A computer simulation of the EPI survey strategy.

A Monte Carlo simulation study was designed to evaluate the sample survey technique currently used by the Expanded Programme on Immunization (EPI) of the World Health Organization. Of particular interest was how the EPI strategy compared to a more traditional sampling strategy with respect to bias and variability of estimates. It was also of interest to investigate whether the estimates of population vaccination coverage were accurate to within 10 percentage points of the actual levels. It was found that within particular clusters, the EPI method was particularly sensitive to pocketing of vaccinated individuals, but the more traditional method gave more accurate and less variable results under a variety of conditions. However, the stated goal of the EPI, of being able to produce population estimates accurate to within 10 percentage points of the true levels in the population, was satisfied in the artificially created populations studied.

Evaluation of a standardized survey design proposed for use in epidemiological research on AIDS.

A Monte Carlo model simulating actual populations was employed to evaluate the precision in estimation of a standardized sampling method proposed by the Global Programme on AIDS of the World Health Organization, for general use in collecting population-based data on HIV seroprevalence. It appears that in real populations, where there is likely to be 'pocketing' of infection, the proposed methodology will generally fail to provide estimates accurate to within 1% of the true population value. However, if the primary objective of a particular survey is to construct confidence intervals that include the true population HIV seroprevalence rate, then this survey method appears to be a reasonable choice. This study also suggests that selection of only one adult per household improves the precision of resulting estimates. However, since selection of only one adult per household would require that more households be visited, any gain in precision would need to be weighed against the likely increase in cost of household visits.

Estimation and projection of adult AIDS cases: a simple epidemiological model.

Many HIV/AIDS (acquired immunodeficiency syndrome) models have been developed to help our understanding of the dynamics and interrelationships of the determinants of HIV (human immunodeficiency virus) spread and/or to develop reliable estimates of the eventual extent of such spread. These models range from very simple to very complex. WHO has developed a simple model for short-term projections of AIDS, details of which are presented here along with results obtained using the model to estimate and project AIDS cases for the USA, sub-Saharan Africa, and south/south-east Asia. WHO has also developed, based on the model described in this paper, a computer program (Epi Model), which will enable the user to easily change the values of any of the variables required by the WHO model.

PIP: To aid developing countries in short-term program planning, the World Health Organization (WHO) has devised a model capable of estimating likely trends and numbers of cases of acquired immunodeficiency syndrome (AIDS) over the succeeding 3-4 years. A human immunodeficiency virus (HIV) point prevalence estimate is used in combination with both the estimated year in which HIV transmission became widespread and the HIV infection curve during the epidemic period, and then these data are used to calculate annual cohorts of HIV-infected adults. The projected number of AIDS cases is obtained by multiplying each of these annual cohort estimates by the progression rates from infection to clinical AIDS. Basic to the WHO model is the assumption that cumulative HIV infections follow a sigmoid curve; it is further assumed that the distribution of the HIV infection over time will be skewed, with a long right tail. Application of this model to data from the US, sub-Saharan Africa, and South/Southeast Asia revealed some differences with prevailing estimates. According to the WHO model, the cumulative HIV incidence in the US by 1990 should be slightly less than the lower range (1 million cases) of the Centers for Disease Control estimate. For sub-Saharan Africa, the model yielded estimates of a cumulative total of 700,000 adult AIDS cases by the end of 1990 and over 2 million bases by 1994, with 1.75 million cumulative deaths--estimates that exceed official statistics by 10 times. Finally, the WHO model projects at least 60,000 AIDS cases in South/Southeast Asia by 1994, which, again, greatly exceeds official estimates. WHO has developed a computer program, Epi Model, based on this model that enables users to change any of the relevant variables.

Immunization coverage surveys: methodological studies in Indonesia.

PIP: A field study was conducted in rural and urban areas of Indonesia to quantify the relative costs of surveys using the standard Expanded Program on Immunization (EPI) methodology or using 7 randomly selected households within each cluster. Specifically, the objectives were: to assess whether immunization schedules were being adhered to in the study areas; to quantify the differences in survey time and cost for a rural and urban setting between the standard EPI survey method and the statistically rigorous approach, whereby 7 starting households were selected randomly within each cluster; and to assess the applicability of Lot Quality Assurance techniques as a managerial tool in immunization programs. The 2 areas of study were the Gianyar "kapupaten"(district) of Bali and the "timur"(east) municipality of Jakarta. In both places reliable population data were available from the 1980 census and, in Bali, these had been continually updated through regular reporting of births, deaths, and migrations. Gianyar district was chosen because it is rural, has a fairly good immunization reporting system, up-to-date household lists, and is logistically convenient. Both types of surveys. the standard EPI method and the SRS method in which a randomly selected starting household is used for each child, were conducted in each of the study areas during September and October 1986. Children's ages were calculated, in months, with respect to the 1st day of the survey. The age and the date of immunization of each child were then used to compute the age at immunization in months. In Gianyar, 207 children aged from 15 months to less than 24 months had their immunization status properly recorded using the EPI survey method; 209 children were successfully recorded by the SRS method. Using the EPI method, it was estimated that 75.4% of the children had received 8 immunizations. Use of the SRS method gave an estimate of 86.6%. In Jakarta, 207 children were properly surveyed by the epi method, with an overall complete immunization rate of 25.1%. The SRS method, which properly surveyed 209 children, gave a rate of 24.4%. The survey results show that the majority of the children were not vaccinated according to the national immunization schedule. Failure to follow the schedule also was independent of coverage levels. Individual clusters can be correctly classified according to their immunization level using Lot Quality Assurance methods.^ieng

Hepatitis B surface antigen (HBsAg) subtypes in Uganda. A preliminary report.

Hepatitis B surface antigen positive sera for 34 African residents in Uganda were analysed for antigenic determinants and ethnic groups. Twenty four of the subjects belonged to Bantu ethnic groups and the rest to non-Bantu ethnic groups. Nineteen of the Bantu had antigenic subtype adw and only one of the non-Bantu had subtype adw. The association between subtypes and ethnic groups was found to be statistically significant (p less than 0.01). The results in this preliminary report agree with reports from West Africa of a high preponderance of hepatitis B surface antigenic subtype ayw among the non-Bantu West Africans.

Childhood Hodgkin's disease in Uganda: a ten year experience.

Between 1967 and 1977, 48 patients with Hodgkin's disease under 16-years-old were treated with MOPP chemotherapy alone at the Uganda Cancer Institute because radiotherapy facilities are not available. Thirty-eight percent had early stage disease (stages I-IIIA). Prolonged first remissions were achieved in 74% of 42 complete responders. Of 11 patients who relapsed, 5 had prolonged second remissions induced by MOPP. Three patients were lost to follow-up and 15 of the remaining 45 died: 12 of these from progressive Hodgkin's disease, 2 from unrelated causes and 1 from Burkitt's lymphoma after 4 months remission from Hodgkin's disease. Acturial survival for all patients is 67% (75% for stages I-IIIA and 60% for stages IIIB-IV). Treatment complications included Herpes zoster and gynaecomastia. The latter is probably related to gonadal dysfunction. All stages of childhood Hodgkin's disease can be successfully managed with MOPP chemotherapy alone.

Multiple infections in cases of cervical cancer from a high-incidence area in tropical Africa.

The presence of several infections was determined in tissue and serum samples from 34 cases and 23 controls seen in 1984-85 at Mulago Hospital in Kampala, Uganda. When assessing single infections, association with cervical cancer could be shown for 5 agents, namely by Southern blot assay for human papillomavirus types 16 and 18 (HPV), and by serological tests at varying levels of antibody titres, for herpes simplex virus type I and/or 2 (HSV), cytomegalovirus (CMV), Epstein-Barr virus, viral capsid antigen (EBV-VCA), and Chlamydia trachomatis (CLT). Due to interaction, HSV and CMV were associated with cervical cancer only when infection by both of these agents was demonstrable. In the assessment of the simultaneous presence of these 5 infections, moderately high antibody titres were taken as the cut-off point for infection by HSV, CMV, EBV-VCA, and CLT. This showed that 3 and 4 infections at a time were seen in the majority of the cases in contrast to the controls with essentially no more than 2 such infections. A linear trend in the rise of risk for cervical cancer was noted with increasing number of infections.

Long-term experience with Burkitt's lymphoma in Uganda.

The cumulative results and long-term follow-up of all patients with Burkitt's lymphoma treated at the Uganda cancer Institute Kampala are reported. The annual admission rate is 29. The tumor patients commonly present with jaw swelling (72%), abdominal swelling (56%) and central nervous system involvement (30%). Complete response rate is achieved in a high proportion of patients (81%). About 50% of these relapse, equal numbers relapsing before and after 3 months. The most important factor influencing remission duration and survival is disease stage. Other important factors are treatment protocols and, to a lesser extent, the type of relapse. Central nervous system relapse does not necessarily augur poor prognosis as second remissions and long-term survival can be achieved with appropriate therapy. Presently 25% of all treated patients have survived free of disease well beyond 5 years.

Cerebrospinal irradiation of Burkitt's lymphoma. Failure in preventing central nervous system relapse.

Twenty-two patients with Burkitt's lymphoma in complete remission induced by either cyclophosphamide or a combination of cyclophosphamide, oncovin and methotrexate were randomized to receive or not to receive prophylactic cerebrospinal irradiation. Six of 11 irradiated patients relapsed with tumour of the central nervous system as compared to 4 of 11 controls. Relapse frequency appeared to be related to stage of disease on admission. It is concluded that irradiation does not prevent relapse.

Sample size determination in health studies : A practical manual / Sample size determination in health studies : A practical manual

Well designed studies are vital to provide information for the efficient planning, operation, monitoring and evaluation of health services. For any such study--whether of the efficacy of an immunization programme or of the availability of maternity care--the decision on how large a sample to select from the population in question must take into account the need both to obtain statistically valid results and to avoid unnecessary expenditure of time and resources

This manual has been prepared to provide guidance for health workers and managers responsible for making such decisions, and in particular for those undertaking studies at local or distric level without detailed knowledge of statitical methodology. It presents a variety of situations in which minimum sample size must be determined, including studies to estimate population proportion, odds ratio, relative risk and disease incidence. The illustrative examples of health studies are accompanied by over fifty tables that enable the reader to determine the sample size required without recourse to complicate calculations

The manual is not intended to help the reader decide what type of study, confidence level or degree of precision is most appropriate, nor does it discuss the theoretical basis of sample size determination. It is designed to be used in "cookbook" fashion as a prctical guide to making decisions on sample size once a proposed study and its objetives have been clearly defined (AU)

Determinación del tamaño de las muestras en los estudios sanitarios : Manual práctico / Determinación del tamaño de las muestras en los estudios sanitarios : Manual práctico / Sample size determination in health studies : A practical manual

La realización de estudios bien concebidos es requisito indispensable de la obtención de datos para planificar, operar, controlar y evaluar eficazmente los servicios de salud. En cualquiera de esos estudios, sean sobre la eficacia de un programa de inmunización, sean sobre la disponibilidad de atención materna, la decisón sobre el tamaño de la meustra de población analizada debe responder a la doble necesidad de obtener resultados estadísticamente válidos y de evitar un gasto excesivo de tiempo y de recursos

Este manual se ha preparado para dar orientación a los agentes de salud y los administradores que deben adoptar esas decisiones, en particular los que emprendan estudios a nivel local o de distrito sin conocer bien la metodología estadística. En el manual se exponer diversas situaciones en las que debe determinarse el tamaño de la muestra necesaria sin hacer complicados cálculos

El manual no tiene por objeto ayudar al lector a decidir qué tipo de estudio, nivel de confianza o grado de precisión es más adecuado ni tampoco analiza la basé teórica de la determinación del tamaño de las muestras. Semejante a un "libro de recetas", es un prontuario para decidir sobre el tamaño de la muestra una vez puntualizados el estudio y sus objetivos