Jill Walton,
Clinical Effectiveness Co-ordinator
Graham
Henderson, Consultant in Public Health Medicine
Ross Lawrenson,
Consultant in Public Health Medicine
Correspondence
to Jill Walton please:
Abstract
Objective. To review the types of questions asked by general
practitioners in unobserved routine patient consultations.
Design. A descriptive study of prospectively
collected data.
Setting. East Surrey Health Authority area,
Subjects. General practitioners.
Results. Between October 1996 and March 2000, 500 questions were received by a dedicated evidence-based information service from 139 different practitioners. 26.6% of questions were specifically related to women, 13.2% to children and 3% to men. Questions about women’s health were mainly about pregnancy (particularly drug safety), hormone replacement therapy, contraception, breast cancer and cervical screening. The most common questions relating to children were about vaccination, asthma and otitis media. For men’s health, the most common question area was prostate cancer. The questions were grouped pragmatically into 14 categories that describe the type of questions asked. The largest categories covered drug treatments for specific diseases and other prescribing issues. Further significant clinical categories covered disease management, epidemiology or aetiology, general information or overview requests, risks and associations not related to drugs, differential diagnosis, alternative or complementary therapies and screening tests.
Conclusions. General
practitioners generate a very wide range of clinical questions during the
course of everyday consultations. These questions can be answered rapidly by a
small dedicated team who have search and appraisal skills. The majority of
questions were about drug treatment for specific conditions, supporting the
need for evidence-based information about therapies to be easily available to
general practitioners. Most questions as asked are not framed in a way that
lends itself to a classical evidence-based medicine approach, but they can be
categorised into 14 broad types. Analysis of the cumulative database of
questions can identify those clinical areas where knowledge is felt to be
lacking and where continuing medical education can be directed.
Introduction
The
information needed by general practitioners (GPs) to be able to provide up to
date evidence-based care for their patients is vast and increasing daily. In
1996, Richard Smith estimated that the body of medical knowledge increases
fourfold during a doctor’s lifetime [1].
Access to this knowledge base in general has also improved dramatically
in recent years, since its availability in electronic format means that sources
of information can be rapidly retrieved using a desk top computer. However, most general practitioners today
still have barriers to using this knowledge bank to answer everyday questions
concerning patient management in practice.
Apart from the lack of personal access to the range of electronic
databases, and the skills to navigate through them, these barriers include
managing the information overload [2] and the often rudimentary skills in
critical appraisal of scientific literature. The biggest barrier is finding the
time to go through the whole process [3][4][5]. Studies have shown that general practitioners
are more likely to turn to a text book or to a colleague for an answer than
they are to conduct an electronic search of the literature [6][7].
McColl [5]
concluded that general practitioners, in their quest to practising
evidence-based medicine, would prefer to have improved access to summaries of
evidence rather than to be taught the skills of searching and critical
appraisal for themselves. Ely et al [7] found that the
median time spent in general practice pursuing the answer to a question was I
minute, and they reviewed the evidence that busy doctors need bottom-line
answers that have been digested into quickly accessible summaries, framed in a
way that fits with their own perspective rather than that of researchers.
In order
to try to meet these information needs locally, East Surrey Health Authority
has been supporting a project initiated by local general practitioners called
Merlin GP Information. The aim is to provide general practitioners with
evidence-based answers to day-to-day queries that arise during consultations. The
service is primarily aimed at answering clinical questions, but in order to
encourage question-asking behaviour in general, the service also offers to
respond to more general questions about clinical practice. We do not restrict
the answering service to any particular type of question, nor do we restrict
the format in which questions are framed.
This project has now been running for over three years and we decided to
review the types of questions asked by general practitioners, and to gauge
their satisfaction with the service.
Methods
All practices within the health authority area were supplied with a pad on which general practitioners could write down their clinical questions as they arose during consultations. The service was promoted by the Medical Advisory Audit Group lead for the area (AH), primarily to general practitioners, although other primary health care team members have occasionally sent in questions. The questions are faxed or telephoned to the local GP Link Office in Epsom and logged in a database. A general practitioner and clinical manager reads all the questions and will return to the questioner for clarification if needed. The questions are then sent on to the health authority to a consultant in public health medicine who re-frames the question if necessary in order to focus the searches appropriately. Questions that cannot be answered immediately from medical textbooks, the British National Formulary, or other at-hand sources are passed on to a public health information specialist for a structured literature search. Complex queries about drug therapy or interactions are answered by the Health Authority’s pharmaceutical adviser, who may refer to the Regional Drug Advisory Service.
The search
strategy always looks for evidence that is classifiable according to the
standard evidence hierarchy used in evidence-based medicine [8], for example, a
systematic review of randomised controlled trials would rate more highly than
an observational study, which would rate more highly than a case series. Most
searches begin with the Cochrane Library, then work
through the other collections of research that has been subject to secondary
review by looking at the web sites for the main evaluation centres in the
The
consultant appraises the abstracts and papers and formulates a brief answer in
one or two paragraphs, with the key references as footnotes should the reader
wish to pursue the answer in more detail. The response is reviewed by the
general practitioner manager on the Merlin team before being sent back to the
enquirer. A database of the anonymised questions and
answers is kept at the GP Link Office in electronic format so that they can be
published on the local internet or the web in the future.
Between October 1996 and March 2000, 500 questions were logged and answers provided. We tried to analyse the question bank using an evidence-based medicine framework, with the questions classified into a question type, a population of interest, an intervention or exposure of interest and an outcome. However, few of the questions were constructed in a way that lent itself to such classification. We therefore used a different taxonomy in an attempt to categorise the questions into a framework more relevant to the questioners’ perspective in primary care. Because the breadth of the questions was so varied, this took a very pragmatic approach, a compromise between having an extensive basket of categories and wanting to draw broad conclusions for the purpose of highlighting areas of information gaps. The taxonomy placed all 500 questions into one of 14 categories. The questions were categorised by two of the authors independently, using agreed rules to distinguish between the 14 categories where possible overlap existed, and any differences were resolved by discussion.
Each
Merlin answer is returned to the questioner with a simple evaluation form and a
request to send it back completed to the GP Link Office. The form asks for a
view on the quality and speed of the answer and whether clinical practice is
likely to change as a result. Information from these forms was summarised.
There are
65 practices in
Considering specific populations, 66 (13.2%) questions were specifically related to children, 15 (3%) to men, and 133 (26.6%) to women. The questions about women’s health were mainly about pregnancy (41), hormone replacement therapy (21), contraception (20), breast cancer (12) and cervical screening (12). Common enquiries relating to pregnancy women were about the safety of drugs during pregnancy. The most common questions relating to children were about vaccination, asthma and otitis media. For men’s health, the most common question area was prostate cancer.
Table 1 shows the numbers of questions in each of the 14
categories. Three examples of questions under each category are given in
The next biggest category was about disease management where no specific therapy was mentioned (category 3), covering management regimes, referral protocols, monitoring, lifestyle advice, etc. Category 4 included questions on the genetic susceptibility to certain diseases, cause, incidence and prognosis. There were 32 questions about alternative or complementary therapies, 15 of which were about specific dietary supplements. Enquiries about screening tests fell into two categories – tests for those with and without current symptoms or disease. Together, these categories comprised 8.6% of the total question bank. Screening questions
addressed some common areas – cancers, diabetes, allergies (especially food and nut allergies), and circulatory diseases.
Fifty-seven
evaluation forms were returned during the time frame – a response rate of
11%. The forms were from 34 different
respondents, about a quarter of all the enquirers in total. All but four of the
evaluation forms were from general practitioners. The analysis showed a general
overall satisfaction with the service (Table 2).
The most
recent statistics on a consecutive batch of 204 Merlin questions answered in
1999 show that an average of six questions were answered per week, with a
median turnaround time of 8.5 days. The Merlin team is engaged in the enquiry
service for only about one day of the working week.
The
service provided by Merlin GP Information in the last four years has been used
by 56% of the general practitioners and nearly three-quarters of the practices
in the district. The majority of questions are about treatment, and support the
need for evidence-based information about therapies to be easily available to
general practitioners. There were very few questions on economic aspects of
management (category 13). This is interesting
in that one of the criticisms of evidence-based medicine was that it would be
used to manage costs rather than improve other areas of clinical practice [10].
Many of the questions relate to specific patient groups, particularly children
and women (including pregnant and peri- or post-menopausal women). Very few
questions were asked about the specific health problems of men.
According
to morbidity surveys in general practice, many more patients consult for
respiratory diseases than any other type of disorder [11]. We had only 17 questions (3.4%) that related
to the management of respiratory complaints. Presumably, general practitioners
are comfortable managing respiratory complaints, and it is in other clinical
areas that they need further information. The type of questions may to some
extent reflect the health and socioecomonic make up of the local district –
Few of the general practitioners in
encourage a more structured approach to framing the management question being faced, to allow the people providing the service to narrow their search more closely to the answer needed.
The
strengths of the Merlin GP Information service are felt to lie in good team
work between the appraiser and the searcher, who puts a quality filter on the
studies forwarded, and in the rapid transfer of
information electronically. The answers
provided are written in clear and concise language, from a clinical rather than
an academic research perspective, to meet the needs of the busy practitioner.
The user-friendliness of the answers was reflected in the satisfaction scores
in the evaluation (table 2).
A
particular strength of this study is that, unlike previously published work, it
analyses questions that arose in unobserved everyday clinical settings, where
practitioners were under no pressure to either generate or suppress questions.
The anonymity of the service means that general practitioners feel comfortable
in asking for information in areas where their knowledge is not up-to-date, and
through a local
intranet and paper answers others can share the same knowledge. Analysis of the
cumulative database of questions can identify those clinical areas where
knowledge is felt to be lacking and where repeat questions occur, for example
hormone replacement therapy, oral contraception, food allergies. These areas
can then be targeted for joint training or continuing medical education. The
‘hot topic’ list can also be used to inform other national evidence evaluation
agencies of the areas in which general practitioners need up-to-date
effectiveness information.
A limitation of the analysis lies in the possible ambiguity of question categorisation, resulting from the way in which questions are asked and framed. A few enquiries comprised two or three discrete questions rolled into in one; in these cases we categorised on the basis of the main perceived information need and the answer that had been given. We put questions about risk and associations other than those related to drugs in a separate category (category 6). There was occasionally a blurred boundary between this category and the epidemiology/aetiology category. For example, questions about risk factors for a disease were categorised in the latter (category 4), while the risks of certain interventions (like having breast implants, eating liquorice, skiing while pregnant) were put into category 6.
The
evaluation to date has been rather simplistic, and will suffer from the biases
inherent in a satisfaction survey that relies on postal returns. For example it
is not possible to answer a question if there is no evidence in the literature
to provide the answer. Those who replied mostly felt that the information
supplied was of good or adequate quality. A key issue is whether the answering
service makes a difference to patient outcome.
About 32% of doctors suggested that the answers would change greatly the
way they advised or treated patients, while a further 54% felt that their
practice would be changed to some extent. Further resources would be needed to
track questions through the system to see what changes in practice had in fact
occurred. The Merlin service provides an experiential learning opportunity that
is both self-initiated and relevant, and has immediate practical application.
There is a large investment nationally in the secondary appraisal of research studies, with several agencies conducting similar types of assessments over the last few years (e.g. Cochrane, NHS Centre for Reviews and Dissemination, the Health Technology Assessment programme) . However the questions that general practitioners face every day in their surgeries are seldom those being addressed by these evaluation teams and the answering time frame is far too long. Thus Merlin provides quick, short and easily comprehensible answers to pertinent questions. Currently each question takes approximately 4 hours to search for the evidence, retrieve the abstracts, critically appraise them and formulate a response. These are by necessity “rapid review” responses and must be seen in this light. Answers are dated for the record, but not updated unless the question is asked again.
Similar
schemes to the Merlin GP Information project are being piloted elsewhere in the
country and there is currently a group of people meeting to discuss these
issues and potential solutions under the umbrella organisation NOISE (the
National Organisation for Information Support for Effectiveness). This group is
attempting to share methodologies, develop quality standards, and share the
answers to frequently asked questions. There may be economies of scale to be
made in providing such services region-wide.
1)
Smith R. What clinical information do doctors
need? BMJ 1996;313:1062-1068.
2)
Noone J, Warren J, Brittain M. Information overload:
opportunities and challenges for the GP's desktop. Medinfo 1998;9(Pt
2):1287-91.
3)
Tomlin
Z. Humphrey C, Rogers S. General practitioners' perceptions of effective health
care. BMJ 1999;318:1532-5.
4)
Fahey
T. Applying the results of clinical trials to patients to general practice:
perceived problems, strengths, assumptions, and challenges for the future. British Journal of General Practice 1998;48(429):1173-8.
5) McColl A, Smith H, White P, Field J. General
practitioners’ perceptions of the route to evidence based medicine: a
questionnaire survey. BMJ 1998;316:361-365.
6)
7) Ely JW, Osheroff JA, Ebell MH, Bergus GR,
Levy BT, Chambliss ML, Evans ER. BMJ 1999;319:358-361.
8) Centre for Evidence–Based Medicine (http://163.1.212.5/docs/levels.html)
9) The TRIP data can be found at http://www.ceres.ac.uk/trip/.
10) Grahame-Smith D. Evidence based medicine: Socratic
dissent. BMJ 1995; 310: 1126-7
11) McCormick A, Fleming D, Charlton J.
Morbidity Statistics from General Practice. Fourth National Study, 1991 -1992.
HMSO
|
Category |
Type of question |
Number |
Per cent |
|
1 |
Drugs/treatments for
specific conditions |
71 |
14.2 |
|
2 |
Drug
dosage/toxicity/interactions |
71 |
14.2 |
|
3 |
Disease management |
69 |
13.8 |
|
4 |
Epidemiology/aetiology |
42 |
8.4 |
|
5 |
General
information/overview |
35 |
7.0 |
|
6 |
Risks/associations
(non-drugs) |
35 |
7.0 |
|
7 |
Differential diagnosis |
35 |
7.0 |
|
8 |
Alternative therapies |
32 |
6.4 |
|
9 |
Non-clinical |
31 |
6.2 |
|
10 |
Screening/testing
asymptomatic people |
26 |
5.2 |
|
11 |
Screening/testing
symptomatic people |
17 |
3.4 |
|
12 |
Vaccinations |
15 |
3.0 |
|
13 |
Cost-effectiveness/rationing |
12 |
2.4 |
|
14 |
Evidence base for
time-honoured practices |
9 |
1.8 |
|
|
Total |
500 |
100 |
1. Drugs/treatments for specific conditions
§ What is the evidence for the use of progesterone injections in the first trimester in a woman who has recurrent miscarriages?
§ Why do people get prickly heat and what is the correct treatment hierarchy: anti-histamines, steroids or paludrin?
§
Should patients over 75
years of age be offered secondary prevention with a statin if their cholesterol
is over 5, and their LDC over 3? What is the evidence in favour of this?
2. Drug
dosage/toxicity/interactions
§
What is the interaction between Adalat
(nifedipine) and grapefruit?
§
When prescribing GTN tablets the computer gives an
option between 500 and 300 microgram doses.
Which one should I use for standard angina?
§
Is there any contraindication to taking HRT with
varicose veins?
3. Disease management
§
What is the correct management for a ruptured biceps
muscle?
§
How do you treat erythrasma, and how effective is
the treatment?
§
Polycystic ovary syndrome - what is the
recommendation on long-term follow-up?
4. Epidemiology/aetiology
§
What are the odds of a child inheriting vitiligo
from its mother?
§
What proportion of people with a
sore throat are likely to have a bacterial and what proportion a viral
throat infection?
§
Is there any evidence that nutritional zinc
deficiency is linked to postnatal depression or anorexia nervosa?
5.
General information/overview
§
Information please on pseudo gouty arthropathy.
§
What is Syndrome X and how do you treat it?
§
How does quinine work pharmacologically on leg
cramps?
6. Risks/associations (non-drug)
§
Is there a relationship between helicobacter
infection and rosacea?
§
Is there any link between Down's syndrome and the
thyroid status of the mother whilst pregnant?
§
What are the possible side effects or hazards from
the regular use of photocopy machines?
7. Differential diagnosis
§
When an x-ray report says "osteopenia"
can you ignore it or does it correlate with osteoporosis? Does this result mean
that a bone scan is clinically indicated?
§
Can systemic lupus erythematosus present without a
raised ESR?
§
With a single episode of optic neuritis, what is
the risk of subsequently developing MS and what are the most likely other
causes?
8. Alternative therapies
§
Is there any evidence for the effectiveness of
magnetic therapy?
§
Is there an anti-inflammatory cream made from
chilli? (seen recently on TV)
§
Is there any evidence for the use of vitamin E and
folic acid in dementia?
9. Non-clinical/policy
§
How do we get a list of locums
available or temporary staff to cover holidays?
§
Where can we get a TENS machine on the NHS for a
patient with a Sutton address?
§
Brother and sister would like to know if they have
the same father by DNA testing. What is the procedure
and the cost?
10. Screening/testing asymptomatic
people
§
Patient aged 59 whose mother died of cancer of the
rectum aged 83. Does she have an
increased risk of developing rectal cancer and should she be screened?
§
All new patients here have a "New
Patient" medical which includes an examination of urine for albumen and
sugar. Is this a complete waste of time
in patients who are below 50, not diabetic and not complaining of anything?
§
Well man age 60 with no symptoms. Is there any
screening value in running a PSA test or in doing a digital rectal exam?
11. Screening/testing symptomatic
people
§
Is there a serology test for food allergies, and
is it evidence based?
§
38 yr old lady has relapsing
multiple sclerosis. Are there any investigations, e.g. MRI scan,
that would have predictive value for future relapses?
§
Is there a skin prick test for allergies,
particularly asthma?
12. Vaccinations
§
Is there a vaccine for genital herpes simplex?
§
When giving Hepatitis B vaccine does it make any
difference to the immune response where you site the infection (arm, buttock, thigh)?
§
Is the risk of side effects reduced by giving the
components of the MMR vaccine separately?
13. Cost-effectiveness/rationing
§ Is evista more effective at restoring bone mass in known osteoporosis than alendronate? Which is more cost-effective and which more clinically effective?
§
What is the therapeutic and cost difference
between quinine bisulphate and quinine sulphate?
§
In patients with ischaemic heart disease who are allergic to aspirin, is it justifiable to use
Clopidogrel as a replacement agent for secondary prevention, bearing in mind
that it is far more expensive?
14. Evidence base for time-honoured
practices
§
Is there any evidence that bimanual pelvic
examination at the time of cervical smear testing is of any value?
§
What is the validity of clinical breast
examination as a screening process?
§
What medical evidence exists, if any, to support
the routine 6 week post-natal check?
Table 2. GPs’ evaluation
of the Merlin GP Information service (n=57)
|
Yes No |
|
|
Did the reply answer your
question? |
|
|
Fully Partially Not at all |
|
|
Will this information change the
way you advise/treat patients? |
|
|
Greatly A little Not at all |
31 8 |
|
Was the speed of the answer..? |
|
|
About right Too slow (no response) |
|
|
What were your views on the
quality of information received? |
|
|
Very good Adequate Poor |
|
*Department
of Public Health
Epsom,
**Anne
Hollings, General practitioner and GP Link Office Clinical Manager
Epsom
Correspondence
to Jill Walton please: