<p>This thesis assesses some indicators of quality for maternity care in Tanzania, using antenatal management of anaemia and hypertension and emergency obstetric care as focal points. The care of pregnant women consecutively enrolled in antenatal care (n=379) was observed and compared with quality standard criteria. From a tertiary level labour ward 741 cases of eclampsia were identified and their antenatal care analyzed. A health systems analysis was performed for 205 cases of pregnancy complications at district level.</p><p>There was inadequate equipment and drugs, inadequate staff knowledge and motivation, and incorrect measurements for investigating anaemia and hypertension in pregnancy. Hospital incidence of eclampsia at tertiary level was 200/10,000 live births, and was not modified by antenatal care. The quality observed in the antenatal programme indicated little impact on either anaemia or hypertensive complications. Compliance with obstetric referral was only 46% and all four observed maternal deaths occurred due to transport problems. The proposed process indicators for essential obstetric care were inadequate to assess the quality of care on a district level. </p><p>There is a need to address structural weaknesses, to motivate health workers and to improve training on quality improvement. More research is warranted on indicators for obstetric needs, accessibility and referral system.</p>

This thesis assesses some indicators of quality for maternity care in Tanzania, using antenatal management of anaemia and hypertension and emergency obstetric care as focal points. The care of pregnant women consecutively enrolled in antenatal care (n=379) was observed and compared with quality standard criteria. From a tertiary level labour ward 741 cases of eclampsia were identified and their antenatal care analyzed. A health systems analysis was performed for 205 cases of pregnancy complications at district level. There was inadequate equipment and drugs, inadequate staff knowledge and motivation, and incorrect measurements for investigating anaemia and hypertension in pregnancy. Hospital incidence of eclampsia at tertiary level was 200/10,000 live births, and was not modified by antenatal care. The quality observed in the antenatal programme indicated little impact on either anaemia or hypertensive complications. Compliance with obstetric referral was only 46% and all four observed maternal deaths occurred due to transport problems. The proposed process indicators for essential obstetric care were inadequate to assess the quality of care on a district level. There is a need to address structural weaknesses, to motivate health workers and to improve training on quality improvement. More research is warranted on indicators for obstetric needs, accessibility and referral system.

Introduction Chaque année, 289 000 femmes décèdent des complications reliées à la grossesse et à l’accouchement, et 2.9 millions de nouveau-nés décèdent avant d’atteindre 28 jours de vie. La quasi-totalité (99%) des décès maternels et néonataux ont cours dans les pays à revenu faible et intermédiaire (PRFI). L’utilisation des services obstétricaux essentiels, incluant l’assistance qualifiée à l’accouchement (AA) et les services postnataux, contribue largement à la réduction de la morbidité et de la mortalité maternelle et néonatale. Il est donc essentiel d’évaluer les déterminants et les inégalités de couverture de ces services, en vue d’informer l’élaboration de politiques et de programmes de santé dans les PRFI. Objectifs 1. Étudier systématiquement les déterminants et inégalités socioéconomiques, géographiques et démographiques dans l’utilisation des services de santé postnataux dans les PRFI. 2. Évaluer l’effet de la politique de subvention des frais aux usagers introduite au Burkina Faso en 2007 sur les taux d’utilisation de l’assistance qualifiée à l’accouchement, en fonction du statut socioéconomique (SSE). Méthodes 1. Nous avons réalisé une revue systématique sur l’utilisation des services postnataux dans les PRFI, en fonction des déterminants socioéconomiques, géographiques et démographiques. Notre étude incluait une méta-analyse de l’utilisation des services selon les quintiles de SSE et le milieu de vie (urbain vs. rural). 2. Nous avons utilisé un devis quasi-expérimental. Les sources de données consistaient en deux sondages représentatifs (n=1408 et n=1403), conduits respectivement en 2008 et 2010 auprès de femmes des districts sanitaires de Houndé et de Ziniaré au Burkina Faso, en plus d’une enquête sur la qualité structurelle des soins offerts dans les centres de santé primaire. Nous avons utilisé des modèles de régression de Poisson, multi-niveaux et segmentés, afin d’évaluer l’effet de la politique de subvention sur les taux d’AA. Nous avons estimé des ratios et différences de taux d’incidence ajustés, en fonction du SSE et du temps écoulé depuis l’introduction de la subvention. Résultats 1. Les estimés de ratio de cotes (RC) agrégés (IC 95%) pour les femmes de SSE élevé (5e quintile ou Q5), Q4, Q3 et Q2 (référence : quintile le plus pauvre, Q1) étaient respectivement : 2.27 (1.75 – 2.93); 1.60 (1.30-1.98); 1.32 (1.12-1.55); et 1.14 (0.96-1.34). La méta-analyse a aussi démontré un gradient d’utilisation des services postnataux entre les femmes urbaines et rurales : RC (IC 95%) = 1.36 (1.01-1.81). L’évaluation narrative a par ailleurs identifié une différence dans la couverture de services selon le niveau d’éducation. 2. Pour les femmes de faible SSE, le taux d’AA était 24% plus élevé (IC 95% : 4-46%) immédiatement après l’introduction de la subvention, en comparaison au taux attendu en l’absence de ladite subvention. L’ampleur de l’effet a diminué dans le temps, correspondant à des estimés (IC 95%) de 22% (3-45%) à 6 mois, 20% (1-43%) à 12 mois, et 17% (-4-42%) à 24 mois après l’introduction de la subvention. La force d’association variait selon les strates de SSE, l’effet le plus prononcé étant observé au sein du SSE le plus faible. Conclusions 1. L’utilisation des services postnataux demeure inéquitable selon le SSE et l’accessibilité géographique aux formations sanitaires dans les PRFI. 2. Notre étude suggère que l’introduction de la subvention des frais aux usagers au Burkina Faso résulte en une augmentation soutenue dans le taux d’assistance qualifiée à l’accouchement, particulièrement chez les femmes de faible SSE. Cette évidence scientifique devrait alimenter l’élaboration de programmes de santé materno-infantile, en plus de guider la planification de politiques et le renforcement des systèmes de santé des PRFI. / Background Each year, 289 000 women die from complications related to pregnancy, childbirth or the postnatal period, and 2.9 million newborns decease before reaching 28 days of life. The near totality (99%) of maternal and neonatal deaths occur in low- and middle-income countries (LMICs). Utilization of essential obstetric care services including skilled birth attendance (SBA) and postnatal care (PNC) largely contributes to the reduction of maternal and neonatal morbidity and mortality. There is a strong need to assess the determinants and inequalities in coverage of SBA and PNC services, to inform health policy planning. Objectives 1. Systematically assess the socioeconomic, geographic and demographic inequalities in PNC services utilization in LMICs. 2. Evaluate the effect of Burkina Faso’s 2007 user-fee subsidy policy on SBA rate across socioeconomic status (SES) strata. Methods 1. We conducted a systematic review of the association between PNC services utilization and key determinants, including a meta-analysis of PNC use across socioeconomic status quintiles, and place of residence (urban vs. rural). 2. We used a quasi-experimental design. The data sources were two representative surveys (n=1408 and n=1403) carried out in 2008 and 2010, respectively, of women from Houndé and Ziniaré health districts of Burkina Faso, and a survey of health centres assessing structural quality of care. Multilevel segmented Poisson regression models were used to assess the effect of subsidy on SBA rate. We estimated adjusted rate ratios and rate differences as a function of time and socioeconomic status level. Results 1. The pooled odds ratio (OR) (95% CI) estimates for highest SES women (quintile 5, Q5), Q4, Q3 and Q2 (reference: poorest quintile, Q1) were respectively: 2.27 (1.75 – 2.93); 1.60 (1.30-1.98); 1.32 (1.12-1.55); and 1.14 (0.96-1.34). Meta-analysis also showed a PNC utilization divide between urban and rural women: OR (95% CI) = 1.36 (1.01-1.81). Narrative assessment of studies identified a gradient in PNC coverage across education levels. 2. For low-SES women, immediately upon the introduction of the subsidy policy, the rate of SBA was 24% higher (95% CI: 4-46%) than expected in the absence of subsidy policy introduction. The magnitude of the apparent effect decreased over time, with the corresponding estimates (95% CI) being 22% (3-45%) at 6 months, 20% (1-43%) at 12 months, and 17% (-4-42%) at 24 months after the policy introduction. Furthermore, the magnitude of the association varied across SES strata, with the apparent effect being most pronounced in the low SES stratum. Conclusions 1. PNC utilization remains inequitable across socioeconomic status and geographic access to health facilities in LMICs. 2. Our study suggests that introduction of user fee subsidy in Burkina Faso resulted in sustained increase in the rate of SBA, especially among low-SES women. This evidence should inform maternal and child health programmes and guide health policies and health care systems in LMICs.