‘A treatise on all the bad habits of mankind’: Major Greenwood and the political economy of epidemiology in early 20th-century Britain

Introduction

The last chapter of Major Greenwood’s groundbreaking 1935 textbook, Epidemics and Crowd Diseases, is an early attempt to explain the epidemiology of cancer. Describing a ‘sickness which increases with the spread of material civilisation’ (Greenwood, 1935, p. 361), Greenwood ends the book with a question as to the limits of the discipline that he had done so much to establish:

Among crowd evils always with us, whooping-cough is not less deadly than measles, while traffic accidents (which surely come within my definition of a crowd disease) are a good deal more deadly. The trouble is that when one enlarges the definition of epidemiology one is theoretically committed to a treatise on all the bad habits of mankind. (Greenwood, 1935, p. 377; original emphasis)

This is not, however, a biography of Major Greenwood, although another perhaps deserves to be written. ² Instead, through three chronological sections, this article employs Greenwood’s writing and the trajectory of his career to consider the growing political-economic utility of epidemiology prior to the more significant social interventionism of the welfare state. ³ In the first section, we see Greenwood’s population-level perspective develop out of eugenics and the eugenic analysis of medical and vital statistics. We see the intellectual origins and social context that shaped these ideas, as well as their initial sources of patronage. The second section illustrates how Greenwood’s career gathered momentum during and after the First World War, with epidemiology turned to understanding and improving the conditions of military and industrial labour. Finally, as the government’s go-to medical statistician and chair of Epidemiology and Vital Statistics at the London School of Hygiene and Tropical Medicine (LSHTM), Greenwood’s professorial career illustrates the newfound importance of epidemiological reasoning for both British and British-imperial government. Bridging the theoretical study of disease and its practical application within state apparatuses, the approach to epidemiology that developed under and around Greenwood reflected and complemented the political economy of medicine and public health during these years. Prior to the advent of the National Health Service in 1948, ill health was costed more with a mind to human capital and potential output than the cost of care. Here, population-level analyses and a probabilistic approach to the health of groups rather than individuals found favour because they promised broad and economical interventions in pursuit of the national health. As national health became more firmly understood as a fundamental determinant of national success, epidemiology provided a means towards more healthful and productive populations. This – the political economy of early-century epidemiology – was a notable departure from that which came before and emerged after.

The growing governmental utility of epidemiology during this period was made possible by intellectual developments in the field – changes that Greenwood both contributed to and effectively translated for the policymakers he came to advise. Already upset by the discovery of ‘carrier status’ in the 1890s, the influenza pandemic in 1918 illustrated the limitations of contemporary public health and demanded the reorientation of epidemiology away from bacteriology explanations for disease incidence and spread (Mendelsohn, 1998, p. 306; see also Amsterdamska, 2005). Perhaps understandably, however, the histories of epidemiology covering the years after the First World War – the period in which ‘epidemics became complex’ (Mendelsohn, 1998) – have tended to focus on how understandings of infectiousness and virulence developed within infectious disease research (see Amsterdamska, 2005; Fine, 1979). This research included the development of more advanced models for predicting the spread of infectious disease, including by Kermack and McKendrick (1927) in Britain and by Reed and Frost in the USA (see Engelmann, 2021a). Elsewhere, epidemiological complexity came via the study of pathogens, through recognition of pathogenic diversity and greater understanding of their capacity to change (Anderson, 2004; Engelmann, 2021b; Méthot, 2016). This, however, has somewhat overlooked the concurrent development of population-level research as a higher-order approach to epidemiology. Of particular interest here is the strand of epidemiological thought that emphasised the relative robustness of the host, rather than the relative pathogenicity of the agent or vector of disease (see also Mendelsohn, 2001).

This reaction to epidemiological complexity, pioneered in no small part by Greenwood, applied modern statistical methods to an older integrative understanding of health as influenced by social, demographic and environmental pressures (see e.g. Coleman, 1982). Embracing the analysis of vital and social statistics alongside biomedical data, this new approach to epidemiology drew from and contributed to the history of demography (Ittmann, 2003) and ‘population thinking’ in medicine (see Ariew, 2008); bore some relevance for infectious disease work in terms of population-level resistance to infection; and offered insight into what would later be understood as the social determinants of disease. Focusing on the vigour of a given population rather than the vigour of a given microbe, it was this higher-order epidemiology that found purchase and patronage amidst the industrialism, militarism and imperialism of Britain’s early 20th century. Unlike research charting the trajectory only of infectious disease, Greenwood’s more integrative epidemiology came at a time when the physical capacity of labouring human bodies was seen as increasingly relevant for Britain’s national ambitions.

Eugenics, demography, and the political economy of epidemiology before the First World War

Born in 1880, Major Greenwood had wanted to study history but, following his father’s wishes and in his father’s footsteps, instead enrolled at the London Hospital Medical College. Here, he was drawn to the physiology laboratory, where, for the physiologist Leonard Hill ‘and a little later’ for Karl Pearson, he recalled developing ‘an almost schoolgirl passion’ (quoted in Hogben, 1950, p. 139). After reading Pearson’s Grammar of Science at the age of 19, Greenwood could ‘henceforth envisage medicine as a career of endless opportunity for measurement and for mathematics – with humane studies – his first love’ (Hogben, 1950, p. 140). Given this background, it is perhaps unsurprising that Greenwood came to see mathematics as a means to detail the physiological effects of social space, and to consider health as a product of social history. In this respect, and as has been noted elsewhere (see Mendelsohn, 1998, p. 309), from the beginning of his career Greenwood worked to integrate modern analytical statistics with the older holism of the pre-bacteriological tradition. However, his early work also reflects the economic anxieties and political prescriptions that flourished prior to the First World War. Under the influence of Pearson and Arthur Bacot – the entomologist and Fabian, and Greenwood’s dear friend – Greenwood’s early epidemiology worked in support of the gradualist middle-class socialism of this milieu. ⁴ Central to these politics at this period, and to their application in medicine, was eugenics (see MacKenzie, 1976, 1979). Eugenics offered a compelling explanation for disparities in gross mortality and morbidity as well as a programme to improve health at the population level as part of the broader engineering of society. In this vein, Greenwood’s early epidemiology included work to distinguish eugenic and dysgenic physiologies while also attending to differential fertility as an epidemiological and political good.

Already by the turn of the 20th century, in Britain the widespread employment of statistical inference had made medicine ‘a more multifaceted notion’ (Matthews, 1995a, p. 31) than in continental Europe where, following Louis Pasteur and Robert Koch, laboratory methods and bacteriology defined medical science. Here, long-standing efforts to collect and consider social and vital statistics included work to understand how health shapes population and population shapes health. This tradition can be traced at least to the close of the 18th century and Thomas Robert Malthus’s 1798 An Essay on the Principle of Population. The Principle’s basic assumption – that whenever a population increases beyond the bounds of the environment, the ‘positive checks’ of war, famine and disease curtail ‘overpopulation’ – is rarely written into histories of epidemiology. However, by promoting the idea that there was an optimal population on which the health of the nation might rest, the Principle backgrounds the subsequent development of enumeration as a biopolitical tool (see e.g. Hacking, 2015).

Despite criticism from the more numerate 19th-century social scientists, like Adolphe Quetelet or William Farr, the tenor of Malthusianism carried into the 20th century as a political economy of population, promoted variously by radicals, feminists, colonialists and eugenicists (Hall, 2000). The most notable of the eugenecists, Francis Galton employed evolutionary theory alongside social statistics in support of the idea that humanity might be improved through selective breeding (see Porter, 1986, pp. 231–314). To this end, eugenicists argued for greater scientific and governmental consideration of differential fertility (MacKenzie, 1976). As with many early eugenicists, Galton was critical of the ‘neo-Malthusians’ who, arranged into the Malthusian League from 1877, argued that the social disruption of overpopulation should be mediated via birth control. Fearing that any generalised endorsement of fertility reduction would be taken up disproportionately by the wealthy and well educated, Galton’s eugenics instead encouraged population growth amongst more ostensibly desirable groups while simultaneously ‘checking’ the ‘process of evolution … pushing towards the bad’ (Galton, 1891, p. 12; see also Klausen & Bashford, 2010). During the 1880s, Galton developed regression and correlation to prove relationships initially between anthropometric measurements, and then broader social problems, including poverty, crime and disease (Porter, 1986, pp. 136, 296). The driving aim of Galton’s research – the advancement of statistics in furtherance of eugenics – was picked up from the 1890s by Pearson. Mathematical statistics flourished under Pearson’s direction, and his development of biometrics significantly influenced anglophone epidemiology. As Galton before him, Pearson’s understanding of disease was notable for its consideration of multi-causality, albeit consistenly emphasising that the ‘hereditary constitutional factor is immensely more important than the infection factor’, as per his famous analysis of tuberculosis (Pearson, 1912, p. 12). Although he came to resist their more reductive assertions, Greenwood was an inheritor and advocate of both Pearsonian statistics and the British eugenic tradition.

In 1904, following a heated debate in the medical press regarding the bacteriologist Almroth Wright’s ‘Opsonic Index’ – an attempt to quantify an individual’s reaction to infection – Pearson (1904, p. 1667) wrote in the British Medical Journal that there was a ‘crying need for a more exact treatment of statistics in medical science’. Then coming to the end of his medical training, Greenwood had already been corresponding with Pearson for two years and had resolved ‘to be an honest soldier under your [Pearson’s] flag’ (quoted in Matthews, 1995a, p. 35). This same year, Greenwood (1904) published his first article in Pearson’s journal, Biometrika. Taking data from the London Hospital’s postmortem records, here Greenwood correlates the weights of various internal organs. Following Pearson, his conclusions emphasise a ‘constitutional’ basis for disease. Finding ‘the individuals who die of disease in adult life to be more variable and less highly correlated in their organs than the “healthy”’, and working from the assumption ‘that the hardiest and most prolific groups have the least coefficients of variation’ (p. 73), Greenwood’s (1904) first foray into medical statistics offered postmortem evidence for eugenic and dysgenic physiologies.

Away from the London Hospital, the quantitative challenge to bacteriology and the importance of eugenics in British medicine was complemented by a deeply pessimistic prognosis for the health of the British public. The ‘condition of England’ question, a driving concern of 19th-century social science, had swelled to incorporate the use of anthropometry and articulate a crisis of ‘physical deterioration’ (see Searle, 1971; Soloway, 1982). Between 1845 and 1901, the height requirements for British military recruits had to be dropped by six inches; the inauspicious performance of the British military during the South African War of 1899–1902 was likewise attributed to physical disparities between British and Boer soldiers. In response to fears regarding Britain’s declining ‘national efficiency’, in 1904 the government established the Inter-Departmental Committee of Physical Deterioration. The explanations for physical deterioration included worsening diets and the malign effects of urbanisation. Yet the sanitarian reforms central to Victorian public health had significantly reduced mortality from epidemic disease, even within urban slums. Clearly, then, infectious disease was not the only cause of epidemic physical decline. Eugenicists suggested that these sanitarian interventions had actually disrupted the eugenic check of infant mortality, whereas Edwardian advocates of ‘preventative medicine’ agitated for more state involvement against the environmental causes of disease (see Porter, 1991). Legislation was soon passed both to improve living conditions and to understand their epidemiological effects. This included the National Insurance Act of 1911, which led to the establishment of the Medical Research Committee in 1913 and, a year later, its department of medical statistics, in which Greenwood would soon become a central figure (Higgs, 2000, p. 328). The statistical analysis of physiology that had captured Greenwood’s imagination 10 years prior had, on the eve of the First World War, gained and the explicit support of the state.

In the intervening years, Greenwood had been quietly working, first in his father’s general practice, then as a demonstrator in physiology at the London Hospital Medical College. He had also been honing his knowledge of statistics, taking Pearson’s course at University College, teaching a little himself, publishing sparingly and, in 1908, setting up a Statistics Laboratory at the London Hospital (see Farewell & Johnson, 2014). It was this laboratory which caught the attention of Charles James Martin, then director of the Lister Institute and, at the start of 1910, Greenwood was appointed head of the Lister’s new Department of Medical Statistics, the first such position in the country. Although the Lister was intended to be a British Institut Pasteur, and had previously focused primarily on bacteriological research, Greenwood seems to have been given the job because of his public willingness to check the research being produced by British bacteriologists and, in particular, the research being produced by Almroth Wright. ⁵ It may, in fact, be no coincidence that his first publications at the Lister worked to systematically dismantle the statistics found in the 1898–1899 Indian Plague Commission (Greenwood, 1910), ‘the report of which is practically his [Wright’s] own work’ (letter from Greenwood to Pearson, quoted in G. B. Greenwood, 2016, p. 59). The growing demand for statistical rigour from the Lister and then the Medical Research Committee might ultimately be seen as a reaction against the often incoherent use of statistics in medicine, as epitomised through this period by Wright’s Opsonic Index (see Matthews, 1995a). The conventions of analytical statistics, by contrast, promised neutrality, order and new standards. This Greenwood also suggests in a 1909 address to the Pathology Section of the Royal Society of Medicine. After detailing the varied mathematical problems with Wright’s calculations, his speech concluded with the prediction that ‘the time will come when such co-operative work is not the exception but the rule, to the benefit of both kinds of investigator’, bacteriologist and biometrician (Greenwood, 1909, p. 155).

Alongside these scientific benefits, it seems likely that there was some dawning recognition of the administrative potential of analytical statistics too, especially for those large organisations directing, co-coordinating and funding research, like the Lister and the Medical Research Committee. Requiring little more than space, staff and a few calculating machines, biometric analyses offered a relatively inexpensive way to police much more expensive research agendas. Indeed, throughout this period, Greenwood’s arguments against bacteriological programming in public health typically turned to economics, to questions of cost, waste and the allocation of finite public finances (see e.g. Mendelsohn, 1998, p. 309; Squire, 1911, p. 46).

Often, as in Wright’s case, the problem with medical data was not so much due to ‘errors of technique’ – ‘experimental or observational error … owing to want of skill’ – as ‘mathematical errors … owing to the fact that the sample may not typify the characters of the population from which it is drawn’ (Greenwood, 1909, p. 146). A persistent concern throughout Greenwood’s career, the fundamental problem of population sampling and skewness limited the explanatory potential of a great deal of medical data. Conversely, by working to consider more complete populations, the epidemiological analysis of social and vital statistics had some potential to transcend these limitations. By dealing with death and disease at the population level, these sources of data also had the greatest epidemiological, eugenic and administrative potential. Already by 1908, an anonymous article in the British Medical Journal – almost certainly written by Greenwood (see Farewell & Johnson, 2014) – spoke admiringly of the recent German and Dutch publications that mobilised administrative data to correlate infant mortality with feeding practices, seasonality and class (Infantile mortality, 1907).

Soon, Greenwood’s work at the Lister also moved towards the epidemiological analysis of vital statistics and gross mortality. In 1912, for instance, he published an analysis of the relative influence of poverty and fertility on infant mortality. Tapping into eugenic and neo-Malthusian concerns that demographic interventions might be necessary to address the national burden of ill health, here he argues that differential birth rates were a potential threat to morbidity and mortality at the population level and over the long term:

Suppose we sorted out our population into two mutually exclusive classes, one comprising the tuberculous, the criminal, the mentally ill-balanced … the other class being formed of the rest of the population … Also let us suppose that the fertility of the first class is greater than that of the second. What consequences may be expected to ensue? If our tuberculous and insane stocks are so in virtue of an innately low resisting power towards the harmful elements of a common environment, we shall hardly be surprised if their offspring are in turn less fit to combat the ordinary dangers to which infants are exposed than are the children of normal persons; in other words there would appear to be some correlation between fertility and infant mortality. (Greenwood, 1912, pp. 289–290)

After the war, eugenics became increasingly associated with the extreme right wing of British politics. Groups like the Eugenics Education Society – of which Greenwood was an active member, teaching on their courses throughout the 1910s – shifted from political campaigning to scholarly debate (MacKenzie, 1976, p. 518). Greenwood’s eugenics soon also softened. However, as a way to explain health differentials at the population level, and as a political-economic project aimed at the betterment of the national population, eugenic reasoning was foundational for Greenwood’s epidemiology. His enduring interest in population dynamics is one way in which his work ‘wed modern statistical methods to neo-hippocratism’ (Mendelsohn, 1998, p. 309). This should also be read as the result of early-century political anxieties around demographic fitness and national health. Greenwood’s lifelong interest in vital statistics speaks to the influence of eugenics in 20th-century epidemiology and to the tangled history of demography and epidemiology in the period before their crystallisation as distinct disciplines. ⁶ Attention to population dynamics and demographic data offered the greatest potential for either hygienic or eugenic intervention and, once employed at the Lister and more proximate to government, Greenwood’s ideas were articulated with greater mind to policy. Here, statistical analysis offered both epidemiological insight and administrative direction. Although always advocating for broad, sometimes radical, improvements in public services and state support, that such measures would positively influence national health was little more than an ‘elementary truism’ (Greenwood, 1912, p. 298). ⁷ What statistics instead offered was a way to judge ‘the amount of success they [the various interventions] may be expected to achieve under present conditions of society’ (Greenwood, 1912, p. 287), and to assess the degree to which the most pressing epidemiological concerns – like infant mortality – were ‘within the field of administrative control’ (Greenwood, 1912, p. 291). By this point, and more so in the coming years, the political affordances and economic efficiencies offered by statistical analysis were being more widely recognised across the British medical system. It was, however, his particular affinity for the realpolitik of state medicine – as well as his ability with numbers – that left Greenwood uniquely positioned for a long career alongside government.

Human capital, population dynamics, and the political economy of epidemiology during and after the First World War

The First World War was a turning point in the history of epidemiology. The influenza pandemic that spread rapidly with demobilisation in 1918 could not be easily explained by contemporary bacteriology, stimulating greater complexity in epidemiological reasoning (see Amsterdamska, 2004; Mendelsohn, 1998). Even before this, however, the war itself had fundamentally changed the orientation of British epidemiology. Enduring questions about Britain’s national efficiency were compounded by the challenges of a total war. Increasingly, the health and physical capacity of workers on the home front came to be seen as expressly relevant to the war effort, with statistical inquiries used to direct improvements in the human capital necessary for the war economy. As the biopolitical relevance of medical statistics gathered pace, so did Greenwood’s career. Beginning the war teaching military hygiene and inspecting camp sanitation as part of the Royal Army Medical Corps, Greenwood was quickly seconded to the Ministry of Munitions, spending the rest of the conflict working on questions of health and labour wastage in munitions factories. The peacetime utility of Greenwood’s approach was soon also recognised, first by the newly established Ministry of Health, where in 1919 he was made Senior Statistician, and then, from 1928, by his appointments as a professor at LSHTM and director of the Medical Research Committee’s Statistical Department. Although also a notable actor in the interwar development of infectious disease modelling, it was the epidemiology of industrial illness that began Greenwood’s civil service career and consolidated the political-economic utility of epidemiology. Drawing on advances in the theoretical and mathematical modelling of populations, as before the war, Greenwood’s epidemiology demanded fuller consideration of demographic indices and their social determinants. This approach furthered critiques of first bacteriological and then eugenic reductionism. While this contributed to the development of increasingly complex epidemiological models, it was the more practical application of epidemiology – and its ready implementation in aid of British interests at home and abroad – that drove the institutionalisation of the field into the 1930s.

A student of vital statistics in the 19th-century tradition, and an intellectual inheritor of Galton and Pearson, for Greenwood bacteriology had always been somewhat limited in its explanatory potential. In a 1916 letter to Nature, written before the onset of the influenza pandemic, he had argued that ‘it is high time that epidemiology is extracted from its present humiliating position as the plaything of bacteriologists’ (Greenwood, 1916). By this point, however, Greenwood had also started to become sceptical of hard-line eugenic reasoning in medicine, finding that Pearson’s dogmatism stifled consideration of other determinants of health and disease (Mendelsohn, 1998, pp. 308–310). In 1913, Greenwood broke from Pearson’s laboratory at University College following a heated disagreement over the suggestion that nutrition had nothing to do with a child’s academic achievement, describing this conclusion as ‘simply SHIT from beginning to end’ (letter to Yule, quoted in Kevles, 1986, p. 337). Instead, as submarine attacks stifled imports and the shift to wartime production compromised British agriculture, Greenwood actively pursued the integration of epidemiology with the new science of nutrition (see e.g. Greenwood & Thompson, 1918).

Over the course of the war, as food became increasingly relevant for military strategy, the biopolitical relevance of nutrition grew exponentially (see e.g. Smith, 2009). Yet, as first with bacteriology and then eugenics, Greenwood was resistant to the more reductive conclusions emanating from the subsequent explosion of biochemical nutrition research (see Petty, 1989). Sometime later, in a letter to the Medical Research Committee’s Walter Fletcher, Greenwood explained:

My emotion boils over quite as much as yours, but not because the powers that be don’t read the latest news from the research front about vitamins or the specific dynamic action of protein, but because people are told to use their common sense and economically stopped from doing so. This subject is a morass of dull trifles, which researchers are far too scientific and superior to notice … I submit that the people of this country are in no pressing need of any kind of preaching, what they are in need of is an economic organization enabling them to do what they for the most part knew suited them, before the Health or the Medical Research Council existed. It seems to me utterly untrue that by some simple, unexpensive, change in diet, any serious improvement can be effected. ⁸

As also apparent in the growing interest in nutrition, it was the shift to wartime production that generated excitement over the environmental causes of illness. As Blayney’s (2019) history of ‘industrial fatigue’ has recently explained, ‘Britain’s entry into the First World War, and the demands it placed on domestic industry, made the question of the limits to the body’s productivity a question of national political – perhaps even existential – significance’ (p. 319). Following the ‘shell crisis’ of May 1915, when a public outcry followed the revelation that British armaments factories were struggling to keep up with demand from the front, the health of factory workers became increasingly relevant to the war effort. By September 1915, then prime minister David Lloyd George had created the Health of Munition Workers Committee, which included amongst its varied membership the physiologists Leonard Hill and Walter Fletcher, both of whom were also involved in the administration of the newly established Medical Research Committee (Blayney, 2019, p. 320). Greenwood was seconded to the Ministry of Munitions the following February and, although not a permanent member of the Health of Munition Workers Committee, he contributed to the statistical analysis of committee reports as well as to its peacetime analogue, the Industrial Fatigue Research Board. From 1916, Greenwood became the government’s go-to source for statistical analyses of physiology and bodily efficiency. During the war alone, he contributed to government investigations on the physiological effects of alcohol, including its effects on factory work; the relative nutritional needs of different workers; the most efficient cadence for marching soldiers; food consumption amongst factory workers; accident-proneness amongst factory workers; and labour wastage amongst female factory workers (for details, see Farewell & Johnson, 2014, 2016; G. B. Greenwood, 2016).

The culmination of Greenwood’s work on industrial epidemiology was a 1921 textbook, written with another member of the Health of Munition Workers Committee, Edgar Leigh Collis. The Health of the Industrial Worker is both an articulation of Greenwood’s integrative approach to epidemiology and a stark conflation of epidemiological and economic logics. Here, ‘the goal of the economist – output – can be best attained through the same agencies as allow the medical man to obtain his objective – health’, whereas ‘industrial fatigue is a drag on the wheels of industry impeding the course of progress’ (Collis & Greenwood, 1921, pp. 79, 80). Influenced by the eugenic analysis of pathophysiology and vital statistics, similar mechanistic considerations of bodily efficiency are also readily found in Greenwood’s earlier publications (see e.g. Greenwood, 1912, p. 298). However, it was over the course of the war, as ‘work became the fundamental telos of the body’ (Blayney, 2019, p. 327), that industrial health provided a new social role for physiology. The pressures of wartime production also illuminated the economic relevance of population-level epidemiology as applied away from infectious disease. As the Health of Munition Workers Committee concluded in 1916, the study of industrial fatigue should not consider ‘the individual, taken at any one moment’, but ‘his life history, his heredity, his family, his domestic life, his personal habits and customs, his home as well as his workshop’ (quoted in Blayney, 2019, p. 327). By considering the aggregated effect of individual physiological stresses, it was the science of epidemiology that emerged as the principle means to assess and gradate the economic effects of ill health.

After the war, Greenwood was soon brought into the Ministry of Health, which, as per its second annual report, was now fully invested in ‘statistical science as an aid to sound medical administration and as an instrument of research’ (Ministry of Health, 1921, p. 154). Although clearly pleased that his ideas were finding traction, Greenwood remained concerned about the standard of statistical literacy and the misapplication of statistics. ‘What God really meant me to do’, he wrote to Fletcher in 1920,

was first to move to [the National Institute for Medical Research in] Hampstead and valet old [Leonard] Hill for a few years at the same time developing a little training centre in statistic-physiologico-industrial work at Cambridge to prevent the young hopefuls of the Industrial Fatigue Board wasting so much of their own and other people's time (and money) over crudities. ¹¹

In this same article, Greenwood and Yule (1920) offer ‘a general solution of the problem of the distribution arising when the chance of a happening is affected by antecedent success or failure’ (p. 276). In doing so, they had doubtless been influenced by Ronald Ross’s (1916) ‘a priori’ modelling of ‘happenings’, inclusive of epidemics. Although employing markedly different approaches, the Ross and the Greenwood and Yule articles both employed probability theory to develop some predictive capacity in epidemiology. Less well known than his Nobel-prizewinning research on the malaria parasite, Ross’s probabilistic modelling of various ‘social happenings’ immediately struck a chord with Greenwood. The same year, he wrote that:

modern [epidemiological] researches fall into two classes. On the one hand those directly or indirectly inspired by the epoch-making discoveries of Professor Karl Pearson in the theory of mathematical statistics; on the other, the independent investigations of Sir Ronald Ross. (Greenwood, 1916, p. 243)

While Pearsonian statistics offered valuable analytical tools, Ross worked towards a more integrative model which utilised demographic indices to illustrate and structure epidemic patterns:

he looked at the problem as one of transfer, viz., of mutual interchange between groups of affected and unaffected individuals, an interchange complicated by the subjection of each group to certain rates of natality, mortality, emigration and immigration. (Greenwood, 1916, p. 244)

Despite such praise, Ross (1929, p. 637) came to complain that few people had shown much interest in his expansive model of happenings, aside, that is, from Greenwood in Britain and Alfred J. Lotka (1912, 1919) in the USA. ¹² Although never substantively employed by a university (but rather by an insurance broker), Lotka was nevertheless integral to the development of demography between the wars, pioneering an inductive approach to the study of population that was not reliant on empirical data. Extending the analysis of life tables beyond the study of mortality, Lotka developed age-specific fertility rates in order to add dynamism to demographic forecasting. This was something of a reaction against the biologists then also studying the question of population. Popularising the logarithmic curve as a utilitarian ‘biological law’ of population growth, Raymond Pearl had considered the USA’s slowing population growth in broadly Malthusian terms, as a population nearing its biological limit (see Pearl & Reed, 1920). By contrast, Lotka and his colleague at the Metropolitan Life Insurance Company, Louis I. Dublin, formulated a non-Malthusian theory of population that emphasised depopulation prior to stabilisation, rather than overpopulation prior to collapse (Dublin & Lotka, 1925; see also Kingsland, 1985; Ramsden, 2002). Greenwood had undoubtedly been following this debate and perhaps saw the computational biologists and the vital statisticians as contributing to either of the ‘two classes’ of research that held value for the study of epidemiology. In 1927, for instance, Greenwood had suggested to the Medical Research Committee that Pearl might take Pearson’s chair on his eventual retirement: ‘with Raymond Pearl at University College the statistical future would be plain sailing … [although] not nearly of K.P.'s [Karl Pearson’s] class he is quite certainly the next best man alive on the eugenics line’. ¹³ Some years later, in a discussion at the Royal Statistical Society, Greenwood likewise hailed the way that

Lotka gave a new direction to those interested in theory and new tools to those who wished to make arithmetical studies … my own students are left in no doubt as to whom I consider to be the real population expert in the statistical field! (Greenwood, 1937, p. 413)

On either side of the Atlantic, Ross’s mathematical modelling of happenings was apt to find favour with Greenwood and Lotka. Both appreciated a theorisation of epidemics that foregrounded population dynamics as of fundamental importance to the spread of disease. ¹⁴ However, and unlike at the start of the century, this was not a period of pronounced demographic anxiety, and the study of population dynamics held limited political-economic importance. ¹⁵ Where vital statistics were of more practical use – and could be made even more useful in collaboration with other data and through the proper application of correlation or regression – was the population-level analysis of health and disease. Although always interested in mathematical modelling and always ready to emphasise the theoretical limitations of any data, Greenwood’s career progressed largely because of the administrative potential of analytical statistics. This utility drove the institutionalisation of epidemiology in Britain, at the Medical Research Committee, the Ministry of Health and then LSHTM. Greenwood was also well aware of this. As he explained in 1931:

the scientific output of the [LSHTM] Division of Epidemiology and Vital Statistics … requires an intimate liaison with the statistical and epidemiological work of the Medical Research Council, the Industrial Health Research Board and the Medical Department of the Ministry of Health … No student ought to leave us academic-minded, in the bad sense, he should never be an Athenian blockhead, for he should know how the world's work is and must be carried on. (Greenwood, 1931, p. 543)

Here, it is important to recognise that ‘the world’s work’ at the time – or at least for those in the orbit of LSHTM – often related to the administration of Britain’s global empire, and that LSHTM’s approach to epidemiology worked towards the maintenance of healthful and efficient working populations both at home and in Britain’s colonies. As with other forms of biomedical research, the colonial extension of epidemiology tended to prioritise the same pursuit of physical efficiency as industrial-fatigue research had a decade earlier. ¹⁷ Indeed, as Hirsch and Martin explain in LSHTM’s accounting of its own colonial history, its involvement in colonial research encouraged

greater efficiency in the economic exploitation of the natural and human resources of the colonies. This, rather than the medical care for colonised populations was the main concern of colonial governments and is reflected in the School’s syllabus, the staff hired, and the research conducted. (Hirsch & Martin, 1960, p. 134)

Model populations, modelling populations: Experimental epidemiology and its imperial extension

From the middle of the 1920s, Greenwood’s earlier disinterest in bacteriology started to soften, largely due to a long collaboration with William Topley. Alongside Topley, who was made Chair of Bacteriology at LSHTM in 1928, the same year Greenwood took up his professorship, Greenwood spent much of his later career focusing on the development of experimental methods in epidemiology, introducing infections into mouse populations and detailing their spread. As they note in one of their earliest collaborative articles (Greenwood & Topley, 1925), this work was demanded because ‘the great majority of investigators have been so preoccupied with the individual that they have neglected the herd’ (p. 56). Amsterdamska (2004) has described this as the development of a ‘view of epidemics as distinct biological phenomena, disturbances in a multifactorial equilibrium’ (p. 490). In this respect, mouse models offered a way to test Greenwood’s earlier thinking around the population-level determinants of physiological resilience. Such an approach bore relevance not just for infectious disease research but also for the more general study of health at the population level. It is, however, also important to recognise that the development of experimental epidemiology at LSHTM coincided with and complemented the massive extension of biomedical research throughout the British Empire (see e.g. Clarke, 2013; Moore, 2016). Where mouse models granted insight into the development of the epidemic curve in an isolated population, the diversity of the colonised world provided a ‘living laboratory’ (Tilley, 2011) and access to an enormous diversity of disease environments. Although distinct, these evolving methodologies employed the manufacture, manipulation and surveillance of model populations in pursuit of epidemiological reason. Both approaches found favour under the remit of LSHTM.

Initially, in the years immediately after the First World War, Topley’s (1919) solo experiments with induced typhoid epidemics in mice sought to show that ‘an increase in the pathogenicity of the specific parasite is an essential factor in the rise of epidemics’ (p. 3). Over the course of his experiments in the next few years, however, he started to question the importance of changes in bacterial virulence (Topley, 1922). By 1923, he was explaining instead that

we are led, indeed, to think of the relation between a potentially harmful parasite and its animal hosts as an unstable equilibrium – an equilibrium which may be readily upset by migration, or by such changes in the ways of living or of transit as will result in freer and more intimate mingling of units of population. ¹⁸

Although an important antecedent to ‘ecological understanding’ within infectious disease research, this approach largely avoided the ‘germ’s-eye view’ (Joshua Lederberg, quoted in Anderson, 2004, p. 40) necessary for the later microbial ecologies of disease that sought to integrate disease, environment and evolutionary processes. Greenwood and Topley’s experiments would, instead, go on to upset the equilibrium in other ways – through inoculation or variations in diet (e.g. Greenwood et al., 1931; Topley et al., 1931). Yet diet seemed to offer a minor and inconsistent guard against infection; inoculation, likewise, provided only partial and sometimes temporary protection. Ultimately, and as per his previous critiques of narrow or technical solutions to the structural and social causes of illness, Greenwood (1935) came to the conclusion that experimental epidemiology showed that ‘the trouble was a continuance of membership of a herd exposed to risk’ (p. 75; original emphasis). Returning to the logics of efficiency and economy that had driven his war work, these experiments illustrated that ‘it is a great deal better to provide clean houses and food than to pre-immunise people against the possible consequences of dirty houses and food’ (Greenwood, 1935, p. 75; original emphasis). For infectious disease modelling, the primary finding of relevance was that ‘a circulation of non-immunes seems to be a necessary condition of epidemicity’ (Greenwood, 1932, p. 39). In other words, the immigration of healthy non-immunes was seen to reinvigorate epidemics that had otherwise finished, the rest of the population having died or recovered but still carrying the infection: ‘in order to bring a herd illness to a definitive end, the introduction of new members must be wholly stopped’ (Greenwood, 1935, p. 70). This finding offered further proof that infectious disease research could not rely on clinical data alone. Demographic data was also essential.

While the Greenwood–Topley period of close collaboration initiated the explicit consideration of a host population’s susceptibility to disease, Greenwood always recognised the limits of these methods, or that ‘experimental epidemiology has the same relation to human epidemiology that … simple Latin prose has to the compositions of Cicero’ (Greenwood, 1935, p. 69). Unable to exert experimental controls on human populations, the extension of this research demanded detailed epidemiological surveys of a diversity of host populations. Beginning around this time – the early 1930s – large-scale surveys mobilised biometrics to detail correlations between health, diet, and various other social and demographic variables (see e.g. Rowett Research Institute, 1955). The institutional orientation of LSHTM facilitated the imperial extension of such work, offering ready access to the ‘natural experiments’ found in the vast variability of the British Empire. Prior to his appointment at LSHTM, Greenwood had asked Andrew Balfour, the School’s first director, for a ‘list of the islands the vital statistics of which would in your judgment repay investigation from the standpoint of migration studies’. ¹⁹ These were envisaged as the potential human translations of the mice models. Balfour, however, was not hugely optimistic, because it ‘is a little difficult to say in which of them the vital statistics can be absolutely relied upon’. ²⁰ In order to address such shortcomings, in the early 1930s Greenwood’s Division of Epidemiology and Vital Statistics began to provide future imperial administrators with training in Medical Statistics and Racial Hygiene, a course that included ‘census taking and vital registration … among primitive or illiterate peoples’, and emphasised the necessity of ‘elementary cultural anthropology’. ²¹

Greenwood, however, had little academic interest in British colonies. Rather than any moral or political opposition, this, it seems, was because of the paucity and poor quality of the data collected by Britain's imperial governments. ²² Indeed, while at LSHTM, he advised the Colonial Office (Higgs, 2000, p. 335); taught many cohorts of colonial administrators; represented the University of London in at least one Imperial Social Hygiene Congress (Universities and colleges, 1939); and commented encouragingly on his colleagues’ research into colonial statistics. It is this last example that offers the greatest insight into Greenwood’s position. In the foreword to a 1944 book titled Vital Statistics and Public Health Work in the Tropics by P. Granville Edge – a friend and military man who had also taught in the Division of Epidemiology and Vital Statistics – Greenwood explains how he hopes readers will ‘have their courage strengthened by this book … because that is the virtue a vital statistician working under a tropical sun chiefly needs’ (Edge, 1944, p. v). Here, courage was required to deal primarily with the problems of the data, with the fact that, despite ‘generations’ of British government in many colonies, ‘the vital book-keeping … is far cruder than that of the home country three hundred years ago’. This was especially troublesome given ‘what a smaller nation has achieved’, with Dutch colonial statistics conferring both ‘admiration and also a sense of national humiliation’ (Edge, 1944, pp. vii–viii). ²³

Implementing Greenwood’s ideas overseas and meeting the demand for better vital statistics was met in various ways by various imperial administrations. In the Gambia, for instance, from the 1940s, the first director of the Medical Research Committee unit spent 30 years producing detailed demographic profiles for the entire population of four coastal villages (see Kelly, 2015, p. 306). Elsewhere, and at the same time, other surveys employed more advanced mathematical modelling to make up for the lack of vital registration. The East African Medical Survey, for instance, surveyed thousands of individuals in six locations across Kenya and Tanzania, collecting ethnographic and demographic data, alongside stool, urine, blood and skin samples. The Survey also employed Bill Brass – who would later join Greenwood’s old department as LSHTM’s first professor of demography – in order to extrapolate vital statistics from the limited fertility and mortality data that could be gathered from cross-sectional questionaries. ²⁴

Greenwood had ostensibly retired from LSHTM by the time these projects started, although he never really did, dying at the age of 69 as he rose to speak at a meeting of the Cancer Research Committee (G. B. Greenwood, 2016, p. 164). However, at least within the British sphere of influence, innovations in the production of vital statistics developed out of the institutions that he had done so much to develop. For a time, through the late 1940s and 1950s, Greenwood’s influence endured in ‘social medicine’, a loosely defined discipline that sought to apply statistical methods to ‘the study of all environmental agencies, living and non-living, relevant to health and efficiency’ (Crew & Hogben, 1948). ²⁵ However, even by the end of the 1940s, domestic epidemiology had begun to move away from the data-led, population-level studies and interventions that Greenwood had routinely argued for. Instead, epidemiology was shifting to focus on individual bodies and lifestyles, individuating both the risk of contracting or developing a disease and the responsibility for its mitigation or cure (see Phillips, 2025). These changes precipitated and complemented the decline of social medicine during the 1960s, the increase of consumerism within British health care, and the demands of an increasingly ‘autonomous’ patient public (see Coulter, 2002; Pickstone, 2003; Porter, 1997, pp. 97–119). Generally speaking, these political-economic developments were absent from countries dealing with the challenges of post-colonial health provisioning. Again generally speaking, incoming post-colonial governments also tended to eschew the technocratic statistical targeting of ill health, focusing instead on economic redistribution, public health programmes and the broad extension of primary health care (see Nott, 2025). How and where Greenwood’s ideas survived deserves fuller consideration. However, the tenor of Greenwood’s epidemiology – the generation of population-level data and the pursuit of broad population-level interventions – persisted in externalist analyses of health and well-being in the Global South: in development studies, tropical medicine, medical demography and international health; in the analyses of the United Nations, Food and Agriculture Organization, and World Health Organization; and in specialist institutions, often with imperial beginnings, like Greenwood’s old department at LSHTM.

Conclusions, speculations, and directions for further research

The 1934 LSHTM annual report describes the activities of the Division of Epidemiology and Vital Statistics as follows:

Mr. E. Lewis-Fanning has collaborated with Professor Greenwood in various inquires, and is at work on his analysis of the statistics of crimes of violence in this country … Greenwood himself has continued to work on experimental epidemiology … he has also given much attention to various official or semi-official inquiries of interest to the Ministry of Health concerned with nutrition, national physique, the putative effect of unemployment on health. ²⁶

Increasingly over the first decades of the 20th century, and especially during the First World War, national health came to be seen as essential for national interests. Here, the integrative epidemiology promoted by Greenwood, and the rationalisation offered by analytical statistics, held significant political-economic value. Prior to free universal health care, the statistical analysis of population-level data helped maximise interventions in support of public health. More fundamentally, by offering a means towards more healthful and productive populations, Greenwood’s epidemiology complemented early-century demands for human capital. This integrative understanding of health – as well as his active translation of medical statistics for successive administrations – helped elevate epidemiology into a biopolitical tool with broad governmental utility. We may see this, for instance, in the application of epidemiology to ‘crimes of violence’ at LSHTM. As early as 1934, epidemiology had begun to gather policy potential beyond the strict confines of medicine, and although warning against epidemiology becoming ‘a treatise on all the bad habits of mankind’ (Greenwood, 1935, p. 377), Greenwood’s work precipitated precisely this slippage.

Yet Greenwood himself was eminently cautious in his conclusions. His enthusiasm for statistics was, for instance, caveated by the warning that they would always be partial and apt to mislead, explaining in the 1935 textbook Epidemics and Crowd Diseases that

any intelligent man who has had a medical training can think of a hundred possible factors of epidemicity; but … he may be too ignorant to choose well, and, having chosen, may still be unable to isolate what he desires to study … With the help of the method of multiple correlation some progress may be made … but the journey is slow and difficult. (Greenwood, 1935, p. 66)

Here we are reminded that, as often in the history of science, British epidemiology was shaped as much by the political and economic potential of its analyses as by the opinions and interests of its practitioners. Greenwood’s feted career is, then, testament to both his ideas and their political-economic utility at this moment. Prior to the development of the National Health Service, but alongside growing political concern for the national health, Greenwood’s population-level perspective helped isolate the most efficient and cost-effective interventions. In this respect, Greenwood’s articulation of epidemiology and its employment in pursuit of national interests might be seen as an antecedent of the post-war ‘economization of life’ (Murphy, 2017) or the ‘economic gaze’ (Sridhar, 2011) of late-century global health. Although it will take more work to trace Greenwood’s contribution through the 20th century, an analogous approach to epidemiology emerged at the century’s end, after rather than before the heyday of nationalised medicine. This – the ‘population health approach’ – also draws on the analysis of social and vital statistics; also pursues population-level interventions against prevalent poor health; and was also exercised in the context of inequitable or often inaccessible health care. Defined as an attempt ‘to maximise overall health outcomes and minimise health inequities at the population level’ (Kindig, 2007, p. 158), population health coalesced into the 1990s in response to the limitations of the individualistic, risk-based epidemiology that had come to dominate western epidemiology following the Second World War (see Krieger, 2011; Szreter, 2003). Population health also rose to meet the demands of the ‘neo-liberal’ reorientation of politics, economics and society that had been underway since the late 1970s (see Harvey, 2005). Based on the principles of market freedom, low taxation and reduced state involvement in social security, neo-liberalism has contributed to increasing inequality in terms of income and health. Prioritising cost savings and economic efficiencies, the population-health approach evolved to address health inequities without entirely contradicting the small-state ideology of the late 20th century.

Although finding purchase in very different times and places, a focus on health at the population level promises efficient and cost-effective solutions for improving the health of populations that are unserved or underserved by extant medical infrastructures. In the early 20th century, Greenwood’s epidemiology gathered momentum in the absence of nationalised health care but alongside nationalist and imperialist demands for better labour. Likewise, the population-health approach emerged at the century’s end amidst and against the neo-liberal turn away from state responsibility and towards market mechanisms in health care. Although often critical of constrained and economical health provisioning, by offering a means to understand and mitigate the adverse health effects of industrialisation or free-market fundamentalism, such an approach also facilitates economising logics in medicine (see Nott, 2025). Indeed, although Greenwood routinely called for state intervention against the social and environmental causes of disease, his ideas found purchase and his career flourished at least in part because his work complemented the political economy of the period. Amidst the industrialism, militarism and imperialism of Britain’s early 20th century, the value offered by the elucidation of any such structural determinants of illness derived from the relationship between health and bodily productivity in the aid of national interests and in the pursuit of capital.