completed, even though he or she will be registered with Civil Acts Registration Bodies, in medical documentation this will be registered as an abortion. It is not clear whether similar rules apply in other countries. Note: Laws and decrees governing live birth and stillbirth definition and registration are referenced in more detail in Appendix II. 9 Another document, issued by Russian Ministry of Health in 2002 to further explain “Statistical counting of children born alive before 28 weeks of gestation are completed and weighing less than 1000 grams who survive perinatal period (168 hours)” (see Appendix II) states that any pregnancy which ends before 28 weeks of gestation are completed is classified as a miscarriage (abortion). If the infant survives seven days, a medical certificate of birth is issued and the infant is registered with the Civil Acts Registry as liveborn. However, such a pregnancy will still be classified in the documentation of the health facility as a miscarriage (abortion) and not a delivery, and will be recorded as a “termination of pregnancy (before 28 weeks)”, or as a case of “medical assistance to pregnant women”. It will not be included in the number of “Total deliveries” recorded by the health facility.6 The WHO definition of live birth was “adopted” in similar fashion in Armenia, Moldova, Ukraine and Kyrgyzstan. Thus, in these countries, in practice the shift from the Soviet to the WHO definition of live birth was partial in the case of both medical and civil registration statistics.7 What difference does the Soviet definition make to infant mortality statistics? Anderson and Silver (1986) have produced perhaps the most widely cited account of the possible impact on the official infant mortality rate if countries using the Soviet definition switched to the WHO definition. According to their estimates (based on data on infant survival for the US in 1960), the infant mortality rate would jump by between 22 and 25 per cent if the birth weight and gestation criteria in the Soviet definition were made consistent with WHO criteria. Table 4 shows that in the case of four countries where the change in definition was made in the early 1990s, and where infant mortality rates were recalculated, this estimate may still have some validity, but variation between countries and over years is considerable. In Poland, the infant mortality rate increased consistently by a fifth. In Moldova, the annual increase ranged from 5 to 17 per cent (however, as Table 3 notes, the new definition in Moldova retained several elements of the Soviet definition). In Latvia (where data, however, are presented for only one year) the infant mortality rate jumped by 40 per cent. Where other post-communist countries would fit within this range of increases is difficult to say, but the impact of the Soviet definition on infant mortality rates clearly remains important. 6 If the premature infant is stillborn or survives less than 7 days, the mother is only entitled to sick leave “under general conditions”. She is entitled to maternity leave if the infant survives 7 days or more. 7 It is worth noting that while most advanced industrialized countries progressively adopted the WHO definition of live birth from the 1960s onwards, in these countries too, adoption of the WHO definition has often been slow and partial. During the early 1990s, some Central European countries surpassed several Western countries in terms of conforming to the WHO definition. See Gourbin and Masuy–Stroobant (1995). 10 Table 4: Estimated increase in infant mortality rate caused by switch from Soviet to WHO definitions (per cent increase) Estonia Latvia Moldova Poland 198ource: Data supplied to MONEE Project by National Statistical Offices. 3. MIS-REPORTING OF INFANT BIRTHS AND DEATHS As noted in the previous section, under the Soviet definition, a birth must fulfil several criteria in order to be counted as a ‘live birth’. Velkoff and Miller argue that if an infant in the Soviet Union died and there was doubt regarding any of the criteria, “it would have been to the benefit of the hospital or the clinic to err on the side of stating that the infant was extremely premature, or intentionally to mis-report the birth as a stillbirth, because infant mortality rates were one criterion used to evaluate hospitals and clinics.” (1995, p.243). Many experts agree that mis-reporting of infant deaths was widespread in the Soviet Union, and that it has continued in many of its successor states (for example, see Buckley, 1998; Yeganyan et al., 2001; Centers for Disease Control and Prevention and ORC Macro, 2003). Suspicions have also been raised in relation to some countries in South Eastern Europe, but little hard evidence has been presented (Serbanescu, Morris and Marin, 2001; Rechel and McKee, 2003). The literature notes two types of mis-reporting that reduce the infant death count: 1. The classification of live births (even according to the Soviet definition) as stillbirths or miscarriages – referred to in the citation above of Velkoff and Miller. 2. The recording of deaths of infants aged under one year as occurring when the infant was aged over one year.8 8 A third type of mis-reporting sometimes mentioned is the classification of deaths with ‘sensitive’ causes, for example intestinal infections, as being caused by other less sensitive factors, for example