Headache yesterday in Russia: its prevalence and impact, and their application in estimating the national burden attributable to headache disorders

The Global Burden of Disease Survey 2010 (GBD2010) ranked tension-type headache (TTH) and migraine as the second and third most common diseases worldwide [1]. More importantly, these primary headache disorders are associated with disability, reduced quality of life, public ill health and high economic burdens on both individual and population levels. Migraine has been recognized as the seventh highest among specific causes of disability worldwide [1, 2]. The costs of headache disorders are enormous: estimated in 2012 at well in excess of EUR 100 billion per year in the European Union [3].

In the last 20 years a large number of population-based studies of headache prevalence have been performed, especially in the United States of America and Western Europe [4]. Some have also estimated headache-attributed burden, and those assessing socioeconomic impact have employed questionnaires such as the Migraine Disability Assessment (MIDAS) instrument [5] or its derivative, the Headache-Attributed Lost Time (HALT) questionnaire [5, 6]. The former, and in its original form (HALT-90) also the latter, depend upon enquiry relating to the three months prior to the interview, an approach subject to the limitations and errors (and possible biases) of recall over such a long period [7]. The consequences of these limitations, errors and possible biases need to be considered, and an alternative approach that allows this is enquiry into headache occurring on the day before the interview (“headache yesterday”; HY). Recall of HY and its impact is likely to be highly reliable. While what happened yesterday is not indicative of headache-attributed burden in individuals, at population level it is. Such an approach does, however, require a high number of respondents, since far fewer people have headache yesterday than the number who have headache in 3 months. Furthermore, diagnosis cannot reliably be based on a single headache episode; unless a respondent can report that HY is typical of a recurrent headache that has been diagnosed, its nature will not be known.

A question subset concerning HY [8, 9], has been included into several recent large epidemiological studies supported by Lifting The Burden and conducted within the Global Campaign against Headache, including those in China [10], India [11], Pakistan [12], Nepal [13], Zambia (unpublished) and countries of the Eurolight project [7].

Here we report the analysis of HY in Russia. We apply this analysis in making an estimate of the national socioeconomic burden attributable to headache disorders, and compare this to the estimate based on 3-months’ recall [14].

The methods of the study have been described in detail previously [15] and are only summarised here.

This is the third published analysis of headache yesterday (HY) worldwide, following those in China [10] and eight countries of Europe [7], and the first in Russia. It is of some interest to know the proportion of a population afflicted by headache on any day, but the true value of this population-based enquiry lies in its almost complete detachment from the recall errors to which traditional burden-of-headache surveys are so strongly subject. This approach was pioneered by Lifting The Burden in its Global Campaign against Headache [18, 19] and is included in the Headache-Attributed Restriction, Disability, Social Handicap and Impaired Participation (HARDSHIP) questionnaire recently developed for population-based studies of headache [8, 9].

In our study, 14.5% of participants, aged 18–65 years (i.e., of working age), reported HY. Extrapolated to the entire population, this means that approximately every seventh person of working age in the country suffers from headache on any and every day. This finding is quite similar to the 15-19% reported from 15 countries in Europe [7], although these data from the Eurolight study were not entirely population-based. In absolute numbers, in Russia’s total population of 142 million, there are 99.8 million adults aged 18–65 years [20] among whom, on any and every day, there are 14.5 million with headache.

Comparing these data with those previously reported in our 1-year headache prevalence study [15], we found by prediction the same result – 14.6% of the adult population with headache every day. Interestingly, in both earlier published studies of HY, similar internal consistency was seen between reported HY and predicted 1-day prevalence of headache based on recalled frequency during the preceding 3 months [7, 10]. In our study, the fact that well over 90% of participants reported HY to be typical of their most bothersome (and therefore diagnosed) headache allowed us to analyze HY according to headache type. For each – episodic migraine, episodic TTH and headache on ≥15 days/month – quite similar consistency was observed.

Headache on ≥15 days/month is highly prevalent (10.5%) in Russia [15]. In accordance with this, the various causes of headache on ≥15 days/month accounted for almost 50% of cases of HY, confirming their importance as contributors to population ill health and disability in Russia [14]. Almost 52% of those with HY took acute headache medications, which extrapolates to 7.5 (14.5 × 0.52) million people doing so every day. Almost half of these (48%, extrapolating to 3.6 million people) used combination analgesics, medications probably closely associated with medication-overuse headache (MOH) [21]. In Russia, medication overuse was also reported by two thirds of survey participants with headache on ≥15 days/month [15], and their diagnosis was probable MOH.

The main purpose of this survey was a correct evaluation of headache-attributed burden. The mean duration of HY of 6 hours was relatively short – reflecting probably the effect of treatment – but in line with duration reported in the European study [7] and slightly more than was found in China [10]. From this mean duration and the prevalence of HY, we calculate that 3.6% of the population aged 18–65, or 3.6 million adults, have headache at any one time. This proportion is quite similar to the 4.0% in Europe (6) but double the 1.8% reported in China [10]. In terms of pain intensity, the mean was 2.1 on a scale of 1–3, expected to be disabling. A quarter of participants (24.8%) described severe headache, and approximately the same proportion (22.9%) reported that their daily activities were reduced to nil by HY (Table 2). In Russia’s 99.8 million people of working age [20], this means about 3.5 million people so affected every day. Since the indirect costs make up by far the greater part of all headache-attributed costs [3, 22], and in order accurately to evaluate headache-attributed economic burden, we concentrated on the working population. In employed participants, the prevalence of HY was lower (11.9%) without meaningful difference between those who had a workday (11.6%) or a free day (12.4%) yesterday, indicating that the phenomenon of “weekend headache” has a minor effect at population level. Above, we calculated a 4.0% lost productivity in the workforce on every workday because of headache (see Results). Since the total working population aged 18–65 years in Russia is 65.6 million [20], this means 2.62 million people unproductive every day or a loss of 2.62 million person-years/year. This number is approximately 70% higher than the 1.54 million person-years/year estimated through recall over the last 3 months [14]. Even more time was lost to HY from household work or leisure. Household work is generally more easily abandoned or delayed than paid labour, so this finding was expected, and has been reported in previous studies [14, 23].

Although in Table 1 we present HY according to diagnosis (based on participants’ assertions in 93.5% of cases that HY was typical of their diagnosed most bothersome headache), because of the diagnostic uncertainty we did not go beyond this to attribute burden to each headache type. In terms of prevalence, approximately half of cases of HY were one or another of the causes of headache on ≥15 days/month and a quarter were each of migraine and TTH. Burden would depend, in particular, on how disabling each headache type tended to be, with, on average, migraine expected to be most and TTH least. Headache on ≥15 days/month is variable in its effect; although chronically disabling, it does not necessarily cause a high level of disability because people develop coping mechanisms for it. But it is worth recognizing that about two thirds of headache on ≥15 days/month in Russia (as in many other countries) is probably MOH (14), an avoidable condition. Which, with this further and explicit demonstration that headache not only causes much ill health in Russia but costs the country a substantial proportion of its gross domestic product, raises the question of what should be done about all this. The answer is urgent government investment in adequate structured health care for headache, supported by educational initiatives aimed both at health-care professionals and at the general public (who include people affected by headache) (13). The irony, while this does not happen, is that it would almost certainly be cost-saving [24].

The strengths of our study were several. Enquiry into headache yesterday is a powerful approach to population-based burden-of-headache studies, virtually eliminating recall error and any bias that may follow from it. Through face-to-face interviews we achieved a comparatively low non-participation rate of 25.7%, and could avoid the previously described bias of “headache today”: participants with headache on the day they receive the questionnaire by post may defer responding until headache-free, and then spuriously report HY [7]. Since household visits were unannounced and potential respondents were not informed about the survey topic, we believe we had little interest bias (due only to interruption of interviews after they had commenced). Our study sample matched the country population demographically owing to multi-stage cluster sampling. As to limitations, we did not specifically enquire into absenteeism from work yesterday, estimating lost productivity from reports of what had been done in relation to what had been planned. Arguably this served a more conservative analysis; certainly it could not lead to an overestimate of lost productivity at work.

In conclusion, using HY as a method of enquiry free from recall error, we confirm a very high level of headache-attributed burden in Russia. We repeat our call to the government for action [14], observing that correct action would almost certainly be cost-saving [24].