Levels and trends of demographic indices in southern rural Mozambique: evidence from demographic surveillance in Manhiça district
© Nhacolo et al; licensee BioMed Central Ltd. 2006
Received: 04 May 2006
Accepted: 30 November 2006
Published: 30 November 2006
In Mozambique most of demographic data are obtained using census or sample survey including indirect estimations. A method of collecting longitudinal demographic data was introduced in southern Mozambique since 1996 (DSS -Demographic Surveillance System in Manhiça district, Maputo province), but the extent to which it yields demographic measures that are typical of southern rural Mozambique has not been evaluated yet.
Data from the DSS were used to estimate the levels and trends of fertility, mortality and migration in Manhiça, between 1998 and 2005. The estimates from Manhiça were compared with estimates from Maputo province using the 1997 National census and 1997 Demographic and Health Survey (DHS). The DHS data were used to estimate levels and trends of adult mortality using the siblings' histories and the orphanhood methods.
The populations in Manhiça and in Maputo province are young (44% <15 years in Manhiça and 42% in Maputo); with reduced adult males when compared to females (all ages sex ratio of 78.7 in Manhiça and 89 in Maputo). Fertility in Manhiça is at a similar level as in Maputo province and has remained around 5 children per woman, during the eight years of surveillance in Manhiça. Although the infant mortality rate (IMR) in Mozambique has decreased during the last two decades (from 148 deaths per 1000 live births in 1980 to 101 in 2003), it has remained stable around 80 in Manhiça during the surveillance period. Adult mortality has increased both in Manhiça (probability of dying from ages 15 to 60 increased from 0.4 in 1998 to 0.6 in 2005 in Manhiça, from 0.3 in 1992 to 0.4 in 1997 in Maputo province and from 0.1 in 1980 to 0.6 in 2000 in Mozambique). Consequently, the life expectancy decreased from 53 to 46 in Manhiça and from 42 years in 1997 to 38 in 2004 in Mozambique. Migration is high in Manhiça but tends to stabilise after the movements of resettlement that followed the end of the civil war in 1992.
The population under demographic surveillance in Manhiça district presents characteristics that are typical of southern rural Mozambique, with predominance of young people and reduction of adult males. Labour migration and excess adult male mortality are the major factors for the reduction of adult males. Mortality is high and only infant mortality has started to stabilise while adult mortality has increased, and as consequence, life expectancy has decreased. The Manhiça DSS is an adequate tool to report demographic measures for southern rural Mozambique.
In Mozambique as well as in many other sub-Saharan African countries, people are born and die before being formally registered and most of demographic data are obtained using census or sample survey including indirect estimations. However, problems with surveys and census include errors such as omissions of live births that died immediately, deaths from households that disappeared after death, multi reporting of individuals and other events when the concept of household is mixed with that of family, which is likely in most African countries, and particularly, the fact that the data are often out of date [1, 2].
Since 1996, a Demographic Surveillance System (DSS) has been running in Manhiça district, Maputo province, Southern rural Mozambique. A DSS aims to produce high quality estimates, but how good is it on reporting demographic indices that are typical of Mozambique has not been evaluated yet. This paper aims to assess the extent to which this DSS yields demographic measures that are typical of southern Mozambique by comparing it's demographic indices with the 1997 National Census and the 1997 Demographic and Health Survey (DHS) data.
The data comes from different sources namely (i) the Demographic Surveillance System (DSS) in Manhiça district, Maputo province, carried out by the Manhiça Health Research Center (CISM); (ii) the Demographic and Health Survey (DHS) carried out in Mozambique by the National Institute of Statistics and the Ministry of Health in 1997 and (iii) the IInd National Census carried out in Mozambique also by the National Institute of Statistics in 1997.
Population characteristics in Manhiça DSS (1998–2005), Maputo province (1997–2005) and Mozambique (1997–2005)
DSS, mid yr pop.
DSS, sex ratio*
DSS, household size**
DSS, % female headed
Maputo prov., sex ratio*
Mozambique, sex ratio*
The DHS data were accessed from the Macro International website. In brief, women aged 15–49 years were interviewed on the numbers of brothers and sisters born to the same mother and their corresponding vital status. Similarly, children less than 15 years of age were asked on the survival of their parents.
The analysis comprises a description of the basic demographic measures such as sex ratio, population pyramid, mean household size, age-specific fertility rates, abortion ratio, age-and-sex-specific mortality rate including the probability of dying between ages 15 and 60 years (45q15), life expectancy at birth and migration rates from the DSS. Rate ratios, 95% Confidence Intervals (CI) and P-values of mortality and migration comparing males to females, after controlling for age in Manhiça were calculated. A Poisson regression model using the year as covariate was used to evaluate the linear trends of infant mortality, adult mortality and migration in Manhiça. The coefficient for the "linear" model estimates the change in the incidence per year increase. The significance of the linear effect was evaluated using a Likelihood Ratio Test comparing the null model with the "linear" model. In order to evaluate if the linear is sufficient to explain the differences by year, a model using the year as categorical covariate was used. These estimates were compared, where applicable, with the estimates from the DHS and Census for Maputo province and Mozambique as a whole. Estimates from DHS include 45q15 using the direct siblings' survival method (based on the respondent's number of siblings ever borne to the same mother and their survival status at the time of the survey), and the orphanhood method (based on the proportions of people whose parents have died). These methods are described elsewhere [1, 4].
Some of the weaknesses of this analysis include (i) differences in sizes of the areas compared (Manhiça DSS area is very small compared to Maputo province and it covers only 9.9% of the Maputo province population); (ii) differences in time-periods compared (Census = 1997, Manhiça = 1998–2005 and DHS = 1997 with time location of adult deaths rolling back to 3–15 years before the survey); (iii) the inability to measure the statistical significance of the differences or similarities observed when comparing the indices from different sources. Whereas rate ratios, 95% CI, Poisson regression for trend analysis and P-values were calculated for the Manhiça DSS data, this was not possible with the census data because the National Institute of Statistics does not deliver dataset at individual level and the indices were calculated from published data in tables. Although the DHS data was at individual level, the indirect methods used did not allow for testing the significance of the differences or similarities between its indices and that of other sources.
Results and discussion
2.1. Population characteristics
Villages in Manhiça typically comprise a loose conglomeration of compounds separated by garden plots and grazing land. Houses are simple, with walls typically made of cane (84%) and conventional material (15%). About 83% of households have latrines. In towns, houses are often grouped into family compounds and surrounded by grass fences. Towns grew substantially during the civil war in the 1980s as displaced people looked for refuge . Water comes mainly from community wells. With the exception of the center of the towns of Manhiça, Maragra, Palmeira and Ilha Josina, which has a 24-hours public electricity service, the rest of the area relies on more traditional systems for lighting. These characteristics are common in many other areas of southern rural Mozambique .
The population under surveillance has increased in part due to the expansion of the study area, but the sex ratio, mean household size and the percentage of households headed by females remained constant, indicating that the newly incorporated areas were very similar to the older ones (table 1). The population growth rate within the DSS between 1997 and 2001 was 3.8%, which is 65% higher than the rate in Maputo province in 2002 (2.3%) . The national population growth rate for 1997 and 2003 was 1.7% and 2.3% respectively .
Fertility in Manhiça DSS, 1998–2005; Maputo province and Mozambique
DSS, Abortion ratio*
DSS, Still birth ratio*
DSS, Total Fertility Rate (TFR)
Age-specific fertility rate is a measure that shows the variation of fertility by age of women; is obtained by dividing the number of children born to mothers in specific age by the number of women in that age. The patterns of fertility by age are similar, but women in Manhiça appears to start childbearing at very young ages and stop earlier than in Maputo province and Mozambique as a whole.
Mortality in Manhiça 1998–2005, Maputo province 1997–2002 and Mozambique, 1997–2004
Infant* M. Rate
Under five Mort. 5q0
Maternal M. ratioa
Matern M. ratio
Migration is a very important component of population change in Mozambique, but little is known about its levels and trends. Migration is the determinant of population change more difficult to study. The difficulties start with defining who is a migrant and, since migration involves two distinct places (origin and destination), one need to take into account the distance or boundaries between these places, the duration of the stay and the intention. The census and sample surveys, which are the common sources of data, track only part of these movements .
In Manhiça an out-migrant is a resident who leaves the DSS area for three or more months. Immigration occurs when a person enters the area to live for at least three months (include return migration). Should he move from a household to live in another household for three months or more within the DSS area he or she is regarded as internal migrant. The definition of migration in Manhiça is based on a combination of the intention of the move and the duration of the stay to avoid the ambiguity on temporary absence and temporary visits, which complicates the definition of migration events. Thus, anyone who comes into the DSS to visit his relatives, rather than to live but who stays for three months or more he is counted as an immigrant. Conversely, any person who reportedly comes to live in the area is registered as an immigrant resident immediately, and keeps this status in the database even if he subsequently leaves within three months. Irrespective of his/her economic contribution, anyone who never stays within the area for more than three months is not counted as resident. All the vital events and migration are linked to the individual by the Permanent Identification Number, which allowed estimates of rates by dividing the number of events by the number of person-years of observation. The migration rates for Maputo province were calculate by National Institute of Statistics  using the census data, comparing the reported place of residence five years before the census (1992), one year before the census (1996) and the place of residence at census time (1997).
The preferred destinations for emigrants from Manhiça are the Maputo city (35.6%), followed by South Africa (19.6%), other districts within Maputo province (19%) and other provinces. Immigrants are mostly from Maputo City (40.3%), Maputo province (19.8%), South Africa (15.6%) and other provinces.
The population under demographic surveillance in Manhiça district presents characteristics that are typical of southern rural Mozambique, with predominance of young people and reduction of adult males when compared to females. This reduction of adult males may represent the impact of the historical pattern of migration in southern rural Mozambique, where a work-related flow of males to Maputo city and South Africa has been kept for centuries. Despite efforts carried out to improve the health services in Mozambique, mortality is still high and only infant mortality has remained stable while adult mortality has increased, and the life expectancy has worsened over the last two decades. The pattern of mortality by age and sex in Manhiça is typical of developing countries that are heavily affected by the AIDS epidemic. The Manhiça DSS is an adequate tool to report demographic measures for southern rural Mozambique.
The Spanish Agency for International Cooperation funds the running costs of CISM. During the recent years, various studies were funded by different sources, including de World Health Organization, United Nations Children's Fund, INDEPTH, Hospital Clinic of Barcelona (University of Barcelona, Spain), the Spanish Ministry of Health, and the Eduardo Mondlane University, Mozambique. CISM is the first peripheral research centre of Mozambican Ministry of Health. It is being developed under the premises of the cooperation program between Mozambique and Spain. A collaborative agreement, further developed between the Ministry of Health, the Faculty of Medicine at Eduardo Mondlane University, and the Hospital Clinic of Barcelona, ensures the running of the centre.
Special acknowledgments to the Measure ORC Macro, for accepting the use of the 1997 Mozambican DHS for this analysis. We also thank the community in Manhiça for allowing the surveillance to take place.
- Timaeus I, Garaham W: Measuring adult mortality in developing countries: a review and assessment of methods. CPS Working paper N0 88-4. 1988, London, UK, Centre for Population Studies. London School of Hygiene and Tropical Medicine. University of London.Google Scholar
- Blacker J: The estimation of adult mortality in Africa from data on orphanhood. Population Studies. 1977, 31,1.:Google Scholar
- Alonso P, Saute F, Aponte J, Gomez-Olive F, Nhacolo A, Thompson R, et al: Manhiça Demographic Surveillance System, Mozambique. Population and health in developing countries. 2002, Ottawa, Cairo, Dakar, Montevideo, Nairobi, New Deli, Singapore., International Development Research Centre (IDRC), 1: 295-308.Google Scholar
- Timaeus I, Zaba B, Blacker J: Estimation of adult mortality from data on adult siblings. Brass Tacks Essays in Medical Demography A tribute to the memory of Professor William Brass. Edited by: B Z and J B. 2001, London and New York, Athlone Press, 43-66.Google Scholar
- (INE) INE: Características sócio-económicas das comunidades rurais em Moçambique, 2002/3. Relatório final. 2003, Maputo, Mocambique, Instituto Nacional de EstatisticaGoogle Scholar
- (INE) INE: Anuário estatístico 2002, província de Maputo. 2003, Maputo, Mocambique, Instituto Nacional de EstatisticaGoogle Scholar
- (INE) INE, (MISAU) MS: Moçambique, inquérito demográfico e de saúde 2003. 2005, Maputo, Moçambique, Instituto Nacional de Estatística (INE), Ministerio de Saude and Measure DHS +/ Macro InternationalGoogle Scholar
- (INE) INE: Mulheres chefes de agregados familiares em Maputo Cidade. Cifras e realidades. 1999, Maputo, Mocambique, Instituto Nacional de Estatistica, Serie: Estudos No 4:Google Scholar
- WHO: Unsafe abortion: Global and regional estimates of the incidence of unsafe abortion and associated mortality in 2000. 2004, [http://www.who.int/reproductive_health/publications/unsafe_abortion_estimates_04/index.html]4thGoogle Scholar
- Newell C: Methods and models in demography. 1998, London, Great Britain, Belhaven PressGoogle Scholar
- (WHO) WHO: The World Health Organization Report 1999. Conquering suffering, enriching humanity. 1999, Geneva, WHOGoogle Scholar
- (WHO) WHO: The World Health Organization Report 1999. Annex. 2000, World Health Organization, Accessed 30.08.02., [http://www.sdusd.org/sdi/issues/health/world-health-report]Google Scholar
- (WHO) WHO: The World Health Organization Report 2000. Mental Health: new understandings, new hope. 2001, World Health Organization, Accessed 30.08.02., [http://www.who.int/whr/2001/en/annex1_en.pdf]Google Scholar
- Gom P, Clark S: Adult mortality in the era of HIV/AIDS: Sub-Saharan Africa. In workshop on hiv/aids and adult mortality in developing countries. UN Population Division New York, 8-13 September 2003. 2003, New York, UN Population Division., accessed 15.06.2006, [http://www.un.org/esa/population/publications/adultmort/CLARK_Paper3.pdf]Google Scholar
- (INE) INE: Atlas sócio-demográfico de Moçambique. 2000, Maputo, Mocambique, Instituto Nacional de EstatisticaGoogle Scholar
- (INE) INE: II Recenseamento geral da população e habitação. Resultados definitivos. 1999, Maputo, Mocambique, Instituto Nacional de Estatistica, [http://www.ine.gov.mz/censo2/Recens.htm]Google Scholar
- Mozambique UNICEF: Mozambique. 2006, Maputo, Mocambique, UNICEF MozambiqueGoogle Scholar
- Census USB: Data Mozambique 1997: Results from the Demographic and Health Survey. 2000, US Bureau of Census, Accessed 23.06.02., [http://www.popcouncil.org/pdfs/sfppdr/sfp/sfp31-1.pdf#Mozambiquedhs]Google Scholar
- Dgedge M, Novoa A, Macassa G, Sacarlal J, Black J, Michaud, et al: The burden of disease in Maputo City, Mozambique: registered and autopsied deaths in 1994. Bulletin of the World Health Organization, 2001. 2001, Geneva, World Health Organization, 79(6): 546-552. Accessed 12.07.02, [http://www.who.int/bulletin/pdf/2001/issue6/vol.79.no.6.546-552.pdf]Google Scholar
- (INE) INE: Projecções anuais da população por província e área de residência, 1997-2010. Moçambique. Serie: Estudos No 2. 1999, Maputo, Moçambique, Instituto Nacional de EstatisticaGoogle Scholar
- Globalis: Globalis-Mozambique (quoting UN common database). 2006, accessed on 02.03.2006., [http://globalis.gvu.unu.edu/indicator_detail.cfm?IndicatorID=138&Country=MZ]Google Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2458/6/291/prepub