Background
The Global Campaign against Headache is conducted by
Lifting The Burden (LTB), a UK-registered non-governmental organization in official relations with the World Health Organization (WHO) [
1]. Its ultimate purpose is to reduce the burden of headache worldwide [
2‐
4]. At its launch in 2004, our knowledge of both the scope and scale of the burdens attributable to headache was extraordinarily imprecise; from very large areas of the world, there were almost no reliable data [
5]. In the 13 years since, LTB has endeavoured to fill these knowledge gaps, undertaking or supporting a series of population-based studies in countries around the world [
6‐
13] using standard methodology and survey instruments [
14,
15]. Findings from these studies have informed various iterations of the Global Burden of Disease (GBD) study [
16‐
18], while providing country policy-makers with local knowledge to guide priority-setting in health care.
In this process, headache disorders have been revealed as the third-highest cause of disability worldwide [
17,
19]. Migraine and medication-overuse headache (MOH) are the major contributors [
18,
19], being highly disabling at individual level; tension-type headache (TTH), while less disabling at this level, is the second most prevalent disorder in the world [
17]. All three of these disorders are therefore important causes of public ill health globally. They place heavy demands on health services, and call upon policy-makers in control of health-resource allocation to give them due attention [
19‐
21]. Worldwide, however, policy-makers have notably failed in the last regard, a situation to which WHO has not only called attention but also urged action to provide remedy [
21].
It is unquestionably challenging to implement health-care solutions for the very large numbers of people affected by these disorders. On the other hand, equally unquestionably, it is not humane to ignore the burdens of ill health and disability that they impose. Neither is it economically sensible: effective treatments exist [
22] and, appropriately used, they are likely to be cost-saving in most economies [
21,
23]. Why does nothing happen [
19,
24]?
There is no clear answer to this question, but more persuasive data appear to be needed at national level; therefore, the Global Campaign’s series of country-based studies continues. For Ethiopia, we have already published prevalence data: 44.9% of participants in our population-based survey (males 37.7%, females 49.9%) described headache in the last year, and 7.1% (males 4.1%, females 9.2%) reported headache during the day before enquiry (headache yesterday [HY]) [
13]. Adjusted for gender, age and urban or rural habitation, the 1-year prevalence of migraine was 17.7%, of TTH 20.6%, of pMOH 0.7% and of other headache on ≥15 days/month 2.5%. The adjusted 1-day prevalence of any headache was 6.4%.
Headache is therefore very common in this country. Here, expressly as a needs-assessment to inform national health policy, we present data from the same survey on the burdens attributable to headache in Ethiopia. Few studies of headache burden are yet available from sub-Saharan Africa (SSA), but this survey follows, and mirrors, a similar one conducted by LTB in Zambia [
10].
Discussion
Our earlier manuscript showed headache to be as prevalent in Ethiopia as elsewhere in the world [
13]: migraine was in fact more common (1-year prevalence 17.7%) than the global estimate from GBD2010 (14.7%), and TTH as common (20.6% versus GBD’s 20.8%) [
16]. Here, estimating the levels of burden arising from the key headache disorders, we find these to be commensurately high.
In summary, symptom burdens affect, particularly, those with migraine or pMOH. The former, we estimated, spent on average 11.7% of all their time in the ictal state, with headache described as moderate to severe (2.6 on the scale 1–3). The consequent disability burden averaged over time was 5.2%. People with pMOH are only 0.7% of the adult population but, as might be expected, carry much more individual disability: 60.2% of their time was spent in the ictal state, a huge loss of healthy time, with headache rated severe (2.95 on the scale 1–3) and a disability burden of 13.4%. Disability from TTH was much lower. Consequential losses of productive time due to migraine were 4.5% from paid work and 5.3% from household work, and, again as expected, much higher at 29.2 and 16.0% from pMOH. At population level (among those aged 18–65 years), we estimated disability from migraine at 0.92%, with 0.80% of paid workdays lost, from pMOH at 0.09%, with 0.20% of paid workdays lost, and from other headache on ≥15 days/month at 0.30%, with 0.22% of paid workdays lost. Total population-level disability was 1.4% and lost paid workdays slightly higher at 1.6%.
These estimates were not all reflected exactly in other measures of burden. The losses of productive time to migraine (4.5% from paid work and 5.3% from household work) days were consistent with the estimated underlying disability level of 5.2%. Productive-time losses to TTH, while smaller (2.0% and 1.2% respectively), nevertheless substantially exceeded the estimated 0.27% disability. This might be explained by the skewed distributions, and a relatively highly-disabled minority with frequent headache; alternatively it suggests the DW of 0.037 attributed to TTH in GBD2013 [
29] is too low. The losses for people with pMOH (29.2 and 16.0%) were also greater than the estimated disability level of 13.4% (the former more than double).
Corroboration of these estimates came from enquiry into HY. This enquiry obviates the potential problem of faulty recall over periods of 3 months [
14]. In fact, the predicted 1-day prevalence of headache based on reported frequency over the preceding 3 months was 6.9% - very close to the 7.1% reported prevalence of HY [
13]. As for the effect of HY on activities yesterday, the lost “output” per person in the entire population of 3.0% was entirely compatible with the averaged lost paid workdays of 1.6%, given that the former included all planned activities (additionally household work, which had similar losses to paid work, and social events).
The obvious comparison to make is with Zambia, where LTB conducted a similar study with the same methodology and questionnaire [
10,
30]. Both Ethiopia and Zambia are landlocked but ethnically and ecologically diverse African countries. Ethiopia, in East Africa, is <20% urbanised, with agriculture accounting for some 85% of the labour force [
31], but its capital, Addis Ababa, sits at 2400 m. It is a low-income country: despite a rapidly-growing economy, its GDP per capita remains one of the lowest in the world, and only 4.7% of this is spent on health (latest estimate from 2011 [
32]). Many of the rural population in particular live in poverty. Its recent history is troubled, including foreign occupation, war (both civil, and with neighbouring Eritrea), genocide and drought. Zambia, much further south, is one of the most highly urbanised countries in SSA (>40%), with sparsely populated rural areas [
33]. Nevertheless, agriculture provides most jobs, although the country’s economy is dependent on copper-mining and vulnerable to fluctuations in copper pricing. It is a lower-middle-income country: its per-capita GDP is 50% higher than Ethiopia’s, with 6.1% spent on health (latest estimate from 2011 [
34]); nonetheless, >60% of Zambians live below the poverty line (latest estimate from 2010 [
35]). It is politically stable.
The two countries therefore have many differences that might affect the prevalence and impact of headache disorders. In fact, in terms of prevalence, there is slightly less migraine (17.7%) in Ethiopia than in Zambia (22.9%; chi-squared = 12.7602;
p = 0.0004), a similar level of TTH (20.6% versus 22.8%; chi-squared = 2.1126;
p = 0.1460), but
much less pMOH (0.7% versus 7.1%; chi-squared = 115.319;
p < 0.0001) [
13,
30]. GBD2015 refuted any notion that headache disorders are associated with poverty [
18,
36]. Rather, in both these countries, pMOH is associated with higher income (ORs of 2.1 in Ethiopia [
13] and 6.0 in Zambia [
30]). The relative poverty of Ethiopia might, therefore, be one explanation of the difference in prevalence of this disorder – quite simply, Ethiopians cannot afford to overuse medications. A second and probably more influential factor is the greater urbanisation of Zambia: again in both countries, pMOH is very strongly associated with urban dwelling (ORs of 6.1 [
13] and 8.6 [
30]). This is likely to be explained by poor rural access to medications.
As for burden, people with migraine in Zambia [
10] were 4.3% disabled overall (Ethiopia 5.4%), losing 6.3% of paid workdays (Ethiopia 4.5%) and 4.2% of household workdays (Ethiopia 5.3%). These are not highly dissimilar findings. In Zambia but not Ethiopia, lost paid worktime exceeded the underlying estimated disability level, which again may reflect the poverty of Ethiopia (people cannot afford to miss paid work). In Zambia, estimated disability from migraine in the entire working population was 0.98%, in Ethiopia very similar (0.92%); but in Zambia, 1.4% of all workdays were lost to migraine, in Ethiopia only 0.80%. People with pMOH in Zambia were 8.3% disabled and lost 7.4% of paid workdays and 5.0% of household workdays [
10]. These estimates are different from those in Ethiopia (13.4%, 29.2% and 16.0%), but, with such different prevalences (7.1% [
30] versus 0.7% [
13]), and numbers in Ethiopia very small, nothing should be made of this. In other words, with similar prevalences of all but pMOH, the modest differences in headache-attributed burden can be explained by geographical, cultural and environmental differences between these two countries. This suggests that, in the absence of reliable data from elsewhere, the findings in these two countries can reasonably be extrapolated to others in SSA – at least in the east and south.
An important part of our purpose was to inform policy, which the population-level findings do very well. Considering first the study limitations, in our earlier paper we recorded our inability to conduct a diagnostic validation [
13]. The diagnostic questionnaire had, however, been employed successfully in multiple other countries and cultures [
15]. Furthermore, the essential messages here for policy purposes relate to headache overall rather than to any headache type, and they are not significantly affected by this limitation. The study had several strengths, also noted previously [
13]. We employed population-based sampling, included diverse regions with a large sample of >2400, applied ICHD-II diagnostic criteria [
26] and used established methodology also tested in numerous other countries [
14].
The policy messages are these. Headache disorders are not only common in Ethiopia but also heavily burdensome. Individual disabilities attributable to all headache types make up a total disability of 1.4% among the entire population aged 18–65 years (compared, incidentally, with 1.6% in Zambia [
10]). It is important to recognise that this population is effectively the working population. Lost paid workdays in this population from headache total 1.6%, slightly less than the 1.9% in Zambia [
10] but substantially more than the 1.1% estimated in a similar LTB study in India [
37]. This is an enormous economic burden in a low-income country, likely to be reflected in national productivity and gross domestic product (GDP).
Conclusions: What is to be done?
At population level, Ethiopia may lose 1.6% of its GDP to headache, but the country has many other health-care problems. Communicable diseases (including HIV) and malnutrition, along with lack of access to clean water for nearly half the population [
38], are high among the causes of ill-health. The major non-communicable diseases (cardiovascular disease [CVD], cancer, diabetes and chronic obstructive pulmonary diseases) have been less well documented but nonetheless also contribute substantially to morbidity, with CVD notably on the rise [
39]. At the same time, doctors are relatively few [
40]. Nevertheless, headache disorders are not only treatable [
22] but cost-effectively treatable [
23]. As in Zambia [
10], WTP would contribute little towards the cost of care, but this should not be a disincentive to finding the solution. Health politicians need to sit down with experts and discuss what must be done to alleviate the headache burden, and how, not just because people in Ethiopia lose much of their health and quality of life to headache but also with the expectation of cost-saving nationally [
21]. WHO has recommended structured headache services with their basis in primary care as the most efficient, effective, affordable and equitable solution [
21], and the model proposed by LTB for Europe [
41], which is highly adaptable, could be reworked to match the health-care infrastructure of Ethiopia.
Acknowledgements
We thank the Regional Health Bureaux, the Department of Neurology at Addis Ababa University, the data collectors and data-entry clerks, and all the people who were interviewed for this study.