The integrated population and health model (PHOENIX)

The rapid decline of mortality especially during the past decades in the less-industrialized countries without a parallel decline in fertility has led to an historically unprecedented rapid increase in population. In order to provide reasonable assessments of future developments, we need insights into the interrelated processes determining fertility and morbidity/mortality levels. For this purpose, an integrated population and health model was developed as part of the TARGETS 1.0 simulation model (Rotmans and De Vries, 1997; Hilderink et al., 1998). This population and health model was elaborated further, resulting in PHOENIX, as described in detail by Hilderink (2000).

The regional population estimates generated by PHOENIX can be used in the model WorldScan, the energy demand and supply model (TIMER) and the Terrestrial Environment System (TES). Many aspects of demography were disaggregated from the original projections For the implementation of the IPCC SRES scenarios presented in this USS.

PHOENIX is a tool to assess future changes (simulation period is 1950-2100) in the population size and structure in relation to the socio-economic conditions and state of the environment for the 17 world regions of IMAGE 2.2.

PHOENIX consists of three components:

• Fertility submodel, determining the number of births by region
• Mortality submodel, simulating the process of exposure to and death caused by several health risks by region
• Population submodel, which includes core demographics for regions using a standard cohort-component model of 100 age groups and the two sexes; it also combines the outcomes of the fertility and mortality submodels.

Population submodel

The population submodel uses an integrated systems approach, in which the results of the fertility and mortality submodels are structured into pressure, state, impact and response (P-S-I-R) modules:

• A pressure module describing the health determinants, which are divided into both socio-economic factors (income and literacy status) and environmental factors (food and water availability, malaria risk)
• A state module simulating fertility behaviour and population dynamics for disease and disease-specific mortality; both of these serve as input to the population module distinguishing sex and age groups
• An impact module describing the quantitative and qualitative aspects of the state module, such as the burden of disease and life expectancy, and the size and structure of the population
• A response module consisting of population policies influencing the fertility behaviour and health policies influencing the disease processes.

The Fertility Submodel

Human fertility is a biological process governed by social, economic, cultural and environmental variables. The effects of these variables on fertility levels are mediated by a set of proximate variables. The relationship between these proximate variables and fertility, which is well understood, forms the core of the fertility model (Bongaarts and Potter,1983).

The main outcome of the fertility submodule is the number of births. The calculation of births is based on the Bongaarts model, which assumes that an average biological maximum total fertility rate of 15.3 children per woman (FERTmax) is reduced by the following four determinants:

• The index of marriage (Cm), based on the average age of marriage and determines the fraction of the reproductive life span spent in stable sexual union
• Index of contraceptives (Cc), which represents the reducing effects of the use and effectiveness of several methods of birth control on reproduction
• Index of postpartum infecundity (Ci), which is defined as the fraction of the fertile life span lost for reproduction because of breast feeding and culturally motivated abstinence
• The index of abortion (Ca), which is a function of the number of induced abortions combined with the fraction of reproductive life span loss because of abortions

The combination of these factors results in the total fertility rate (TFR), which represents the number of children to which a woman has given birth at the end of her fertile period:

The model adopts the perspective that fertility change is the result of a 'modernization' process. Modernization is seen as a complex of interrelated processes of societal change, driven by gross domestic product, (female) literacy and life expectancy at birth. These are combined into the human development index (HDI) (see, for example, UNDP, 2000), which is used as indicator for the modernization process.

The concept of 'human development' represents an extension of a purely economic view on development. The two main characteristics which have been added to the Bongaarts approach are the linkage of the fertility determinants to the level of socio-economic development and the modelling of contraceptive use (Rosero-Bixby and Casterline, 1993).

The Mortality Submodel

The Mortality submodel simulates the number of persons exposed to various health risks and the number of deaths related to these exposures. The health risks associated with the exposed population are based on the broad and proximate health determinants of the health transition (Frenk et al., 1993).

The major health determinant is socio-economic status (SES) (Najman, 1993). The distinction between high and low socio-economic status is derived from the income status and the fraction of the total population that is literate. Further health risks include malnutrition, absence of safe drinking water, occurrence of malaria, habitual smoking and high blood pressure and poor availability of health services. Health risks are clustered into 12 categories on the basis of the empirically estimated contribution to mortality and disease levels in societies, as inferred from international statistics.