This thesis aimed to study the actual and perceived quality of midwifery practices during childbirth at peripheral and central health care levels in two low-income countries, Angola (I-III) and Mozambique (IV-V). Theoretical models interpreting women's and midwives' views have been developed.

Objectives: Study I evaluates midwives' use of an adapted model of the World Health Organization's partograph, a tool used to monitor the progress of labour. Study II describes midwives experiences of working without immediate medical assistance. Study III explores women's perceptions of care-seeking behaviour during childbirth. Study IV observes perinatal midwifery care routines and examine partograph documentation. Study V explores and develops a theoretical understanding of factors perceived to obstruct or facilitate midwives ability to provide quality of perinatal care.

Methods: Study I: A one-group pre-and post-test interventional evaluation of 100 partographs from one peripheral delivery unit. Study II: Semi structured interviews with eleven midwives, analysed in a qualitative process comprising six steps. Study III. Ten focus group discussions with pregnant and non-pregnant women, analysed using the grounded theory technique. Study IV. Pre-and post intervention observation of midwifery care of 702 vs. 616 women during delivery and examination of the partographs. Study V: In-depth interviews with 16 midwives, analysed using grounded theory technique. Educational interventions were designed and applied in study II and IV.

Results: Study I. Significant improvement of documenting was found in seven of the ten variables and more partographs were correctly documented in sample II compared to sample I. Missed transfers increased, however, in sample II. Study II: The midwives experiences were sorted under four main areas: (1) Society/culture (2) Significant Others (3) Personal Self (4) Professional Self. Confidence was felt in the role as autonomous midwives but dependency on various factors such as the partograph, a functional referral system, peer support, community trust and continuous supervision was emphasised. Socio-economic hardships were identified as major stress factors for themselves and the women. Study III: Women seemed compelled to "mould" their care seeking behaviour and four patterns, two 'avoiding' and two 'approaching' institutional care were identified. The salient features of each pattern were found to be "personal courage", [B1]"disempowerment", "discarding traditional practices" and "awareness and emancipation". Study IV. No improvements were found in quality of care following the intervention. Common problems proved to be hypothermia and rare initiation of the graphic part of the partograph, which monitors progress of labour. Study V. A process labelled "changing perinatal care management" emerged, which comprised four dimensions addressing aspects related to i) existing environment ii) midwives' interaction with women in labour, iii) midwifery profession and iv) caring technology in order to improve quality of care. Communication and collaboration were identified as change agents.

Conclusions: The findings in this thesis indicate that midwives' ability to provide quality of maternal and perinatal care in lowincome and post-war affected countries is restricted by organizational, structural, educational as well as attitudinal aspects. Women in need of assistance during childbirth are negatively affected as a consequence of midwives reaction to the various constraints, which at times (Luanda) seems to oblige adverse care seeking behaviour. Midwives recognises the need for change, but change is found to be a slow and complex process, which requires engagement by all levels of the care chain. A model suggesting how to achieve quality of maternal and perinatal care in Safe Motherhood context is presented.