Spatial disparities at death Age-, sex- and disease-specific - - PDF document

1

Spatial disparities at death Age-, sex- and disease-specific mortality in the Belgian districts at the beginning of the twentieth century

Tina Van Rossema,b,c, Laura van den Borrea,b, Patrick Deboosereb and Isabelle Devosc a: PhD Fellow Research Foundation – Flanders (FWO) b: Interface Demography, Vrije Universiteit Brussel, Belgium c: History Department, Ghent University, Belgium Working paper, please do not cite without permission of the authors Abstract At the beginning of the twentieth century, average life expectancy at birth was much lower in Flanders, the northern part of Belgium, than in Wallonia, the southern part of the country. In the literature, this excess mortality is mainly attributed to high levels of infant mortality, caused by bad feeding practices and low quality drinking water. In contrast, little is known about the mortality risks at other ages, their determinants and their regional variability. In this article, we reconstruct age-, sex- and disease-specific death rates for the 41 Belgian districts around 1910. We construct maps classified according to a relative scale to visualise the mortality variations. Our spatial analysis shows that after childhood there was no clear-cut Flemish-Walloon divide in total mortality. High death rates for male and female adolescents, adults and elderly were found in both areas. The inclusion of disease-specific mortality data in our analysis is innovative compared to previous studies. For these rates as well, we note large differences between the Northern and Southern districts, even for infants and children. At young ages, respiratory and airborne infectious diseases were important killers together with a lack of viability and neurological diseases. At adolescent and young adult ages, we

• bserve a large contribution of pulmonary tuberculosis and respiratory diseases, while cardiovascular

diseases became more important with age. We conclude that the majority of the spatial disparities in total and disease-specific mortality crossed the Flemish-Walloon border as well as provincial borders and that the ranking of the districts varied considerably according to age and sex. The diverging spatial mortality patterns between adolescent and adult men and women suggest, moreover, the importance of sex-specific determinants of health and mortality at these ages. Keywords: mortality, causes of death, spatial variation, Belgium, sex differences

• 1. Introduction

The study of spatial differences in mortality has a long tradition in Belgium. Already in 1826, the famous Belgian government official and statistician Adolphe Quetelet noted ‘la prodigieuse difference’ between city and countryside [Quételet, 1826]. His work on urban health penalties was confirmed by later studies for Wallonia, the southern part of Belgium [Eggerickx and Debuisson, 1990; Eggerickx, 2001; Neven, 2000; Oris, 1998]. Whereas many studies in Belgium and beyond have focused on the urban-rural mortality gap, few have examined other spatial levels. Notable exceptions for Belgium are the study of Grimmeau et al. [2010] on the provincial level and our recent studies on urban health penalties [Devos and Van Rossem, 2015; Van Rossem et al., 2017]. During the nineteenth century, there was in Belgium a marked health divide between the Flemish and Walloon

• provinces. Average life expectancy at birth was much lower in Flanders, the northern part of the

country, than in Wallonia [Grimmeau et al., 2010]. Seemingly, the low levels of life expectancy in Flanders were strongly related to high levels of infant mortality, caused by bad feeding practices and low-quality drinking water. The regional variability and the determinants of infant mortality were studied in detail by Masuy-Stroobant [1983] and Debuisson [2001]. However, because the comprehensive measure of life expectancy is extremely sensitive to infant and child mortality, mortality in Flanders could potentially have been lower at other ages. Our research on urban

2 mortality has already demonstrated spatial variation in Belgium between mortality of different age

• groups. High death rates for infants and children did not per se occur in the same towns and cities

where high death rates for adolescents, adults and elderly were noted. Whereas the highest levels of mortality for infants and children were consistently found in Flemish towns and cities, above the age

• f 10 the ranking of Flemish and Walloon cities was mixed. In other words, high death rates for

adolescents, adults and elderly were observed in Flemish as well as Walloon towns and cities [Van Rossem et al., 2017]. In this article, we aim to deliver more insight into the spatial mortality differences within Belgium at the beginning of the twentieth century. Grimmeau et al. [2010] already indicated that their analyses at the provincial level possibly masked interesting patterns at the district level. We therefore present maps classified according to a relative scale with age-, sex- and disease-specific death rates for the 41 Belgian districts around 1910. According to the analyses of Grimmeau et al. [2010], this period largely marks the end of clearly lower life expectancy levels in Flanders. After the 1920s, this spatial clustering gradually blurred1. The advantage of using district data is threefold: (1) There is a larger social and economic homogeneity in districts than in provinces. (2) In contrast to some small municipalities, the level of (disease-specific) deaths in districts is still large enough to derive meaningful statistical results. (3) The collection and digitalization of mortality data for 41 Belgian districts is manageable within a reasonable amount of time. Since Belgium at the time consisted of more than 2,600 municipalities, it would take months to digitalize all the data on population and mortality [Debuisson, 2001]. We use death rates to avoid the disadvantages of the life expectancy measure. The mortality data are derived from the register of vital events Le Mouvement de la Population et de l’Etat Civil2, whereas we made use of the population census of 1910 for data on population size and structure3. We distinguish in this article six different age groups: infants (<1 year), children (1-9), adolescents (10-19), young adults (20-39), middle-aged adults (40-59) and elderly (60-79). Men and women are studied separately because they experience different genetic and biological risks. As women have two X chromosomes, they appear to be more resistant to infectious diseases than men because of X- linked immunoregulatory genes [Waldron, 1983]. Moreover, in contrast to men, women experience many risks related to childbearing. Besides biological differences, the different work tasks of men and women are important. Furthermore, we include disease-specific mortality data in our study, which is innovative for research

• n early twentieth-century mortality risks in Belgium. Belgian mortality data at the aggregate level

are available from 1886 onwards, but 1910 is the first year for which age-specific cause-of-death data are available in Le Mouvement based on the new detailed international nomenclature introduced in

• 19034. The cause-of-death data are extremely useful because most diseases are strongly linked to

specific determinants. Hence, according to Reid and van den Boomen [2015: 310] “by using epidemiological insights into the origin and course of the diseases and conditions that eventually caused death, the complex context determining mortality levels can be, to a certain extent, unveiled”. We distinguish in this article 16 different disease categories. Most of them, such as airborne infectious diseases or enteritis, are strongly related to specific environmental and socio-economic conditions via distinct transmission routes. Although it is not our intention to explain the observed spatial differences in mortality, the maps with cause-specific mortality figures may as such hint at possible determinants.

1 From the 1980s onwards, the pattern is reversed and life expectancies are clearly lower in Wallonia than in Flanders

(Grimmeau et al., 2010).

2 National Archives in Brussels, Statistiques du Mouvement de la Population et de l’Etat Civil, 1909-1910-1911. 3 Statistique de la Belgique, Population. Recensement général du 31 décembre 1910. Bruxelles: Ministère de l’Intérieur. 4 This classification is based on the nomenclature agreed upon during the international conference on the revision of the

nosological nomenclatures organised in Paris in 1900 (Velle, 1985).

3 In sum, the analyses in this article answer three important questions regarding the nineteenth- century Flemish-Walloon divide in life expectancies. These questions are the following: (1) Did every age group experience higher mortality risks in the Flemish compared to the Walloon districts? (2) Were there important differences in the spatial mortality patterns of men and women? (3) Which diseases were responsible for the largest differences in mortality between the Belgian districts? These questions are handled separately in three analysis sections, i.e. in parts 4, 5 and 6. Before we turn to answering these research questions, we briefly discuss in part 2 the industrial and socio- economic background of the Flemish and Walloon districts. In part 3, the data and methods used to create the maps are explained. The article concludes in part 7 with a discussion on the large spatial disparities in Belgium at the beginning of the twentieth century. The absence of a clear geographical clustering of high or low death rates after childhood is particularly notable.

• 2. The Belgian context

At the beginning of the twentieth century, the Belgian territory was divided in 9 provinces and 41

• districts. This division is presented in figure 1. The Flemish districts are situated in the northern part
• f the country and consist of the districts of the provinces of West Flanders, East Flanders, Antwerp

and Limburg, together with the Brabant districts Brussels and Leuven. The southern or Walloon districts are the districts of the provinces of Hainaut, Namur, Liège and Luxembourg, together with the Brabant district Nijvel.

Figure 1 Belgian districts and provinces Source: Map constructed by author

Around 1910, differences in the industrial and socio-economic situation of the Belgian provinces and districts were largely the result of the industrialisation process that took off during the middle of the nineteenth century. Belgium was the second country to enter the Industrial Revolution after the pioneer Great Britain. At first, the process in Belgium was largely concentrated around the sectors of mining, textile and metal [Deneckere et al., 2009]. Especially the Walloon manufacturing belt, situated between the Borinage area in Hainaut and Verviers, boomed due to the presence of rich coal and iron ore mines. Thriving steel, glass and machine-building industries and the export of transport equipment brought success to this area. At the same time the fabrication of woollen fabrics in the area of Verviers was gradually displaced from the workshops of domestic workers to large factories. This led to the emergence of an important textile industry. The other Belgian industrial centre was situated around the Flemish city of Ghent. Large textile factories provided employment to many women and children. However, this mechanisation process led to the collapse

4

• f the rural handmade linen industry. Together with several successive failed agricultural harvests

during the 1840s and the absence of real economic development in West Flanders and in the rural parts of East Flanders, this led to the impoverishment of broad sections of the Flemish population. Economic development was also largely absent in the agricultural provinces of Limburg, Namur and Luxembourg [Deneckere et al., 2009; Devos and Van Rossem, 2015; Leboutte et al., 1998]. The provinces of Antwerp and Brabant took an intermediary position. Commercial and transport-related industries were of particular importance here. The city of Antwerp and its surrounding area underwent a rapid transition as its harbour developed into an international port. The construction of the Brussels-Charleroi canal in 1832 fostered further industrial growth, and created an important axis

• f economic power between Antwerp, Brussels and Charleroi. In fact, many factories settled in the

Brussels suburbs Sint-Jans-Molenbeek and Anderlecht which were crossed by the canal. Chemical industries and the production industry of metals, vehicles and machinery was especially important. In the area around Vilvoorde, also part of the Brussels district, a complex of coke ovens was planted. Because the city of Brussels was established as capital city in 1831, administrative and financial activities were furthermore important in the Brussels district [De Beule, 1994; Devos and Van Rossem, 2015; Leboutte et al., 1998; Lefebvre and Buyst, 2007]. This different evolution of the industrialisation process in the Belgian provinces and districts was largely reflected in their population growth. During the nineteenth century, there was a negative relative population growth in West Flanders, East Flanders, Limburg, Namur and Luxembourg. On the

• ther hand, the industrial activities in Hainaut and Liège and the important commercial and

administrative activities in Antwerp and Brabant attracted many migrants. Most of the population growth took place in large cities, but there was also a sharp increase in communes between 10,000 and 50,000 inhabitants [Devos and Van Rossem, 2015]. This was related to the growing importance

• f suburbs, industrial basins, for instance around Mons and Charleroi, and regional service centres

[Deneckere et al., 2009; Devos and Van Rossem, 2015]. Besides these demographic differences, the industrial developments also provoked a clear-cut divide in relative GDP per capita. Analyses by Erik Buyst [2009] for 1896 have clearly demonstrated low levels of GDP per capita in ‘poor Flanders’ compared to a prosperous situation in Wallonia. The four Flemish provinces of West Flanders, East Flanders, Antwerp and Limburg had a GDP per capita below the Belgian average. The highest levels were found in Liège and Hainaut, while moderate levels prevailed in Namur, Brabant and Luxembourg.

• 3. Data and method

We present in this article age-, sex- and disease-specific death rates of the 41 Belgian districts. They reflect the number of deaths per 1,000 people in 1910. We show the average rates for six age groups: we make for men as well as women a distinction between infants (<1 year), children (1-9), adolescents (10-19), young adults (20-39), middle-aged adults (40-59) and elderly (60-79). In particular, the presentation of the variation in the most important causes of death by age group and sex is innovative. A specific cause was selected if it provoked more than 10% of the age- and sex- specific deaths at the national level5. To highlight the geographical patterns, we created maps classified according to a relative scale. The classes for each map were constructed by using the mean value of the death rates (!) and the standard deviation (s.d.). This resulted in the following classification: …, [!-2.5 s.d. - !-1.5 s.d.[, [!-1.5 s.d. - !-0.5 s.d.[, [!-0.5 s.d. - !+0.5 s.d.[, [!+0.5 s.d. - !+1.5 s.d.[, [!+1.5 s.d. - !+2.5 s.d.[, … The class [!-0.5 s.d. - !+0.5 s.d.[, containing the average mortality level, is coloured in blue. White, yellow and green colours represent lower death rates;

• range, red, purple and black colours represent higher death rates. The larger the deviations from

the mean and the between-district disparity, the more classes were used.

5 The only exceptions to this rule are the presentation of results on maternal mortality and the exclusion of the general

category ‘other causes of death’. Because of the historical relevance of maternal mortality, we present the death rates for 20 to 39 year old women. The percentage of this cause of death for the whole of Belgium amounted for this age group to 9.3%. We furthermore excluded the ‘other causes of death’ from our analysis since this category of unspecified diseases would not improve our knowledge on the determinants of spatial disparities at death.

5 The collection of data on population numbers and mortality, needed to construct the death rates, was facilitated by the digitalisation of nineteenth century mortality and population statistics at the local level in the HISSTER database6. The data on population size were derived from the population census of 19107. It offers detailed age- and sex-specific figures of the population with usual residence for all Belgian provinces, districts and municipalities with 10,000 inhabitants or more. The mortality data were derived from ‘Le Mouvement de la Population et de l’Etat Civil’ 8, the register of vital

• events. From 1886 onwards, Le Mouvement included tables with the total number of de facto deaths

according to sex, age and residence. This information was derived from municipal population registers and civil status registers, of which the municipal government had to make a yearly summary

• n pre-printed forms. For the figures on total mortality, we used data from three years centered

around the census year, i.e. 1909, 1910 and 1911. The disease-specific information, on the other hand, was based on data for 1910 only. Only the registers at census years provide an age-specific classification of cause-of-death mortality9. We clustered the causes of death into 16 categories, mainly based on their aetiology and transmission route. As a result, we categorized airborne, waterborne and other infection diseases, cancer, cardiovascular diseases, a lack of viability, maternal mortality, neurological diseases and urogenital diseases. Because of their historical importance, pulmonary tuberculosis was separated out from the category of respiratory diseases [see also Reid et al., 2015] and enteritis from the category of intestinal diseases. We furthermore distributed the external causes of death into the categories of violent deaths and accidents, and distinguished a general category with ‘other causes of death’. The classification matrix is presented in table A1 in the appendix. The population data and total mortality numbers can be considered fairly reliable. There was since the end of the nineteenth century a well-developed public administration in Belgium, clear guidelines for counting and processing the data, and a thorough supervision by the central government [Preneel, 2010; Vrielinck, 2013]. Internationally, historical demographers and medical historians have, however, highly discussed the quality of historical cause-of-death statistics [for an overview see Van Rossem et al., forthcoming]. The main problems noted are related to (1) the limited medical knowledge of civil reporters, (2) the statement of false causes of death, for instance in the case of socially loaded diseases such as pulmonary tuberculosis, (3) the time-varying nomenclatures and (4) misclassification of causes of death during the conversion from individual death certificates to larger categories at an aggregated level. Consequently, the results of the disease-specific mortality figures have to be interpreted very carefully. The data for Belgium are based on the municipal registers of causes of death that Belgian local governments were obliged to keep since 1851. A uniform cause of death nomenclature for the completion of these certificates was already introduced in 1867 in

• Belgium. Changes of the classification scheme in 1874 and 1903 profoundly facilitated the cause-of-

death registration. The nomenclature of 1903 was the first to be based on an international classification, i.e. on the nomenclature agreed upon during the Statistical Congress in Paris in 1900 [Bracke, 2008; Velle, 1985].

• 4. Spatial variation in age-specific mortality

The first question we handle is whether every age group experienced higher mortality risks in the Flemish compared to the Walloon districts. As can be seen on figures 2 to 13, only for infants and

6 HISSTER. A database of Belgian mortality statistics for the 19th and 20th centuries available at the local and regional level,

Ghent University, Quetelet Center supervised by Isabelle Devos.

7 Statistique de la Belgique, Population. Recensement général du 31 décembre 1910. Bruxelles: Ministère de l’Intérieur. 8 National Archives in Brussels, Statistiques du Mouvement de la Population et de l’Etat Civil, 1909-1910-1911. 9 Six age groups were distinguished in the register of 1910, except for particular causes of death, such as homicides,

accidents and several ‘other causes of death’, i.e. <1 year, 1-6, 7-14, 15-20, 21-49, >50. To obtain disease-specific mortality figures for five-year age groups, we first distributed the number of non-age-specific causes of death over the six age groups according to their size. Next, we calculated for the six age groups the relative importance of each cause of death and applied these percentages to the relevant age-specific total mortality numbers. We hereby assume that the death risk due to a certain cause of death was similar within each of the six original age groups.

6 children younger than 10 years, the death rates were consistently higher in Flanders than in

• Wallonia. The death rates were especially high in districts of West and East Flanders. In specific, the

rates for male and female infants and male children were highest in the coastal Ostend district with respective levels of 332.04, 282.91 and 14.74 per 1,000 (the averages were 169.74, 138.93 and 9.56). For female children, the highest level was noted in Thielt (14.31 compared to an average of 9.10). In the Walloon districts, only very low to moderate death rates were observed. The lowest rates were noted in Philippeville and amounted only to 98.56 and 69.97 for male and female infants, and to 6.05 and 5.51 for male and female children. Previous studies have explained the high infant mortality levels in Flanders through the impact of both breastfeeding habits and environmental factors such as water quality. Children who were breastfeed during the first months of their life had an initial survival advantage. Because of the difficult economic conditions for men in the slowly industrializing West and East Flemish provinces, however, the employment of (married) females was necessary to substantiate the household budget. Most of these working mothers were unable to breastfeed their

• children. The health impact of such early weaning depended largely upon the quality of the cow milk

and artificial nutrition that was supplied to children instead of breastmilk. In Flanders, the water added to the bread soups or porridges of potatoes was usually extracted from artesian walls and of low quality. In contrast, domestic water in Wallonia usually came from fast flowing rivers [Backs, 2003; Cornut, 1999; Debuisson, 2001; Lesthaeghe, 1987; Masuy-Stroobant, 1983]. From adolescent ages onwards, we do not longer observe a Flemish-Walloon divide in mortality (see figures 6 to 13). Districts with high and low death rates were found both in the northern as in the southern part of the country and the ranking differed by age. There were, moreover, large differences in death rates between neighbouring districts, also between those belonging to the same

• province. Figure 8, reflecting the death rates of young adult men, shows for instance that two

districts of the Luxembourg province, i.e. Virton (3.84 per 1,000) and Bastogne (4.22), experienced very low death rates compared to the average (5.04), while a very high rate was noted in Arlon (6.57). In contrast to the results for infants and children, it is not yet clear which specific determinants were responsible for the spatial variation in mortality at older ages. Summarized, the answer to our first question is that around 1910, not every age group in Flanders consistently experienced higher mortality risks than in Wallonia. Hence, in a period when mortality risks at young ages were still very high, life expectancy levels can be misleading and are not designated to understand the mortality experiences of every population group.

• 5. Spatial variation in sex-specific mortality

Next to spatial differences in age-specific mortality, the maps also visualize gender differences in the spatial mortality patterns. Figures 2 to 5 demonstrate that there were no important differences by sex for infants and children. In general, districts with high death rates for boys also displayed high death rates for girls and vice versa for low death rates. From adolescent ages onwards, we notice, however, large differences in the colouring of districts according to male or female death rates (see figures 6 to 13). Male adolescents in the Charleroi district experienced for instance higher mortality rates than the Belgian average (3.02 per 1,000 compared to an average of 2.63), while female adolescents in the same district were in a rather favourable position (2.28 compared to an average of 2.88) (see figures 6 and 7). Vice versa, female young adults experienced extremely high death rates in Tongeren (6.86 compared to an average of 5.34), while male young adults were here confronted with a death rate lower than the average (4.77 compared to an average of 5.04) (see figures 8 and 9). Numerous other examples of such sex-specific differences can be noted when comparing the maps with data on adolescents (figures 6 and 7), young adults (figures 8 and 9), middle-aged adults (figures 10 and 11) and elderly (figures 12 and 13). We assume that the sex-specific mortality variation at adolescent, adult and elderly ages was related to the increasing importance of sex-specific determinants of health and mortality during the life time. There were for instance large differences between the employment activities and associated health impacts of men and women [e.g. Van Rossem et al., 2017]. In addition, adult women experienced many risks related to pregnancy and

• childbearing. We will check in the next section whether the differences in total mortality were also

7 reflected in disease-specific mortality. For now, we can deliver a positive answer to our second question: from the age of 10 onwards, there were important spatial mortality differences between men and women.

Figures 2-13 Death rates (per 1,000) of men and women in the Belgian districts by age group, 1910

Men Women Infants (<1 year) Infants (<1 year) Children (1-9 years) Children (1-9 years) Adolescents (10-19 years) Adolescents (10-19 years) Young adults (20-39 years) Young adults (20-39 years)

8

Middle-aged adults (40-59 years) Middle-aged adults (40-59 years) Elderly (60-79 years) Elderly (60-79 years) Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911

• 6. Spatial variation in disease-specific mortality

Figures 14 to 47 display the disparities in disease-specific mortality of male and female infants, children, adolescents, young and middle-aged adults and elderly. We present only the results for the causes of death explaining more than 10% of the sex-specific deaths in a specific age group at the national level. We will sequentially discuss the results of the six age groups. The results inform us whether there was a Flemish-Walloon and/or gender divide for these age- and disease-specific death rates. The most important causes of death for infants were enteritis or diarrhoeal diseases, airborne infectious diseases such as whooping cough, respiratory diseases including bronchitis and pneumonia and a lack of viability. Digestive diseases, respiratory diseases and airborne infectious diseases are strongly related to malnourishment [Masuy-Stroobant, 1983; Rabb and Rotberg, 1985]. In addition, the two last types of diseases are usually spread by droplet infection, whereby crowding facilitates its spread [Kiple, 1993]. The category ‘a lack of viability’ is not clearly described in Le Mouvement, but we may assume that the description refers to endogenous deaths. In contrast to exogenous deaths, those deaths are not related to external environmental factors, but occur in the first months of life due to genetic diseases, congenital malformations or injuries connected with birth [Pressat, 1969]. The death rates due to enteritis (figures 14 and 15) and airborne infectious diseases (figures 16 and 17) were especially high in East- and West-Flemish districts, hereby partly explaining the high total mortality in these areas. In addition to the high death rates related to respiratory diseases in many West- and East-Flemish districts, this is also the case in the three Antwerp districts (see figures 18 and 19). Figures 20 and 21 furthermore demonstrate that the registration of deaths because of ‘a lack of viability’ mainly occurred in West-Flemish, Liège and Luxembourg districts. The Flemish- Walloon divide does thus not hold for this specific cause of death. In general, we do not notice large disparities between the spatial mortality patterns of boys and girls. Like for infants, airborne infectious and respiratory diseases were important causes of death for

• children. In addition, neurological diseases explain more than 10% of the deaths of male and female

9 children in Belgium in 1910. Children were mainly infected by two of the three types of diseases classified as ‘neurological’, i.e. simple meningitis and tuberculosis of the meninges10. These diseases are usually spread by droplet infection, whereby they are strongly related to crowding [Kiple, 1993; WHO, 2015]. Districts with high death rates due to airborne infectious diseases were mainly clustered in West Flanders (see figures 22 and 23), but for the other causes of death we find districts with high death rates all over the country. There was thus no clear Flemish-Walloon divide. On figures 24 and 25 reflecting death rates due to respiratory diseases, we note a significantly different classification for male and female children (i.e. a difference of more than one level) in some districts such as Leuven and Marche-en-Famenne. For the other causes of death, the spatial variation between men and women was largely confound to one level (or colour) (see figures 22-23 and 26- 27). Pulmonary tuberculosis and respiratory diseases explain an important part of the mortality of adolescent men and women. For women, also the proportion of neurological diseases exceeds our 10% threshold. The development of pulmonary tuberculosis and respiratory diseases is strongly related to dusty living and working conditions in badly ventilated and illuminated houses and workplaces, as well as to an inadequate diet [Kiple, 1993]. Likewise, crowding fosters the development of the neurological diseases of simple meningitis and tuberculosis of the meninges (see supra). The spatial patterns displayed in figures 28 to 32 demonstrate that there were high death rates because of these diseases in Walloon as well as Flemish districts. The results for pulmonary tuberculosis and respiratory diseases, moreover, display a clear difference in the classification of the districts between men and women. This underpins the assumption that the development of these diseases was for a significant part related to working conditions. The most important causes of death for young adults, i.e. pulmonary tuberculosis and respiratory diseases, equal those of adolescents. For young adult women, we also present the spatial dispersion

• f death rates due to childbearing such as puerperal fever. Puerperal fever was strongly related to

unhygienic conditions in hospitals, for instance the use of contaminated instruments or the lacking habit of medical staff to wash their hands. After the introduction of antisepsis, the risk of puerperal fever strongly declined. In most European countries, the introduction of hygienic guidelines in

• bstetrics occurred by the end of the nineteenth century [De Costa, 2002]. We assume that most
• ther fatal conditions related to childbirth were also strongly influenced by maternal care in hospitals
• r at home. The results on pulmonary tuberculosis are similar to those for adolescents (see figures 33

and 34). We note important differences in the colouring of districts between men and women, and high death rates in both Flanders and Wallonia. The results for respiratory diseases, presented in figures 35 and 36, also demonstrate that districts with high death rates were found both in Wallonia and Flanders. The highest rates, nevertheless, always occurred in Flemish districts (coloured in red or purple). Figure 37 depicts the death rates due to maternal mortality. Only one Walloon district, Virton, displayed very high death rates, in comparison to six Flemish districts. For middle-aged adults as well, respiratory diseases were an important cause of death. In addition, the percentage of cardiovascular diseases exceeds for this age group the 10% threshold. In contrast to infectious diseases, cardiovascular diseases mainly affect elderly, whereby they are frequently placed under the general heading of ‘degenerative diseases’. The development of this type of diseases is fostered by unhealthy lifestyle such as tobacco smoking, an unhealthy diet or stress [WHO, 2017]. The spatial patterns of respiratory and cardiovascular diseases differ between men and women, although in most cases the difference does not exceed more than one level. The majority of districts with high female death rates due to respiratory diseases are located in Flanders (see figure 39), while for men there are also an important number of Walloon districts with high death rates (see figure 38). Concerning cardiovascular diseases, no clear spatial clustering is visible for men nor women and high death rates were found in both country parts (figures 40 and 41).

10 The third neurological disease, i.e. congestion, haemorrhage and softening of the brain, mainly affected adults and

elderly.

10 Finally, three categories of diseases exceed the 10% threshold for elderly. It concerns respiratory diseases, cardiovascular diseases and neurological diseases. In contrast to children and adolescents, death due to neurological diseases at older ages was probably mainly ascribable to ‘congestion, haemorrhage and softening of the brain’. Strokes formed a large part of this category of diseases. The main risk factor for strokes is hypertension, which is strongly related to stress. Just as for cardiovascular diseases, an unhealthy diet is also important [Kiple, 1993; Roman, 1987]. For the three diseases, high death rates were observed in the north and the south of Belgium (see figures 42 to 47). Moreover, there were not always high death rates for men in districts with high rates for women and vice versa, although the differences in classification remained limited. In sum, this overview of spatial disparities in disease-specific mortality demonstrates that for most causes of death, districts with high death rates were found in both parts of the country. Only for infants, the highest disease-specific death rates mainly occurred in Flanders. Moreover, for adolescents and adults there were important differences in the classification of the districts by sex.

Figures 14-21 Disease-specific death rates (per 1,000) of male and female infants (<1 year) in the Belgian districts, 1910

Men Women Enteritis Enteritis Airborne infectious diseases Airborne infectious diseases Respiratory diseases Respiratory diseases

11

Lack of viability Lack of viability Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911 Figures 22-27 Disease-specific death rates (per 1,000) of male and female children (1-10 years) in the Belgian districts, 1910 Men Women Airborne infectious diseases Airborne infectious diseases Respiratory diseases Respiratory diseases Neurological diseases Neurological diseases Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911

12

Figures 28-32 Disease-specific death rates (per 1,000) of male and female adolescents (10-19 years) in the Belgian districts, 1910 Men Women Pulmonary tuberculosis Pulmonary tuberculosis Respiratory diseases Respiratory diseases Neurological diseases Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911

13

Figures 33-37 Disease-specific death rates (per 1,000) of male and female young adults (20-39 years) in the Belgian districts, 1910 Men Women Pulmonary tuberculosis Pulmonary tuberculosis Respiratory diseases Respiratory diseases Maternal mortality Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911

14

Figures 38-41 Disease-specific death rates (per 1,000) of male and female middle-aged adults (40-59 years) in the Belgian districts, 1910 Men Women Respiratory diseases Respiratory diseases Cardiovascular diseases Cardiovascular diseases Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911

15

Figures 42-47 Disease-specific death rates (per 1,000) of male and female elderly (60-79 years) in the Belgian districts, 1910 Men Women Respiratory diseases Respiratory diseases Cardiovascular diseases Cardiovascular diseases Neurological diseases Neurological diseases Source: The Population Census, 1910 and Le Mouvement, annual data for 1909-1910-1911

• 5. Conclusions

The spatial analysis of mortality in this article clearly demonstrates that geographical differences in Belgium cannot be simplified to the Flemish-Walloon divide. Only for infants and children below the age of 10 there was a clear distinction, with a substantial higher total mortality in Flanders than in

• Wallonia. Infants and children experienced a particularly large penalty in West and East Flanders.

Because of the sensitivity of the life expectancy measure to mortality at young ages, these outcomes explain the low life expectancies at birth in Flanders at the beginning of the twentieth century. However, for adolescents, young and middle-aged adults and the elderly, low and high mortality were observed in Flemish as well as in Walloon districts. Moreover, for many age groups large differences in mortality were noted among districts belonging to the same province. Hence, the district level seems more suitable for studying spatial mortality than the provincial level. The ranking

• f the districts varied furthermore considerably according to age: for men as well as women, high

death rates of adolescents, adults and elderly did not per se appear in the same districts. Moreover,

16 from adolescent ages onwards, we noted important differences in the ranking of the districts by sex- specific mortality. Besides total mortality, we also examined variation in disease-specific mortality. The high infant mortality in Flanders was mainly explained by high levels of airborne infectious diseases, enteritis and respiratory diseases. Deaths due to a lack of viability were important in Flanders as well as in several southern Belgian districts. The most important causes of death for children were airborne infectious, respiratory and neurological diseases. Except for airborne infectious diseases, already from these young ages onwards, no clear division between Flemish and Walloon districts was noted. Geographical clustering was also not observed for most disease-specific death rates in other age

• groups. At adolescent ages, it concerned deaths due to pulmonary tuberculosis and respiratory

diseases, and for women also neurological diseases. The main killers for young adults were pulmonary tuberculosis and respiratory diseases, and for middle-aged adults respiratory and cardiovascular diseases. Finally, respiratory, cardiovascular and neurological diseases explained an important part of the mortality of elderly. Only for the death rates of young female adults due to maternal mortality we observed a clear disadvantage in Flanders compared to Wallonia. Furthermore, for adolescents and adults the spatial variation of most diseases, especially of pulmonary tuberculosis, differed between men and women. This suggests an increasing importance

• f sex-specific determinants of health and mortality during the life course. Future studies should

study the determinants of these patterns more in detail. To conclude, our study clearly demonstrates that at the beginning of the twentieth century some Belgian districts were more lethal than others, but that their ranking varied considerably according to gender, age and disease. For infants and children, a clear divide in total mortality between the Flemish and Walloon districts could be distinguished. However, already from these young ages

• nwards this gap did not hold for specific causes of death. After age 10, geographical clustering

seems to have disappeared for both total and disease-specific mortality.

• 6. References

BACKS J., 2003, Mortaliteit in Gent (1830-1950). Gent, Maatschappij voor Geschiedenis en Oudheidkunde, 298 p. BRACKE N., 2008, Een monument voor het land: Overheidsstatistiek in België, 1795-1870. Gent, Academia Press, 467 p. BUYST E., 2009, Reversal of fortune in a small, open economy: regional GDP in Belgium, 1896-2000. VIVES - Discussion Paper, 8, pp. 1–28. CORNUT P., 1999, La circulation de l’eau potable en Belgique et a Bruxelles. Enjeux sociaux de la gestion d'une ressource naturelle. Doctoral thesis: Bruxelles, ULB, 297 p. DE BEULE M., 1994, Bruxelles une ville industrielle méconnue. Les Dossiers de La Fonderie, 1, pp. 1- 72. DE COSTA C., 2002, “The contagiousness of childbed fever”: a short history of puerperal sepsis and its

• treatment. Medical Journal of Australia, 177, 11-12, pp. 668–671.

DEBUISSON M., 2001, The decline of infant mortality in the Belgian districts at the turn of the 20th

• century. Journal of Belgian History, 31, 3-4, pp. 497–527.

DENECKERE G. et al., 2009, Een geschiedenis van België. Gent, Academia Press, 315 p. DEVOS I., VAN ROSSEM T., 2015, Urban health penalties. Estimates of life expectancies in Belgian cities, 1846-1910. Journal of Belgian History, 4, pp. 74-109. EGGERICKX T., 2001, La mortalité dans le bassin industriel de Charleroi aux 19ème et 20ème siècles: un handicap socio-démographique récurrent. Espace, Populations, Sociétés, 3, pp. 351–368. EGGERICKX T., DEBUISSON M., 1990, La surmortalité urbaine: le cas de la wallonie et de Bruxelles à la fin du XIXe siècle (1889-1892). Annales de Démographie Historique, 23–41. GRIMMEAU J.-P. et al., 2010, Evolution de la géographie de la mortalité en Belgique 1840-2002. In T. Eggerickx, J.-P. Sanderson (Eds.). Histoire de la population de la Belgique et de ses territoires,

17 Louvain-la-Neuve, Presses Universitaires de Louvain, pp. 613–626. KIPLE K., 1993, The Cambridge world history of human disease. Cambridge, Cambridge University Press, 1176 p. LEBOUTTE R. et al., 1998, Un siècle d’histoire industrielle. Belgique, Luxembourg, Pays-Bas, industrialisation et sociétés 1873-1973. Paris, SEDES, 298 p. LEFEBVRE W., BUYST E., 2007, Enkele verklaringsfactoren voor de regionaal gedifferentieerde industriële ontwikkeling van Vlaams-Brabant tussen 1896 en 1961. Belgisch Tijdschrift Voor Nieuwste Geschiedenis, 37, 1-2, pp. 41–77. LESTHAEGHE R., 1987, The breast-feeding hypothesis and regional differences in marital fertility and infant mortality in the Low Countries during the 19th century - Comments on a debate. IPD working paper, 3, pp. 1-22. MASUY-STROOBANT G., 1983, Les déterminants individuels et régionaux de la mortalité infantile. La Belgique d’hier et d'aujourd'hui. Louvain-la-Neuve, CIACO, 540 p. NEVEN M., 2000, Mortality differentials and the peculiarities of mortality in an urban-industrial population: a case study of Tilleur, Belgium. Continuity and Change, 15, 2, pp. 297–329. ORIS M., 1998, Mortalité, industrialisation et urbanisation au 19e siècle. Quelques résultats des recherches liégoises. In Dix essais sur la démographie urbaine de la Wallonie au XIXe siècle, Bruxelles, Crédit communal, pp. 289–322. PRENEEL M., 2010, Over levendgeboren onwettige meerlingen in Wortel, Luikse emigratie naar Argentinië en epidemische meningitis in Zedelgem. “Le Mouvement de la Population de l’état civil (1841-1976)”: een nauwelijks ontgonnen bron voor historisch-demografische statistiek. In

• T. Eggerickx, J.-P. Sanderson (Eds.), Histoire de la population de la Belgique et de ses territoires,

Louvain-la-Neuve, Presses Universitaires de Louvain, pp. 73–82. PRESSAT R., 1969, L’analyse démographique. Paris, PUF, 321 p. QUETELET A., 1826, Mémoire sur les lois des naissances et de la mortalité à Bruxelles. In Mémoires de l’Académie Royale des Sciences et des Belles Lettres de Bruxelles. Bruxelles, pp. 495–512. RABB T., ROTBERG R., 1985, Hunger and history: the impact of changing food production and consumption patterns on society, Cambridge, Cambridge University Press, 336 p. REID A. et al., 2015, “A confession of ignorance”: Deaths from old age and deciphering cause-of- death statistics in Scotland, 1855-1949. The History of the Family, 20, 3, pp. 320–344. REID A., VAN DEN BOOMEN N., 2015, The faces of death: regional differentiation in cause-specific mortality in the past. The History of the Family, 20, 3, pp. 309–319. ROMAN G., 1987, ‘Cerebral Congestion. A Vanished Disease’. Archives of Neurology, 44, pp. 444-448. VAN ROSSEM T. et al., 2017, Death at work? Mortality and industrial employment in Belgian cities at the turn of the twentieth century. Explorations in Economic History, 26 p. VAN ROSSEM T. et al., forthcoming, The Brussels’ graveyard. Disease-specific mortality of men in the Belgian capital at the beginning of the twentieth century in a comparative perspective. 26 p. VELLE K., 1985, Statistiek en sociale politiek: de medische statistiek en het gezondheidsbeleid in België in de 19de eeuw. Belgisch Tijdschrift Voor Nieuwste Geschiedenis, 16, 1-2, pp. 213–242. VRIELINCK S., 2013, De macht van het getal. Belgisch Tijdschrift Voor Nieuwste Geschiedenis, 43, 2-3,

• pp. 74–113.

WALDRON I., 1983, The role of genetic and biological factors in sex differences in mortality. In A. Lopez, L. Ruzicka (Eds.). Sex differentials in mortality, trends, determinants and consequences, Canberra, Australian National University, pp. 141–164. WHO, 2015, Fact sheet n. 141: Meningococcal meningitis. Available via: http://www.who.int/mediacentre/factsheets/fs141/en/ WHO, 2017, Fact sheet: cardiovascular diseases (CVDs). Available via: http://www.who.int/mediacentre/factsheets/fs317/en/

18

• 7. Appendix

Table A1 Classification of the different causes of death into comprehensive categories

Comprehensive category Diseases Description of the causes of death in Le Mouvement (1910) Accidents Accidents Accidents Airborne infection diseases Diphteria and croup Influenza, Measles, Scarlet fever, Smallpox, Whooping cough Diphtérie et croup Grippe Rougeole Scarlatine Variole Coqueluche Cancer Cancer Cancer et autres tumeurs malignes Cardiovascular diseases Organic diseases of the heart Maladies organiques du Coeur Enteritis Diarrhoeal diseases and enteritis (younger than 2 years) Diarrhée et entérite (au-dessous de 2 ans) Intestinal diseases Alcoholism, acute or chronic, Cirrhosis of the liver, Diseases

• f

the stomach (cancer excluded), Hernia, intestinal obstruction Alcoolisme aigue ou chronique Cirrhose du foie Affections de l’estomac (cancer excepté) Hernies, obstructions intestinales Lack of viability Congenital debility Décès par défaut de viabilité Maternal mortality Other accidents

• f

pregnancy, childbearing and childbirth, Puerperal septicemia Autres accidents puerpéraux de la grossesse et de l’accouchement Septicémie puerpérale Neurological diseases Congestion, haemorrhage and softening

• f the brain,

Simple meningitis, Tuberculosis of the meninges Congestion hémorragie et ramollissement du cerveau Méningite simple Tuberculose des meninges Other causes of mortality Other unspecified diseases and causes of death, Senile debility, Sudden death, Unreported and unknown causes Autres causes de décès Vieillesse (débilité sénile) Mort subite Causes de décès non spécifiées ou mal définies Other infection diseases Intermittent fever, Other infection diseases, Tuberculosis of other organs Fièvre intermittante et cachexia palustre Autres affections épidémiques Autres tuberculoses Pulmonary tuberculosis Tuberculosis of the longs Tuberculose des poumons

19

Respiratory diseases Bronchitis, Pneumonia Bronchite aigue & Bronchite chronique Broncho-pneumonie & Pneumonie Urogenital diseases Nephritis and Bright’s disease, Non-cancerous tumors and other diseases

• f the female genital organs

Néphrite aigue et maladie de Bright Tumeurs non cancéreuses et autres maladies des organes génitaux de la femme Violence Homicides, Suicides, Suspicious cases Homicides Suicides Cas douteux Waterborne infection diseases Cholera, Typhoid fever Choléra Asiatique & Choléra nostras Fièvre typhoïde (typhus abdomional)