Journal of Clinical Epidemiology 55 (2002) 192–196 0895-4356/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S0895-4356(01)00463-2 Interest of LQAS method in a survey of HTLV-I infection in Benin (West Africa) Dismand Houinato a,b , Pierre-Marie Preux a,c, *, Bénédicte Charriere d , Bruno Massit e , Gilbert Avodé a,b , François Denis d , Michel Dumas a , Fernand Boutros-Toni a,c , Roger Salamon f a Institute of Neuroepidemiology and Tropical Neurology, School of Medicine, 2 rue du Dr Marcland, 87025 Limoges, France b Department of Neurology, School of Medicine, Cotonou, Benin c Department of Biostatistic and Medical Informatic, School of Medicine, 2 rue du Dr Marcland, 87025 Limoges, France d Department of Bacteriology and Virology, Dupuytren University Hospital, 2 Avenue Martin Luther King, 87042 Limoges, France e National Directory of Sanitary Protection, Ministry of Public Health, Cotonou, Benin f Institute of Public Health, INSERM U 330, University of Bordeaux II, 136 rue Léo Saignat, 33076 Bordeaux, France Received 25 November 2000; received in revised form 18 July 2001; accepted 2 August 2001 Abstract HTLV-I is heterogeneously distributed in Sub-Saharan Africa. Traditional survey methods as cluster sampling could provide information for a country or region of interest. However, they cannot identify small areas with higher prevalences of infection to help in the health policy planning. Identification of such areas could be done by a Lot Quality Assurance Sampling (LQAS) method, which is currently used in industry to identify a poor performance in assembly lines. The LQAS method was used in Atacora (Northern Benin) between March and May 1998 to identify areas with a HTLV-I seroprevalence higher than 4%. Sixty-five subjects were randomly selected in each of 36 communes (lots) of this department. Lots were classified as unacceptable when the sample contained at least one positive subject. The LQAS method identified 25 (69.4 %) communes with a prev- alence higher than 4%. Using stratified sampling theory, the overall HTLV-I seroprevalence was 4.5% (95% CI: 3.6–5.4%). These data show the in- terest of LQAS method application under field conditions to detect clusters of infection. © 2002 Elsevier Science Inc. All rights reserved. Keywords: Africa; Benin; HTLV-I; Retrovirus; LQAS; Prevalence 1. Introduction Lot Quality Assurance Sampling (LQAS) is useful for local managers to identify small areas of poor performance. This method, which is derived from production line indus- try, is identical to stratified sampling where small samples randomly selected from each lot are used to classify the lot as acceptable or unacceptable according to a preset quality level [1]. A decision of intervention is then made on the un- acceptable units (lots). Recently, some workers have at- tempted to apply these principles to the health sector [2–7]. Human T-cell Leukaemia Virus Type I (HTLV-I) is the etiological agent of Adult T-cell Leukaemia and Tropical Spastic Paraparesis/HTLV-I Associated Myelopathy. This retrovirus occurs in endemic proportions throughout South- western Japan [8] the Caribbean region [9], Sub-Saharan Africa [10], and Melanesia [11,12]. The modes of transmis- sion of HTLV-I are similar to those of blood-borne viruses, such as Human Immunodeficiency Virus and Hepatitis B Virus [13–16]. HTLV-I is geographically distributed in clusters (zones of high prevalence). Several studies have been carried out in Benin on the HTLV-I [17–19]. All of them have been conducted using a cluster sampling method. This method is easy to carry out under field conditions, and provides an estimate of HTLV-I prevalence for a country or a region of interest. However, this survey method is not use- ful for local managers, because it is inadequate to identify areas with high prevalences of HTLV-I. Between March and May 1998 a survey was conducted in 36 communes of the Department of Atacora in the North of Benin using LQAS method. The objectives of the study were to identify areas of high prevalence of HTLV-I and evaluate the feasibility of the application of the LQAS method to attain this objective. 2. Population and methods 2.1. Population and sample size The study was conducted in the Department of Atacora located in the North-Western Benin, bordering the Republic * Corresponding author. Tel.: 33 (0)5 55 43 58 20; fax: 33 (0)5 55 43 58 21. E-mail address: ient@unilim.fr (P.M. Preux)