Abstract
We carried out an epidemiological study of headache in a rural parish in the mountainous region of southern Norway. During a 2-year period from October 1995, 1838 parishioners in the age range 18–65 (or 88.6% of the target group) were examined in a structured interview based on a questionnaire. The questionnaires were validated in two ways: 100 records were re-checked, with a consistency of 98%, and a re-check of 41 parishioners was carried out >2 months after the 1st examination. The details of the latter control study will be reported later. Only one result of the study is given: namely, idiopathic stabbing headache (“jabs and jolts syndrome”) was present in >30% of parishioners.
Many good-sized population studies (1 –5) have followed in the wake of the pioneering studies on headache epidemiology by Bille (6) and Waters (7). They differ vastly in design from one another and also from the present one. A detailed description is therefore given of the design and demography of our study.
The community of Vågå—topography and demography
The present investigation was a large-scale, epidemiological study of headache (n = 1838) in 18–65-year-old parishioners in the community of Vågå and carried out between October 1995 and October 1997 by one solitary neurologist with a life-long interest in headache. The study was recommended by the Regional Ethics Committee at the University of Trondheim and accepted by the “State Data Inspectorate”. All participants had to read some general information concerning the study and to sign a paper emphasizing that they could withdraw at any time. The results of the present investigation will be compared with those in previous investigations at a later stage.
The rural commune selected is situated in the mountainous region of southern Norway. A large part of the highest mountain range in Northern Europe, Jotunheimen, is within its borders. If a peak is defined as a summit towering > 10 m above the neighboring area, there are 55 peaks >2000 m above sea level in Vågå (the highest at 2368 m). Approximately 80% of the area is >900 m above sea level. The lowest point in the parish is at ca. 320 m. The forest border is at ca. 1000 m, farms being situated up to 850 m above sea level. Annual rainfall is ca. 345 mm (the driest area in the country). The winter season is from early-mid-November until the transition March/April at the valley bottom (source: meteorologist Sigmund Hogåsen, Vågåmo, April 1998).
The majority of the inhabitants live at the bottom and on the slopes of the main, u-shaped valley and along Vågå lake. The communal centre is Vågåmo (in the local dialect: Mo'einn), with approximately 1600 inhabitants. There are a couple of other small villages. Farming is still a prevalent occupation (approximately 1/5 of the population, 1990). The percentage of people directly connected with farm life is higher than the national average (1995: 3.68%; 1997: 3.40% of the total number of workers in the country. Source: Norwegian Central Bureau of Statistics), and fishing naturally lower than the national average. Medium-sized, industrial enterprises, such as hydroelectrical power generation, furniture production, and small companies, such as carpentry, printing, etc., are established and important features of this society. Tourism is also a significant factor, with mountaineering, hang-gliding (a world championship has been held at the Norwegian national centre), and rafting.
Although modernism is slowly seeping into this society, tradition and folklore are still highly esteemed. Rural arts and crafts, e.g. wood carving, are kept up, and folk music and folk dances are well attended, but also the old buildings and famous stave churches in this valley are greatly revered. As a matter of fact, the Conservation Centre for Traditional Rural Culture is located within the parish. This adherence to tradition and to the past may have contributed to the relative social tranquillity and stability of the area. The standard of living is commensurate with the national average, and modern domestic comforts are commonplace.
Unemployment during the period of our study was approximately at the national average (2.5% vs Vågå ca. 2.9%; end of 1997). Problems with drugs and crime are at an ebb. Commuting, both long-term and long-distance, has become more extensive in recent decades and clearly influenced the recruitment of parishioners for this study. The population is nevertheless fairly stable, although some migration is taking place, the outward migration slightly superseding the influx (see Table 1).
Vågå parish. Number of inhabitants during the actual period.
The critical date for inclusion was 9 January 1995.
The size of the study group
In January 1995 there were 3907 inhabitants in Vågå, and 2360 were in the age group 18–65 years. However, 285 of them were not accessible (Table 2) for reasons which included migration, military service, mental underdevelopment, dementia, severe disease, and others (see Table 3). The detailed work concerning non-eligibility was carried out mainly by SH, who has an intimate knowledge of the populace. The remaining 2075 represented the target group. Of those accessible, 1838 were studied, i.e. 88.6% (Table 4).
Vågaå study. Demographic data.
Groups/individuals not studied.
Groups A-D represent parishioners that for all practical purposes were unavailable for the study. None of the individuals in groups A/B/C were contacted directly or indirectly. Nor were certain individuals in group F contacted, because their negativity to the project in some way had come to the ears of the health personnel. Some parishioners in group F (a) expressed a clear will to participate — by telephone — or indirectly, through family, but failed to get an investigation owing to various unfortunate circumstances.
Vågå study. Percentage studied.
Reasons for selecting Vågå
Why was such a relatively small, rural commune selected, and in this part of Norway?
The relative stability of the populace was one reason (Table 1).
Just as important was the fact that the people are known for their stoicism and phlegmatic behavior. They have been taught by relatives and fellow parishioners not to worry too much about unpleasant bodily sensations and not to use drugs other than those sold over-the-counter and preferably not even these (“they may be harmful, you know”). This stoic attitude can form the basis of a favorable environment for studying the natural manifestations and development of headaches without the interference of chronic use of drugs (analgesics and so on).
Family members were automatically studied, since they too are parishioners.
The principal investigator (OS) was brought up in this commune. He has a first-hand knowledge of its inhabitants and can address the parishioners in their own dialect.
How could such a high recruitment rate be obtained?
The original plan was to drive around to the farms and village houses and to examine the various individuals on the spot. However, the chief of the health service in the parish (JB) opposed the plan, contending that the “base” had to be an office within the Health Center itself, with access to records, possibility to order tests, etc. In retrospect, this was obviously the correct solution.
There were various phases in the acquisition of individuals, although they overlapped to some extent:
Friends and acquaintances of the family (of OS) and some of their relatives signed up, but this “source” was quickly exhausted (ca. 1.5% of the total examined group).
Many approaches were attempted at an early stage: Advertisements were placed in the Health Center, in shops and in public institutions encouraging people to register. Propaganda advances made on the local radio, in local newspapers and parish newsletters, etc. produced public awareness of the project, and scores or migraine patients committed themselves as a result (ca. 5–6%).
One approach that proved of real value in the early phase was acquisition through the healthcare system itself, the personnel speaking to or literally grabbing hold of the patients, i.e. the regulars: “Have you participated in the headache investigation?” This added around 32%. But even the clientele at the Health Center was limited. Most visitors were in the middle/elderly age group, and many, i.e. those above 65, were automatically eliminated. Some younger parishioners did not even have a health record.
Various groups (“Lions”, female organizations, etc.) urged their members and families to sign up through the various phases of the study, and this clearly was useful (4–5%?).
Special assistants were recruited to acquire people for the investigation. This was not a very efficient way of doing things, but it did add ca. 11%. Further approaches were, therefore, necessary.
“Office-days” in most outlying areas of the parish were tried, adding ca. 1.5%. Some peripheral hamlets are 12–30 km away. This approach could have been expanded if necessary, but the last two approaches (below) provided many recruits.
Lists were provided of workers in factories, shops, and offices. These people were approached and encouraged to enlist. Some 14% of those incorporated were acquired in this way.
Telephone contact lists were provided. These also contained the names of people living together (husband and wife frequently had different surnames, and common laws were often not listed in the telephone directory). Personal telephone contact was made and the scope of the investigations explained and the reasons given as to why it was important that (s)he be examined. This approach secured at least 30% of those examined. It proved possible also to involve many family members through those last two approaches.
Although, of course, all participation was voluntary, a telephone call, or a personal contact, puts some pressure on an individual to answer. “The resistance” of someone who previously has not been able to give this project high enough priority is weakened. This is probably the essence of the matter. Those who still wanted to “think about it”; “come back to it”, “telephone later”, and so on can usually quite safely be deleted from the list. So also can those who fail to meet, without explanation, for a second appointment (Table 3).
In short, the three approaches that proved most useful were:
Acquisition by the health personnel
Acquisition through lists of personnel from factories/institutions with subsequent personal contact
Acquisition through personal telephone contact
This, third, approach was really the ultimate method of getting a really high percentage of recruitment, at least in this area at this time. The “telephone-call”-method is time-consuming and tiring. It cannot be used on a daily basis over a period of years without having some negative— even destructive—influence on the life of an eager-to-get-down-to-business researcher. It can therefore really only be used in the final stages.
It is possibly of some interest that when the principal investigator originally stated officially, at local orientation meetings and in newspapers, that the aim was to examine 70%, preferably 75% of the people, wise people in the parish shook their heads: “there is no chance—they might as well give up before starting”. Fortunately, this was told to him until later.
In this study, 237 parishioners could not be investigated (Tables 3, 4). Most were not motivated. We discovered this, directly or indirectly, through the health personnel. Some parishioners nevertheless did show some interest, but practical difficulties hindered enlistment (e.g. too few and short visits to the parish). In other words, the percentage could have been increased by a few per cent, perhaps to 3–5%, with a combination of patience and time. It became increasingly clear, however, that the “resistance” was strengthening with time; the work was becoming more strenuous, with the daily “gain” now approaching only 1/3 of the gain in the early-mid-phase of the study.
The questionnaire
This study is first and foremost a cross-sectional in form. Headaches occurring earlier in life were also systematically asked about. An inherent weakness of such a study, however, is that the lack of memory precision for events in the distant past is an important factor, especially in the elderly. As for headache diagnoses, the principles laid down by International Headache Society (IHS) (8) have been adhered to, and also those of IASP (9), especially for cervicogenic headache.
Questionnaires to be filled in without the guidance of headache specialists are probably of limited value in headache research in grown-ups, regardless of how elaborate they may be. If there is more than one headache, or the existing headache has changed character, even mildly, the patient or individual will—left alone—get lost even at the first slight bend in the road, with few or no possibilities of returning to the main track later in a fashion that can be relied upon. A structured interview is probably the only way to provide meaningful data, a sine qua non. With >30 years' practice as a headache clinician, we strongly feel that we would not have been able to carry out this study adequately 10 years ago; previous work as a headache neurologist can in many ways be viewed as a preparatory stage for this epidemiological work.
While the questionnaire is based on our routine questionnaire, it is much more detailed. Tried out on a limited number of patients prior to the study, the questionnaire is a 24 page document containing 22 questions regarding social relationship and general health; 99 questions regarding headache, 27 headings (with subheadings) for physical/ supplementary findings/neurological examination and 79 additional questions if >1 headache was suspected. The number of questions/headings for findings totalled 227. In addition, strategically placed, counter-suggestive questions were inserted. At the very beginning, a full examination took 70–75 min. Quite quickly, with a more rational sequence of questioning, examination time could be reduced to ca. 60 min. In cases of “no headache” or only mild tension-type headache, a ca. 45 min examination could suffice. If the time approached 75 min, especially if in the late afternoon/evening after a day at work, not infrequently an extra appointment was made for another day. New appointments were necessary especially if the headache history was complicated; for example, two or more headaches being intertwined. One or more extra appointments were necessary to discuss the outcome of tests, etc., if treatment had been given.
The questions were put to parishioners in stereotypical fashion, i.e. with the same wording, tone, speed, and even with the same body language. Although OS knew the local tongue fairly well as a boy (up to 15), he had forgotten most and had to re-learn it. This dialect is further removed from written Norwegian (“riksmål”, i.e. the Oslo dialect) than Norwegian is from the Swedish. The fact that the parishioners could be addressed in their own tongue created an atmosphere of confidence and goodwill. Those who did not master the local tongue were addressed in the Oslo dialect.
The first headache question was: “Have you at any time been troubled by headache?” A “no” was the usual answer. The next question, therefore, was: “Have you had any headaches in the course of your life?” Not even a “no”-response to this question was a guarantee for headache freedom. The reasoning on the part of the parishioner was usually: “When I compare what I have experienced with what I have heard from others, I reckon my own headache for nothing”. An individual's own opinion as to whether he/she had experienced more than one headache was exploited early, i.e. after the sinusitis and fever headache questions.
Minor modifications were made in a few instances in the way questions were addressed in the early phase of the study, because the revised version was more informative and more easily understood. In the original version, a question was asked concerning nocturnal headaches: “Did he/ she wake up in the night with a headache at times?” If so: “Was the awakening due to the headache?” Two questions were added: “When awakening during the night due to headache, was it then due to the “old” headache (headache also just prior to falling asleep)?” Or: “Was it due to a “new” (“de novo”) headache (no headache, previous evening)?”
The readiness with which the parishioners spoke about their experiences with alcohol was unexpected: they proved for instance to have clear views on whether nausea was caused by the headache and/ or by a “direct” effect of alcohol per se. The scheme had to be adapted accordingly.
Two coexisting headaches may be hard to deal with anamnestically if the individual has preconceived, rigid opinions as to their nature, e.g. if a clear migraine of childhood had “transformed” into a mild migraine in grown-up age, the other features of the headache persisting (and interpreted as a “new” headache, due to tension, by the parishioner). Also, to clarify a certain point, for instance photophobia (if present also outside attacks, and only worse during attacks?) the structured interview had at times to be deviated from, which at times could be quite time-consuming.
The physician has to spend time with each parishioner. He has to let the parishioners feel they can relax, that there is no hurry or rush, and that they are encouraged to give their own opinions, even trifling details, and that they are taken seriously. Important information can suddenly come when the parishioner is about to leave the room—“Come to think of it…”
The most important single feature of an investigator making valuable registrations in this field, besides insight, is attitude. The researcher must have a completely “open” mind, not a mind replete with preconceived ideas and conclusions, trying to “prove” certain points. If a researcher adopts this latter attitude, elements making up the headache diagnosis will readily be interpreted in certain directions. The sum of the diagnostic elements will also be interpreted in certain directions, and the ability to see constellations of symptoms and signs in another, conceivably worthwhile light may be reduced. In these early stages of headache epidemiology research—where we are still without set norms—such an attitude may actually be directly counterproductive.
The examination
A routine, full-scale clinical, neurological examination was not part of the study for the following reasons. These examinations have been an integral part of the work-up in > 10,000 headache patients examined by OS. These referral patients were diagnostically difficult to classify, severe cases and/or difficult to treat. Almost invariably the neurological examination was negative. If otherwise, the clinical neurological examination per se only substantiated suspicions provided by anamnesis. Examples of such a course of events include irradiating pain into the finger(s) in the occasional case of cervicogenic headache, where upper extremity reflex changes (hyporeflexia) have been observed, providing a clue as to the correct level. Only in a very few patients has the complete, clinical neurological examination per se led to other, unexpected abnormal findings.
There were of course a number of symptomatic headache cases among these > 10,000 patients, but their nature was almost invariably unravelled by supplementary investigation, computed tomography (CT), magnetic resonance imaging (MRI) etc. All in all, a thorough clinical examination of neck and head (including cranial nerves) is more revealing than a routine neurological examination.
The grass-roots level individuals presumably would be characterized by generally less severe headache than the aforementioned referral series. The likelihood of making unexpected, abnormal clinical, neurological findings in this clientele would a priori be even less than in our regular referral patients. The possible gain with routine neurological examination had to be weighed against the negative side—a prolongation of the examination by ca. 15 min. Some parishioners felt that even 60 min was a long time. The rumour that 75–90 min was to be spent would probably have had a negative influence on recruitment. At no time was it a consideration to reduce the time of the interview to allow more time for the neurological examination. The interview was too important for that.
The following policy was, therefore, adopted: The cranial nerves were routinely examined (with the exception of visual acuity, ophthalmoscopy and cranial n. no. VIII). Full clinical, neurological examination was only carried out when anamnesis and/or the head/neck examination findings made this necessary. Supplementary studies, i.e. blood tests, X-rays of the skull, sinuses, and cervical spine, and CT (and rarely MRI) of the brain, MRI of the cervical spine, EEG, and blood tests were carried out in suspected symptomatic cases. Nerve blockades could not generally be carried out due to the circumstances.
Control studies
Impact of the non-examined parishioners
In studies of this sort there will always be criticisms of this nature. “The parishioners not examined are probably different from those examined”, and “All those with headache have signed up, but the remainder probably have no/almost no headache and this will ruin the data!” With a low compliance, these are of course valid counterarguments. In the present study, 88.6% were examined, with a ratio between examined/non-examined of 7.77 (Table 4). The impact of non-compliance would, therefore, be limited anyhow. A good idea of headache prevalence in the non-examined fraction can be obtained by comparing parishioners examined before and after certain “landmarks”.
We seriously considered closing the study at 1750 parishioners (i.e. at 84%), because the daily “gain” was steadily falling off. A comparison could therefore be made between those examined prior to and those after this fictitious landmark. Another comparison could be made between those examined before and after the study really was stopped at n = 1809. But how could the study then be restarted? In a last, whole-hearted effort to increase the number, some of the parishioners who had been interested but had been hindered for some reasons were re-contacted approximately 2 months after the conclusion of the study. These were parishioners mainly within the then group of long-distance/ long-term commuters, but a couple of draftees/ students were also included, the gain being 29 extras.
Validation of the robustness of the questionnaires
One hundred records (questionnaires) were rechecked blindly. Blinding was carried out in regard to name, sex (pain in connection with menses), and family history. The sample consisted of 50 consecutive records from the middle plus 50 from the later part of the study (female/male ratio of 0.61). One record was removed because of the difficulties the patient had had in giving a precise history and substituted by another one. The decision to do so was made prior to looking for the previous diagnosis. The re-check was carried out >10 months after the examination of the last parishioner in this group; in other words, so much later that the investigator had no chance of recognizing any case history. If a parishioner could have been identified in some way, that specific record would have been substituted by another one.
There was absolute congruence between the two assessments in 98% of cases, including the “major” headaches (migraine, tension-type headache, cluster headache, and cervicogenic headache at the 2nd digit level of the IHS classification; for cervicogenic headache: IASP criteria (9)) and also the “minor” ones (hangover headache, exertional headache, etc.). In two cases, there was a discrepancy between the first and second assessments, and over the same issue: whether photo- and phonophobia were part of the picture; whether constantly present in a borderline fashion, but increasing to some degree during attacks, off and on. In both cases, the headache was originally headlined as non-classifiable; the indecision had led to a preferred classification as acute tension-type headache in one case (?) and as migraine without aura in the other (?). On reassessment, the diagnosis was changed in both cases to the opposite, but still with question marks.
The test situation for Procedure 1 is easier than in the work by Ganella et al. (10), in which there were four physicians versed in headache diagnostics, independently addressing the same videotape (n = 103). In this way, interobserver bias enters the picture. The high degree of congruency in the present study must not be interpreted as showing that the diagnoses made are necessarily the correct ones. The control study shows only that, done in this way and by this investigator, there is little intrainvestigator deviation. But this investigator can be biased, he may constantly interpret a constellation in the same, personally flavored, slightly wrong way.
In a second control study, 41 parishioners were re-examined blinded as far as diagnosis was concerned. This was a far more complicated situation than the one with the questionnaire check; one main problem was whether for instance the presence/ absence of each of an individual's headaches should be assessed at the two examinations, or whether a special scoring system should be developed. This will be dealt with in detail in another context.
Prevalence of one solitary headache: idiopathic stabbing headache (IHS, 8) or: Jabs and Jolts syndrome (IASP, 9).
No data treatment of the study has so far taken place. No results will, therefore, be revealed, except one: When jabs (stabs) were defined as pain paroxysms of <3 sec duration, the prevalence of this headache seemed to exceed 30%.
Footnotes
Acknowledgements.—
The authors are most grateful to Glaxo-Welcome of Norway for supporting this study. We also wish to thank the staff of the Health Center in Vågå and the various institutions that helped us with recruitment. Last but not least, we thank the people of Vågå who dutifully signed up for this investigation.