Table 1 Definition of variables

Age

Age was categorized into three groups: 0.5–1 year, 2–4 years and 5–14 years

Altitude

Altitude (metres) was categorized into three groups with the following cutoff Low altitude ranging (1128–1264), Medium altitude ranging (1265–1352) High altitude ranging (1353–1654)

Social economic status

SES was created as a weighted sum of data on household possessions and utilities, using principal components analysis and divided into tertiles

Improved housing structure

Is a binary variable where houses are coded as ‘yes’ if they have at least one of the following intact ceiling, brick walls, plastered walls, full cement floors and iron roofs

Household crowding

Number of people sleeping per room. Number of sleepers per room was dichotomized into groups with 1, and > 1 person sleeping in the room

Eaves of the house

Binary variable ‘open’ if household has open eaves or ‘closed’ if household has no open eaves

LLINs’ ownership

Binary variable ‘yes’ if household owns at least one LLIN

Individual LLIN condition

Physical condition of the net used by sleeper. The variable is categorized in. 1/ good condition net (PHI 0–64). 2/ moderate condition net (PHI 65–642) and 3/ torn net (PHI > 643)

Individual LLIN usage

This is a binary variable indicating whether individual is sleeping under a net or not

Household net access

Binary variable ‘yes’ if household owns enough net to cover every sleeping place. Which is number of LLINs owned (regardless if used or not) divided by the number of sleeping places

Indoor residual spraying (IRS)

Binary variable ‘yes’ if household was sprayed in 2013

Livestock present

Binary variable ‘yes’ if livestock are owned

Animal kept inside the house

Binary variable ‘yes’ if animals are sleeping inside the household, this included goats, sheep and cows

Level of education of head of household

Binary variable ‘yes’ if the head of household has at least completed primary education

Moderate to severe Anaemia

Defined as haemoglobin < 8 g/dL