Validation of neonatal tetanus elimination in Zambia by lot quality-assurance cluster sampling/Validation de l'elimination du tetanos neonatal en Zambie a l'aide d'un sondage en grappes pour le controle de la qualite des lots.

Introduction

Zambia has a population of approximately 12 million. According to estimates from the 2001-2002 Zambia Demographic and Health Survey, (1) between 1997 and 2001, the rate of neonatal mortality was 37/1000 births, the infant mortality rate was 95/1000 births and the maternal mortality ratio was 729/100 000 live births.

In order to protect mothers and their newborn babies against tetanus, WHO recommends that tetanus toxoid (TT) vaccine be given to all pregnant women; Zambia follows WHO's recommendations. In 2006, 79% of all pregnant women received a protective dose of TT vaccine. A total of 60% of all deliveries took place in hygienic conditions (administrative data). WHO and UNICEF estimate that in 2006, 90% of births were protected against tetanus.

Between 2001 and 2006, TT supplementary immunization activities (SIAs) were implemented in 18 districts at high risk of neonatal tetanus (NT), targeting about 400 000 women of childbearing age with 3 doses of TT. About 80% of the women targeted received at least 2 TT doses during these SIAs. The number of reported NT cases fell from 130 in 2000 to 37 in 2006.

In August 2007, the Ministry of Health, in collaboration with WHO and UNICEF, carried out an evaluation to determine whether NT had been eliminated in Zambia. Elimination is defined as an incidence of <1 case of NT/1000 live births in every district in a country. A community-based survey targeted 2 districts where children were considered to be at highest risk of NT.

Methods

District selection

Zambia is divided into 72 administrative districts. In preparation for assessing whether NT had been eliminated, district-level data were analysed by representatives from the Ministry of Health, WHO and UNICEF to assess the likelihood of elimination and to identify districts where the risk of NT was highest.

Using data from 2005 (the year for which the most recent data were available), the indicators assessed for each district included: the proportion of pregnant women who had received their second or subsequent dose of TT vaccine (designated as TT2+); the proportion of deliveries occurring in hospitals or other health facilities; whether the district was considered as an urban or rural area; the reported incidence of NT and the proportion of pregnant women who had had at least 1 antenatal care visit. From analyses of these data, the first 3 indicators were found to be the most useful in highlighting the worst-performing districts. Local knowledge about districts was also solicited. Sesheke, a rural district in the Western Province, was assessed by consensus as being the poorest-performing district. Because Sesheke's population was small compared with the required sample size, a second poorly-performing district was chosen to be included in the survey. A total of 2 rural districts that were among the worst performing were considered as the second district: Chama in the Eastern Province and Kaoma in the Western Province. Kaoma District was chosen because of its vicinity to Sesheke, which would simplify the logistic arrangements for implementing the survey.

Sesheke has about 4500 live births annually; Kaoma has about 9500. In 2005 in Sesheke, administrative coverage of TT2+ was 57%; the proportion of deliveries occurring in hospitals or other health facilities was 46%; and 2 cases of NT were reported (rate: 0.45/1000 live births). For Kaoma, the figures were: 57% administrative coverage of TT2+, 31% for the proportion of deliveries occurring in hospitals or other health facilities and 0 cases of NT. By comparison, in Chama, the administrative coverage of TT2+ was 54%; 32% of deliveries occurred in hospitals or other health facilities; and 1 case of NT was reported (rate: 0.21/1000 live births). In these 3 districts, coverage of TT vaccine and facility-based delivery were among the lowest in the country. In addition, in 2003, a survey by Zambia's Central Statistical Office found that the Western Province had the worst poverty in the country. SIAs implemented to provide TT vaccine took place in Sesheke in 2001 and 2002, and again in 2005 and 2006, but they were not implemented in Kaoma.

Survey protocol and forms

The survey method was adapted from a WHO protocol that uses the principles of lot quality-assurance sampling in combination with cluster sampling to determine with high probability whether the mortality rate from NT was <1/1000 live births during the period lasting from 1 to 13 months prior to the survey. Modifications were made to accommodate logistic constraints in Zambia. The sample plan chosen was a single sampling plan designed to survey 1375 live births. It had an acceptance number of 1 death from NT (n = 1375, d = 1); this means that if =1 death from NT was found in the survey, the disease could be considered to have been eliminated. This design and sample size provided similar probabilities of classification to the double sampling plan most often used in surveys in other countries. Live births that occurred 1-13 months before the survey (that is, between 15 July 2006 and 14 July 2007) were eligible for inclusion. In …