Transactions of the Royal Society of Tropical Medicine and Hygiene (2007) 101, 1136—1142
available at www.sciencedirect.com
journal homepage: www.elsevierhealth.com/journals/trst
Hypertension, urbanization, social and spatial disparities: a cross-sectional population-based survey in a West African urban environment (Ouagadougou, Burkina Faso) Ali Niakara a, Florence Fournet b,∗, Jean Gary b, Maud Harang b,c, erard Salem b,c Lucie V.A. N´ ebi´ e a, G´ a
Centre Hospitalier Universitaire Yalgado Ou´ edraogo (CHNYO), Service de cardiologie, 01 BP 3401, Ouagadougou, Burkina Faso Institut de Recherche pour le D´ eveloppement, UR 178 Conditions et Territoires d’Emergence des Maladies, 01 BP 182, Ouagadougou 01, Burkina Faso c Universit´ e Paris 10-Nanterre, Laboratoire Espace, Sant´ e et Territoire, 200 avenue de la R´ epublique, 92 001 Nanterre Cedex, France b
Received 28 April 2007; received in revised form 9 July 2007; accepted 9 July 2007 Available online 13 August 2007
KEYWORDS Hypertension; Urbanization; Inequalities; Risk factors; Socio-economic factors; Burkina Faso
∗
Summary Data show that hypertension has become a public health problem in developing countries. Many studies have reported social disparities among the affected populations, but few of them pointed out spatial disparities within towns. We aimed to show that hypertension could be a good indicator of the medical change that occurs unequally in towns. A cross-sectional survey was done in April and October 2004 in Ouagadougou, Burkina Faso, among 2087 adults over 35 years old in different kinds of urban areas. Social and demographic data were collected and blood pressure was measured. Prevalence of hypertension was 40.2%. Age, body mass index, level of equipment, absence of community integration, absence of occupation, duration of residence over 20 years, protein-rich diet and absence of physical activity were identified as risk factors, but there were social and spatial disparities according to location of housing (parcelled-out or non-parcelled-out areas) and to integration within the town. The high rate of hypertension found in Ouagadougou and the heterogeneity of the risk within the population highlights that social and spatial risk factors have to be taken into account for the prevention of the non-transmissible diseases in countries in full process of urbanization and medical change. © 2007 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.
Corresponding author. Tel.: +226 50 30 67 37; fax: +226 50 31 03 85. E-mail address: fl
[email protected] (F. Fournet).
0035-9203/$ — see front matter © 2007 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.trstmh.2007.07.002
Social and spatial disparities of hypertension
1137
1. Introduction
2.2. Study design
Rapid urbanization involves changes that impact at the health level, particularly in the persistence of infectious and parasitical diseases, and development of chronic and degenerative pathologies. Of these emergent pathologies, hypertension is important because of its high frequency and because of the associated increased risks of cardiovascular diseases (Kearney et al., 2005). The link between hypertension and urbanization has been known for a long time (Cruickshank and Beevers, 1989; Gampel et al., 1962; Poulter, 1989; Scotch, 1963). The risk factors associated with hypertension have been well described (including diet, sedentary lifestyle and stress), but there is marked variability in their expression. In fact, the rapid social and spatial changes that mark developing societies lead to disruptions that can have different effects. For example, in Cuba, men of low educational level have a reduced risk of hypertension, whereas women of low educational level are at increased risk compared with individuals of higher educational level (Ordunez et al., 2005). In addition, it seems that, in contrast to the situation in industrialized countries, in developing countries hypertension is found in the poorest populations as well as in the more prosperous populations (Mendez et al., 2003). The objective of our study was to study hypertension risk in subjects exposed to rapid urbanization in a developing country such as Ouagadougou, Burkina Faso.
A cross-sectional study was done in April and October 2004. The city was stratified in order to identify and select different situations that would be representative of the urbanization processes that occurred inside the capital. Stratification was performed according to the type of housing, which opposes the regular city (parcelled out) to the irregular one (not parcelled out), and the density in habitat: a strong density in habitat relates to the concept of density of population, which is important for health in terms of overcrowding (House et al., 1988). Four strata were identified: (1) parcelled-out dense; (2) parcelled-out non-dense; (3) non-parcelled-out dense; (4) non-parcelled-out non-dense. In each stratum, two areas were selected for the study. Thus, a total of eight areas were studied. Area selection was based on age (old or new) and position (central or peripheral). Houses were randomly selected. In each selected house, the household was eligible only if the head was more than 35 years old (because hypertension risk increases with age) and had lived in Ouagadougou for more than 5 years (to ensure a sufficient exposure to urban lifestyle). Within each eligible household, all adults aged more than 35 years were eligible for inclusion in the study. A sample size of 384 individuals for every stratum was calculated to be adequate for measuring the prevalence of hypertension, assuming an expected prevalence of 10% with a precision of 3%. Results regarding hypertension were analysed according to the criterion of allotment — in both parcelled-out and non-parcelled-out areas. Indeed, the criterion of density in habitat has less importance for non-communicable diseases, so considering the number of hypertensive individuals identified in the different areas could lead to poor significance levels in the statistical tests. The research protocol was validated by the Ethics Committee for Research in Health of Burkina Faso. Each person included in our study formally agreed to take part in the investigation. People found to be suffering on the day of the investigation were assisted financially (e.g. for emergencies, cost of the specialized consultations and the first medical prescription). Blood pressure was measured with an automatic tension meter OMRON M5-I (OMRON HEALTHCARE Europe B.V., Hoofddorp, The Netherlands) in all individuals in a seated position after a 10 min rest (Niakara et al., 2003). Blood pressure was measured three times on each arm, with a 5 min interval to reduce the likelihood of aberrant values. The figures were calculated as follows: for each arm, of the three recorded measurements, the average of the two measurements closest to the mean of the Euclidean distance was calculated. People with systolic blood pressure (SBP) ≥140 mmHg for at least one arm and/or with diastolic blood pressure (DBP) ≥90 mmHg for at least one arm were classed as hypertensive of at least rank 1 (Chalmers et al., 1999). The limits were 160 mmHg for SBP and 100 mmHg for DBP for classification of hypertension of at least rank 2. People who were receiving treatment for hypertension at the time of the investigation were added to the hypertensive individuals. Age, gender, ethnic origin, age at arrival in Ouagadougou (before or after 25 years of age), duration of residence in the city (greater than or less than 20 years), occupation (a remu-
2. Materials and methods 2.1. Study area The study was carried out in Ouagadougou, the capital of Burkina Faso, in West Africa. Today, it is estimated that half the urban population of Burkina Faso lives in Ouagadougou, which has nearly 1 200 000 inhabitants (INSD, 2000). The size of the city has increased particularly rapidly: the urban area has multiplied more than six-fold in 40 years, reaching 230 km2 . Although the city has a relatively homogeneous appearance, particularly owing to the flat landscape, the development of urbanization is unequal. In the centre, old regular districts concentrate multi-family dwellings, whereas at the periphery, there is an increase in both popular and wealthy districts. Irregular districts border the regular ones. In regular areas, the township authority (cadastral services) allots streets and public services, such as tap water, sanitation, electricity and telephone, and parcels of land are allocated to inhabitants. In irregular areas, usually at the city margin, development occurs with no cadastral organization and without public services. Although it seems that settlement can be voluntary in irregular districts, it is generally forced initially due to financial reasons (lack of means to buy or even rent housing in the town centre) and sometimes social reasons (lack of social bonds not allowing access to housing).
1138
A. Niakara et al.
nerated activity or no activity), educational level (primary or less versus secondary or more) and belonging to social networks (active participation versus passive participation or lack of participation) were recorded. Weight and size were measured to calculate the body mass index (BMI; measured in kg/m2 ). If the patient’s BMI was equal to or greater than 25, the individual was classified as being overweight. Based on the information described above and other information on living conditions (food practices, physical activity, hygiene and protection, knowledge, disease representation), the following variables were defined: community integration, which was defined as being indigenous (Mossi) or of recognizing the traditional authority of the head of the district; a protein-rich diet, which was defined as eating fish or meat at least once per day; and physical activity, which was defined as walking or cycling for at least 30 min per day. An index of equipment was created using the scores from a principal components analysis carried out on the following variables: telephone, television, refrigerator, fan, presence or absence of living room furniture, means of transportation (bicycle, moped, motorbike, car). This index was then considered a proxy of the socio-economic status. The approximate prevalence of hypertension was calculated on the whole sample, and was standardized according to age and gender. The reference population was the whole population from the four stratified layers.
2.3. Statistical analysis The 2 test was used for comparisons of prevalence of hypertension or of proportions according to populations categorized by a qualitative variable. The 2 test of tendency was used to assess the evolutionary tendency of the prevalence of hypertension according to ordinal variables (particularly index of equipment). A logistic model of regression was applied to test variables that can affect
Table 1
hypertension. The interactions of these variables with the parcelling-out variable were tested in the model. The ‘significant’ term was used when a statistical test made it possible to reject an assumption with the threshold of 5%. The statistical analyses were carried out using SPSS 13.0 for Windows (SPSS Inc., Chicago, IL, USA).
3. Results 3.1. Study population A total of 2087 subjects was enrolled in the study. The prevalence of hypertension was 40.2%, and 18.5% of these hypertensive individuals were known and already treated. The prevalence of hypertension of rank 2 or more was 16.3%. After standardization according to age and gender, the prevalence was 42.8% in parcelled-out areas and 37.2% in non-parcelled-out areas. The difference was not significant (comparative morbidity figure = 1.15; 95% CI 0.99—1.34]. Table 1 shows the frequencies of the various variables previously reported as potential risk factors for hypertension. The population of the parcelled-out areas was older than that of the non-parcelled-out areas and also tended to be overweight. The apparent difference with gender is related to the fact that we questioned a greater number of women than men. Data from the 1996 population census showed that in the general population the gender ratio was balanced (INSD, 2000). Populations of the non-parcelled-out areas showed significant differences from populations of the non-parcelled-out areas in variables likely to have an effect on hypertension.
3.2. Social and spatial disparities of hypertension In the general population, the prevalence of hypertension increases with age, BMI, socio-economic status, lack of
Characteristics of the study population Total (n = 2044)
Non-parcelled-out areas (n = 972)
Parcelled-out areas (n = 1072)
P-value
Age (years)
45—54 >55
25.4 24.8
24.8 14.8
25.9 33.9
