Background mortality
While the majority of deaths within a short time of parturition can be expected to be direct obstetric deaths, the proportion that is will vary according to the occurrence of
Table 11.2 Stillbirth rates
| Location | |
| 0.0074 | |
| Croatia 1874-80 | 0.0106 |
| Croatia 1881-90 | 0.0131 |
| Croatia 1891-1900 | 0.0188 |
| Croatia 1901-10 | 0.0222 |
| Croatia 1906-10 | 0.0222 |
| Pozega County 1906-10 | 0.0179 |
| Nova Gradiska 1910 | 0.0087 |
| Five Swedish parishes | 0.0298 |
Source: Schofield 1986: Table 9.2.
epidemic disease and the efficacy of obstetrical techniques.
Our approach in graphical analysis is to subtract the estimated risk of BM from that of mortality within the critical time period, to obtain an estimate of net MM. In regression analysis we use background mortality centred on the year of birth as a predictor of maternal death so that estimates of the effects of other predictors are effectively of their influence on net MM.There are several ways to estimate BM. Wrigley et al. (1997: 312 ff) use the mortality rate of husbands in the same critical period. Indeed, deducing the MM rate by comparing male to female mortality has been a common practice (see for example Henry 1987). However, where there are other correlated components to the gender difference, MM will be mis-estimated. For example, where tuberculosis was an important cause of death, female survivorship in the eighteenth—nineteenth centuries was often markedly worse than that of men. Cortes-Majo et al. (1990) have pointed to this specifically; more general discussions of gender differences are found in Ginsberg (1989) and Hammel et αl.(1983). The numbers of males and females with tuberculosis as the reported cause of death in the Slavonian data are 1,640 and 1,622, respectively, almost equal. By contrast, the numbers of males and females suffering some form of violent death (drowning, kicks by horses, falls from trees, murder, gunshot, etc.) were 190 and 91, respectively Thus, using males for BM would result in an underestimate of MM.
We have estimated BM, by calendar year and age of mother, with a risk period from 61 to 730 days after giving birth and for nulliparous women, within two years of marriage. Two years is somewhat shorter than the mean birth interval (2.6 years).23
6.