Table 2.
Logistic regression analysis to examine the effect of recall period on the VA assessment of all cause of death groups in Agincourt, South Africa between 1992 and 2011
| Cause of death | Recall perioda | ||
|---|---|---|---|
| 3–5 months OR (p value) |
6–11 months OR (p value) |
≥12 months OR (p value) |
|
| HIV/TB (n = 3759) | |||
| Crude | 0.96 (0.423) | 1.00 (0.897) | 0.66 (0.000)* |
| Adjustedb | 0.94 (0.339) | 1.01 (0.885) | 0.87 (0.065) |
| Acute respiratory (n = 712) | |||
| Crude | 0.90 (0.287) | 0.74 (0.001)* | 0.78 (0.018)* |
| Adjustedb | 0.93 (0.467) | 0.92 (0.468) | 0.88 (0.340) |
| Other Infectious (n = 405) | |||
| Crude | 1.10 (0.453) | 0.93 (0.564) | 1.40 (0.009)* |
| Adjustedb | 1.07 (0.605) | 0.87 (0.336) | 1.03 (0.862) |
| Neoplasm (n = 609) | |||
| Crude | 0.90 (0.313) | 1.13 (0.165) | 1.15 (0.173) |
| Adjustedb | 0.85 (0.140) | 0.88 (0.247) | 0.99 (0.961) |
| Other NCDs (n = 1429) | |||
| Crude | 1.10 (0.166) | 1.00 (0.990) | 0.91 (0.225) |
| Adjustedb | 1.10 (0.181) | 1.04 (0.584) | 1.07 (0.466) |
| Neonatal (n = 187) | |||
| Crude | 0.97 (0.876) | 0.74 (0.102) | 1.69 (0.004)* |
| Adjustedb | 1.17 (0.504) | 1.49 (0.114) | 2.58 (0.000)* |
| External (n = 641) | |||
| Crude | 0.95 (0.712) | 1.20 (0.074) | 1.68 (0.000)* |
| Adjustedb | 0.99 (0.936) | 1.14 (0.326) | 1.27 (0.097) |
| Indeterminate (n = 3140) | |||
| Crude | 1.12 (0.085) | 1.09 (0.119) | 1.56 (0.000)* |
| Adjustedb | 1.12 (0.089) | 1.06 (0.425) | 1.03 (0.701) |
OR odds ratio
* significantly different from reference group (p < 0.05)
a <3 months recall is the reference group
b Odds ratios were adjusted for: year of death, season (wet/dry), age, education, nationality and place of death. Only in the neonatal group, age was excluded from the model