Skip to main content

Table 2 Multilevel linear models for stunting prevalence (2 years time-lag)a

From: Factors behind the success story of under-five stunting in Peru: a district ecological multilevel analysis

Dimension of the predictor variables

Predictor Variables

Time-adjusted regression coefficient

95% CI

p

Time and confounder-adjusteda regression coefficient

95% CI

p

 

Time (year)

−1.728

−2.056 to −1.399

0.000

−0.234

−0.862 to 0.395

0.466

Box A: Social determinants

log GDP per capita (USD/person)

−10.712

−17.174 to −4.250

0.001

-

-

-

Unmet basic needs (at least 1) % of people

0.187

0.070 to 0.305

0.002

0.209

0.108 to 0.310

0.000

Poverty line (% of people below)

0.416

0.302 to 0.529

0.000

0.196

0.073 to 0.319

0.002

Gini coefficient for income

0.050

−0.204 to 0.305

0.699

-

-

-

Urbanization (% of urban population)

−0.518

−0.626 to −0.409

0.000

−0.220

−0.374 to −0.066

0.005

Box B: Non-health sector variables

Years of schooling of women (median)

−2.839

−3.662 to −2.017

0.000

−0.928

−1.938 to 0.082

0.072

Improved water source (% of households)

−0.005

−0.097 to 0.088

0.921

-

-

-

Total fertility rate

1.241

−0.849 to 3.332

0.245

-

-

-

Cash Transfer Programme Coverage (No. fam/1000 rural pop)

0.001

−0.033 to 0.036

0.944

-

-

-

Box C: Health sector variables

SIS utilization (Number of attendance/u5 pop)

−0.585

−1.528 to 0.359

0.224

-

-

-

Per capita expenditure on child health activities (USD/u5 pop)

0.004

−0.009 to 0.016

0.569

-

-

-

Density of human resources for health (per 10,000 pop)

−0.147

−0.454 to 0.161

0.351

-

-

-

 

Composite coverage index (CCI)

0.166

−0.020 to 0.352

0.080

0.277

0.095 to 0.458

0.003

  1. aUnits of analyses are 168 (24 departments × 7 years). Variables in each group are adjusted for all other variables in the same group or above