Background: The challenge of research in resource-poor settings remains a profound concern and is closely linked to African social development. Work of this thesis spans the end of apartheid and first decade of the democratic era in South Africa, along with emergence of the HIV/AIDS pandemic. It also covers the founding decade of the INDEPTH Network. Aims: Through appraising health and population research in a rural southern African sub-district over the past decade, to evaluate the utility of health and socio-demographic surveillance in rural African settings for: • capturing the dynamics of health, population and social transitions • supporting a mix of research designs, and • contributing to policy and programme development and evaluation. To extend this appraisal by examining the multi-site opportunities offered by the INDEPTH Network. Methods: Work was sited in the Agincourt sub-district, a heavily populated border area of rural north-eastern South Africa. Health and socio-demographic surveillance, introduced in 1992, involved prospective follow-up of the entire sub-district population of 70,000 people (including some 30% Mozambican immigrants) who lived in 11,700 households and 21 villages. Annual census rounds systematically updated household membership and recorded all vital events (births, deaths and migrations) since the previous census. A maternity history was asked of women of reproductive age and a verbal autopsy carried out on all deaths registered. The resulting ‘data and research platform’ – a core feature of all INDEPTH field sites – provided data for computation of trends in vital events and supported an extensive interdisciplinary project portfolio. The population under surveillance can be disaggregated into cohorts selected by age, sex or other criteria. Analyses are possible at multiple levels (individual, family/household or neighborhood) and can include socioeconomic factors. Findings: The Agincourt community experienced a serious worsening of mortality among most age-sex groups, rapidly declining fertility to near replacement level, and changing patterns of labour migration. This resulted in major changes in population structure and household composition. The rising burden of chronic disease involved both chronic infectious illness (HIV/AIDS and tuberculosis) and non-communicable disorders (such as stroke and related vascular disease). The burden of illness requiring chronic care increased disproportionately to that needing acute care. Potential contributions of field sites based on health and socio-demographic surveillance to local and national health policy are considerable yet remain underexploited. Interpretation: Rural South and southern Africa is in the midst of multiple, interrelated transitions with implications for health, social and development sectors. Health and socio-demographic surveillance systems are effective research instruments that can capture the rapidly-changing dynamics of health and social transitions in developing settings. Similarly, they can support a range of observational and intervention study designs including policy evaluations. The INDEPTH Network should boost much-needed comparative research; yet singly, and as a group, many of these sites have yet to fulfil their undoubted potential.

Kap. 1 schätzt den Effekt sozial gemischten Wohnens auf Resilienz gegenüber epidemischen Schocks. Anhand von Gesundheitsberichten und Berufsdaten aus Stadtverzeichnissen assoziiere ich die Verbreitung der Cholera während der 1866er Epidemie mit einem Maß für soziale Diversität für ca. 12200 Häuser Berlins. Diversere Häuser erleben mit höherer Wahrscheinlichkeit mindestens einen Fall, sind aber auch erfolgreicher bei der Eindämmung weiterer Fälle. Zur kausalem Interpretation nutze ich exogene Variation, die sich aus den geometrischen Eigenschaften der Gebäude ergibt. Ich zeige, dass Exposition gegenüber Außenkontakten und gemeinsamer Zugang zu Leitungswasser in gemischten Mietergemeinschaften die Inzidenzeffekte teilweise erklären. Kap. 2 evaluiert, ob die Cholera als Katalysator für städtische Was¬ser-infrastrukturreformen fungierte. In einer Fallstudie Berlins im 19. Jahrhundert zeige ich, dass die Interpretation der Cholera durch Miasma- und proto-epidemiologische Theorien der prä-bakteriologischen Ära ineffiziente, kontraproduktive Wasserwirtschaftsreformen inspirierten, was die Sterblichkeit für einige Zeit erhöhte. Das gängige Narrativ eines durch epidemische Schocks „erzwungenen“ sanitären Aufbruchs vermittelt ein irreführendes Bild der westlichen Volksgesundheitsgeschichte. Kap. 3 zeigt, dass Leitungswassernetze ohne Kanalisation geringen gesundheitlichen Nutzen stiften. Mittels Wasserspülung schwemmen Individuen Krankheitserreger in Rinnsteine, Grundwasserleiter, Straßen und offene Gewässer. Entlang dieser Abwasserströme lebende Nachbarn werden zusätzlichen Gesundheitsrisiken ausgesetzt, die durch den Anschluss der Abfallverursacher an eine Kanalisation neutralisiert werden. Mittels eines Flussrichtungsmodells schätze ich die Abwasser-Exposition für alle Gebäude Berlins in 1875/1880. In einer Differenz-in-Differenzen-Regression zeige ich, dass die negativen externen Effekte der Leitungswassernutzung dessen direkte Vorteile im Aggregat teilweise aufheben. / Chapter 1 estimates the causal effect of mixed-income housing on resilience to epidemic shocks. Using detailed health reports and occupational data from town directories, I relate cholera incidence to a social diversity measure at the level of Berlin’s roughly 12,200 buildings during the 1866 pandemic. Mixed tenant communities are more likely to experience an initial case, but also more successful in containing further in-house spread. To establish causality, I exploit exogenous variation from building lots’ geometric properties in an instrumental variable approach. I find that increased exposure to outside contacts and shared tap water access partly explain the effects. Chapter 2 evaluates whether cholera functioned as catalysts for the efficient reform of urban water infrastructure. Studying 19th century Berlin, I find that cholera’s conception through miasmatist frameworks and the proto-epidemiological tools of the pre-bacteriological era inspired inefficient and counterproductive approaches to water management and potentially deepened the mortality penalty for a certain time. This suggests that the popular interpretation of a sanitary awakening enforced by epidemic shocks paints a misleading picture of Western public health history. Chapter 3 tests a mechanism explaining why cities yield little health benefits from tap water if they do not simultaneously construct sewers. Individuals use the pressurized water supply to flush pathogens from their local environment, thus feeding additional waste to gutters, groundwater acquirers, streets and open water bodies. Neighbors living along the resulting waste flows bear indirect costs, only neutralized once waste emitters connect to sewers. Using a flow direction model based on Berlin’s elevation profile, I estimate waste flow trajectories and exposure for all buildings in Berlin in 1875/1880. In a difference-in- differences approach, I find that tap water’s negative external effects partly offset its direct benefits.