Table 2.
Univariate and multivariate analysis of under-five mortality
| Univariate | Multivariate | |||
|---|---|---|---|---|
|
|
|
|||
| Variable | HR | 90% CI | HR | 95% CI |
| District | ||||
| Kilombero (Ref) | ||||
| Ulanga | 0.97 | 0.86–1.10 | ||
| Gender | ||||
| Female (Ref) | ||||
| Male | 1.22 | 1.09–1.37 | 1.16 | 1.01–1.35 |
| Age at start of study | 0.49 | 0.45–0.52 | 0.35 | 0.29–0.43 |
| Euclidean distance | ||||
| < 5 km (Ref) | ||||
| > 5 km | 1.08 | 0.96–1.21 | ||
| Networked distance | ||||
| < 5 km (Ref) | ||||
| > 5 km | 1.19 | 1.06–1.33 | 1.17 | 1.02–1.38 |
| Mothers age | ||||
| < 20 (Ref) | ||||
| 20–29 | 0.59 | 0.51–0.69 | 0.93 | 0.74–1.17 |
| 30–39 | 0.59 | 0.50–0.69 | 1.21 | 0.89–1.64 |
| > 40 | 0.32 | 0.22–0.45 | 1.00 | 0.611–1.64 |
| Mother education | ||||
| None (Ref) | ||||
| Primary | 0.91 | 0.75–1.10 | ||
| Post primary | 0.56 | 0.26–1.19 | ||
| Parity | ||||
| 1 (Ref) | ||||
| 2 to 3 | 0.65 | 0.55–0.77 | 0.61 | 0.52–0.83 |
| 4+ | 0.69 | 0.59–0.81 | 0.56 | 0.52–0.91 |
| Multiple/single birth | ||||
| No(Ref) | ||||
| Yes | 2.99 | 2.46–3.63 | 2.91 | 2.27–3.74 |
| Mother death | ||||
| No(Ref) | ||||
| Yes | 4.41 | 3.34–5.82 | 5.87 | 4.11–8.40 |
| Birth interval to previous sibling | 1.00 | 0.99–1.00 | ||
| Birth interval to next sibling* | 0.91 | 0.90–0.92 | ||
| Death of preceding sibling | ||||
| No(Ref) | ||||
| Yes | 4.5 | 3.42–5.91 | 1.90 | 1.37–2.65 |
| Season | ||||
| Rainy(Ref) | ||||
| Dry | 0.95 | 0.85–1.07 | ||
| Wealth quintiles** | ||||
| Poorest (Ref) | ||||
| Poorer | 0.86 | 0.71–1.02 | ||
| Poor | 0.88 | 0.73–1.04 | ||
| Less poor | 0.91 | 0.76–1.09 | ||
| Least poor | 0.86 | 0.72–1.02 | ||
| HHH occupation | ||||
| Farming (Ref) | ||||
| Causal worker | 0.85 | 0.64–1.12 | ||
| Business | 0.95 | 0.80–1.12 | ||
| Gov't employee | 1.09 | 0.90–1.33 | ||
| Other | 0.79 | 0.59–1.05 | ||
| Family size | ||||
| 1–2(Ref) | ||||
| 3+ | 1.36 | 1.22–1.52 | 1.30 | 1.11–1.52 |
*
Not included in multivariate because of small numbers.
**
Wealth Index showed significant association with child (1–4) mortality though no association existed with infant mortality in the analysis.