UgleMedFak logo

The Utility of Research in General Practice
By Hogne Sandvik

blueNovember 10th 1995 I defended my doctoral thesis. The judging committee asked me to deliver a test lecture on the 'utility of research in general practice.' While preparing this lecture I asked for help from internet colleagues, members of the lists FAMILY-L, NAPCRG-L, and GP-UK. In gratitude for all help I have put the entire lecture here. Your comments are welcome.

(slide 1) The given topic for this lecture is The Utility of Research in General Practice. Since there are no controlled studies on this, I will give a subjective lecture. Some may even find it polemic. I have chosen a rather broad approach, thereby also risking a somewhat superficial approach.

(slide 2) I have found the topic difficult to delimit. What is utility? Should we consider it an outcome measure, a dependent variable? Is it possible to measure the utility in terms of improved health? Or should I concentrate on the utility experienced by the physician/scientist himself?

Maybe I should demonstrate that research on selected patients will not always be transferable into general practice? Therefore, it is not only useful, but quite necessary to perform research in general practice. What can the general practice researcher add to the scientific community?

However, the given topic also entails a suspicion. Why has the committee chosen this subject? Are they nourishing prejudices that research in general practice only concentrates on banalities and produces obvious answers? Is it possible that a cardiologist would have been given an equivalent theme: The Utility of Research on Heart Diseases?

(slide 3) As I said, there are no controlled studies that has measured the utility of research in general practice. Nevertheless, there are many who have expressed their opinion on the subject.

I have used several sources in the preparation of this lecture. I have looked into the history of medicine in order to find examples of research in general practice. Even though we are newcomers in the universities, analytical thinking was not completely unknown to the old district physicians.

I have used MEDLINE to get an impression of the extent of research in general practice. The number of papers is a common measure of utility. I have also used internet and presented the topic in a few electronic conferences for general practitioners. This has created a very useful global brainstorm on the subject.

The last three years I have been a member of the Board of General Practice Research, a very useful observation post. The Board grants scholarships to general practitioners who want to conduct small research projects, and also assesses the quality of research protocols for drug trials in general practice.

(slide 4) During this lecture I will show you some slides with a selection of comments I have received through the internet, from foreign colleagues. Lots of good advice and good wishes!
(slide 5) I have chosen this outline:

First, I have used MEDLINE to get an impression of the extent of publishing in general practice.

Then I shall discuss what is special with research in general practice. How is it different from other kinds of medical research? I shall demonstrate that research in general practice is not only useful, but also necessary. In this section I'll use some examples from the history of medicine.

Next, I shall present some more recent, Norwegian experiences, mainly from the Board of General Practice Research.

Finally, I will take a look into the future.

(slide 6) This figure shows how many of the MEDLINE references are indexed on three different medical fields. Any MESH subcategories are also included in the numbers. The MEDLINE database consists of several files, each containing 1-2 million references.

You can see that these three subjects together make up less than 1% of all the references. Even alternative medicine has more references than family practice.

But even 0.3% of 7 million references are quite a lot, and as you can see, there is an increasing tendency in more recent years.

(slide 7) Here I have compared the three Nordic medical journals. We see how many general practice papers they have published during recent years. As expected, the Swedish Lakartidningen is trailing behind. Traditionally, general practice has had a weak position in Sweden. It seems reasonable that active publishing reflects active research, which again reflects the position of general practice in the three countries. Perhaps it is a bit surprising that the Danish Ugeskrift hasn't published more papers from general practice.
(slide 8) The same phenomenon is illustrated in this slide. First, I have identified all references indexed on a specific country. Thereafter, I have checked how many of these are from general practice.

I guess there is no country where general practice has a stronger position than in England. And there is no country in which the general practice research activity is higher than in England. One might ask what is cause and what is effect, but I believe that the general practice research activity in England has contributed to the central position of general practice in the National Health Service.

(slide 9) We come to the next section of this lecture: In what respect does general practice research differ from research on more selected patients? How can the general practitioner make the most of his/her unique position to do research that is not possible in a hospital?
(slide 10) Who else than the general practitioner may perform research on the common conditions? How often does a hospital physician see patients with common colds, otitis, tonsilitis, cystitis, tension headache, insomnia, moderate hypertension etc.

The general practitioner meets early and incomplete diseases. He/she follows the patient throughout the process. This provides unique possibilities for research on predictive and prognostic factors.

(slide 11) Patients in a hospital have passed through a filter, a referral filter. They are not representative for the usual patient with otitis or hypertension.
(slide 12) In 1896 Johan Nielsen (general practitioner in the coastal district of Gulen, western Norway) wrote in his annual report that one should keep in mind that it is the most serious cases that are hospitalized. In the university hospital he had been taught that most patients with pleuritis subsequently developed lung tuberculosis. But according to Nielsen's experience this seemed not to be the case in Gulen. During 9 years only one of his numerous pleuritis patients developed lung tuberculosis.

Nielsen understood the meaning of referral (spectrum) bias. Research results from the university hospital may not be valid in general practice.

(slide 13) This colleague does not agree. Like many others he thinks that general practitioners should avoid research areas already covered by hospital based colleagues. There are many who think that general practitioners should restrict themselves to process research, that is research on the consultation, communication, health service research.

I think we should take a broader approach. This is not about duplication. The selection mechanisms are so strong that it is necessary to conduct new research in selected areas of general practice. Guidelines developed in hosptals may be even harmful if they are applied uncritically in general practice.

(slide 14) This phenomenon is also illustrated in this figure. Basically, it represents the whole population. During let's say one month approximately 75% will experience an illness or injury. Only some of them will consult a general practitioner who takes care of most of them himself, the rest being referred to a specialist or to a hospital. Ultimately, a few end up in a university hospital.

A strong selection process controls this stream of patients, and that fact should entail consequences for the generalizability of research results. Some of the research conducted in the university clinic may not be generalized to general practice.

However, we should also keep in mind that the help seeking behaviour in the population also constitutes a strong selection. We cannot generalize findings in our own practice to the whole population. How many of the illnesses experienced by people are known to the health service?

(slide 15) A very thorough work on this subject was performed by Alf Habel in 1889. Habel was a district physician in the rural district of Lindaas, western Norway. He did an epidemiological study on measles.

I doubt if this research had been possible nowadays. Habel instructed all memebers of the local board of health to register all measles cases in their part of the district. At the end of the year he counted the cases, and produced this table. A total of 1204 cases were noted, 90% children. 20 died. But most important in this respect: Only 59 of these patients had consulted the doctor, a 5% consultation rate!

And on this tiny fraction the official statistics were based. According to these statistics measles were 5 times as frequent in the city of Bergen as in the countryside. This difference may be ascribed to differing doctor densities (5 times as high in Bergen).

(slide 16) The epidemiology of general practice was investigated in the first major projects conducted in modern Norwegian general practice research. This was an important beginning, to document the unique characteristics of the general practice epidemiology.

It would be reasonable to expect that the education took notice of this, preparing the students for the reality they were to meet after their exam. However, such research was not valued by the medical establishment at that time. Maybe Dr Bentsen was ahead of his time.

(slide 17) I will go into some more detail on another research project that has been conducted in general practice, by a district physician. I do this to illustrate several aspects of the advantages of the genral practitioner with regard to research: continuity, intimate knowledge of families and social relations. The example will also show a pragmatic attitude towards science, an attitude which sometimes is necessary in general practice. This is a patient in the Leprosy Nursery in Bergen.
(slide 18) The Leprosy Nursery may be seen to the left, Lungegaardshospitalet on the right. Both were institutions for leprous patients. This picture was probably taken in 1860, a few years before Armauer Hansen started his work here.
(slide 19) In 1873 Armauer Hansen discovered the leprosy bacillus. He did so by histological studies of the leprous nodules. Hansen declared that leprosy was a contagious disease.
(slide 20) In outer Nordhordland (western Norway) district physician Thomas Collett voiced scepticism towards the research conducted by Armauer Hansen. Collett believed epidemiological methods should be used to disclose the cause of leprosy. He recommended the use of the protocols that the general practitioners kept on all lepers living in their district.
(slide 21) And so he did. He studied the protocol from outer Nordhordland, and developed this table. This he did in 1884, 29 years after the registration started.

He organized the patients in this table, separated them into different parishes, relative status, marital status, noted if they had leprous children etc. Collett had been a district physician for over 20 years by this time, and he made use of all his local knowledge.

He had registered 164 patients. Most of them (124) had leprous relatives. 85 were married, but in only two marriages were both husband and wife leprous. He also noted that leprosy was most prevalent in the more affluent parishes. Thus, it did not seem that diet and hygiene played an important role.

(slide 22) Collett concluded that heredity must play a major part in the development of leprosy. Why else was this disease restricted to a few families? And if this was a contagious disease, why did almost all husbands or wifes of leprous patients avoid the disease?

We cannot accuse Armauer Hansen of being wrong, but it should be noted that the leprosy bacillus is a nontoxic, intracellular parasite. Infection alone is not a sufficient factor for developing leprosy. Modern research has shown that genetic susceptibilty is a necessary factor.

Interestingly, the general practitioner Thomas Collett could suggest this by using his own notes, combined with an extensive knowledge of his community and a well developed analytical mind.

(slide 23) However, the pragmatic general practitioner chose to use the contagion theory in his public health work. In fact, he propagated a theory in which he himself had little faith! The reason for this was that the local people realized the benefit of hygienical improvements when they were threatend by contagious diseases. Thus, in order to improve the general hygienical standard in his distric, Collett chose to compromise. Heritable diseases was of no use in this respect - they only entailed fatalism and indifference.
(slide 24) We continue with the third section of this lecture. Let's take a look at some more recent Norwegian experiences.
(slide 25) Hospital based medicine expanded enormously during the first half of our century. This happened at the expense of primary care. In the 1950s Norwegian general practice was in a deep crisis.

But then came a change. And the government or the universities are not to be thanked for this shift. The Norwegian Medical Association paid for and established the first university departments of general practice. In 1974 came the Board of General Practice Research as a co-operation between the university departments and the Medical Association. In 1983 we got the Norwegian Society for General Practice and in 1986 the National Research Council established a programme for primary care research.

(slide 26) The newly established university departments put in an application with the Medical Association for the foundation of an arrangement with short term scholarships for general practitioners who wanted to do research. They argued that this would prevent the escape from reality and theorizing that threatens all university departments without connection to daily practice. So this is another utility of research in general practice: A remedy to avoid theorizing! ;-)
(slide 27) The traditional conflict between academy and practice is probably more pronounced in general practice than in other clinical disciplines. Research does not merit much in general practice, and thus follows little research interest and tradition.

Also, it takes longer time before research results influence clinical habits in general practice. I said there are no controlled studies of the utility of research in general practice. But some have demonstrated that general practitioners seldom change their working habits as a result of new research results.

In a way this has to do with the utility of research. Research without practical consequences is less useful. Maybe the researchers have posed the wrong questions? Maybe they are not able to communicate their results in a sensible way?

(slide 28) Well, we got this arrangement with short term scholarships for general practice research in 1976. Again, the Medical Association put up the money, and still does so.

And when we talk about the utility of general practice research, I think these fellows are the best example. They have been evaluated twice, in 1985 and 1995.

The number of scholarships has increased over the years. Now 36 months are granted every year. If we apply the traditional utility measure, publications, we find that more than half of the fellows published their work in medical journals.

(slide 29) But there is another measure of utility that may be just as important, the significance experienced by the individual research fellow. Many have expressed that the scholarship allowed them to take a break from daily routines. It feels good to be able to spend a lot of time on a limited field of interest. Usually, this is not possible for general practitioners.

Many have experienced a growing research interest as a consequence of the scholarship. Approximately 15 have later continued with large scale research projects. Most find that they have become more analytical and critical. They have developed a more active attitude towards their professional practice, and can better evaluate the significance of other's research. By doing research we become better doctors. Therefore, a compulsory basic course in research has been included in the education for general practice specialists in Norway.

I think maybe this is the most important utility of doing research, what we call research training.

(slide 30) A clinical field that produces good research gets respect and is appreciated by the medical establishment, at least in the universities. This gives an opportunity to make one's mark on the education. But here we are talking about research on a higher level.

The programme for primary care research that was started by the National Research Council in 1986 was primarily aimed at general practitioners with projects on a thesis level. In 1990 the programme was evaluated by an external group of experts. They were rather critical. Although the programme had created a lot of activity, they found few international publications, and they expected that few of the fellows would make it to thesis level. The experts also criticized the administration of the programme, and recommended that future follow up of the fellows ought to take place in the university departments.

The Research Council was not satisfied with the result, and ended the programme rather sudden at the end of 1990. But this is not the end of the story.

(slide 31) Most of the fellows continued their projects and published their findings. Here you can see the status in 1994. 15 fellows were paid by the Reseach Council, 22 had other funding sources but followed the programme's research education.

At the beginning of 1994 15 fellows had defended their thesis or submitted it. Almost as many were still pursuing their prosjects. I don't have the most recent numbers, but I know that several others have followed during the last year also. I think this illustrates that many general practice research projects are very time consuming. Especially the collection of data may take much longer time than what is usual for laboratory researchers.

(slide 32) However, the research that involves most general practitioners, are drug trials conducted by pharmaceutical companies. Here is an example of this kind of research, a typical multicentre study. 50 doctors participated and 366 patients, each with four visits. This large study has been conducted very professionally.

The aim was to study two different doses of an antihypertensive agent, 10 mg and 20 mg. You can see that the higher dose induced a larger reduction in blood pressure, 3 mmHg systolic and 1 mmHg diastolic. And these differences are in fact significant. But the authors conclude that the differences are too small to be of clinical significance.

But what has been going on here? 50 doctors and nearly 400 patients? Was it necessary to include so many? Why did they not include a sufficient number of patients in order to be able to reveal interesting differences? If 5 mmHg is defined as a clinical interesting difference one would have to include only a fraction of this number of patients. I call this statistical overkill, some would even say it is unethical. Anyhow, I think it is relevant to ask what is the utility?

There are much worse examples than this, studies whose only purpose is the marketing of a new drug. One important task for the Board of General Practice Research is to weed out these studies. A typical example would be a company who wants to perform a 'surveillance study' - thousands of hypertensive patients are to be given their new and expensive drug, and they are to be followed with regular and extensive controls in order to measure effect and side effects. Totally open studies with no control groups.

Of course, we should not turn down all drug trials, but it is a problem that so much resources are tied up in this type of research. There are more interesting research questions in general practice than which tablet lowers the blood pressure most.

(slide 33) In the period 1988-90 the Board of General Practice Research evaluated 21 industry protocols. In 1995 we tried to find out what happened with these studies. We have been especially concerned that the results are published. It turned out that half of them were published in medical journals. In fact, this was somewhat better than we had anticipated.

But again I would stress this problem: Too much resources are tied up with drug trials. This research potential could be put to better use in other types of research.

(slide 34) Let us take a look at the future possibilities.
(slide 35) Colleagues abroad have established their own research networks. They co-operate on projects with skilled supervision. And most important - they decide themselves the subject of their research!
(slide 36) As usual, England is the most developed country in this respect. They already have many general practice research networks. The government contributes funding, the National Health Service spends 1.5% of their budget on research. More of this money is now being channeled into primary care research.

I think we should pursue this course also in Norway. We should build networks that are able to perform good multicentre studies. We could exploit the market in which the industry now revels.

In Norway we have a stable, surveyable population. I can hardly think of any country in the world better suited for long term studies. Perhaps Norwegian general practice could take on the task of evaluating the effect of nonpharmacological treatment of hypertension?

(slide 37) We are coming towards the end of this lecture and I would like to broaden the perspective. General practice research may also be trail blazing with regard to methodolody. New research methods are developed, such as qualitative methods, hitherto more or less unknown in other medical disciplines. This is undoubtedly a utility aspect, first and foremost to science itself.

The fact that qualitative methods hold so much promise in general practice is related to the epidemiology of general practice. A lot of the patients in general practice do not have traditional diseases with specific causes. We meet symptoms, complaints, illnesses - and these are different from individual to individual. A treatment that worked well with one patient may produce quite different results in a similar patient. Qualitative methods take better care of this aspect than traditional quantitative methods.

(slide 38) However, research becomes more difficult when we leave traditional diseases and concentrates on nonspecific and complicated illnesses. There may be a danger that research results will be just as individual and different as the patients themselves.
(slide 39) I shall now venture further out on the icy surface. There is something called paradigms.

A paradigm is a universally accepted way of reasoning, a model of understanding. When Hippocrates rejected the old magical medicine and claimed that diseases had natural causes, he represented a paradigmatic shift, a scientific revolution.

We may also consider the discovery of the bacteria as a paradigmatic shift. For the first time one could identify specific causes for diseases. This biomedical paradigm is still valid, and an important task for research has been to identify the specific causes of different diseases.

However, this paradigm does not fit very well in general practice. Many of the illnesses presented by our patients do not have simple, specific causes. They are multifactiorial, a mixture of biological and psychosocial causes.

Therefore, some has predicted that general pactice will be the arena for a new paradigmatic shift. There is talk about a biopsychosocial model or a cybernetic model. Cybernetics is the theory of self organizing systems. And undoubtedly, there are many who remain healthy, in spite of all odds!

(slide 40) Maybe we should concentrate more of our research on what keeps people healthy? And what is this unspecific placebo effect that seems to have an effect on everything?
(slide 41) Hippocrates called it vis medicatrix naturae, we call it the placebo effect.

Thank you for your attention!

Your comments are welcome.

blue Home page Department
blue Home page University

Department of Public Health and Primary Health Care, last updated 20.11.95

Hogne.Sandvik@isf.uib.no