Table 1 All variables in final mixed-effect regression model

TB incidence

The number of new cases of TB/quarter/population of the province, in x province (x = 1,2, … 63) at y quarter (y = 1–4, in a given year), LOESS detrended

Urban

Proportion of people living in urban compared to rural areas per province per year, LOESS detrended

Sex

Proportion of male to female per province per year, LOESS detrended

Population density

Number of individuals per province divided by the geographical area of the province, per province per year, LOESS detrended

Literacy

Percentage of literate individuals in those ≥15 years of age per province per year, LOESS detrended

Young

Percentage of population under 20 years of age per province per year, LOESS detrended

Old

Percentage of population ≥ 65 years of age per province per year, LOESS detrended

Poverty

Ratio of the number of people whose income falls below ½ the median income of the country, per province per year, LOESS detrended

Hospitals

Number of hospitals per province per year, LOESS detrended

HIV

Prevalence of HIV at 31st December each year, per province, LOESS detrended

migration

Rate of in migration into a given province, per province per year, LOESS detrended

quarter

1–4, beginning in January grouping into 3 month blocks.

Influenza-Like-Illness

Number of cases of influenza-like-illness/population of the province, per province per quarter, lagged by two quarter, LOESS detrended

Average temperature

average temperature per province per quarter, lagged by two quarters, LOESS detrended

Absolute humidity

average absolute humidity per province per quarter, lagged by two quarters, LOESS detrended

Rainfall

summed rainfall per province per quarter, lagged by two quarters, LOESS detrended

Sunshine

summed hours of sunlight per province per quarter, lagged by two quarters, LOESS detrended

Latitude

Latitude value of centroid point of province

(1|province)

Province as a random effect.

(1|year)

Year as a random effect