Content uploaded by Piers D Mitchell
Author content
All content in this area was uploaded by Piers D Mitchell on Oct 14, 2021
Content may be subject to copyright.
1
Accepted, Unformatted version of:
Robb, J., Cessford, C., Dittmar, J.M., Inskip, S., Mitchell, P.D. (2021) The greatest health problem of
the Middle Ages? Estimating the burden of disease in Medieval England. International Journal of
Paleopathology 34: 101-112.
The Greatest Health Problem of the Middle Ages? Estimating the Burden of Disease in Medieval
England
Abstract
Objective: To identify the major health problems of the Middle Ages. Bubonic plague is often
considered the greatest health disaster in medieval history, but this has never been systematically
investigated.
Materials: We triangulate upon the problem using (i) modern WHO data on disease in the modern
developing world, (ii) historical evidence for England such as post-medieval Bills of Mortality, and (iii)
prevalences derived from original and published palaeopathological studies.
Methods: Systematic analysis of the consequences of these health conditions using Disability
Adjusted Life Years (DALYs) according to the Global Burden of Disease methodology.
Results: Infant and child death due to varied causes had the greatest impact upon population and
health, followed by a range of chronic/infectious diseases, with tuberculosis probably being the next
most significant one.
Conclusions: Among medieval health problems, we estimate that plague was probably 7th-10th in
overall importance. Although lethal and disruptive, it struck only periodically and had less
cumulative long-term human consequences than chronically endemic conditions (e.g. bacterial and
viral infections causing infant and child death, tuberculosis, and other pathogens).
Significance: In contrast to modern health regimes, medieval health was above all an ecological
struggle against a diverse host of infectious pathogens; social inequality was probably also an
important contributing factor.
Limitations: Methodological assumptions and use of proxy data mean that only approximate
modelling of prevalences is possible.
Suggestions for Further Research: Progress in understanding medieval health really depends upon
understanding ancient infectious disease through further development of biomolecular methods.
Key words: plague; infectious disease; infant death; tuberculosis; DALYs; medieval health; Global
Burden of Disease
2
1. Introduction
When asked to identify the greatest health problem of the Middle Ages, most people –
scholars and public alike – would probably name bubonic plague. The Black Death (1347-9 in Britain;
1346-53 in Europe) was the most notorious epidemic in history; when it struck, it killed between a
third and a half of the people of Europe. This initiated the Second Pandemic, a relentless series of
smaller plague epidemics which kept much of Western Europe in demographic stagnation for
over three centuries (Benedictow, 2004; Platt, 1997). In many standard histories of the Medieval
period, bubonic plague is virtually the only health condition mentioned. But what was medieval
health really like? If, like some medieval saints, you could shield the population from specific
sicknesses or harm, what would you target?
What was the biggest health problem of the Middle Ages? Modern public health specialists
consider questions of this kind frequently, and they are important questions. For modern health,
they allow planning of public health strategies. For ancient times, such questions are not merely
historical trivia. One goal of both bioarchaeology and history is to understand how people lived –
what their experience of health and illness was and what life risks and problems they suffered.
Moreover, asking such questions may shed new light on the real impact of disease on society and
history. It goes without saying that this is a challenging problem to study. It requires unaccustomed
methods, and it gives an unaccustomed kind of answer, a tendentious thought experiment rather
than a positive differential diagnosis or significant p-value. But it is worth attempting, if only to
expand our imagination and to understand conditions of medieval life.
In this paper, we attempt to estimate the relative consequences of various health issues in
medieval society, focusing particularly upon England between 1200 and 1500. To do so, we adapt
the method of estimating Disability Adjusted Life Years employed by the World Health Organisation
and the Global Burden of Disease project (Mathers, 2017; Murray and Lopez, 1996) . This
methodology involves many assumptions and estimated parameters when applied to modern
populations; applying it to medieval populations involves even more. We end up triangulating
between various sources of data, all problematic in their own ways. But, just as with modern
populations, this method may supply a general order-of-magnitude answer that can give a broad
outline of conditions and provoke new research questions about the history of health.
2. Medieval health: the setting
The conditions of health in High and Late Medieval England are well-known and can be
summarised briefly (See Dyer, 2002 for overview) (although we focus here upon England between
1200 and 1500, the analysis and discussion are broadly applicable to later medieval Europe as a
whole). Before the Black Death epidemic of 1348-9, England had an overall population of between
three and five million people; after the Black Death, the population fell to around 50-60% of that
level (Broadberry et al., 2015). More than 90% of the population lived in villages or in the
countryside rather than in towns, and most of them worked as agricultural labourers, either on their
own holdings or as tenants of a manor. There was a small class of urban craftspeople and merchants,
who may have lived in somewhat different health conditions (Rawcliffe, 2013), and an even smaller
upper class, with fewer than 1000 knights in England and a few dozen aristocrats; 2-5% of adult
males were members of the clergy.
The population was relatively young, with a high birth rate and high subadult mortality.
Medieval demography involves complex reconstruction, as the earliest systematic records of births
3
and deaths date to the 16th century, but estimates for life expectancy at age 25 based on sources
such as tenancy records range from 20 to 25 years (Jonker, 2003; Razi, 1980). The population
reached its peak during the climatic optimum of the 12th and 13th centuries, and some scholars have
suggested that England was near its carrying capacity around 1300, with widespread conversion of
woodland and pasture to farmland (Campbell, 2016). Landowners extracted surplus from the
working poor in the form of labour, produce and animals, and the church extracted tithes as well.
Much of the population lived in or close to poverty. As the climate worsened into the Little Ice Age,
the 14th century was dogged by famines following crop failure. Perhaps 10% of the population died
during the catastrophic Great Famine of 1315-1317 (Slavin, 2019), as exacerbated by outbreaks of
other human and animal diseases. Such conditions may have exacerbated the death toll of the Black
Death (DeWitte, 2015).
Conditions were favourable for the spread of diseases of all kinds. Both in towns and in rural
areas, people lived in crowded housing, with entire families often living in one or a few rooms. Close
contact with livestock was part of everyday experience. Particularly in towns and cities, water
sources were often unclean, and sanitary facilities were minimal. Very few people anywhere had
access to formal medical care . The few educated physicians in society were normally based in
towns or attached to clerical institutions or wealthy patrons; their learning and interests were often
principally theoretical in any case. To alleviate suffering, most people turned to experienced
laypeople within their own communities dispensing traditional remedies; their services ranged from
extracting diseased teeth to providing herbal medicines, prayers and charms (Rawcliffe, 1995).
Understanding medieval health involves triangulating between several sources, all
informative but problematic in some ways.
● Modern epidemiological studies of health in less developed regions gives a rough sense of
some of the problems historic populations may have faced, both in their range of health
problems and in the suffering these cause. However, they cannot be read as simple proxies
for medieval health, both because they often concern groups living in very different social
and environmental settings and because contemporary factors such as antibiotics, public
health knowledge and anaesthetics have a worldwide reach, spanning “developed” and “less
developed” regions.
● The history of medicine helps us to understand what health problems medieval people
recognised, how they understood them, and what interventions were available to address
them (Rawcliffe, 1995) , but provides little quantitative evidence about actual health
conditions.
● Archaeology has revealed exceptional sites such as the Smithfield (London) plague pits
where thousands of victims of the Black Death were buried in mass burials (Grainger et al.,
2008), but it provides little information on less dramatic health circumstances.
● Historical sources provide abundant evidence of exceptional health events, notably
epidemics, but far less on everyday conditions. For Britain, the earliest systematic health
records date to the 17th century; extrapolating such records back to medieval times requires
caution (Mitchell, 2017). Before then, historical sources afford mostly anecdotal evidence for
specific complaints and causes of death (for instance, we know that dysentery was a common
problem for armies on military campaigns, but we know almost nothing about how often and
in what ways it affected ordinary people). The more systematic textual sources from the
period 1200–1500 all relate to specific types of deaths or conditions that were described in
texts, such as suicide, epilepsy or insanity, particularly those that attracted legal attention
such as coroners’ inquests into sudden or unnatural deaths (Butler, 2015; Hanawalt, 1986;
4
Seabourne and Seabourne, 2001). How accurate and representative of society as a whole they
are is unknown and it is difficult to draw concrete prevalence data from them.
● Palaeopathological study of skeletons provides essential information (Aufderheide, 2011;
Buikstra, 2019; Waldron, 2009). However, besides interpretive complications from ever-
present sample biases and the “osteological paradox” (Wood et al., 1992),
palaeopathological study of skeletons is informative only for conditions which leave traces
upon bone. This includes many traumatic injuries and degenerative conditions and some
important chronic diseases such as tuberculosis and leprosy, but it omits most infectious
diseases, many cancers, and metabolic, neurological and cardiovascular conditions. Even
conditions which leave skeletal signs may leave ambiguous evidence; for instance, less than
10% of tuberculosis cases involve bony changes. Ancient DNA studies are beginning to
transform the history of infectious disease by revealing pathogen DNA directly. The most
notable example is Yersinia pestis (bubonic plague) itself (Bos et al., 2011 and subsequent
literature), which leaves no skeletal traces and is only now yielding information on plague
victims not buried in plague pits (Cessford et al., 2021). However, such studies have not yet
been done systematically enough to provide prevalence of such conditions.
Piecing these sources together to draw an overall picture of medieval health is challenging,
but they outline several important parameters. First, medieval people suffered from a broad gamut
of health problems (for an important synthesis, cf. Roberts and Cox, 2003) including infectious
diseases, cancers, sudden deaths probably related to strokes and coronary problems as well as a
wide range of accidental and trauma. Palaeopathological evidence shows conditions such as
tuberculosis, leprosy and traumatic injuries to have been common. Secondly, infectious disease was
probably the most important general single category of killer. Here, the 17th-19th century London
Bills of Mortality (Marshall, 1832; Millar, 1759) may give a proxy. They portray a densely crowded
urban population dating to several centuries after the medieval period that is the focus of this
article. Their latter half represents the early industrial era, rather than predominantly rural,
agricultural population of medieval England; the effects of pollution, hunger and industrial work may
be greater. The Bills of Mortality were only records of what the people of the time chose to record as
cause of death. They were restricted to medical diagnoses contained within their understanding of
health and disease at that time, and the people compiling them were unaware of many of the
conditions we think of today as causing death. Furthermore, a physician may never have been
involved in recording choosing the cause of death. Importantly, such past social diagnostic labels do
not directly equate to our modern understanding of disease, even if the terms used may appear
similar (Mitchell, 2011). Our observations nevertheless outline several important parameters. They
give a rough order-of-magnitude picture of what killed many people before the health advances of
the later 19th-20th centuries, and particularly before the era of antibiotics, vaccinations and public
health campaigns. Other than old age and infant and child deaths (themselves often due to
infectious disease, but usually categorized in the Bills of Mortality by the age of the victim), past
terms used that we think probably refer to infectious diseases occupy all of the top places for cause
of death. Not counting plague, infectious diseases cumulatively account for at least half of all deaths.
Among them, tuberculosis is probably the single most important disease, while other plausible
causes include dysentery, cholera, whooping cough, measles, scarlet fever, malaria, chest infections,
and smallpox. In historical London, all of these categories caused at least twice as many deaths as
the next most significant cause (abortive or stillborn foetuses). Even given the imprecision of
identifying specific medical causes from such historic records, and the different living conditions
between medieval and early modern times, it seems likely that acute infectious disease would have
claimed at least half of medieval people. In contrast, palaeopathologically-identifiable conditions
5
such as leprosy, syphilis and fractures rank relatively low. Modern killers such as cancers, diabetes
and coronary disease were likely not conditions they were in a position to understand or diagnose,
and may be hidden under “aged”. Thirdly, infants and children were probably the commonest
victims. In the Bills of Mortality, for instance, their deaths are simply listed under the age of death
(“Chrisomes” [children dying within a month of baptism, “Infants”, and “Teeth[ing; e.g. children
dying around the age of teething]”). Most of these were probably caused by infectious diseases, but
obstetric problems were also significant. It is hard to believe that the outlook for the very young
would have been much better in the medieval period.
Table 1. Causes of death listed in London Bills of Mortality (1657-1758) (Millar, 1759)
1
. Causes which
describe only non-specific symptoms ascribable to many causes (such as “convulsions”) have been
omitted, and complaints of similar aetiology which are often listed together have been combined
under general rubrics. “Reported causes” is as given at the time, and “possible modern medical
interpretation” is approximate only. The 20 commonest causes are listed, as well as some less
frequent ones for purposes of comparison.
Rank
Reported Causes
Possible modern medical interpretation
Combined
percentage
of
recorded
deaths
1
Consumption and Tissick
Tuberculosis and other pulmonary
infections
15.50
2
Fever, Malignant Fever, Scarlet
Fever, Spotted Fever, Purples
Varied bacterial and viral infectious
diseases
9.30
3
Flox, Smallpox, Measles, Chicken
Pox
Smallpox, measles, chicken pox and
related viral diseases
7.44
4
Aged
Varied degenerative conditions of aging
6.98
5
Chrisomes, Infants, Teeth
Varied causes of death in newborns,
infants and children up to the age of
teething, probably mostly from infectious
diseases
5.95
6
Colick, Gripes, Wind, and Twisting
of the Guts
Varied infectious gastroenterological
conditions
5.21
7
Ague, Calenture and Fever
Varied fever-related infectious conditions,
including malaria
5.03
8
Dropsy and Tympany
Conditions where fluid collects in the
chest and thorax
3.91
9
Plague
Yersinia pestis (the 1665 epidemic, the
only one during the period for which there
are Bills of Mortality)
3.02
10
Abortive and Stillborn
Death of foetus before or during delivery
2.61
11
Childbed and Miscarriage
Death of mother during childbirth
1.08
12
Rickets
Rickets
1.01
13
Asthma and Tissick
Asthma and breathing diseases
0.75
1
Data were recategorized from figures provided online by https://www.curiousgnu.com/yearly-bills-of-
mortality-1657-1758)
6
14
Apoplexy and Suddenly
Collapse and sudden death, potentially
including strokes, heart attacks and
pulmonary embolus
0.66
15
Jaundice
Jaundice
0.43
16
Gangrene, mortification,
putrefaction
Gangrene and related necroses
0.42
17
Rising of the Lights
Pulmonary diseases -- asthma,
emphysema, pneumonia, etc.
0.41
18
Tissick
Chronic cough, presumably from chronic
pulmonary conditions
0.41
19
Quinsy, canker, sore mouth,
thrush
Various infections of the mouth and
throat
0.41
20
French pox and other venereal
diseases
Venereal diseases
0.34
23
Cancer and related complications
Cancer and related complications
0.28
26
Accidents
Accidents
0.25
27
Bleeding, Bloody Flux, Scowring,
and Flux
Dysentery, typhus, and similar complaints
0.24
28
Stone, gravel, and strangury
Kidney-stones and related problems
0.24
29
Chincough, Cold, Cough, Hooping
Cough
Colds, influenzas, whooping cough
0.23
30
Melancholy, grief, frenzy, hysteria,
vapours, megrims
Mental illnesses and perhaps related
neurological conditions
0.21
31
Sores, ulcers, bruises, broken
limbs
Infected injuries
0.20
35
Suicide
Suicide
0.13
36
Gout
Gout
0.13
39
Scurvy
Scurvy
0.11
42
Violence
Violence
0.07
43
Executed
Execution
0.07
46
Enlarged liver
Liver disease
0.04
53
Fractures
Fractures
0.01
55
Leprosy
Leprosy (and some skin diseases?)
0.01
57
Diabetes
Diabetes
0.01
3. Assessing the “burden of disease”
The most widely recognised methodology for assessing the impact of disease on a
population is that employed in assessing the “Global Burden of Disease”, used to inform World
Health Organisation (WHO) and governmental policies. This methodology converts disease patterns
into DALYS (Disability Adjusted Life Years), which measure how much life and life quality is lost to a
particular illness. DALYs involve two components: Years of Life Lost (YLL) and Years Lived with
Disability (YLD) (Mathers, 2017). “Years of Life Lost” count how much of an expected lifespan has
been lost: if a health condition kills a person at age 30 when they otherwise would have lived until
70, they have lost 40 years of life. “Years Lived with Disability” estimates the quality of life lost. If a
person acquires a health condition which reduces their ability to function and enjoy normal life by
50% and which lasts for 20 years, they have an estimated 10 Years Lived with Disability. The “burden
7
of health” ascribable to a specific condition is simply the sum of Years of Life Lost to it and the Years
Lived with Disability it causes for all individuals within a particular population.
The “burden of disease” methodology has been criticised, both for various methodological
assumptions such as whether it takes into account interaction between health conditions and for
philosophical issues. For example, the method assigns weightings for how much a given disease
reduces a sufferer’s life quality (e.g. malaria is considered to reduce one’s life quality 19.1% and
thus is given a standard weighting of .191). Such weightings are based upon cross-cultural surveys
and regularly updated, but they assume that one can specify a uniform, cross-culturally valid
parameters for how much a particular ailment impacts someone’s ability to function socially (Metts,
2001; Nord, 2013). Even when applied to living populations, the GBD methodology is an approximate
tool for relative estimation, not an exact science. However, DALYs are useful in comparing how much
different kinds of conditions affect human life. For example, because their basic parameter is years
of life, they highlight how long-term conditions which affect childhood have a greater effect on a
group’s aggregate health than conditions which affect old age. They also highlight the cumulative
effect of long-term, chronic, often invisible conditions such as back pain and depression. They help
to draw a broad approximate picture of an overall healthscape in a way which a more exact study of
individual conditions cannot.
Applying the DALYs method to ancient health is fraught with challenges. The only previous
use of this methodology in bioarchaeology (Stodder, 2016, 2017) uses it for different purposes than
we do here. Stodder points out that the DALYs approach helps to highlight the human cost of
medical conditions. Conceptually, she uses it as a tool for building knowledge outward from
palaeopathology; she considers conditions which can be inferred from skeletal evidence, and uses
DALYs to estimate their impact on a sufferer’s experience. She also aggregates YLD scores to
compare the burden of disease between communities. Here, we are interested in considering
historical health patterns more globally, and hence we consider health conditions not evidenced
skeletally as well. Calculationally, we also take a different route. As Stodder (2016) points out, we
usually cannot know at what point in a human life an injury or medical condition arose and how long
the sufferer would have lived without it; she thus opts for a prevalence-based snapshot of health.
We start from aggregate prevalences as well, but try to develop some expectation of how health
may have varied over the lifespan (see below). Both approaches require assumptions and
extrapolations, as does the DALYs method itself when applied to living populations; both methods
provide valid ways of investigating rather different questions. Here, as a conceptual experiment, we
think it is worth grasping the nettle of methodological assumptions and seeing what can be learned
about medieval health. We acknowledge the limits of the approach freely: virtually all of the
necessary parameters can be estimated only in the most approximate way, based on general
demographic, medical, or historic comparisons (see below). As with all models, there is the risk of
“garbage in, garbage out”, although we think the results are reasonably robust and not sensitive to
small to moderate differences in modelling. The exercise should not be understood as attempting to
model specific conditions with true precision; it attempts to sketch in a broad panorama of health, a
useful, order-of-magnitude heuristic for exploring the historical burden of disease.
3.1. Identifying relevant health conditions and their parameters
To begin with, what health conditions are relevant, and how did they behave? Here we have
to mosaic different sources. For some conditions, palaeopathology gives us reliable information
about prevalences and medical consequences; for others, we must rely upon historical information,
8
comparisons with later periods, or modern comparisons. As we cannot model every possible cause
of death, we work here with a list of about twenty conditions which are known to have affected
medieval populations and which have life-changing effects, as well as a few which are of particular
interest to historians and palaeopathologists. We do not include ailments which are known to have
existed but remain medically unidentified (such as “sweating sickness”). We include a few non-
medical conditions – warfare, famine and social inequality, as these contributed to poor health and
mortality.
Infectious diseases are a challenging category to model. As noted above, they were probably
the most significant general cause of death, and any overview which omits them would be seriously
distorted. At the same time, there is little solid information to work with. Few of them leave skeletal
markers. Modern epidemiological statistics are strongly influenced by successful campaigns of
prevention over the last two centuries which have eliminated or limited deadly scourges such as
smallpox, measles, whooping cough, scarlet fever, cholera, and (for some parts of the world)
tuberculosis. Before the 18th century, historical data is scanty and is limited by our ability to equate
historical disease terms with modern pathogens. Here, we use London Bills of Mortality simply to
gauge the relative significance of fuzzy, umbrella categories which group multiple pathogens with
similar symptoms together.
To apply the GBD methodology, for each condition, the key parameters are:
a) Its overall prevalence in the population
b) What proportion of cases result in the sufferer’s death? What proportion of cases result in
disability?
c) How much disability does it impose upon the sufferer? Is this lifelong or for a limited
duration?
d) How does it affect people of different ages?
Grounds for estimating prevalence and other parameters are given in Table 2, and specific
parameters are given in Tables 3 and 4. Parameters (a) and (b) are estimated differently for each
condition according to a combination of skeletal, historic and modern comparative sources (Table 2).
Parameter (c) is derived from weights assigned to each condition for the 2004 Global Burden of
Disease study (World Health, 2004), with adjustments for medieval settings or when a GBD
parameter is not available (Table 3). For parameter (d), each condition was assigned to a template
weighting its effects for different age categories, based upon modern medical knowledge and the
same sources as (a) and (b) (Table 4). Attempting to model the age-related incidence of each
condition in detail would go beyond the resolution of the available information, and in Table 4 we
simply provide a rough categorisation of whether each one affects all ages uniformly, or affects sub-
adults, adults, or older people preferentially.
These parameters represent the overall population of medieval England. Of course, as
skeletal data show, some conditions have different prevalences for males and females. Moreover,
death in childbirth is specific to women, and death in combat principally affected males. However, as
the goal here is to represent the overall population, we here use a mid-sex mean for such conditions.
Similarly, the prevalence of episodic conditions is averaged over the overall time span; for instance,
if an epidemic disease struck every 10 years, and when it did, it killed an average of 10% of the
population, its overall prevalence per year would be 1%.
9
Table 2. Assessing historical prevalence for the conditions modelled. Information on contemporary
causes of death in different regions and economic strata is from WHO estimates
(https://www.who.int/healthinfo/global_burden_disease/estimates/en/). Bills of Mortality data are
based upon the data cited above (Table 1). Skeletal data come from Roberts and Cox (2003) and
from ongoing results from the “After the Plague: Health and History in Medieval Cambridge” (ATP)
project (Cambridge University).
Condition
Notes
Infant and child death
In modern developing areas, infant and childhood mortality (due largely
to infectious diseases such as pneumonia, malaria, diarrhoea, measles,
but also to obstetric issues and other causes of infant death) comprises
15-20% of people between 0-5 years (Chao et al., 2018); it would likely
have been higher in premodern contexts with less medical knowledge
and treatment. In Bills of Mortality, explicitly labelled deaths of infants
and children up to about 3 years comprised ca. 5-10% of deaths in
historic London, but this is a considerable underestimate; “convulsions”,
for instance, the most frequently named cause of death (22% of all
deaths), refers principally to deaths of very young children. Infant death
in medieval populations was high, particularly from infectious disease in
urban settings (Lewis and Gowland, 2007). Here, we estimate,
conservatively, that ca. 20% of the population died before age 10.
Tuberculosis and other
bacterial pulmonary
infections
Tuberculosis is very common in British history (Roberts, 2002). The WHO
figure in low-income economies today is 4.2% of deaths, though this
reflects the impact of medical care, particularly antibiotics. Bills of
Mortality suggest that 10-20% of deaths in historic London may have
been related to “consumption” (mostly tuberculosis, though possibly
including some other pulmonary infections); by 1913 this had dropped to
0.3%. Both WHO figures and Bills of Mortality do not include additional
cases which did not cause death. Of tuberculosis cases, only a fraction
have bony involvement (estimates range from 3-5% (Jaffe, 1972) up to
12% (Roberts and Manchester, 1995; Roberts and Bernard, 2015).
Prevalences in archaeological samples vary from 0 to 21% according to
both the sample and the methods used (Roberts and Cox 2003: 232;
11.5% ATP data); the true prevalence would have been many times
higher, though not all of these cases may have caused death. Here we
estimate that about a third (30%) of medieval people may have had TB,
and it would have been fatal in 30% of these cases. Its GBD weighting is
.271.
Fevers of various kinds
This includes scarlet fever, formerly an epidemic disease with up to 25%
mortality, as well as many other, less distinguishable fevers;
cumulatively, in Bills of Mortality, such conditions accounted for 9.3% of
recorded London deaths. Prevalence is estimated here at 25% and
mortality at 25% of these cases, with a GBD weighting of .23.
Diarrheal and
gastrointestinal infections
Including dysentery and cholera as well as other diseases, these remain
an important cause of death in modern low-income countries,
accounting for 7.2% of deaths today. In Early Modern England, stomach
and intestinal infections accounted for 5.2% of deaths. Prevalence is
estimated here at 25% and mortality at 25% of these cases, with a GBD
weighting of .105.
10
Viral pulmonary infections
The principal pathogens here are measles, smallpox, cowpox,
chickenpox, and similar viruses spread through coughing. Only measles
remains an important modern pathogen today, but these accounted
cumulatively for 7.4% of London deaths (this may also include some
deaths from typhus, a bacterial condition which also results in a spotty
rash). Prevalence is estimated here at 25% and mortality at 25% of these
cases, with a GBD weighting of .152.
Hansen’s Disease (Leprosy)
Leprosy (Hansen’s disease) is known to have been more common in
medieval England than in the 17th-18th centuries , as well as culturally
important (Rawcliffe, 2006; Roberts, 2020). The highest prevalence today
does not reach 1% (WHO: .578% for India in 1983). Skeletally, it is
present in 3.5% of medieval English samples (Roberts and Cox 2003:271),
though this sample includes several leprosaria; ATP data are 1.7%. As
many cases would not have skeletal changes (Roberts and Manchester
2005), a prevalence of 10-20% (here: 10%) is plausible; most affected
would have lived among the general population and displayed few
symptoms. It can cause extensive disability and may contribute
eventually to risk of death but is rarely lethal in itself (estimated 2% of
cases). Its GBD weighting is .15.
Malaria
Malaria contributes 4.6% of deaths in modern low-income countries, but
many of these are in tropical or sub-tropical environments. It was
formerly important in Europe, including England. About 5% of London
post-medieval deaths were attributed to “agues, fevers and calentures”.
Gowland and Western (2012) do not specify prevalences, but argue that
malaria was endemic in low-lying, coastal and marshy areas of England,
including some densely populated areas; Dobson (1980) shows that in
marshy areas it contributed significantly to mortality. Malaria is modelled
here as having a prevalence of 5% (averaged over all areas of the
country), a 25% lethality, and a GBD severity weight of .191.
Plague (yersinia pestis)
Here we consider only the prevalence of lethal cases; medically, 70-90%
of cases would have been lethal, and no medieval figures are available
for people recovering from plague. Our DALYs estimate thus does not
include Years Lived with Disability for plague recoverees, but this would
not be a large figure, as it did not confer lasting disability. The death toll
from the Black Death in England was ca. 50% (Dewitte and Kowaleski,
2017). Later epidemics typically had 5-10% mortality, with outliers of 15-
20% (1362, 1665); many were localised and statistics may refer to their
epicentre rather than nationally. In the period studied here (1200-1500),
there were no plague outbreaks before the Black Death in 1348. Here we
estimate one outbreak every five years between 1350-1500, with an
average nationwide mortality of 10%, plus outliers of 50% mortality
(1348-9) and 15% mortality (1362, 1665). These figures, probably erring
on the high side, yield an average mortality of 12.5% during 30 plague
years and 1.25% over all 300 years.
Cancers
Cancer prevalences are conjectural. Bills of Mortality ascribed less than
1% of deaths to cancers, though the term was not used as it is today. In
modern populations, cancers are a major cause of death only in high-
income countries; they do not amount to 1% of death in low-income
countries. Palaeopathological data gives rates of 0.17-1.79% for various
forms of cancers (Roberts and Cox 2003:280-1). When skeletal remains
are systematically screened radiographically, the prevalence is
11
considerably higher (10-15% in adults (Mitchell et al., 2021)) and cancers
may have been much more frequent in medieval times than usually
supposed. However, this is based upon adults only; prevalences in the
overall population would be lower depending upon the population age
structure, and their lethality is unknown; they may also have caused
significant chronic morbidity. For medieval populations, somewhere
between 1 - 3% of deaths would be reasonable. Here, we consider only
lethal cancers and model their overall prevalence in line with WHO
figures for developing countries as 1%, affecting older individuals
preferentially.
Dental disease
Dental disease (including caries, ante-mortem tooth loss, periodontal
disease, and abscesses) is so common as to be normal, both today and in
medieval times. 52.6% of medieval samples have dental caries, and
26.7% have abscesses (Roberts and Cox 2003:258-260). Here we
estimate that 60% of adults and 30% of the overall population suffered
from dental disease in some form. Dental disease is rarely lethal unless
sepsis develops, but it can contribute to risk of death (Li et al., 2000) and
this would have been much more common before the availability of
antibiotics (here we estimate 1% lethality). The GBD severity of .081
is probably too low to represent the era before effective dentistry and
anesthesia, and we have increased it to .150.
Gout
Gout rarely causes death today; 0.13% of deaths in London Bills of
Mortality were ascribed to it. Worldwide prevalence data is not very
useful, as gout relates strongly to diet and environment. Skeletal and
other evidence suggest that a range of 10-30% for older males, and an
overall population prevalence of 1-4% (ATP data: 2% for all adults), is
plausible for medieval England. Here gout is modelled as affecting 1% of
the total population. It can cause significant pain and disability (GBD
severity weighting .132).
DISH
DISH (Diffuse Idiopathic Skeletal Hyperostosis) is included here as a
condition of interest to palaeopathologists. Little is known of its possible
prevalence in non-modern, non-Western populations other than through
skeletal research. Palaeopathological data show a range of 0.2-11.1% in
medieval samples (Roberts and Cox 2003:246; ATP data 9.3%). Based
upon palaeopathological data, a prevalence of ca. 10-30% in males over
40 might be reasonable, with a limited occurrence in older women;
averaged over the population, a prevalence of 2-4% (here 3%) is possible.
DISH is almost never fatal; its severity here is estimated as minimal (.01),
as the GBD does not publish a weighting for it and it is largely
asymptomatic.
Rickets
Rickets (a childhood skeletal manifestation of Vitamin D deficiency) is
found skeletally in ca. 0.19-3.63% of medieval cases (Roberts and Cox
2003:247-248; 4.4% ATP data). In London Bills of Mortality, about 1% of
deaths are ascribed to rickets, but this reflects post-medieval urban
conditions (limited exposure to sunlight, for instance). Here Vitamin D
deficiency is modelled at 3%, preferentially affecting subadults, and
lethal in 20% of cases. The GBD does not assign a severity to rickets; in
light of consequences such as bowing of limbs, it is modelled here as a
chronic condition (5 year duration) at .3 severity.
Scurvy
Scurvy (Vitamin C deficiency) is of interest as a skeletally visible
deficiency disease. Its prevalence is strongly related to diet. Medieval
12
skeletal samples (ATP data) reveal prevalences of skeletal signs of scurvy
on the order of 5%; it may have been seasonal rather than chronic. It is
modelled here as of 3% prevalence, more likely to affect subadults, and
lethal in 10% of cases (commensurate with its overall toll in the Bills of
Mortality). The GBD does not assign a weighting to it, but .1 is
commensurate with other moderate deficiency disorders.
Strokes
Strokes are included here as a major cause of death in modern societies.
Prevalence of stroke in medieval people is conjectural. In modern low-
income countries, stroke accounts for ca. 5% of all deaths. Bills of
Mortality list “apoplexy and suddenly” as causing .66% of deaths, but this
may not correspond entirely with stroke, perhaps including coronary
conditions as well. Both sources only list lethal cases. Here, we model
stroke as having an overall prevalence of 4%, lethal in 50% of cases, with
the remainder causing long-term disability of .266 (standard GBD
weighting).
Back pain
Back pain is included here as it is a major contributor to DALYs in modern
populations. It is non-lethal and episodic, but can limit activity severely
(GBP weighting for back pain .061; for chronic intervertebral disk pain
.120). Prevalence data for medieval groups is conjectural, but much of
the population performed strenuous manual labour at times.
Palaeopathological data suggests that 48% of medieval adults (ATP data)
had at least one Schmorl’s node, indicating disk damage; Schmorl’s nodes
have been linked to back pain (Kyere et al., 2012; Mattei and Rehman,
2014), though many may have been asymptomatic and back pain has
many other causes. Here, we estimate a prevalence for episodic lower
back pain of 40% among adults (similar to today), spread uniformly
among adult age categories (20% among overall population), with a GBD
weighting of .061.
Osteoarthritis
Skeletally, osteoarthritis is one of the most common pathologies,
occurring to some degree in most older adults. Here we model it as non-
lethal, found in most adults over 40 (ca. 12.5% prevalence for the overall
population), and with no, slight or moderate experiential consequences
(weighting .05; GBD weighting of .126-.129 is for well-developed cases in
hip or knee, but most skeletally observed cases are less severe and may
have been minimally symptomatic).
Fractures
Fractures include traumatic injuries to bone of all kinds. Palaeopathology
provides the best guide to prevalence. In many medieval British
collections, trauma is seen in around 10% of adults (Roberts and Cox
2003:238.39-11.1%; over 20% in ATP data (Dittmar et al., 2021). It is
assumed that subadults had a similar prevalence. Overall prevalence
here is estimated at 20%. The great majority are healed, and fractures
and broken limbs are a negligible cause of death in the Bills of Mortality;
here, it is assumed that they will prove fatal in only 2% of cases, but may
cause lasting disability. GBD disability weights vary for different kinds of
fractures, and range from .1 to .44 while healing and from 0 to .47 for
lasting disabilities. Here, a relatively modest weighting of .05 represents
the fact that most skeletally observed fractures healed without
complication.
Childbirth
If each birth had a risk of maternal death of between 0.05% and 2% (1%
in ca. 1800: Chamberlain, 2006), and a woman gave birth to 5-8 offspring
during her reproductive years, a woman’s overall risk of death in
13
childbirth would be between 2.4-14.9%. A mid-range estimate of 1% risk
and 6 births gives a risk for women of ca. 5.8% -- very similar to historical
estimates of 6% (Podd, 2020 p. 127). As reproductive-age women made
up ca. 21% of the population, prevalence of death in childbirth averaged
over the population is estimated at 1.2%, a general level consistent with
Bills of Mortality figures and with historic data (Roberts and Cox
2003:252-254: 0.1-1.67% range).
Famine
“Famine” includes widespread hunger due to crop failure (rather than
chronic poverty). Here we consider only deaths due to famine, rather
than suffering due to episodes people survived. This is hard to estimate
with any certainty. Small localised famines occurred regularly, but would
have had lower death counts; widespread, severe ones were less
common (the 1315-1320 famine, with ca. 10-20% overall mortality
(Slavin, 2019) was exceptional). Here, we assume a more or less serious,
more or less localised famine every 10 years, resulting in 3% mortality
nationwide (if anything, this is probably an over-estimate); averaged over
300 years, this gives a prevalence of 0.3%.
War
The toll of war here includes only direct casualties, not collateral civilian
deaths, famine caused by pillaging, etc, particularly as much of England’s
medieval warfare happened abroad. Between 1200 - 1500, England was
actively at war in France, Scotland, Ireland, or in civil conflict
approximately every 2-3 years. Armies were small, with ca. 50,000 men
in the field at most, and normally many fewer. If one fifth of these were
dead or disabled casualties to injury or disease each year, annual
casualties of ca. 10,000 in a total population of ca. 2.5 million would give
a prevalence in war years of ca. 0.4%; at 40 war years per century, the
prevalence averaged over all years is 0.16%. It is assumed that war
casualties were predominantly young to middle-aged males, and that
half died, while half incurred long-term disability (severity weight .25, on
a par with long-term serious injuries).
Mental illness
The historical cost of mental health is entirely conjectural, but we include
it here because it is a major contributor to DALYs today. Medieval people
recognised and understood mental health conditions differently to today,
but conditions such as “melancholy” are attested, as are things perhaps
corresponding to manias, schizophrenia, anxiety and neurological
problems. In the London Bills of Mortality, 0.21% of deaths are ascribed
to “melancholy”, “grief”, “frenzy”, “vapours”, “hysteria”, “vertigo”
and/or “megrims” [migraines]; a further 0.13% are ascribed to suicide
(total 0.35% for mental illness and suicide). If we estimate that only 5% of
cases were severe enough to result in death or suicide, this might yield
an overall prevalence of perhaps 7%. Here, for purposes of discussion,
mental illness is modelled at 7% prevalence, with 5% lethality and a two-
year illness of .35 severity (WHO weight for moderate depressive
episode) in non-lethal cases.
Social inequality
Social inequality is a major contributor to health problems, both today
and in historic societies. The effect of social inequality upon medieval
health is impossible to estimate with any certainty, and here we hazard
an exploratory guess merely to underline its significance. We assume
that (at least) 10% of the population lived in poverty resulting in periodic
lack of food, shelter, clothing, cleanliness, or safety severe enough to
expose them to additional risks and compromise their immune system’s
14
ability to ward off infection. We assume that dire poverty was a lifelong
condition affecting all ages equally. Inequality, malnutrition and poverty
would have amplified many other risks. However, for simplicity, we treat
poverty here as a free-standing social disability on a scale with a
moderate illness, injury or chronic, painful condition (severity weighting
.15).
Table 3. Health conditions modelled and their parameters. Age pattern is modelled as affecting all
ages uniformly unless there is a major medical pattern to the contrary. For conditions which heal,
duration is estimated as a discrete episode; for lifelong conditions, duration is estimated as life
expectancy at the age incurred. Disability weights are from GBD 2004 (World Health, 2004), unless
otherwise noted in Table 2.
Condition
age
pattern
overall
prevalenc
e
% lethal1
non-
lethal
duration
Disability
weightin
g
(severity
for non-
lethal
cases2
Infant and child death
sub-adult
0.200
1.00
Tuberculosis
uniform
0.300
0.30
5.00
0.271
Fevers
mostly
sub-adult
.25
.25
1
.23
Diarrheal and GI infections, including
dysentery, cholera, and others
uniform
.25
.25
1
.105
Viral pulmonary infections, including
measles, smallpox, and others
uniform
.25
.25
1
.152
Leprosy
uniform
0.100
0.02
Lifelong
0.15
Malaria
uniform
0.050
0.25
1.00
0.19
Yersinia pestis
uniform
0.0125
1.00
Cancers
old adult
0.010
1.00
Dental disease
old adult
0.300
0.01
2.00
0.15
Gout
old adult
0.010
0.10
5.00
0.132
DISH
old adult
0.030
0.00
Lifelong
0.01
Rickets
sub-adult
0.030
0.2
5.00
0.30
Scurvy
uniform
0.030
0.1
1.00
0.10
Strokes
old adult
0.040
0.50
Lifelong
0.266
Back pain
adult
0.200
0.00
5.00
0.061
Arthropathies
old adult
0.125
0.00
Lifelong
0.05
Fractures
uniform
0.200
0.02
1.00
0.05
Mental illness
uniform
0.070
0.05
3.00
0.35
Childbirth
middle
adult
0.012
1.00
Famine
uniform
0.003
1.00
War
middle
adult
0.0016
0.50
Lifelong
0.025
15
Social inequality
uniform
0.100
0.00
Lifelong
0.15
1 % of cases resulting in disability is equal to (1 - % lethal). 2 Disability weight is a number between 0
(no effect upon lifestyle) and 1.0 (death).
Table 4. Weightings assigned to age categories (figures are arbitrary weights which give the
relative probability of incurring a condition at each stage of the lifespan).
Age pattern
0-10
10-20
20-30
30-40
40-50
50-60
60-70
70-80
80-90
90-100
Uniform
1
1
1
1
1
1
1
1
1
1
Sub-adults only
2
1
0
0
0
0
0
0
0
0
Sub-adults mostly
5
3
1
1
0
0
0
0
0
0
Adults-only
0
.3
1
1
1
1
1
1
1
1
Middle-aged mostly
0
1
5
5
5
2
0
0
0
0
Age-progressive
0
0
1
2
3
4
4
4
4
4
Older adults only
0
0
0
0
1
3
3
3
3
3
3.2. Calculational methodology
In calculating DALYs, the first step is to model the demographic regime of medieval England,
as the lifetime effects of illness and injury depend upon when it strikes and what the victim’s life
expectancy is at that point. The demographic regime is calculated here using model life tables. Coale
et al.’s West model, level 3 is used; this is a general-purpose life table based on varied sources from
Western Europe, but with much higher infant and child mortality than modern Europe (Coale et al.,
1983, appendices p. 43). In this model, life expectancy at birth for females is 25.0 years, for males,
22.8 years; life expectancy at age 20 for females is 31.3 years, for males 29.4 years. Here male and
female values are averaged, and age categories are summed into 10-year intervals. As standard for
life table modelling, this is calculated for a stationary population of 100,000 births and deaths per
year. This is used to derive an overall age distribution and life expectancies at each age (Table 5).
Table 5. Demographic parameters of modelled population.
Age category
0-10
10-20
20-30
30-40
40-50
50-60
60-70
70-80
80-90
90-
100
Life expectancy
21.4
36.9
30.3
25.0
19.9
14.8
10.0
6.2
3.5
1.7
Proportion of
population
0.235
0.197
0.172
0.142
0.110
0.078
0.044
0.015
0.002
.000
N in stationary
population of
100,000
23553
19777
17243
14218
11079
7884
4486
1567
185
3
The next step is to apportion disease to people within age classes. Estimating real age-
specific incidences is not possible with the information at our disposal. Instead, we follow a simple
proxy method here. Using the overall prevalence of each disease or health problem (Tables 2, 3), we
calculate how many cases of it might be expected in a population of 100,000 people. Then we
16
distribute these cases among the age categories according to the population’s age distribution
(Table 5). This distribution is weighted according to how the disease affects people of different ages
(Table 4). For example, if a condition afflicts people of all ages equally (the “uniform” age pattern),
cases are allocated to each age category in direct proportion to how many people are in it, so that
the prevalence for each age category is the same. If a disease affects “sub-adults mostly”, cases are
allocated to each age categories so that the prevalences of 0-10, 10-20, 20-30 and 30-40 are in
proportions of 5, 3, 1 and 1. In the resulting distribution, the total number of cases across all age
categories sums to the overall population prevalence, and the relative prevalences in each age
category are in proportion to how the disease affects people of that age.
Using the number of cases per age interval, Years of Life Lost for any age group is calculated
as:
YLL = (total number of cases of a condition in the age group) x (proportion of cases which
are lethal) x (standard life expectancy in an age group)
For each age group, Years Lived with Disability is calculated as:
YLD = (total number of cases of a condition in the age group) x (percentage of cases which
are not lethal) x (duration of resulting effects) x (severity of resulting effects).
Next, both Years of Life Lost (YLL) and Years Lived with Disability (YLD) are summed across
age categories to yield a population total. Finally, YLL and YLD are added together to obtain a global
estimate of DALYs (Disability Adjusted Life Years).
As with the Global Burden of Disease analyses for modern populations , modelling
illness in this way involves simplifications and assumptions. For example, the methodology
assumes that a condition either kills a sufferer or afflicts them; cases where a disease first afflicts
somebody for a long period and then kills them (which is common, for example, with tuberculosis)
are simply counted as instantly lethal. It is assumed that the proportion of cases which kill the
sufferer is constant across age categories, as is the severity weight assigned to a disability. The
methodology takes no account of interactions between conditions, and it also assumes that a non-
lethal illness does not bias life expectancy (e.g. that people who become disabled at age 30 will live
as long as people who do not). The question raised by such approximations is not whether they
simplify real-world health; they undoubtedly do. It is whether adding further complexity to the
model to take them into account would incur more analytical liability in terms of further
assumptions than it would solve in terms of accuracy, and whether such considerations change the
major outlines of the results. As modellers often say, all models are wrong, but some are usefully so.
4. Results
Though we have tried to be systematic and informed by evidence, this is a thought
experiment or range-finding exercise; it does not aim to provide precise results . But it is the
first attempt to use DALYs to answer real questions about historical health, and even such an
exercise can usefully sketch the general outline of things. The general picture which emerges is both
intuitive and robust. How different conditions contribute to the overall burden of disease is marked
enough that even if specific conditions are modelled differently than we have here – if we model the
prevalence of tuberculosis at 40% instead of 50%, or of cancers at 5% instead of at 1%, or make
leprosy lethal in some cases, for instance – the major results remain similar.
17
So what was the biggest health problem of the Middle Ages? The answer is straightforward
(Table 6; Figure 1, Figure 2, Figure 3): infant and child death, followed by various infectious diseases,
of which the most significant single one was probably tuberculosis (cf. Waldron, 2009 90). This
picture is familiar to modern public health in developing areas, but it is radically different than the
images derived from either historical sources or palaeopathology. Plague (Yersinia pestis) was
important, but it was in a second rank of conditions, perhaps half as significant as the top killers.
When epidemics occurred, they killed many people; but no other epidemic was as lethal as the
1348-9 Black Death, and there were many years when plague was not present. Considering the later
medieval period as a whole, a person’s chances of dying from plague were vastly lower than of dying
from less visible, common infectious diseases which were always present. Other than old age and
infant and child deaths (themselves probably overwhelmingly due to infectious disease, but
categorized by the age of the victim), infectious diseases occupy all of the top places for cause of
death. Not counting plague, infectious diseases cumulatively account for around half of all deaths.
Among them, tuberculosis is probably the single most important one. But all of these categories
cause at least twice as many deaths as the next most significant cause.
In the ruthlessly utilitarian logic of the “burden of disease” methodology, chronic conditions
weigh more heavily than episodic conditions, and conditions affecting the young weigh more heavily
than conditions affecting the old, as they contribute disproportionately to both years of life lost and
to years lived with disability. Such conditions dominate the Years of Life Lost (Figure 2) and the
Disease-Adjusted Life Years (Figure 1). They are clearly what the WHO would target to improve
medieval life most cost-effectively. But to explore other dimensions, other measures may be more
relevant. For the misery inflicted upon sufferers who continued to form part of society, Years Lived
with Disability may be more relevant (Figure 3). If the goal is to improve life for survivors, rather than
to avert death, the targets shift entirely. The biggest factor is social inequality and poverty, which
highlights a hitherto undiscussed but important focus of research – health inequality in the past.
Other major contributors to the experience of poor health include leprosy, tuberculosis, fractures,
strokes, back pain and dental pain.
Table 6. Results (Years of Life Lost, Years Lived with Disability, and Disability Adjusted Life Years).
Results are modelled for a stationary population of 100,000 people, using age distribution and
disease parameters outlined in Tables 2-5.
Condition
YLL
YLL rank
YLD
YLD rank
DALYs
total rank
Infant and child death (varied
causes)
519782
1
0
519782
1
Tuberculosis
225492
2
28455
3
253947
2
Fevers
167486
3
4313
10
171799
3
Viral pulmonary infections
156592
4
2850
12
159442
4
Diarrheal and GI infections
156592
4
1969
13
158561
5
Leprosy
5011
17
36380
2
41391
6
Social inequality
0
20
37582
1
37582
7
Strokes
27580
9
7366
7
34946
8
Fractures
10022
12
24554
4
34576
9
Malaria
31318
6
713
14
32031
10
Yersinia pestis
31318
6
0
31318
11
Childbirth
31231
8
0
31231
12
18
Rickets
16079
11
3600
11
19679
13
Cancers
19093
10
0
19093
14
Mental illness
8769
13
6983
8
15752
15
Dental disease
5728
16
8910
6
14638
16
Osteoarthritis
0
20
11933
5
11933
17
Scurvy
8039
14
270
18
8309
18
Famine
7516
15
0
7516
19
Back pain
0
20
6100
9
6100
20
War
2082
18
521
16
2603
21
Gout
1379
19
594
15
1973
22
DISH
0
20
414
17
414
23
Table 7. Causes of mortality in England, top 20 causes of Years of Life Lost to disease: 1200-1500
(this study) and 2016 (Schmidt et al., 2020 Figure 2)
Rank
1200-1500
2016
1
Infant and child death (varied causes)
Ischemic heart disease
2
Tuberculosis
Lung cancer
3
Fevers
Stroke
4
Viral pulmonary infections
Alzheimer disease
5
Diarrheal and GI infections
COPD
6
Malaria
Lower respiratory infections
7
Yersinia pestis
Colorectal cancer
8
Childbirth
Breast cancer
9
Strokes
Self-harm
10
Cancers
Other cardiovascular
11
Rickets
Pancreatic cancer
12
Fractures
Prostate cancer
13
Mental illness
Esophageal cancer
14
Scurvy
Other neoplasms
15
Famine
Cirrhosis/ alcohol
16
Dental disease
Stomach cancer
17
Leprosy
Leukemia
18
War
Neonatal preterm birth
19
Gout
Congenital defects
20
Social inequality
Brain cancer
Table 8. Morbidity: Years lived with disability in England, 1200-1500 (this study) and 2016 (Schmidt
et al., 2020 Figure 5)
Rank
1200-1500
2016
1
Social inequality
Back and neck pain
2
Leprosy
Skin diseases
3
Tuberculosis
Migraine
4
Fractures
Sense organ diseases
5
Osteoarthritis
Depressive disorders
6
Dental disease
Anxiety disorders
7
Strokes
Falls
19
8
Mental illness
Oral disorders
9
Back pain
Asthma
10
Fevers
Other musculoskeletal problems
11
Rickets
Drug use disorders
12
Viral pulmonary infections
Diabetes
13
Diarrheal and GI infections
Bipolar disorder
14
Malaria
Osteoarthritis
15
Gout
Schizophrenia
16
War
Other mental disorders
17
DISH
Stoke
18
Scurvy
Autism spectrum
19
Upper respiratory infections
20
Other cardiovascular
20
Figure 1. The burden of disease: Disability Adjusted Life Years
Figure 2. The burden of disease: Years of Life Lost
21
Figure 3. The burden of disease: Years Lived with Disability
5. Discussion: The Four Horsemen Revisited
What do DALYs mean for historians of health? These results have several important
implications. Most obviously, we need to dethrone plague as the poster child for medieval
health problems. Plague is visible, dramatic and famous; but it is merely the tip of the iceberg.
Across the later Middle Ages as a whole, your risk of dying from infant and childhood diseases was
probably 15-20 times higher than your risk of dying of plague, and your risk of dying from any of half
a dozen endemic bacterial and viral infections was 4-8 times higher. Paradoxically, histories tend to
be silent on such diseases both because they are invisible and because they are omnipresent;
historical evidence bearing on infectious disease is almost entirely lacking, and it is often assumed
that they are simply always present at more or less constant levels, making them a static background
rather than the eventful figured ground of history. Palaeopathologists tend to work narrowly within
the purview of what can be seen skeletally. One value of this exercise is to highlight how medieval
health looks when evaluated within the same framework used to assess health in modern real-world
situations. Medieval people lived in a sea of pathogens which assailed them continually. Particularly
when their immune systems were compromised by hunger, poverty, or other illnesses, they died in
droves. Cumulatively, infectious diseases formed a force as omnipresent, powerful and invisible as
gravity, causing grief, moulding settlement patterns, deciding the fate of enterprises such as
campaigns of warfare, and shaping demographic regimes.
Fear is rarely commensurate with the mathematical realities of risk. The cultural logic of
health shows this clearly. The Four Horsemen of the Apocalypse of St. John of Patmos were Famine,
War, Plague, and Death. Consulting the numbers, medieval people should instead have feared
Infections, Infections, Infections and Death. The gap is not merely because they had no knowledge of
bacteria or viruses; if they did not know about Mycobacterium tuberculosis, they were certainly
22
familiar with dying from tuberculosis. Rather, they assumed that death is inevitable and out of
human hands – a realistic view, given their very limited ability to intervene effectively on most
serious medical conditions. Instead, they may have had defined health differently. Understanding
medieval health on its own terms would be an important project, but a different one than applying a
cross-cultural definition of health. Medieval health may have incorporated a lower benchmark of
what constituted physical well-being, and the spiritual was undoubtedly important; “health” may
have included one’s spiritual as well as physical health, and death was routinised and wrapped in
complex deathways. Medical conditions such as tuberculosis which allowed death to be foreseen
and coped with ritually caused pain and grief, but could be seen as part of the order of things. An
abrupt death that prevented one from having last rites and dying a Heavenward “good death” may
have been much more feared (Ariès, 1974; Binski, 1996). The same may be true for conditions such
as leprosy that were physically disfiguring and/or carried moral valences as well as health
consequences (Roberts, 2020).
In terms of overall health regimes, modern England provides an interesting, if not entirely
surprising contrast (Table 7, Table 8). Obviously, Tables 7 and 8 cannot be taken too literally: modern
and medieval figures are based upon entirely different classifications, methods and sources of data.
Modern data access much information we cannot know for the medieval period; conversely, the
modern GBD subdivides cancers and mental health conditions much more finely but does not
include categories such as “social inequality”, though perhaps it should. However, such comparisons
do reveal a macro-shift in the human world of health, from one dominated mostly by infectious
diseases to one dominated mostly by longer-term system breakdowns (heart disease, cancers,
neurological conditions of age) and social/mental health issues. The medieval regime probably
typifies a specific human world of health which contrasts with both the dispersed, mobile worlds of
hunter-gathers and the urban worlds of the last two centuries. The world of ancient and historic
civilisations, with high population density, widespread poverty among stressed populations, and
little sanitation or medical capacity, was dominated above all by an unrelenting battle between
humans and pathogens. If we have inherited robust immune systems today, we may owe them
largely to the cumulative selective filter of this historical epoch. The fact that infectious disease does
not figure largely in the “burden of disease” in the modern West shows how much our health world
has changed in the last century. The historic battle against pathogens has perhaps has now been
shifted to the proxy war of evolving pathogens and antibiotic/ antiviral resistance.
Turning to evolutionary implications, we do not mean to imply that epidemics such as the
Black Death may not have historic or evolutionary effects, but what these are may depend upon
other, structural factors much more than simple mortality. To use an ecological comparison, plague
epidemics are like wildfires: they sweep over a landscape dramatically in a very short period, leaving
scenes of devastation and traumatised populations behind them. Yet the landscape recovers, plants
evolve adaptations to burning or means of recolonising quickly, and a few years later there may be
very little sign that the wildfires ever happened – other than the adaptations embedded in the
organisms themselves. In contrast, conditions such as tuberculosis seem much more comparable to
climate change: they are omnipresent, they change gradually in small increments, and it is easy to
not notice or to disregard their impact, but they form a constant pressure which may shape the
course of change much more profoundly. Thus, for instance, how much economic havoc an epidemic
wreaks will depend as much upon the structure of the economic system as on the total mortality the
disease caused. The Black Death caused psychological trauma and social disorientation, but it caused
neither mass starvation, mass unemployment or mass poverty; in today’s society, an epidemic of
much smaller scope might quickly lead to both. This is because most production and consumption
were highly local, giving basic productive systems a self-sufficient, cellular resilience; in modern
23
systems where food, goods and services are produced through complex technologies and far-flung,
specialised supply chains, a smaller disruption can have much broader consequences. Where the
Black Death seems to have had notable effects, the mechanism seems twofold. First, social systems
are typically designed to cope with average conditions, perhaps with some safety margin, rather
than with extreme conditions. A city may cope with a thousand deaths spread evenly across a year;
if a hundred deaths happen in one week, the system breaks down, people die unshriven and are
buried in mass pits. These changes are likely to be traumatic, but in most cases normality would
reassert itself quickly once the crisis passes. Secondly, the Black Death’s more lasting changes seem
principally to have resulted from gross changes to overall population levels; for instance, readjusting
the balance between people, land and wealth, and between agricultural and pastoral uses of land.
Again, what is unique about epidemics may not be the overall levels of change so much as its abrupt
nature, which causes punctuated rather than gradualistic change.
Finally, this exercise has methodological implications for studies of ancient health. Skeletal
studies contribute important information on many health problems, as well as how people
understood and responded to them – all important subjects of study. But, using DALYs as a measure
of a health problem’s human importance, skeletal sources are relatively mute upon many major
existential problems of health, particularly infectious diseases. Instead, this work suggests several
new horizons for research. Methodologically, the next round of breakthroughs in the evolutionary
and social history of health are most likely to come from molecular studies – the proteomics,
genomics, metabolomics of both humans and disease organisms -- which can give us insight into
how humans interacted with pathogens. Conceptually, it may shift the focus from merely
documenting the presence of disease in the past (whether skeletally, textually, or molecularly) to
thinking about how disease interacts with social factors – for instance, about health inequality in the
past and present.
Acknowledgements
The authors would like to thank all the members of the “After the Plague” project, including
Ruoyun Hui, Toomas Kivisild, Christiana Scheib, Sarah-Jane Harknett, and Bram Mulder, for
discussion of these issues. We thank Ileana Micarelli, Mary Anne Tafuri and Lorna Tilley for the kind
invitation to take part in the EAA session this was originally presented in, and Lorna Tilley for
detailed comments on the manuscript. All deficiencies remain our own.
This work was supported by the Wellcome Trust (Award no 2000368/Z/15/Z).
Bibliography
Ariès, P., 1974. Western attitudes toward death from the Middle Ages to the present. Johns
Hopkins University Press, Baltimore.
Aufderheide, A., 2011. The Cambridge Encyclopedia of Human Paleopathology. Cambridge
University Press, Cambridge.
Benedictow, O., 2004. The Black Death, 1346-1353: the complete history. Boydell, Woodbridge.
24
Binski, P., 1996. Medieval death: ritual and representation. British Museum Press, London.
Bos, K.I., Schuenemann, V.J., Golding, G.B., Burbano, H.A., Waglechner, N., Coombes, B.K., McPhee,
J.B., DeWitte, S.N., Meyer, M., Schmedes, S., Wood, J., Earn, D.J.D., Herring, D.A., Bauer, P., Poinar,
H.N., Krause, J., 2011. A draft genome of Yersinia pestis from victims of the Black Death (vol 478, pg
506, 2011). Nature 480, 278-278.
Broadberry, S., Campbell, B., Klein, A., Overton, M., van Leeuwen, B., 2015. Population, in:
Broadberry, S., Campbell, B., Klein, A., Overton, M., van Leeuwen, B. (Eds.), British economic growth,
1270-1870. Cambridge University Press, Cambridge, pp. 3-45.
Buikstra, J.E., 2019. Ortner's identification of pathological conditions in human skeletal remains (3rd
edition), Academic Press, London.
Butler, S.M., 2015. Forensic medicine and death investigation in medieval England. Routledge,
London.
Campbell, B., 2016. The great transition: climate, disease and society in the Late-Medieval world.
Cambridge University Press, Cambridge.
Cessford, C., Scheib, C.L., Guellil, M., Keller, M., Alexander, C., Inskip, S.A., Robb, J.E., 2021. Beyond
plague pits: using genetics to identify responses to plague in medieval and post-medieval
Cambridgeshire. European Journal of Archaeology in press.
Chamberlain, G., 2006. British maternal mortality in the 19th and early 20th centuries. J. R. Soc.
Med. 99, 559-563.
Chao, F.Q., You, D.Z., Pedersen, J., Hug, L., Alkema, L., 2018. National and regional under-5 mortality
rate by economic status for low-income and middle-income countries: a systematic assessment.
Lancet Glob. Health 6, E535-E547.
Coale, A.J., Demeny, P., Vaughan, B., 1983. Regional model life tables and stable populations, 2nd
edition. Academic Press, New York.
DeWitte, S.N., 2015. Setting the stage for medieval plague: Pre-black death trends in survival and
mortality. American Journal of Physical Anthropology 158, 441-451.
Dewitte, S.N., Kowaleski, M., 2017. Black Death Bodies. Fragments 6, 1-37.
Dittmar, J.M., Mitchell, P.D., Cessford, C., Inskip, S.A., Robb, J.E., 2021. Medieval injuries: Skeletal
trauma as an indicator of past living conditions and hazard risk. American Journal of Physical
Anthropology 175, 626-645.
Dobson, M., 1980. March fever: the geography of malaria in England. Journal of Historical Geography
6, 357-389.
Dyer, C., 2002. Making a living in the Middle Ages: The people of Britain, 850-1520. Yale University
Press, New Haven.
25
Gowland, R.L., Western, A.G., 2012. Morbidity in the marshes: Using spatial epidemiology to
investigate skeletal evidence for malaria in Anglo-Saxon England (AD 410-1050). American Journal of
Physical Anthropology 147, 301-311.
Grainger, I., Hawkins, D., Cowal, L., Mikulski, R., 2008. The Black Death cemetery, East Smithfield,
London. MOLA, London.
Hanawalt, B.A., 1986. The ties that bound: peasant families in medieval England. Oxford University
Press, Oxford.
Jaffe, H., 1972. Metabolic, degenerative, and inflammatory diseases of bones and joints. Lea and
Febiger, Philadelphia.
Jonker, M.A., 2003. Estimation of life expectancy in the Middle Ages. Journal of the Royal Statistical
Society Series a-Statistics in Society 166, 105-117.
Kyere, K.A., Than, K.D., Wang, A.U., Rahman, S.U., Valdivia-Valdivia, J.M., La Marca, F., Park, P., 2012.
Schmorl’s nodes. European Spine Journal 21, 2115-2121.
Lewis, M.E., Gowland, R., 2007. Brief and precarious lives: Infant mortality in contrasting sites from
medieval and post-medieval England (AD 850-1859). American Journal of Physical Anthropology 134,
117-129.
Li, X., Kolltveit, K.M., Tronstad, L., Olsen, I., 2000. Systemic diseases caused by oral infection. Clin
Microbiol Rev 13, 547-558.
Marshall, J., 1832. Mortality of the Metropolis: A Statistical View of the Number of Persons Reported
to Have Died, of Each of More Than 100 Kinds of Disease and Casualties Within the Bills of Mortality,
in Each of the Two Hundred and Four Years, 1629-1831 ... : Accompanied with a Variety of Statistical
Accounts, Illustrative of the Progress and Extent of the Amount Expended for the Maintenance of
the Poor. J. Marshall, London.
Mathers, C., 2017. WHO methods and data sources for global burden of disease estimates 2000-
2015. WHO, Geneva.
Mattei, T.A., Rehman, A.A., 2014. Schmorl's nodes: current pathophysiological, diagnostic, and
therapeutic paradigms. Neurosurgical Review 37, 39-46.
Metts, R., 2001. The fatal flaw in the Disability Adjusted Life Year. Disability and Society 16, 449-452.
Millar, A., 1759. A collection of the yearly bils of Mortality from 1657 to 1758 inclusive. A. Millar,
London.
Mitchell, P.D., 2011. Retrospective diagnosis and the use of historical texts for investigating disease
in the past. International Journal of Paleopathology 1, 81-88.
Mitchell, P.D., 2017. Improving the use of historical written sources in paleopathology. International
Journal of Paleopathology 19, 88-95.
Mitchell, P.D., Dittmar, J.M., Mulder, B., Inskip, S., Littlewood, A., Cessford, C., Robb, J.E., 2021. The
Prevalence of Cancer in Britain Before Industrialization. Cancer 127, 3054-3059.
26
Murray, C.J.L., Lopez, A.D., 1996. The Global burden of disease : a comprehensive assessment of
mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. World
Health Organization, Geneva.
Nord, E., 2013. Disability weights in the Global Burden of Disease 2010: Unclear meaning and
overstatement of international agreement. Health Policy 111, 99-104.
Platt, C., 1997. King Death: The Black Death and its aftermath in late-medeival England. UCL Press,
London.
Podd, R., 2020. Reconsidering maternal mortality in medieval England: aristocratic Englishwomen, c.
1236-1503. Continuity Change 35, 115-137.
Rawcliffe, C., 1995. Medicine and society in later medieval England. Sutton, Stroud.
Rawcliffe, C., 2006. Leprosy in medieval England. Boydell, Woodbridge.
Rawcliffe, C., 2013. Urban bodies: communal health in late medieval English towns and cities.
Boydell Press, Woodbridge.
Razi, Z., 1980. Life, marriage and death in a medieval parish: economy, society and demography in
Halesowen 1270-1400. Cambridge, Cambridge University Press.
Roberts, C., 2002. Tuberculosis in Britain: its history and palaeoepidemiology. Antropologia
Portuguesa 19, 101-119.
Roberts, C., Cox, M., 2003. Health and Disease in Britain : from Prehistory to the Present Day. Sutton,
Stroud.
Roberts, C., Manchester, K., 1995. The archaeology of disease, 2nd ed. Cornell University Press,
Ithaca.
Roberts, C.A., 2020. Leprosy: Past and present. University of Florida Press, Gainesville.
Roberts, C.A., Bernard, M.C., 2015. Tuberculosis: A biosocial study of admissions to a children's
sanatorium (1936-1954) in Stannington, Northumberland, England. Tuberculosis 95, S105-S108.
Schmidt, J.C., Fox, S., Hughes, A., Betts, A., Flowers, J., Newton, J., Clark, S., 2020. The burden of
disease in England compared with 22 peer countries: a report for NHS ENgland. Public Health
England, London.
Seabourne, A., Seabourne, G., 2001. Suicide or accident - self-killing in medieval England - Series of
198 cases from the Eyre records. British Journal of Psychiatry 178, 42-47.
Slavin, P., 2019. Experiencing famine in fourteenth-century Britain. Brepols, Turnhout.
Stodder, A., 2016. Quantifying morbidity in prehispanic Southwestern villages, in: Herhahn, C.,
Ramenofsly, A.F. (Eds.), Causation and Explanation: Demography, Movement, and Historical Ecology
in the Prehistoric Southwest. University of Colorado Press, Boulder, pp. 121-135.
27
Stodder, A., 2017. Quantifying Impairment and Disability in Bioarchaeological Assemblages, in:
Byrnes, J.F., Muller, J.L. (Eds.), Bioarchaeology of Impairment and Disability: Theoretical,
Ethnohistorical, and Methodological Perspectives. Springer, New York, pp. 183-200.
Waldron, T., 2009. Palaeopathology. Cambridge University Press, Cambridge.
Wood, J.W., Milner, G.R., Harpending, H., Weiss, K., 1992. The osteological paradox: problems in
inferring prehistoric health from skeletal samples. Current Anthropology 33, 343-370.
World Health, O., 2004. Global Burden of Disease 2004 Update: Disability weights for diseases and
conditions. WHO, Geneva.
