Table 5.
Results of univariate and multivariate logistic regression analysis with a backwards elimination procedure
| Variables | Control | Cases | Univariate | Multivariate | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| mean | (s.d.) | mean | (s.d.) | Crude OR (95% CI) | P value | OR (95% CI) | P value | |||
| Social contact score level 1 | 19·5 | (3·5) | 20·5 | (3·0) | 1·09 | (1·01–1·18) | P = 0·024 | 1·09 | (1·00–1·19) | P = 0·043 |
| Social contact score level 2 | 20·6 | (7·3) | 24·1 | (6·7) | 1·07 | (1·03–1·11) | P < 0·001 | 1·07 | (1·03–1·11) | P = 0·001 |
| Social contact score level 3 | 7·1 | (3·1) | 7·6 | (3·1) | 1·05 | (0·97–1·14) | P = 0·266 | |||
| Household size | 5·3 | (2·3) | 5·0 | (2·0) | 0·93 | (0·82–1·05) | P = 0·253 | |||
| Other variables | Control, n (%) | Cases, n (%) | ||||||||
| Wealth quintile (asset index) | ||||||||||
| 1 | 40 | (20·1) | 25 | (27·8) | 1·00 | |||||
| 2 | 40 | (20·1) | 20 | (22·2) | 0·80 | (0·38–1·67) | ||||
| 3 | 40 | (20·1) | 16 | (17·8) | 0·64 | (0·30–1·38) | ||||
| 4 | 40 | (20·1) | 17 | (18·9) | 0·68 | (0·32–1·45) | ||||
| 5 | 39 | (19·6) | 12 | (13·3) | 0·49 | (0·22–1·12) | ||||
| Wealth score continuous | 0·75 | (0·57–0·97) | P = 0·029 | |||||||
| Educational level* | ||||||||||
| High | 113 | (56·8) | 49 | (54·4) | 1·00 | |||||
| Low | 86 | (43·2) | 41 | (45·6) | 1·10 | (0·67–1·81) | P = 0·711 | |||
| Food shortage† | ||||||||||
| No | 128 | (64·3) | 47 | (52·2) | 1·00 | 1·00 | ||||
| Yes | 71 | (35·7) | 43 | (47·8) | 1·65 | (1·00–2·73) | P = 0·052 | 2·03 | (1·17–3·52) | P = 0·012 |
| Sex | ||||||||||
| Female | 116 | (58·3) | 41 | (45·6) | 1·00 | |||||
| Male | 83 | (41·7) | 49 | (54·4) | 1·67 | (1·01–2·76) | P = 0·045 | |||
| Age (years) | ||||||||||
| <10 | 22 | (11·1) | 6 | (6·7) | 1·00 | |||||
| 10–19 | 65 | (32·7) | 20 | (22·2) | 1·13 | (0·40–3·17) | P = 0·819 | 1·38 | (0·48–4·00) | P = 0·554 |
| 20–29 | 27 | (13·6) | 21 | (23·3) | 2·85 | (0·98–8·30) | P = 0·054 | 4·07 | (1·33–12·47) | P = 0·014 |
| 30–39 | 32 | (16·1) | 15 | (16·7) | 1·72 | (0·58–5·12) | P = 0·331 | 2·41 | (0·77–7·57) | P = 0·132 |
| 40–49 | 35 | (17·6) | 11 | (12·2) | 1·15 | (0·37–3·56) | P = 0·805 | 1·50 | (0·47–4·87) | P = 0·491 |
| ⩾50 | 18 | (9·1) | 17 | (18·9) | 3·46 | (1·13–10·61) | P = 0·030 | 5·17 | (1·56–17·11) | P = 0·007 |
| Total | 199 | (100) | 90 | (100) | ||||||
OR, Odds ratio; CI, confidence interval; s.d., standard deviation.
*
Educational level: low, highest educated person in the household had <6 years of schooling; high: highest educated person in the household had ⩾6 years schooling.
†
There was a recent period of food shortage reported (in the year before the interview).