Table 1 Components of health-determining factors (HDF) used in this study

Housing quality

A higher monthly rent for the house (0.6), owning a gas burner (0.8), better construction materials (0.5), and also better-quality fuel for cooking (0.9) contributed to a higher quality of housing. Hence, the first component was named 'housing quality', which explained 6.3% of the variance.

Population density

'Population density' explained 5.2% of the variance and was constituted through a higher number of family members (0.8), a higher number of persons sharing the same meals (0.7), and a higher number of persons living in the same room (0.7).

Smoking behaviour

Not smoking cigarettes (0.8), not smoking inside the room (0.8), and a small number of family members who smoke (-0.7) were correlated with the component 'smoking behaviour', which explained 4.8% of the variance.

Access to basic services

A large distance to the nearest park area (-0.8), a short distance to the nearest river (0.5), owning an electric fan (0.6) and a better water supply (0.5) correlated with this component, which we named 'access to basic services'. Owning an electric fan was regarded as a substitute for having an energy supply. This component explained 4.7% of the variance.

Household wealth

This component was correlated with owning a radio (0.6), owning a TV (0.6), owning a tape/CD/VCD (0.7) and the number of rooms (0.5). Having more rooms increases, for example, the capacity to generate income by renting out extra rooms. All of these variables indicate the household's material wealth, so the component was named accordingly. It explained 4.3% of the variance. Note that e.g., the wealth index used by Rutstein and Johnson [66] contained more factors than that of our study.

Natural environment

Larger amounts of vegetation 100 metres around the households (0.8), lesser amounts of surface water 100 metres around the households (-0.6) and longer distances to the nearest major street (0.7) were correlated with the component 'natural environment', which explained 4.3% of the variance.

Flood non-affectedness

Longer distances to the nearest river (0.5), whether the area was regarded as flood non-affected (0.7), and whether the area had an adequate drainage system (0.7) were correlated with this component, which explained 4.1% of the variance in the data.

Job satisfaction

Fewer working hours (-0.4), liking one's job (0.7), and not thinking that the job is harmful to one's health (0.8) were correlated with the component 'job satisfaction', which explained 4% of the variance.

Environmental health knowledge

Thinking that polluted, stagnant water and garbage near one's house could spread disease and increase the risk of poor health (0.8) and that air pollution is bad for one's health (0.6) were found to be correlated with this component. We named it 'environmental health knowledge', as it reflects awareness that the environment can affect one's health. This component explained 3.9% of the variance.

Income generation

A larger number of family members earning income (0.7), higher monthly family income (0.7), having a job contract (0.4), and working more hours a day (0.2) were correlated with the component 'income generation'. This component explained 3.7% of the variance.

Sanitation

A better toilet facility (0.7) and a better garbage disposal attitude (0.6) were correlated with the component 'sanitation', which explained 3.6% of the variance.

Housing sufficiency

'Housing sufficiency' was correlated with sufficient light in the house (0.6), whether the room was used for other purposes aside from living (0.7), and whether the room was regarded as sufficient for one's family (0.5). This component explained 3.6% of the variance.

Housing durability

Two variables, namely the household had a refrigerator (0.7) and the house was considered to be permanent (0.8) were correlated with a component called 'housing durability', which explained 3.5% of the variance.

Personal health knowledge

'Personal health knowledge' reflected awareness of personal behaviour and related health effects. Thinking that smoking tobacco is bad for one's health (0.7) and that physical exercise can be good for one's health (0.7) was correlated with this component. It explained 3.4% of the variance.