GPs frequently feel unable to advise their patients in this

respect, presumably in part through a lack of appropriate knowledge or information

regarding the therapies. As the popularity of complementary therapies grows, there is

a corresponding need for GPs to have access to more information or training

regarding their possible benefits and disbenefits to patients. This need has been

expressed by GPs themselves in recent studies (6,24) .

In conclusion, this study has demonstrated that complementary therapy provision is

widespread in English general practices although the average level of provision to

patients within individual practices remains low. This suggests that complementary

therapies are acceptable to a growing proportion of GPs, and that the level of activity

reported reflects a stage in the development of the provision of complementary

therapies in general practice, rather than current demand by patients.

ACKNOWLEDGEMENTS

The authors wish to thank the FHSAs for their help and co-operation and to express

their gratitude to all the GPs who somehow found the time to complete our

questionnaire and return it to us. We would also like to thank Gwyneth Askham and

Sheila Bray for their contributions to the study and to the production of the report. The

Medical Care Research Unit is funded by the Department of Health. The opinions

expressed in this report are, however, those of the authors alone.

References

1. Smelskyj A. The qualifications and geographical distribution of practising osteopaths in England, Scotland and Wales. Complementary Medical Research, 1992; 6: 1-8.

2. Thomas KJ, Fall M et al. A methodological study to investigate the feasibility of conducting a population-based survey of the use of complementary health care. Final Report to the Research Council for Complementary Medicine. 1993.

3. BMA. Alternative Therapies (BMA, London, 1986).

4. BMA. Complementary Medicine (BMA, London, 1993).

5. Department of Health. Press Release H91/600. Stephen Dorrell clarifies the position on alternative and complementary therapies. December 1991.

6. Cameron-Blackie G. Complementary Therapies in the NHS. Birmingham. National Association of Health Authorities and Trusts. 1993.

7. Calnan M, Cant S, Williams S and Kellovan A. Involvement of the primary health care team in coronary heart disease prevention. BJGP, 1994; 44: 224-228.

8. Chew C A, Wilkin D and Glendenning C. Annual assessment of patients aged 75 years & over: GPs’ & practice nurses’ views and experiences. BJGP, 1994; 44: 263-267.

9. Sibbald B, Addington-Hall J, Prenneman D and Freeling P. Telephone & postal surveys of GPs: methodological considerations. BJGP, 1994; 44: 297-300.

10. Robinson G, Beaton S and White P. Attitudes towards practice nurses - survey of a sample of GPs in England & Wales. BJGP, 1993; 25-29.

11. Hallam L. Patient access to GPs by telephone: the doctor’s view. BJGP, 1992, May; 186-189.

12. Smith LFP. Provision of obstetric care by GPs in the SW region of England. BJGP, 1994; 44: 255-257.

13. Kerr MP. Anti depressant prescribing: a comparison between GPs & psychiatrists. BJGP, 1994; 44: 275-276.

14. Middleton J. Written lists in the consultation! Attitudes of GPs to lists and the patients who bring them. BJGP, 1994; 44: 309-310.

15. Pell JP and Alcock J. Monitoring anticoagulant control in general practices. BJGP, 1994; 357-358.

16. Hobbs FD. Fear of aggression at work among GPs who have suffered a previous episode of aggression. BJGP, 1994; 44: 390-394.

17. Petchey R. Exploratory study of GPs’ orientation to general practice & response to change. BJGP, 1994; 44: 551-555.

18. Whynes DK and Reed G. Fundholders’ referral patterns & perceptions of service quality in hospital provision of elective general surgery. BJGP, 1994; 44: 557-560.

19. Budd C, Fisher B, et al.. A model of co-operation between complementary and allopathic medicine in a primary care setting. BJGP, 1990; 40: 376-378.

20. Logan J. The Blackthorn Trust: expanding a national health practice. Complementary Therapies in Medicine, 1994; 2: 3; 154-158.

21. Burns K and Lyttleton LK. Osteopathy on the NHS; one practice’s experience Complementary Therapies in Medicine, 1994; 2: 4; 200-204.

22. Heberlein TA and Baumgartner R. Factors affecting response rates to mailed questionnaires: a quantitative analysis of the published literature. American Sociological Review, 1978; 43;4; 447-462.

23. Consulting population calculated as 50% of adult GP consultations in the past two weeks. Source: General Household Survey 1993. London HMSO. Population of England source: Annual Abstract of Statistics 1994. London HMSO.

24. Perkin MR, Pearcy RM and Fraser JS. A comparison of the attitudes shown by general practitioners, hospital doctors and medical students towards alternative medicine. Journal of the Royal Society of Medicine, 1994; 87: 523-525.