In cluster randomized trials (CRTs), groups are randomized to treatment arms rather than individuals while the outcome is assessed on the individuals within each cluster. Individuals within clusters tend to be more similar than in a randomly selected sample, which poses issues with dependence, which may lead to underestimated standard errors if ignored. To adjust for the correlation between individuals within clusters, two main approaches are used to analyze CRTs: cluster-level and individual-level analysis. In a cluster-level analysis summary measures are obtained for each cluster and then the two sets of cluster-specific measures are compared, such as with a t-test of the cluster means. A mixed model which takes into account cluster membership is an example of an individual-level analysis. We used a simulation study to quantify and compare power and bias of these two methods. We further take into account the effect of missing data. Complete datasets were generated and then data were deleted to simulate missing completely at random (MCAR) and missing at random (MAR) data. A balanced design, with two treatment groups and two time points was assumed. Cluster size, variance components (including within-subject, within-cluster and between-cluster variance) and proportion of missingness were varied to simulate common scenarios seen in practice. For each combination of parameters, 1,000 datasets were generated and analyzed. Results of our simulation study indicate that cluster-level analysis resulted in substantial loss of power when data were MAR. Individual-level analysis had higher power and remained unbiased, even with a small number of clusters.

Unique characteristics of cluster randomized trials (CRTs) complicate the interpretation of standard research ethics guidelines. Variable interpretation by research ethics committees may further complicate review and conduct. An international web-based survey was administered to corresponding authors of 300 randomly sampled CRT publications. We investigated ethics review and consent practices, investigator experiences with ethics review, and the perceived need for CRT-specific ethics guidelines. The response rate was 64%. Ethics review and consent were under-reported in publications. Ethics approval was obtained in 91%, and consent from individual and cluster level participants in 79% and 82% of trials. Consent varied by level of experimental intervention, data collection, and cluster size. Respondents cited variability among ethics committees (46%), and negative impacts of ethics review on their studies (38%). The majority perceived a need for ethics guidelines (73%), and guidance for ethics committees (70%). CRT-specific ethics guidelines are required to ensure practices meet ethical standards.

Unique characteristics of cluster randomized trials (CRTs) complicate the interpretation of standard research ethics guidelines. Variable interpretation by research ethics committees may further complicate review and conduct. An international web-based survey was administered to corresponding authors of 300 randomly sampled CRT publications. We investigated ethics review and consent practices, investigator experiences with ethics review, and the perceived need for CRT-specific ethics guidelines. The response rate was 64%. Ethics review and consent were under-reported in publications. Ethics approval was obtained in 91%, and consent from individual and cluster level participants in 79% and 82% of trials. Consent varied by level of experimental intervention, data collection, and cluster size. Respondents cited variability among ethics committees (46%), and negative impacts of ethics review on their studies (38%). The majority perceived a need for ethics guidelines (73%), and guidance for ethics committees (70%). CRT-specific ethics guidelines are required to ensure practices meet ethical standards.

Unique characteristics of cluster randomized trials (CRTs) complicate the interpretation of standard research ethics guidelines. Variable interpretation by research ethics committees may further complicate review and conduct. An international web-based survey was administered to corresponding authors of 300 randomly sampled CRT publications. We investigated ethics review and consent practices, investigator experiences with ethics review, and the perceived need for CRT-specific ethics guidelines. The response rate was 64%. Ethics review and consent were under-reported in publications. Ethics approval was obtained in 91%, and consent from individual and cluster level participants in 79% and 82% of trials. Consent varied by level of experimental intervention, data collection, and cluster size. Respondents cited variability among ethics committees (46%), and negative impacts of ethics review on their studies (38%). The majority perceived a need for ethics guidelines (73%), and guidance for ethics committees (70%). CRT-specific ethics guidelines are required to ensure practices meet ethical standards.

Les taux de césarienne sont en constante croissance dans le monde. Ces dernières années, on assiste à une augmentation sans cesse des taux hospitaliers d’accouchement par césarienne dans les pays à faible ressource malgré les recommandations de l’Organisation mondiale de la santé de ne pas dépasser 10 à 15 %. En Afrique au sud du Sahara, en particulier au Sénégal et au Mali, des politiques de subvention de la césarienne ont été introduites à l’échelle nationale à partir de 2005. Ces mesures ont contribué à l’augmentation de l’utilisation des services de maternité et des taux de césarienne. Si l’accessibilité financière à la césarienne a été améliorée substantiellement, la qualité n’a pas toujours suivi. Or, une augmentation trop importante des taux de césarienne peut avoir des effets négatifs sur la santé maternelle et périnatale. Par exemple, l’augmentation des taux institutionnels de césarienne au dessus de 10% en Amérique latine était associée à une augmentation de la mortalité maternelle et périnatale hospitalière. Trois ans après la mise en œuvre des politiques d’exemption des césariennes, le Programme Gesta international (PGI) a été mis en œuvre pour améliorer la qualité des soins obstétricaux dans 23 hôpitaux de référence au Sénégal et au Mali. Ce programme se basait sur l’audit clinque et la formation médicale continue du personnel sur les pratiques optimales en matière de soins intrapartum. Un essai contrôlé randomisé en grappe (essai QUARITE) a été mise en œuvre en 2007-2011 pour tester l’effet du PGI sur la mortalité maternelle hospitalière au Sénégal et au Mali. Initialement prévu pour améliorer la qualité des soins intra-partum, je me suis posé la question de l’efficacité de ce programme sur la pratique et les résultats de la césarienne. Les résultats de notre étude montrent que le PGI a permis de réduire l’évolution des taux de césarienne institutionnels dans les hôpitaux du groupe d’intervention comparativement à l’évolution dans le groupe contrôle. Par ailleurs, nous avons trouvé que le PGI a été plus efficace, en terme de réduction de la mortalité maternelle, parmi les femmes césarisées que parmi celles qui ont accouché par voie vaginale. Nous avons donc recommandé que des programmes d’amélioration de la qualité des soins soient mis en œuvre pour accompagner les politiques de subvention en cours dans la plupart des pays en Afrique au sud du Sahara et limiter ainsi l’utilisation excessive des césariennes dans ces pays. / Cesarean rates are rising steadily worldwide. In recent years, there has been an increasing cesarean rates in low-resource countries despite the World Health Organization recommended to not exceed 10-15%. In Senegal and Mali free cesarean policies were implemented nationally since 2005 and have contributed to increase the access to cesarean section. Access to cesarean deliveries has been improved substantially but quality of care has not always followed. However, excessive increase in cesarean section rates can have negative impacts on maternal and perinatal health. In Latin America, Asia, and Africa, several studies have shown an intrinsic risk of maternal and neonatal mortality associated with cesareans regardless of the initial health status of the mother or fetus. For example, the increase in hospital-based cesarean rates above 10% in Latin America was associated with an increase risk of maternal and perinatal mortality.Three years after the implementation of cesarean sections free policies, The Advances in Labour and Risk Management (ALARM) international program was implemented to improve the quality of obstetric care in 23 referral hospitals in Senegal and Mali. This program was based on maternal death review and staff training on best practices for intrapartum care. A randomized controlled cluster trial (QUARITE trial) was implemented in 2007-2011 to assess the effectiveness of the ALARM international program on in-hospital maternal mortality in Senegal and Mali. Initially planned to improve quality of Emergency Obstetric and Neonatal Care (EmONC), we assumed that this program was also effective on the quality of cesarean delivery.Results showed that the ALARM international program slowed down the trends of hospital-based cesarean rates in the 23 participating centers of the intervention group compared to the changes observed in the control group. Furthermore, we found that the program was more effective on maternal mortality among women who delivered by cesarean section than among women who delivered vaginally. We recommend that quality improvement strategies should support free cesarean policies to limit the excessive use of cesarean delivery.

<p><strong>Background:</strong> Syncope is estimated to account for 1% to 3% of emergency department (ED) annual visits in North America. Although most potential causes of syncope are benign and self-limited, others are associated with serious morbidity and substantial mortality. Recent efforts have focused on prospective identification of ED patients with syncope who are at high risk for early serious adverse outcomes in an attempt to hospitalize them at their first visit to the ED.</p> <p><strong>Objective: </strong>The purpose of this thesis is to describe the methodological issues related to the design of a study to determine whether the Revised Strategy of Syncope Diagnosis in the Emergency Room at the General Hospital Structured Care Pathway (RESASTER-SCP) is superior to usual care in identifying patients at low risk for serious adverse outcomes presenting to the ED who can be safely discharged home. <strong></strong></p> <p><strong>Design and Methods: </strong>A cluster randomized trial will be conducted with EDs (16 teaching and 46 non-teaching general hospitals) as the unit of randomization and patients presenting with syncope (TLOC) as the unit of analysis. Study participants will be followed at 1, 3, 5, and 12 months after the intervention (RESASTER-SCP vs. usual care) has been applied in the ED. Intention to treat analysis will be used. The analysis will be conducted at the individual level using proportions. Alpha level will be set at 0.05 with a power of 0.80 for the primary outcome.<strong></strong></p> <p><strong>Conclusion: </strong>This thesis describes some of the methodological issues concerning the design of a cluster randomized trial to determine whether or not RESASTER-SCP is superior to usual care in identifying patients presenting with syncope to the ED who can be safely discharged home.</p> / Master of Science (MSc)

Contexte : Etant donné la forte progression de la prévalence du surpoids et de l'obésité à l'adolescence ces dernières décennies, la prévention est devenue une priorité de santé publique internationale. Objectif : L'essai PRALIMAP (PRomotion de l'ALIMentation et de l'Activité Physique) évalue l'impact de trois stratégies de prévention du surpoids et de l'obésité - « Education », « Environnement », « Dépistage » - chez les adolescents en milieu scolaire, sur la corpulence, les connaissances, attitudes et comportements des adolescents vis-à-vis de la nutrition (alimentation et activité physique). Méthode : Les trois stratégies ont été affectées à 24 lycées de la région Lorraine (Nord-est de la France) sur 2 années (en classe de seconde et de première) par randomisation en grappe (=lycée), selon un plan factoriel 2x2x2. La stratégie Éducation consistait en des cours et des travaux de groupe sur la nutrition et une fête nutritionnelle annuelle. La stratégie Environnement consistait à améliorer l'offre nutritionnelle des lycées et organiser une fête nutritionnelle annuelle. La stratégie Dépistage consistait en un repérage des adolescents en surpoids ou obèses suivi d'une prise en charge adaptée collective. Les données ont été recueillies à 3 reprises : à l'entrée en classe de seconde (T0), de première (T1) et de terminale (T2). L'indice de masse corporelle (IMC), son z-score et la proportion d'adolescents en surpoids ou obèses étaient les critères de jugement principaux. Les critères de jugement secondaires étaient les connaissances et comportements nutritionnels. Les différences d'évolution T0-T2 des critères de jugement entre les 12 lycées ayant bénéficié d'une stratégie et les 12 lycées témoins de cette stratégie a été analysée à l'aide de modèles mixtes. Une évaluation du processus a permis d'estimer la dose réelle d'intervention de chaque stratégie dans chacun des lycées. Résultats : La stratégie Dépistage a entrainé une amélioration de la corpulence : augmentation moins importante de l'IMC avec la stratégie Dépistage (+0,6 vs +0,7, p=0,0303), diminution plus importante du z-score de l'IMC (-0,9 vs -0,5, p=0,0173) et de la prévalence du surpoids et de l'obésité (-2,3% vs -0,6%, p=0,0386). La stratégie Education a entrainé une amélioration de certains comportements nutritionnels : augmentation plus importante de la proportion d'adolescent suivant la recommandation nutritionnelle sur les féculents (3,6% vs -0,7%, p=0,0357) et du nombre d'activités physiques pratiquées (+0,02 vs -0,10, p=0,0047). La stratégie Environnement a amélioré les comportements et les connaissances nutritionnels : diminution moins importante de la proportion d'adolescents suivant la recommandation nutritionnelle sur le nombre de repas hebdomadaire (-4,5% vs -8,5%, p=0,0101) et amélioration plus importante du score de connaissances nutritionnelles (+1,9 point vs +1,0 point, p=0,0094). Des différences dans la réalisation des activités et la participation ont été mises en évidence suite à l'estimation de la dose d'intervention et peuvent expliquer certains résultats observés. Conclusion : Une stratégie de dépistage structurée en milieu scolaire apparaît bénéfique sur la réduction du surpoids et de l'obésité. L'addition d'activités d'éducation nutritionnelles dans le cursus n'induit pas d'effets sur la corpulence à court terme. La modification de l'environnement nutritionnel scolaire semble avoir un effet sur les indicateurs intermédiaires de connaissances et comportements nutritionnels / Background: Given the increasing prevalence of youth overweight and obesity in the last decade, prevention as become an international public health priority.Objective: The aim of The PRALIMAP (PRomotion de l'ALIMentation et de l'Activité Physique) trial was to evaluate the 2-year effectiveness of three strategies - « Education », « Screening » et « Environment »- aimed at preventing overweight and obesity among adolescents in high school setting on body size and nutritional knowledge and behaviours. Method: PRALIMAP was a school-based randomized controlled trial beginning in 24 state-run high schools (clusters) in Lorraine (north-eastern France). Each study high school was assigned to receive or not, over a 2-year period (grades 10 and 11), each of the three prevention strategies according to a 2x2x2 factorial school randomization. The prevention strategies were: ?education? (development of nutritional knowledge and skills), "environment" (creation of favourable environment by improving availability of dietary items with a good nutritional quality and physical activity), and "screening" (detection of overweight and obesity and, if necessary, adapted care management). The follow-up consisted of three visits: at the entry of grade 10(T0), grade 11(T1) and grade 12(T2). Body mass index (BMI), BMI z-score and prevalence of overweight and obesity were the main outcomes measures. Secondary outcomes measures were nutritional knowledge and behaviours. Comparisons of changes (T2-T0) of outcomes measures between each strategy schools and their control were carried out using a three-level hierarchical mixed model. A process evaluation allowed estimating an intervention dose really received by adolescents. Results: The 2-year change of anthropometric outcomes was more favourable in the 12 screening high schools as compared with the no-screening ones: a lower increase in BMI (+0,6 vs +0,7, p=0,0303), a greater decrease in BMI z-score (-0,9 vs -0,5, p=0,0173) and a greater decrease in overweight / obesity prevalence (-2,3% vs -0,6%, p=0,0386). The education strategy resulted in improved nutritional behaviours: a greater increase in achievement of starchy food guidelines (3,6% vs -0.7%, p=0.0357) and physical activity practice (+0,02 vs -0,10, p=0,0047). The environment strategy resulted in improved nutritional attitudes: a lower decrease in achievement of number of weekly meals guidelines (-4.5% vs -8.5%, p=0.0101) and greater increase in nutritional knowledge (+1,9 point vs +1,0 point, p=0.0094). Some differences in activity implementation and participation were highlighted and can explain some of the results observed. Conclusions: The screening strategy is an effective way to prevent, at two years, overweight and obesity among adolescents in a high school setting. Nutritional education added to the curriculum is not effective in the short term on body size. The school nutritional environment modification is slightly associated with improved nutritional knowledge and behaviours

<p>Promotion of exclusive breastfeeding (EBF), (giving breast milk only without any solids or liquids), has proved to be very challenging in the South African context, although this infant feeding practice has been found to protect babies against diarrhoea and respiratory tract infections and to carry a lower risk of HIV infection than mixed feeding (breastfeeding combined with formula or solids). Study design: The PROMISE-EBF study is a multi-country cluster randomised trial to examine peer support to promote exclusive breastfeeding in Africa. For the South African site in the PROMISE-EBF study, three sites, Paarl, Rietvlei and Umlazi, were selected because of their different geographic settings and each site operated as a separate stratum for cluster selection and randomisation purposes. The clusters were then randomised into intervention and control arms making a total of 17 clusters in each arm. HIV positive and negative women in the intervention arm received support on their choice of infant feeding from the peer supporters who visited them at their homes while the women in the control group only received the standard infant feeding counselling and support provided by health care&nbsp / workers at health facilities. Data collection: Mothers were interviewed at recruitment during the antepartum period to establish eligibility, obtain informed consent and data on socio-economic status. Home visits were scheduled for data collection by trained data collectors at 3, 6, 12 and 24 weeks after birth. Analysis of results: This mini-thesis was a secondary analysis of the PROMISE-EBF data focusing on the South African data only. The data was adjusted for clustering and analysed using SAS. Comparison of variables between the intervention and control groups within sites was done. Results: A significant difference, regarding counselling and infant feeding practices, was observed among all women who received peer support compared to those who received the standard antenatal counselling, with more women in the intervention group (20.5%) practising EBF than those in the control group (12.8%) by Week 3. When the women‟s HIV status was considered, more than 65% of HIV positive and 40% of HIV negative women practised MF and EFF (giving formula milk only with no breast milk) throughout the study, respectively, regardless of the group they were in. For women who hadintended to practise EBF at recruitment, 33% in the control group and 20% in the intervention group actually practised EBF by Week 3. Regarding disclosure and feeding choice, 77.4% of women who had disclosed their HIV status actually practised MF versus 8.6% who practised EBF by Week 3.Conclusion: Community peer counselling should be strengthened as the results from this study showed that a high percentage of women who practised EBF were those who had received counselling, irrespective of their HIV status. The high percentage of HIV positive women who practised high risk feeding, despite receiving infant counselling, is of concern. Disclosure of the women‟s HIV status did not translate to them practising low risk infant feeding methods, which may suggest that there are other issues that determine the women‟s choice of infant feeding.</p>

<p>Promotion of exclusive breastfeeding (EBF), (giving breast milk only without any solids or liquids), has proved to be very challenging in the South African context, although this infant feeding practice has been found to protect babies against diarrhoea and respiratory tract infections and to carry a lower risk of HIV infection than mixed feeding (breastfeeding combined with formula or solids). Study design: The PROMISE-EBF study is a multi-country cluster randomised trial to examine peer support to promote exclusive breastfeeding in Africa. For the South African site in the PROMISE-EBF study, three sites, Paarl, Rietvlei and Umlazi, were selected because of their different geographic settings and each site operated as a separate stratum for cluster selection and randomisation purposes. The clusters were then randomised into intervention and control arms making a total of 17 clusters in each arm. HIV positive and negative women in the intervention arm received support on their choice of infant feeding from the peer supporters who visited them at their homes while the women in the control group only received the standard infant feeding counselling and support provided by health care&nbsp / workers at health facilities. Data collection: Mothers were interviewed at recruitment during the antepartum period to establish eligibility, obtain informed consent and data on socio-economic status. Home visits were scheduled for data collection by trained data collectors at 3, 6, 12 and 24 weeks after birth. Analysis of results: This mini-thesis was a secondary analysis of the PROMISE-EBF data focusing on the South African data only. The data was adjusted for clustering and analysed using SAS. Comparison of variables between the intervention and control groups within sites was done. Results: A significant difference, regarding counselling and infant feeding practices, was observed among all women who received peer support compared to those who received the standard antenatal counselling, with more women in the intervention group (20.5%) practising EBF than those in the control group (12.8%) by Week 3. When the women‟s HIV status was considered, more than 65% of HIV positive and 40% of HIV negative women practised MF and EFF (giving formula milk only with no breast milk) throughout the study, respectively, regardless of the group they were in. For women who hadintended to practise EBF at recruitment, 33% in the control group and 20% in the intervention group actually practised EBF by Week 3. Regarding disclosure and feeding choice, 77.4% of women who had disclosed their HIV status actually practised MF versus 8.6% who practised EBF by Week 3.Conclusion: Community peer counselling should be strengthened as the results from this study showed that a high percentage of women who practised EBF were those who had received counselling, irrespective of their HIV status. The high percentage of HIV positive women who practised high risk feeding, despite receiving infant counselling, is of concern. Disclosure of the women‟s HIV status did not translate to them practising low risk infant feeding methods, which may suggest that there are other issues that determine the women‟s choice of infant feeding.</p>

Magister Public Health - MPH / Promotion of exclusive breastfeeding (EBF), (giving breast milk only without any solids or liquids), has proved to be very challenging in the South African context, although this infant feeding practice has been found to protect babies against diarrhoea and respiratory tract infections and to carry a lower risk of HIV infection than mixed feeding (breastfeeding combined with formula or solids). Study design: The PROMISE-EBF study is a multi-country cluster randomised trial to examine peer support to promote exclusive breastfeeding in Africa. For the South African site in the PROMISE-EBF study, three sites, Paarl, Rietvlei and Umlazi, were selected because of their different geographic settings and each site operated as a separate stratum for cluster selection and randomisation purposes. The clusters were then randomised into intervention and control arms making a total of 17 clusters in each arm. HIV positive and negative women in the intervention arm received support on their choice of infant feeding from the peer supporters who visited them at their homes while the women in the control group only received the standard infant feeding counselling and support provided by health care workers at health facilities. Data collection: Mothers were interviewed at recruitment during the antepartum period to establish eligibility, obtain informed consent and data on socio-economic status. Home visits were scheduled for data collection by trained data collectors at 3, 6, 12 and 24 weeks after birth. Analysis of results: This mini-thesis was a secondary analysis of the PROMISE-EBF data focusing on the South African data only. The data was adjusted for clustering and analysed using SAS. Comparison of variables between the intervention and control groups within sites was done. Results: A significant difference, regarding counselling and infant feeding practices, was observed among all women who received peer support compared to those who received the standard antenatal counselling, with more women in the intervention group (20.5%) practising EBF than those in the control group (12.8%) by Week 3. When the women's HIV status was considered, more than 65% of HIV positive and 40% of HIV negative women practised MF and EFF (giving formula milk only with no breast milk) throughout the study, respectively, regardless of the group they were in. For women who hadintended to practise EBF at recruitment, 33% in the control group and 20% in the intervention group actually practised EBF by Week 3. Regarding disclosure and feeding choice, 77.4% of women who had disclosed their HIV status actually practised MF versus 8.6% who practised EBF by Week 3.Conclusion: Community peer counselling should be strengthened as the results from this study showed that a high percentage of women who practised EBF were those who had received counselling, irrespective of their HIV status. The high percentage of HIV positive women who practised high risk feeding, despite receiving infant counselling, is of concern. Disclosure of the women's HIV status did not translate to them practising low risk infant feeding methods, which may suggest that there are other issues that determine the women's choice of infant feeding. / South Africa