Demographic forum 1999, Oslo, 10-12 June


Session 1 - The role of purposive behaviour in mortality reduction

Bruce Fetter

Genes, Memes, and Contingency: The Effects of Purposive Action on Mortality Reduction

Understanding the role of human initiative in mortality reduction requires a knowledge of both the natural world, which determines the limits of purposive behavior, and human agency. Richard Dawkins' notions of memes - cultural artefacts for the implementation, transmission, and preservation of ideas - can be applied to the four major institutions which have historically provided health care: families, healers, charities, and governments. Health-preserving practices by these institutions also required the formulation of effective policies and access to the resources necessary to implement them.


Chris Wills

What can we expect from Disease Organisms in the Future?

As recently as the 1960s, it was generally assumed that infectious diseases in the developed world were rapidly becoming a thing of the past. Antibiotics could deal with bacterial infections swiftly and effectively, and viral diseases such as polio were succumbing to universal immunization programs. In the last two decades of the century, however, the AIDS epidemic, the appearance of Ebola and other emerging viral diseases, and the recent resurgences of older diseases such as tuberculosis and dengue hemorrhagic fever, have changed this picture. As a result, there has been an overreaction in the press and from the public. A wave of books and films have suggested that we are headed for a kind of Armageddon, triggered by some unstoppable disease emerging from the rainforests or from the noisome depths of the New York subways.

What is real and what is hype? The dangers from our pathogens are real, but many of them are not as great as has been imagined and others, perhaps some of the most serious, have been ignored.

Let me deal first with the obvious dangers. AIDS, quite properly, bulks very large as the most dangerous of the new and emerging diseases.

The source of the HIV-1 virus, which causes the majority of the cases of AIDS around the world, remains a mystery, but the less severe HIV-2 almost certainly made the jump to humans from sooty mangabeys kept as pets in West Africa, and the jump has occurred a number of times. Eventually, the natural reservoir of HIV-1 will be discovered. It may or may not turn out to be a primate, but it will certainly prove to be an African mammal.

This disease is very difficult to deal with, because the virus destroys part of its host's immune system, mutates too rapidly for an effective antiserum to be developed, has a decade-long latent period, and is sexually transmitted. One cannot imagine a worse possible combination of features from the standpoint of the control of an epidemic. Progress has been made in delaying the onset of the disease, but treatment is expensive, indefinitely prolonged, and demands the strictest compliance. The rapidly-growing epidemic in Africa and South Asia can only be slowed by widespread education, changes in sexual habits, and the eventual development of an effective vaccine.

Given all this, is AIDS the agent of doom for our species? Probably not. Genetic resistance to the virus is now becoming understood at the molecular level, and resistance genes are turning out to be surprisingly widespread. It is possible that they have been driven up in frequency by unrelated diseases from the past, such as bubonic plague. WHO statistics suggest that the rate of increase of the disease, even in Africa and South Asia, is beginning to slow, in part the result of the dissemination of information about the disease and in part due to many other factors. And AIDS must be kept in perspective. The total death rate from the disease, while very high, continues to be dwarfed by the millions of needless and preventable deaths from tuberculosis, pneumonia and diarrheal diseases throughout the less developed world. AIDS is a severe public health problem, but it is far from being the most severe that currently faces our species.

Other emerging viruses are likely to pose problems in the future, but I think the situation may be ameliorated by two factors. First, outbreaks of diseases like Ebola and Machupo that cause immediate and severe symptoms can be discovered, followed, and contained relatively promptly. Ebola, while it luckily spreads with great difficulty among humans, remains impossible to treat, but immune plasma and ribovirin have proved effective in treating Machupo. The animal reservoirs of these diseases are also slowly being discovered, which will help to control them.

New AIDS-like viruses with long incubation periods pose the worst danger. It is often suggested that if, unlike AIDS, these new viruses were to be capable of being spread readily by simple contact or by vectors such as mosquitoes, this would result in a severe and perhaps unstoppable plague. And yet, as has been pointed out by many authors, the sexual transmission of AIDS is the greatest difficulty that we face in controlling the epidemic. Any political difficulties that are posed by physical quarantine of infected individuals, or destruction of mosquito and other vector populations, would be far less severe. When faced with such a disease we will not fall quite so badly afoul of Victorian morality and various political agendas as we have in our attempts to control the AIDS epidemic.

The second ameliorative factor, ironically, is the rapid destruction of the untouched tropical ecosystems from which new diseases are likely to appear. Within our own lifetimes, the last wild places on Earth will disappear or will become isolated, fragmented and irreversibly simplified ecologically. Emerging diseases must have some place to emerge from.

Far more threatening, in my view, are the resurgent diseases. Tuberculosis poses the greatest problem, but many other diseases are beginning to spread as a result of increased travel and trade and changing climate patterns. Dengue hemorrhagic fever, particularly severe in children, is spreading as the Aedes mosquito vectors of the disease spread.

In Cuba in 1981, the disease affected a third of a million people and caused 158 deaths. The most widespread potential vector in the U.S., Aedes albopictus, has so far remained free of the virus, but there was a recent report of infected albopictus from Mexico. The disease is poised to become a major health problem in the warmer parts of the developed world. Our response, of course, will be to modify the environment drastically: for example, mosquito abatement programs in the tropics and subtropics will increase in intensity and will unavoidably result in collateral environmental damage.

One very important reason that so many diseases are reappearing is emerging resistance. Treatments that used to be effective are losing their potency. Disease organisms that were essentially controllable in the past are sometimes now uncontrollable even in a hospital setting.

Multidrug-resistant gram-positive organisms are emerging that are resistant to even the newest drugs such as vancomycin. Compounding the problem is the widespread use of antibiotics in animal feed and of antibacterial soaps in households and hospitals. Antimalarials are losing their effectiveness even as their side effects grow - mefloquine, the last really effective antimalarial on the market, is often discontinued by patients and travelers before the full course of treatment because of unpleasant neurological side-effects.

What can be done? New drugs are in the pipeline, but often, as with the antimalarials, side effects increase in severity as more side groups are added to the molecules. Widespread immunization, sometimes mediated directly through foodstuffs that have been genetically engineered to make bacterial and viral antigens, is a real possibility, but here too emerging resistance is likely to be a problem. Perhaps the most promising avenue of research is that opened up by a complete understanding of the genomes of various pathogens. More and more pathogens have been completely sequenced - the latest important advance is the complete syphilis spirochete genome, and the entire Plasmodium genome will soon be available.

As a result many opportunities will be opened up to design new antibodies, new drugs that can block metabolic pathways that are found the pathogen and not in its host, and compounds that inhibit cell receptors and prevent the pathogens from invading. With luck, and at considerable cost, we should be able to keep ahead of our pathogens.

Finally, let me raise a concern that may turn out to dwarf all the others. One little-noticed but remarkably general observation about human pathogens is that they are often genetically defective, sometimes in quite dramatic ways. Yersinia pestis, responsible for bubonic plague, cannot live anywhere but in a host cell - it carries many defective genes and even its Krebs cycle is defective. The spirochete responsible for syphilis is famously defective, with a genome only a third the size of those of its free-living relatives. Much of the virulence of the smallpox virus, luckily now almost extinct, is due to a defective interleukin-1 beta receptor gene, which results in a damagingly high fever in its host. Its much milder relative the cowpox virus has an intact gene.

Why this trend? There is probably an evolutionary reason - the human species has invaded so many new environments, and we have changed all of these environments so much, that our pathogens have been forced to adapt very quickly. Most short-term evolutionary adaptations tend to involve loss rather than gain of function. Of course, this gives us an opportunity for control - we must look for these points of weakness in our pathogens and exploit them.

It is, however, this very property of pathogens that may pose the greatest danger to our own survival. As our species grows in numbers, the resulting overpopulation and overexploitation of resources are obvious dangers, and have certainly been agonized about at length. But a little-known consequence of our destruction of the environment is that we are exposing many other animals and plants to rapidly-changing

environmental conditions. Their pathogens, too, will be evolving rapidly, and it can be expected that they will become more virulent as a result. To make things worse, most of the crops and animals on which we depend are genetically depauperate and therefore have heightened susceptibility to such pathogens. And this trend is unlikely to be confined to our domesticated organisms. As endangered species of animals and plants throughout the world are forced to adapt to changing circumstances they, too, will lose genetic variability and become more susceptible to pathogens that at the same time are becoming more virulent. This evolutionary feedback loop will accelerate the loss of many species that are already endangered, and lead to the endangerment of many others. The effect of all these losses on our own survival could be incalculable.

We can, in short, defend ourselves against our own pathogens. But it will be very difficult for us to defend the rest of the animals and plants on the planet against a wave of newly-virulent pathogens that we ourselves have inadvertently created!


Peter Sköld

Transition Models in the post-McKeown Era

The paper discusses the changes of historical demography in relation to the epidemiologic transition since the influential works of Thomas McKeown were presented during the 1970s. Terminology is stressed as one reason for confusion. Areas of research that were previously neglected by scholars have achieved more interest during the last two decades. Cultural variables, attitudes and mentality are among the most important.

Two examples are presented to illustrate the importance of cultural factors and public attitudes. The native reindeer herding people in northern Scandinavia, the Sámi, experiences much lower smallpox mortality than the rest of Sweden. The explanation was that their religion gave them a different understanding of the disease, leading them to flight.

The other example show higher marriage-ages for persons previously infected by smallpox. The pock-marked faces had a negative effect in the marriage market.

Vaccination was very successfully implemented in Sweden. The explanation has to do with the Swedish mentality, which did not rise protests towards the government easily. Moreover the decision to make the clergy key persons in the organization of vaccination and the law of compulsory vaccination in 1816 were important factors.

Historical demographers need a deeper knowledge concerning so many factors involved in the epidemiologic transition. Each discovery must be put into a context, which helps us to understand the relativity involved. It must be discussed what is possible to generalize and what is not. By doing so we will improve not only our fundamental knowledge, as a base for research, but also help to improve the possibilities for those who will discuss international long-term trends.


Sheila Ryan Johansson

Was Human Mortality 'Natural' in Early Modern Europe?

The mortality transition remains the most important event in modern mortality history. But it was not a conventional historical event with a clear beginning and end. It is and was a fuzzy event with dimly perceived origins; and even now it may or may not be continuing in the developed countries. The mortality transition is not about a permanent escape from death, but the gradual postponement of most deaths to old age, which produced a sustained rise in life expectancy at birth levels. How this postponement occurred remains the subject of intense controversy, partly because it is not even clear whether or not summary measures of mortality like life expectancy at birth should be used for explanatory purposes. Most problematic is whether or not the mortality transition should be treated as one macro level transition involving whole sets of countries, or individual countries, or smaller and more homogeneous subnational populations in those countries. All that is clear is that new research strategies are needed which "problematize" the mortality transition as an event before proceeding to explain it.

A major theme of this paper was whether or not mortality patterns were entirely "natural" in early modern Europe, before the mortality transition began, and what is involved in conceptualizing mortality as 'natural' in periods before sustained declines in mortality are first observed. Problematizing the concept "natural" mortality leads to a close examination of the role of knowledge in pre-modern mortality patterns, and a re-examination of how knowledge was created and applied to the management of mortality before national level mortality transitions began.

Session 2 - Developing countries

Michael Bracher

Breastfeeding in central Java

No abstract submitted


Negussie Taffa

Adolescent Sexual and Reproductive Health: Review of Current Facts, Programmes and Progress since ICPD

Adolescents make up a significant proportion of any population. Owing to its untouched resource potentials, this segment forms the future hope of that nation as well. However, the sexual and reproductive health services provided to this segment now by Governments don't seem to meet that expectation. This is often due to one or a combination of the following reasons: their large number, their predominantly healthy looking appearance using the morbidity oriented disease burden estimates, general failure to understand the needs of young people or due to multitude of competing priorities for a limited resource. Nonetheless, today more than ever, societies and their leaders must be sensitised by the situation in which the young people are living and attempt to fulfil their generation's responsibilities. To mention a few of these situations, the current health, social/psychological and economic impacts of the HIV/AIDS pandemic and teenage pregnancies might suffice. Many young people in developing countries are also suffering from lack of self-esteem and future hope, they are unemployed, victims of different forms of violence and abuse, or are obliged to live with harmful habits like smoking, drug abuse and alcoholism. Moreover, the changing family and social environment, the early age to sexual maturity, and general tendency towards delayed age at marriage, reduce the age of sexual debut and increase the risk for exposure to diseases and unwanted pregnancy.

Of all the controversial issues that were dealt with during the International Conference on Population and Development (ICPD) in 1994, the topic of adolescent sexuality was one of the most contentious. After heated debates, consensus was eventually reached that in general stressed the rigSfs of adolescents. It was the first time during an international conference that the reproductive rights of adolescents were agreed on, in the meaning of the right to reproductive health and to services that are suited to the particular needs of young people.

This paper aims to review current facts on adolescent sexual and reproductive health (ASRH), gather lessons learned from implementing various ASRH progranirnes and projects and attempted to see the progress made on the sexual and reproductive health rights and services for young people five years after ICPD. Desk review of published and unpublished documents, activity reports and various articles was made to get the required information. Since a wider scientific rigor was not mandatory, a critical review on methodological aspects of various articles was not emphasised. Moreover, the review was not exhaustive of completed works during the last 2-3 years, and less emphasis was put on activities implemented in the 80s.

Today, adolescent sexual and reproductive health rights are largely known to most relevant stakeholders, including policy makers, service providers, communities and family members. There are evidences that endeavours on advocacy, information dissemination, and limited service delivery are making a difference to win the political interest of Governments and make a shift in resource allocation. Some of the fruits are the introduction of new legislation, revision/refining of legal restrictions on young people's sexual and reproductive health rights and passing out new policies that help them benefit.

Pertaining to research outcomes, it has now been established that sex education for young people will encourage abstention, delays age at first sex, promotes the utilisation of services and will generally help young people lead a responsible sexual life. However, effective and sustainable programme strategies required to reach adolescents, especially the most vulnerable groups using integrated approaches, remain the subjects of experiment for years to come.

Nevertheless, until all the scientific evidence comes in, we have enough experience to work on thanks to the awakening initiative of the ICPD. For instance, encouraging results from projects that have worked in partnership with the young people, parents and the community at large are increasingly appearing to be attractive for experts and programme owners. Programmes that target young adolescents, out-of-school youth, male adolescents, and projects that have job creation and vocational training components are showing significant results, although it's too early to conclude. More and more components of ASRH are being recognised but the common five areas outlined up to now, are: information provision, acquisition of general life skills, counselling, health services and creation of supportive family and community environment. Innovative project financing mechanisms are being implemented, and there are examples of private sector involvement. Regional and national expert and management networking has provided the chance for learning from one another and devise new strategies to tackle emerging problems.

We will finally look forward to the XPD + 5 review, in the hope of reminding the international community and nations to meet the recommendations drawn from our assessment of the status of ASRH and programmes during the next decade or so.


Helge Brunborg

Mortality impact of HIV/AIDS in Africa

The AIDS epidemic is different from other epidemics in many respects: long latency period; death is certain when infected; relatively low infectivity; infection usually not visible; and inverse social gradient. The first cases were diagnosed around 1980, and the virus was identified in 1984. Since then there has been a dramatic growth. Today more than 33 mill. persons are infected, and 2 ½ mill. died from AIDS in 1998. The current spread of the epidemic is particularly rapid in Southern Africa.

Sources of data on HIV/AIDS mortality and prevalence in Sub-Saharan Africa

Data on mortality are generally lacking, particularly on adult mortality. The coverage of the civil registration system is too low to yield reliable mortality estimates.

The cause of death is, of course, even more underreported, especially deaths due to AIDS (for medical and social stigmatisation reasons). Mortality data from sample surveys can yield fairly reliable estimates of infant and child mortality. It is much harder to obtain data on adult mortality but new and promising indirect methods have been developed, notably the sibling and orphanhood methods:

Orphanhood method: Proportion of respondents by age with living mothers and fathers. Weakness: indi viduals orphaned at young age are often not reported, which causes a downward bias. Another bias is caused by vertical transmission of HIV

Sibling method: Date of birth and age at death (if dead) of brothers and sisters. Weakness: too low mortality estimates 7-13 years before the data were collected. High HIV prevalence causes a downward bias.

Data on size and composition of the population, which are needed to estimate the total number of HIV-positive person, etc., can usually only be derived from census data, which are collected every 10 years - at the most.

The quality of census data is usually too poor to yield satis factory mortality estimates. Other sources such as population registers are usually not available.

The most common source of prevalence data is the Sentinel Surveillance System, which is based on anonymous, unlinked testing, especially of pregnant women, blood donors and people with STDs who come for diagnosis and treatment to STD or primary health care clinics. All of these groups are biased in one way or another.


The most common are population projections of the standard cohort-component model, which are used by United Nations and U.S. Bureau of the Census. They depend on the HIV/AIDS assumptions, which often some from a special model for forecasting HIV incidence and prevalence, such as Epimodel or iwgAIDS.

There are also special HIV/AIDS simulation models, which encounter great modelling problems and data needs. Particularly sensitive are the assumptions on factors like latency period (period from infection to full-blown AIDS), and the infection rate, which is heavily influenced by cofactors such as proportions of women and men with STDs/genital ulcers, proportions using condoms, and sexual mixing patterns. The epidemiological often do not include demographic variables like age and sex, whereas the demographic models include very little detail, if any, on how the epidemic is spreading.

The situation in Southern Africa

The spread of the epidemic has been spreading very rapidly in Southern Africa, including Botswana, which is now the worst affected country in the world. The first AIDS case was diagnosed in December 1985 and in April 1988 there were 43 known cases of full-blown AIDS and 176 known HIV-positive persons in Botswana. However, in 1998 there were an estimated 216,000 HIV-positive persons in the country, of a total population of 1.5 million.

Unlike other countries, the HIV prevalence in Botswana is not the highest in the capital, but in the second largest city in the Northeast, Francistown, where antenatal clinic data for 1997 yielded a 42.9% seroprevalence. This is the highest proportion recorded world-wide for pregnant women in any major city. Infant and child mortality has increased significantly in many countries, especially in urban areas, including Zambia.

The probability of surviving from age 15 to 60 has been estimated from sample survey data using the sisterhood method, showing significant increases from the pre-AIDS period, and yielding, as the highest, 67 per cent for Zambia (Timæus 1998).

The life expectancy at birth has been declining dramatically in many countries. The estimates are fairly sensitive to the assumptions made, especially when the epidemic peaks, however. UN (1998) and US Bureau of the Census (1999) give somewhat different estimates, for example 47 and 40 years, respectively, for Botswana 1995-2000. Both of these figures are, however, substantially lower than 1991 Census estimates of 61.7 years, and the UN estimate for 1995-2000 of 67 years, if there were no HIV/AIDS.

Concluding remarks

Are there any reasons for optimism? Yes, interventions seem to have worked in Uganda and Thailand and there are signs of behavioural changes in Zambia.

Session 3 - Open session

Gigi Santow

Do Gravidity and Age affect Pregnancy Outcome?

Fetal loss has generally been found to vary with gravidity, previous experience of fetal loss, and maternal age, but the literature is divided on the reasons for these associations. In this paper we examine pregnancy histories obtained retrospectively from a nationally representative one-in-one-thousand sample of women in Australia aged 20 to 59 years. The relations of fetal loss ratios with both gravidity and previous outcome are consistent with heterogeneity of risk over the study population and a stopping rule, whereby high-risk women undertake more pregnancies than low-risk women to achieve the same number of live births. Evidence is presented that elevated loss ratios in the teens indicate not higher risk but a selection for short gestation intervals, while loss ratios beyond the mid-thirties do not point unequivocally to a substantial incresae in risk at the older reproductive ages.


Nico Keilman & Arve Hetland

Predictive intervals for age-specific fertility and mortality

The purpose of the paper is to give a precise account of uncertainty attached to future age-specific birth and death rates. Mutivariate ARIMA-models are used to predict the parameters of the Gamma curve for fertility, and the Heligman-Pollard curve for mortality. Given such predictions, age-specific birth and death rates for future years can be computed, together with their confidence intervals. These are needed in order

Session 4 - The 19th century as the crucial transition period

Joseph M. Prince & Richard H. Steckel

Tallest in the World: Native Americans of the Great Plains in the Nineteenth Century

Historians often portray Native Americans as merely unfortunate victims of European disease and aggression, with lives in disarray that followed the arrival of Columbus and other explorers or conquerors. The data we analyze on human stature show, in contrast, that some Native Americans such as the equestrian Plains nomads, were remarkably ingenious

and adaptive in the face of exceptional demographic stress. Using anthropometric data originally collected by Franz Boas, we show that the Plains nomads were tallest in the world during the mid-nineteenth century.

We link this extraordinary achievement to a rich and varied diet, modest disease loads other than epidemics, a remarkable facility at reorganization following demographic disasters, and egalitarian principles of operation. The analysis provides a useful mirror for understanding the health of Euro-Americans.

After a long period of regrettable neglect, Native American history has finally emerged as a growth industry. Inspired in part by a desire for inclusiveness in the study of the past and also by retrospectives associated with the quincentennial of Columbus' arrival in the Americas, scholars have increasingly sought to understand the contours of the Native American experience.

Most Americans are familiar with the broad outlines of Native American history: settlement via the Siberian land bridge some 12,000 to 20,000 years ago followed by widespread geographic diffusion over the ensuing millennia, with the largest concentrations of population ultimately located in Mesoamerica. The arrival of disease-bearing European colonizers, bent on conquest and religious conversion, decimated the aboriginal populations they encountered. In the United States, the population of Native Americans reached a nadir with the reservation period near the end of the nineteenth century.

In telling the story of Native American history, scholars have been ingenious in assembling evidence from archeological and linguistic sources and from eyewitness accounts and oral histories. Inevitably, though, the available information could only be called thin, in part because few written records were ever created until the reservation era. Thus, new information and techniques are always welcome in this field of study, but substantial gaps in our knowledge are likely to persist.

This paper sheds light on the well being of Great Plains equestrian nomads using a source now familiar to economic historians, human stature. Originally collected near the end of the nineteenth century by Franz Boas and his assistants for ethnographic purposes, the present sample contains measurements of 1,485 adult individuals, predominately males, from 8 tribes. The paper offers suggestions for coping with small samples of widely disparate ages by correcting for growth as young adults and for shrinkage at very old ages. After discussing the sample and ecological conditions and tribal ways of life, we consider possible explanations for the remarkable finding that the equestrian nomads of Great Plains were the tallest in the world during the nineteenth century. In addition to shedding light on the exceptional but neglected achievements of Native Americans, the analysis provide a useful mirror for study of influences on health more generally.


Patrice Bourdelais

New cultural reasons to fight against death during the 19th century

After having evaluated the principal characteristics of the policy of public health in France during the 19th century, this contribution is centered on the principal cultural reasons which can account for the individual and collective evolution of the attention given to the questions of health. Four topics are proposed in order to organize the comparison between the stories of our various countries.

First, the laicization of the attitudes in front of life involved a vast movement of request for care of the population as early as the mid 18th century, and later directed the French medical policy during last decades of 19th century (the republicans thought medical policy and assistance would make it possible to remove from the Church one of the means of influence that it held on the population).

The second topic relates to the construction of a historical horizon of eradication of the infectious diseases, born with the discovery of Jenner in the intellectual context of the Enlightenment, then reinforced by the progress of science throughout the 19th and 20th centuries (vaccinations, bacteriology, sulphamides, antibiotics, methods of investigation). The shared opinion was that progress of knowledge was going to make it possible to identify the germ responsible for each disease and then to fight against it thanks to tracing and vaccination.

The third topic is one of the most interesting and also more difficult to treat. It is a question of analyzing the construction of new "intolerables", clearly related in France to the intellectual movement which develops around the French Revolution and of the assertion of human rights and the rights of the citizen. The "intolerables" multiply during the 19th century, as can be ascertained from new laws and legal measures.

The recognition of the individual, of the subject, constitutes the fourth dimension of this approach. It is essential to the recognition of the misfortunes of the individual disease and at the same time it draws up obstacles vis-a-vis the application of any regulation.

It is not a question here to undervalue the other factors of the evolution of public health, the modification of the epidemiologic landscape or the economic interests or policies of the dominant classes, but to insist on a so far underestimated aspect of the history of Public health.


Robert Woods

Ever since 1799 The Fundamental Conditions for Sustained Mortality Decline

In the summer of 1799 Malthus visited Scandinavia. Part holiday, part field excursion for the collection of facts and observations the surviving travel diaries provide an important account of the social and demographic circumstances of especially Norway two hundred years ago.

It is not, of course, that Malthus's description is either full or entirely accurate, but what is significant is the extent to which his field observations altered the emphasis in An Essay on the Principle of Population. The Norwegian adventure helped to tip the Malthusian analysis of European population away from the positive check which dominated the First Essay (1798) towards the preventive check of the Second Essay (1803), that is from mortality to fertility via nuptiality.

Ever since 1799 historical demographers have been inclined to follow; to think of the positive check as 'natural' while the preventive check involved intentional behaviour as well as the regulation of social institutions. But there are also good reasons for considering the role of human intervention, purposive action, in the secular decline of mortality during the last two hundred years.

One way to begin such a consideration and the re-focusing of research would be to list the fundamental conditions for sustained mortality decline. The following provides an initial example.

A. The challenge of urbanization and the public health movement B. Parliamentary democracy with the extension of the electoral franchise and its effect on the development of social intervention for collective health security C. The continuous rise and application of scientific knowledge especially in engineering and medicine Now in the late twentieth century it is possible to observe the combined effects of A+B+C in the rich countries of the world, although there are many populations in which mortality has been brought down rapidly virtually entirely by C alone. But among the European populations the decline of mortality has been a rather more drawn out process and there each of A, B and C has had its bearing on the process. Among Europeans the following sequence might be appropriate:

A > A+B > B+C.

This sequence appears to have been accompanied by parallel changes in focus and even ways of measuring mortality. For example, (1) from the community/environment to the individual and the socioeconomic model of health, (2) from a concern for levels and trends in mortality (the e(0) approach associated with A) to one emphasizing inequalities (the SMR approach encouraged by B).


Simon Szreter

Mortality, Health and Politics in the Era of Global Economic Change, 1500-2000

No abstract submitted

Session 5 - New policies, new bureaucracies

Antoinette Fauve-Chamoux

Changing politics towards abandoned children at the end of the Ancien Regiome and the responsibility of a French hospital administration on infant mortality

A true welfare system for foundlings and abandoned children was implemented in the late eighteenth century (January 1779) by the French royal administration. The King wanted the abandoned children to become in charge of local administrations and no more to be transported to Paris. This paper examines first how the Hôtel-Dieu of Rheims adapted to this new law. The second part is dedicated to the origin of foundlings in Champagne during the following years and to main changes related to the new social system; the third part traces the low chances of survival of the abandoned children in Rheims (from a set of longitudinal data concerning 800 babies) and how it was linked to the choice of nurses by the hospital administrators. The date of weaning was also purely the effect of an administration decision and the financial calendar had great consequences on infant and child mortality. The study is finally placed in a European socio-cultural context.


Virginia De Luca

The part of some civil servant action in the decrease of the mortality. The French case (1880-1920).

The state politics of health cast some officials for a part. It's necessary to know exactly their action in the decrease of the mortality because it shows us a concrete action of the State, close to the populations. Generally, historians study the action of the State by learning the legislative measures of health and the evolution of its financial intervention in the politics of health.

The questions are: how civil servants participate to the decrease of the mortality? Are they only applying the legislative measures or do they try to innovate? If they propose to extend the law, what are their motivations? Which level of state intervention do they present ?

In the French case, we study the intervention in the fight against mortality of the inspectors of Assistance publique and of the physician-inspectors. The period is the beginning of the Third Republic, i.e. the years 1870, till the first world war.

Les inspecteurs des enfants assists are government officials. In each department, one, two or three professionals supervise the respect of the children in care legislation. In 1895, there are two hundred inspectors. Progressively, they oversee all the health and social security services. So, in 1904, they become the Assistance publique services' inspectors : the title of their occupation is appropriate. The legislative texts about this occupation are explicit : their main mission is administrative. This professionals have to do some rounds of inspection : they go in each family who take in a child in care. They verify if the infant, the child or the adolescent is in a good health, if their guardians don't overburden the child in care with work. In many departments they bring up to date the list of the persons in receipt of national assistance and they verify if there is no defrauder. They are administrators.

The inspectors of Assistance publique take concerted action with the physicians inspectors. The physician visit the child in care or person in receipt of national assistance, the poor old people or the poor disabled. The visit are forgone, particularly for the young children but sometimes the inspectors of Assistance publique demands to the doctors to visit a person when he suspects that the person in receipt of national assistance is sad. The visits are paid and indemnified by the Assistance publique 's service. This service is financed by the three levels of politic administration: the commune, the department and the State. So the inspectors of Assistance publique have to supervise and manage; the physicians-inspectors have to diagnose. The administrative shift is necessary to purpose new measures.

Indeed, all physicians-inspectors of the department transmit their diagnosis to the inspector of the Assistance publique. The administrator has to centralise the information on the health of the persons in receipt of national assistance specially the children and the old. He does a synthesis of this information. Sometimes, he calculates some mortality rates to analyse the situation of the health's population. This synthesis and his proposition to decrease the mortality of the poor are transmitted to the high rank administrator in the department: the prefect. So, the intervention of this professionals take place in the communal level and in the departmental level. The physician-inspector diagnoses and the inspector of Assistance publique negotiates with the politicians. Two questions seems particularly interesting and can suggest international comparison.

First: the question of the role's division between two professionals, the doctor and the administrator. This division needs a clear definition and demarcation of the competence of each professionals because their have to legitimate et valorise their own action in the decrease of the mortality and to prove that they are expert in the question of the poor health.

Second: the question of the shift and the politic administration level of intervention. This intervention level create different social and administrative configuration. So the stakes are different and the work of each professionals is ratify in accordance with different evaluation criterions.


Catherine Rollet

Constructing Infant's health. International Conferences on Infant's and Children's health 1880-1920

Certainly, each industrialised country, European or American, has developed his own culture on infant's and children's health and has set up original shemes of social and medical structures to fight infant mortality and to protect infant's life. A lot of studies are now available which show the diversity of experiences between nations : Great Britain (Dwork, 1987), the United States (Meckel, 1990, Preston and Haines, 1991, Klaus, 1993), France (Rollet, 1990), Australia (Smith, 1988), Belgium (Masuy-Stroobant, 1983 and Debuisson, 1993), etc.

But a common culture began to be discussed and shared by the nations participating to those conferences which were hold especially in Europe from the middle of the XIXth century (Rollet, 1997). Some studies have focused on the inter-war period (Rooke and Schnell, Roemer, 1994) but nothing is consecrated to the period before the First World War. Now, one of the first medical conference, perhaps the first one in Europe, which was interested by chilhood, was hold in Brussels in 1876. But it is mainly after 1880 that childhood became one of the official themes of Congresses of Demography and Hygiene (1887, 1889, 1894, 1898, 1900...) and of Congresses of Welfare (1889, 1896, 1900, 1906, 1910...). Moreover, this part of the whole population became the unique subject of some Congresses : in 1883, was hold in Paris the first International Congress devoted to Infant's Welfare, on the presidence of a parisian advocate, Maurice Bonjean (Rollet, 1995).

In this paper, from a corpus of International Congresses on Protection of Children hold in Europe during the period 1880-1920, we would like to examinate what kind of questions were debated first, what influences were exerted on and what kind of recommendations or decisions were taken during this period. We observe that the first predominant points of view on infant health which have legislative and medical consequences were followed by emphasis on the role of families, of mothers, of women, especially expressed by anglo-saxons countries. Focusing on the representation of infants and children as physical bodies ans parts of the Nation, these first Conferences have an history which illustrate first, evolution of ideas on links between State, families and individuals, secondly, the reality of local or national experiences, thirdly, the position of forces between countries. It is the combination of these three factors which explain the evolution of debates on infant's health from the 1880's until First World War.


Jenny Wood

Mortality and the annual reports of the Registrar General for England and Wales: The construction, categorization, significance and use of cause of death

No abstract submitted

Session 6 - The importance of residence

Erik Nymoen

Mortality and residential history

In many mortality studies, death has been related to characteristics of the individual or the environment just before death. Migration, for example, violates this assumption. In a life history perspective, (cause of) death is to be seen as the end point of a sequence of passages through environments and experiences of various types, which over time influences the individual to develop a disease and finally to die. Many contemporary causes of death differ from the previous leading causes of death by their lengthy (in time) aetiologies - often spanning decades. The importance of individual longitudinal data is enhanced by aetiological theories emphasising the effects of different sequencing of individual experiences of living conditions and exposures. This study extends the analysis of regional mortality by taking account of earlier residential experiences such as region of birth and possibly later places of residence.


Michael Haines

The Urban Mortality Transition in the United States, 1800-1940.

In the United States in the 19th century, as in Europe in that era,there was a substantial mortality penalty to be living in urban places. By 1940, that urban penalty had been largely eliminated, and it was healthier, in many cases, to reside in a city than in the countryside. Part of the study of the great mortality transition in the United States is related to this phenomenon.

A significant problem with the history of mortality in the United States stems from the paucity of good statistical information on levels, trends, and differentials. It is possible, however, using a variety of sources and demographic estimation methods, partially to reconstruct the course of mortality in the United States from 1800 onwards and, more particularly, to provide some insight into differentials. When census data, vital statistics, local records, and genealogical data are culled for what they can reveal, the outlines appear.

In the early 19th century, the United States was a relatively low mortality regions by the standards of Western Europe. It was not particularly urban (only 6.1% in 1800), a crude death rate in the range of 20-25 per 1,000 population would not have been unusual. The low mortality was remarked upon by Thomas Robert Malthus [1798, pp. 104-106]. Mortality was likely lowest in New England and rose as the latitude moved further south. Such evidence as we have (mostly for New England the Middle

Atlantic states) does indicate a substantial urban penalty. For example, the Jaffe and Lourie [1942] life tables for 1826/35 show that the expectation of life at age 10 (e(10)) was 51.0 years for 44 smaller New England towns, whereas it was 46.0 for Salem, MA and New Haven, CT (medium-sized cities) and 35.9 years for Boston, New York City, and Philadelphia. By 1900 within the Death Registration Area (the six New England states, New York State, Pennsylvania, Michigan, Indiana, and the District of Columbia), the e(0) for urban whites was 46 years, while it

was 54.7 years for rural whites [Glover, 1921]. Estimates of child mortality for the whole United States based on indirect estimates using the 1900 Public Use Micro Sample of the census find that mortality in urban areas was 13% above the national average, while it was 8% below the national average in rural places [Preston and Haines, 1991, Table 3.1]. These estimates apply to about 1894. These differences had decline to approximately 6% above and below the national average respectively by the 1910 census [Preston, Ewbank, and Hereward, 1994, Table 3.2]. Fro the Death Registration Area of 1900, urban-rural differentials in e(0) for white males decreased from 10.0 years in 1900/02 to 7.8 years in 1909/11 and to 2.6 years in 1939 for the whole United States [United Nations, 1953, p. 62]. Higgs [1973] estimated that urban mortality was 50% higher than rural mortality in the 1880s, and that the urban penalty had dropped to 21% by the period 1910/20. Condran and Crimmins [1978, 1980] and Crimmins and Condran [1983] found that the rural-urban mortality difference was already diminishing in the 1890s, and that the urban penalty was largely due to tuberculosis, diarrheal diseases, and several other infectious, communicable diseases.

This paper will look at the phenomenon of the urban mortality transition over the period 1800 to 1940 using a variety of sources. Particular attention will be paid to the 19th and early 20th centuries, when we know considerably less and before many of the most heralded public health innovations had come into play. Using some new data, reanalyzing old data, and looking at the public health and medical literature will provide clues as to the relationship of public health (broadly defined) to the urban mortality transition.


Øivind Larsen

Mortality on maps - some methodological considerations

Data for some selected epidemic diseases in Norway 1868-1900 have been compiled from the health statistics, and are presented on maps on county level. As there are important changes in both population and disease pattern in this period, visualization by means of maps may simplify the interpretation of what really happend to the health situation.

A health concern index was calculated, based on a comparison of the contemporary verbal description of the prevailing health conditions to the hard facts told by the figures of statistics. This index has also been prepared on county level and is presented on maps.

A limited edition of the atlas was printed in July 1999, and work is going on to have it published. 1

1. Larsen Ø. Epidemic diseases in Norway in a period of change: An atlas of some selected infectious diseases and the attitudes towards them 1868-1900. Preliminary version. Oslo: University of Oslo, Institute of general practice and community medicine, 1999. xviii + 545 pp.

Session 7 - Mortality experinces: infants, children, women, old people

George Alter & Michel Oris

The Role of Family Behavior in Mortality Decline: A case study from rural Belgium

A central problem in historical research on mortality decline is our inability to observe simple but essential actions, like washing hands, that might have important effects on health and mortality. This study uses inter-household differences in infant and child mortality to look for these health-promoting behaviors. If lower mortality was due to personal behaviors learned in childhood, it should carry over to mortality at older ages. Thus, if lower infant mortality was due to better hygiene, we should also see lower adult mortality among children who grew up in homes with fewer infant deaths. Population registers from the Belgian commune of Sart allow us to reconstruct continuous histories of mortality for families over almost the entire nineteenth century. Although Sart was very poor, mortality in childhood and adulthood declined substantially during this time, but there was almost no change in infant mortality. An analysis of sibling groups of 5 to 8 persons does not support the view that differences among families were caused by health-related behaviors learned in childhood. There was an association between infant mortality and deaths of children between ages 1 and 15. However, infant and child mortality of siblings does not predict adult mortality. On the contrary, adults from families with more infant and child deaths had lower mortality after age 15. This may have been due to immunities acquired from early exposure to diseases or from the selection of individuals with greater resistance to disease. Our results do not imply that health-promoting behaviors were transmitted within families during the nineteenth century.


Jan Sundin

Age, gender and the mortality decline. Some Swedish evidence

No abstract submitted


Marie C. Nelson

Childhood Mortality in 19th Century Sweden: The Case of Norrbotten, a High Mortality County

The mortality decline in 19th century Europe is composed of a number oactors such as urbanization and industrialization, the development of medical care, policy measures, as well as the pandemics of childhood diseases rampart at that time, this paper focuses on an area of particular interest. A preliminary analysis of the aggregate material at the county level has shown that high rates of childhood mortality were persistent in Norrbotten, particularly among older children, throughout the period. The task of this paper is to explore the mechanisms at work e.

A look at the causes of death shows that in the inland region the major killers of children were infectious diseases, such as scarlet fever and diphtheria. Interesting also is the significance of accidental deaths among children, most of which were due to drowning. The town of Luleå, while having some bouts with infectious diseases, showed a steadily increasing dominance of tuberculosis, respiratory diseases and epidemics of meningitis as causes of death among children.

The final answers to the questions concerning this increase in childhood mortality will certainly not provide us with monolithic explanations. This study only illustrates the complexity of the forces at work in Sweden in the late 19th century.


Robert Millward & Frances Bell

Infant mortality and links with mothers' health in England and Wales 1870-1914

This paper reports some preliminary findings on the links between mothers' health, fertility and infant mortality in Britain in the late nineteenth and early twentieth centuries. Historical demographers have painted a rich picture of the spatial dimensions of mortality with the highest rates recorded in environmentally deprived areas. In our own work over the last few years food availability has been introduced to capture some of the factors related to resistance to disease. Infant mortality differs from that of other age groups in its overwhelming dependence on what the mother passes on in the immune system, the foetus and in the quantity and quality of breast feeding. These are difficult to measure and we have developed proxies in the form of measures of mothers' health, for example of female mortality in the child bearing age range. Thus the overall framework in which we are working comprises models of : i) Infant mortality as a function of environmental conditions, food availability, fertility and mothers' health. ii) Mothers' health as a function of fertility, the environment and food. iii) Fertility as a function of child survival rates, child employment and schooling characteristics, mothers' heath, cultural factors. The preliminary findings reported in this paper relate to (i) above and to a sample of 36 towns in England and Wales over the period 1860-1914. The data sources allow a distinction between infant deaths from diarrhoea and other sources of death and the key findings so far are : a) Environmental variables like population density and housing have their main impact on diarrhoealdeaths with a negligible impact on other infant deaths. b)Some of the direct effects of fertility on infant mortality can equally be captured by measures of family size in so far as infant mortality is affected by the number of susceptibles in the home and the strain on food resources from large families. We do however find a residual important role for fertility in that high levels reduce the amount of time which the mother can allocate to infant care. c) Mothers' health plays a powerful role in the variation of infant mortality across towns and over time even when, as here , one controls for the level of housing, population density, fertility and food availability. It has a strong impact on the food variable, reducing its role in diarrhoeal deaths and eliminating it completely for non-diarrhoeal deaths. This is consistent with other UK evidence that infants who died from non-diarrhoeal sources are more likely to have been breast fed ( and hence dependant on the mother ) than infants who died from diarrhoea ( who relied more on artificial foods). These results go a long way to explaining the large fall in infant mortality in the textile towns and in suburbia both of which witnessed significant decreases in fertility and in female deaths from TB as well as improvements in food and housing. It also explains the very small decline in infant mortality in the mining areas and the more or less average decline in rural areas. Further work is currently underway to estimate separate models of mothers' health and of fertility with the long term aim of integrating them all into a consistent picture.

Session 8 - The costs and the risks

Daniel M. Fox

Science, money and policy for reducing mortality

Policy makers and researchers in health care in most industrial countries have had strained relationships with each other in recent decades. Most researchers in health care and health services have not accorded sufficient attention to what policy makers, and the researchers who study them, know about how policy is made and implemented. Three generalizations, grounded in evidence, are fundamental to understanding policy making in health: 1) New policy to protect or improve health is usually made on the basis of consensus among powerful people about assumptions that underlie and guide policy; 2) Policy makers rarely have a mandate from voters and powerful interest groups to accord the highest priority to achieving lower mortality or improved health status; 3) Professionals in politics and public management have solid reasons to be wary of claims by researchers and their advocates that they ought to change or create particular policies.


Louis P. Cain & Elyce J. Rotella

Death and spending: Urban mortality and municipal expenditure on sanitation

U.S. urban death rates from typhoid, dysentery, and diarrhea declined dramatically in the first 30 years of the 20th century as U.S. cities increased their expenditures on water supply, sewage disposal, and refuse collection. In this paper we address two questions: (1) How big was the payoff-how big a decline in mortality did cities get in exchange for the money they spent on sanitation? and (2) What explains the large increase in expenditures on sanitation, specifically did cities respond to shocks from an increase in waterborne death rates by increasing their expenditures?

We use both quantitative and qualitative data to answer these questions. Data from 48 large cities are used to estimate the impact on mortality of expenditures on water, sewers, and refuse collection for the period 1899-1929. Articles from the weekly journal Engineering News provide information about what motivated cities to undertake expenditures on public health measures.

Regression analysis shows that the estimated impact on mortality of expenditures on water, sewers, and refuse collection are large. Over this period a 1% increase in expenditures on sanitation (capital and current expenditures on sewers, water, and refuse collection) led to a nearly 3% decline in deaths in the average sized city.

Cities had a complex set of motivations for engaging in the large expenditures on sanitation that they undertook in this era. We find evidence that many cities in did respond to an episode of high mortality by changing their expenditures on water, sewers and refuse collection. But, only part of the increase of municipal expenditures on sanitation can be seen as responses to mortality crises. Other factors, including demonstration effects from other cities and incrementalism in city budgeting, are important determinants of municipal spending on sanitation.


W.R.Lee & J.P.Vögele

The benefits of federalism? The development of public health policy and health care systems in nineteenth-century Germany and their impact on mortality reduction.

Within a European context there was a substantial degree of variance in the historical development of health care systems in terms of administrative, the allocation of executive authority, the formal representation of medical expertise, and specific policy priorities. To some extent, this was a result of the nature of state power and the historical embeddedness of administrative systems. In the case of Germany, political power remained fragmented throughout the nineteenth century: a great deal of government policy was federal, not national. Even after political unification in 1871, new Reich institutions, such as the Imperial Health Office, were denied executive powers and federalism continued to define the nature and objectives of state policy. Even the introduction of a uniform cause-of-death classification system proved difficult to achieve. The German states, therefore, provide a useful framework for examining the significance of specific health care systems and federal policy initiatives on contemporary mortality trends.

This paper examines the structure of medical administration in individual federal states in terms of the allocation of executive responsibility between different levels of government and assesses the extent to which policy-making was determined by civil servants or directly influenced by medical practitioners. Attention is focused on a number of policy areas, in particular smallpox vaccination, water supply, and infant welfare programmes, in order to establish the extent to which health outcomes were affected by differences in government strategies and the relative efficiency of state administrative systems. State support for breastfeeding campaigns in the first decade of the twentieth century was arguably misplaced and had little impact on infant mortality trends. By contrast, in relation to both smallpox vaccination and water-supply improvements, policy differences at the federal level had a visible impact on health outcomes, just as different systems of public health administration affected the operational efficiency of disease control legislation and other aspects of health care provision.


Øystein Kravdal

Social inequalities in cancer survival

Social differentials in survival from 12 common types of cancer were assessed by estimating a mixed additive-multiplicative hazard model on the basis of individual register and census data for the whole Norwegian population. The excess all-cause mortality among cancer patients compared with similar persons without a cancer diagnosis was significantly related to education, occupation and income. It was, on the whole, about 15 % lower for men and women who had completed a post-secondary education than for those with only compulsory schooling, taking age, period and registered differences in tumor characteristics and stage at the time of diagnosis into account. The data do not provide clear indications of whether differences in host factors, such as co-morbities and immune functions, or differences in treatment and care are primarily responsible for these inequalities in cancer survival.