Community health worker knowledge and management of pre-eclampsia in southern Mozambique

Background: Mozambique has drastically improved an array of health indicators in recent years, including maternal mortality rates which decreased 63 % from 1990-2013 but the rates still high. Pre-eclampsia and eclampsia constitute the third major cause of maternal death in the country. Women in rural areas, with limited access to health facilities are at greatest risk. This study aimed to assess the current state of knowledge and the regular practices regarding pre-eclampsia and eclampsia by community health workers in southern Mozambique. Methods: This mixed methods study was conducted from 2013 to 2014, in Maputo and Gaza Provinces, southern Mozambique. Self-administered questionnaires, in-depth interviews and focus group discussions were conducted with CHWs, district medical officers, community health workers’ supervisors, Gynaecologists-Obstetricians and matrons. Quantitative data were entered into a database written in REDCap and subsequently analyzed using Stata 13. Qualitative data was imported into NVivo10 for thematic analysis. Results: Ninety-three percent of CHW had some awareness of pregnancy complications. Forty-one percent were able to describe the signs and symptoms of hypertension. In cases of eclampsia, CHWs reported to immediately refer the women. The vast majority of the CHWs surveyed reported that they could neither measure blood pressure nor proteinuria (90 %). Fewer reported confidence in providing oral antihypertensives (14 %) or injections in pregnancy (5 %). The other community health care providers are matrons. They do not formally offer health services, but assists pregnant women in case of an emergency. Regarding pre-eclampsia and eclampsia, matrons were unable to recognise these biomedical terms. Conclusions: Although CHWs are aware of pregnancy complications, they hold limited knowledge specific to pre-eclampsia and eclampsia. There is a need to promote studies to evaluate the impact of enhancing their training to include additional content related to the identification and management of pre-eclampsia and eclampsia.

This is a component of multi-country national cluster randomized control trial (cRCT) implemented in Nigeria, Mozambique, Pakistan and India, the CLIP (Community Level Interventions for Pre-eclampsia) study ({“type”:”clinical-trial”,”attrs”:{“text”:”NCT01911494″,”term_id”:”NCT01911494″}}NCT01911494). In Mozambique the cRCT is implemented in two provinces, namely Gaza and Maputo, in the southern part of the country (Fig. 1). This study aims to evaluate a community based intervention consisting in measure of blood pressure and proteinuria, clinical management of severe pre-eclampsia with metildopa and magnesium sulphate, ability to timely refer women to the nearest health facility and thus contribute for the reduction of maternal morbidity and mortality due to pre-eclampsia and eclampsia. Map of the study area Southern Mozambique is geographically diverse, with coastal regions as well as large areas of landlocked agricultural land. Maputo Province includes the capital city, Matola, located 10Km west of Maputo city, the country’s capital. Maputo province has a population of 1.098 million. Gaza Province has a total population of 1.362 million [20]. In general, provinces are divided into districts, administrative posts (AP), localities and neighbourhoods. Each AP covers roughly 500–2,000 inhabitants. The APs included in this study area (Calanga, Maluana, Ilha Josina Machel, Três de Fevereiro, Magude, Messano, Macia, Xilembene, Chissano, Mazivila, Chongoene, Chicumbane and Chibuto) are largely impoverished rural areas where the predominant occupations are agriculture, livestock rearing, informal trading, migrant labour (mainly to South Africa), handicrafts, and work in private sugar and rice processing farms. Residents of these APs are mostly of the Changana ethnic group and speak a local dialect of the same name (for more details see Table 1). Study site characteristics Source: Unpublished data from demographic surveillance, health facility assessment (2014) and INE (2007) This study is based on a formative research exercise conducted in preparation to the CLIP trial. The formative research comprised a mixed methods design, a detailed description of these methods is presented elsewhere [21]. Data collections was conducted based on forms and guides, which were developed centrally by the study coordination team, used in the other countries where the study had been previously conducted (Nigeria, India and Pakistan), and adapted to the local context of Mozambique. Quantitative data were collected through self-administered questionnaires completed by CHWs. The qualitative data were obtained through focus group discussions with matrons, and in-depth interviews with CHWs supervisors, district chief medical officers and Gynaecologists-Obstetricians. While CHWs, CHW supervisors, and district medical officers from all study area were eligible to participate, matrons were drawn from selected AP, namely Ilha Josina Machel-Calanga, Três de Fevereiro, Messano and Chongoene. Data collection was complemented by a desk review of existing documents regarding involvement of CHWs in maternal and child health such as policies, guidelines, reports and manuals. Self-administered questionnaires targeted all active CHWs within the study. Recruitment was done through contacts with the health facility to which they are linked. Data collection was conducted either in the health facility where each CHW reports or at the house of the CHW. Focus groups were conducted with matrons. This group was selected because it is also considered as community based alternative point of care for pregnant women and it was important to explore their views and practices regarding pregnancy complications (including high blood pressure and convulsions), pregnancy management (antenatal care and treatment provision) to gain an understanding of the role of matrons in the context of expansion of the maternal health care at community level. As matrons are not formally linked to the health facility and are not formally organized as a group therefore there is no clear ways to systematically identify and track them, their total number in the study area is unknown. After being identified with the assistance of neighbourhood chiefs they were invited to participate in the study. Focus groups were conducted either at the círculos (the usual community gathering location), or at the community leaders’ house, as groups could easily be convened in these locations. In-depth interviews, which involved all CHW supervisors and district medical officers from the study area were conducted one-on one in the work place of the respondents. The entry points were the district level medical officers themselves who in turn identified the CHWs supervisors. As there are no Gynaecologists-Obstetricians in none of the selected districts, they were identified through the Associação Moçambicana de Ginecologia (AMOG) – The Mozambican Gynaecologists and Obstetricians Association. It was conducted to obtain information about existing CHWs and their distribution among study sites, their training profile and scope of work. A variety of documents were reviewed, and both published and unpublished information accounted for this exercise. Most of the published documents were downloaded from the Mozambique Government portal. These included formal policy documents and other official documents such as community involvement strategies, CHW training programmes, monitoring and evaluation manuals, and meetings’ minutes and reports. Data were collected between October 2013 and May 2014. Questionnaires were designed to obtain information concerning CHW preparedness, knowledge and reported skills to manage pregnant women and to perform home-based basic treatment for women with pre-eclampsia. For the purpose of this study we assessed the following warning signs: high blood pressure, hemorrhage and convulsions. The questionnaire included 33 items on a five-point Likert scale. This format was regarded as appropriate to assess CHWs knowledge, attitudes and practices of CHW and level of confidence regarding maternal health care provision, and compare findings not only among all CHWs’ within the study area, but also eventually across the countries involved in the study. In addition, one open-ended question for respondents’ comments or additional information was included. Depending on the number of CHWs per community, individual or collective briefing sessions were held to provide instructions on how to fill it, and when required further clarification was given in Changana. Five trained local social science research assistants were available for clarification when required. The questionnaire took on average 20 min to be completed by participants. Focus groups discussion were used to explore the views of matrons regarding pregnancy complications (including high blood pressure and convulsions), pregnancy management (antenatal care and treatment seeking), and existing health care delivery practices. Based on the FGD guide, earlier mentioned the same trained local social science research assistants facilitated the discussions, which took on average 60 min and were audio recorded. In-depth interviews were conducted with Gynaecologists-Obstetricians, district medical officers and CHW supervisors to allow further probing on pertinent issues, such as their opinions regarding CHW’ ability to identify warning signs in pregnancy, manage pregnancy complications, and their proficiency to administer medications. These interviews were conducted by two social scientists. Interviews lasted between 30–60 min and were conducted in the workplace of participants. Field notes and audio recordings were taken at the time of in-depth interviews. All data collection was led by a Mozambican social scientist assisted by 5 social science research assistants employed by CISM. These researchers were selected due to their familiarity with the local socio-cultural context, the research topic and their relevant qualitative and quantitative data collection expertise. Team members were fluent in Portuguese and Changana, included both male and female, and had no prior relationship with the participants. The data collection and analyse strategy was overseen by the study PI and co-PI. Information obtained through the desk review was systematized and summarized to extract relevant information regarding CHWs history, role and challenges with regards to maternal and child health care that was already part of the scope of work of CHWs. All data captured through questionnaires were sent to the Manhiça Health Research Center (CISM) for data entry and management using REDCap [22]. Double data entry was completed in all questionnaires. The presence of social science research assistants during the self-administration of the CHWs questionnaires helped to maximise the data integrity. Before it was sent to the data Center, the study team members made a revision of each questionnaire while in the field. The failures to validation rules and double data entry discrepancies were checked through queries that led to confrontation with the paper forms. Outliers and missing values were also checked. Data was then exported to Stata 13 (Stata Corp., College Station, Texas, USA) for further statistical analysis. The demographic characteristics of CHWs and the study variables of interest are presented using descriptive statistics (absolute and relative frequencies, ranges, averages and quartiles). The exact logistic regression based odds-ratio and its 95 % confidence intervals were used to describe the association between the self-reported ability of the CHWs to recognise warning signs in pregnancy with their demographic characteristics (age, sex, education and years of experience). Given the sample size no multivariate analysis was attempted. Focus group discussions and in-depth interviews were digitally recorded using Olympus AS-2400 PC® recorders. Together with the open-ended question from CHWs questionnaire they were transcribed verbatim by the same team members who conducted data collection. While in-depth interviews data and that from the open-ended question of the CHW self-administered questionnaire were collected and transcribed in Portuguese. The focus group discussions were held in Changana and translated to Portuguese while being transcribed. Quality control of transcripts was ensured by listening to audio recordings and comparing them against the transcripts to confirm accuracy. The qualitative data were analysed using NVivo version 10.0 (QSR International Pty. Ltd. 2012). A thematic analysis approach was taken (see Fig. 2). The coding structure (based on free nodes, branched nodes, attributes and some pre-determined queries) was developed in advance based on the study objectives through a collaborative discussion between researchers at CISM and University of British Columbia (UBC). Themes were subsequently adjusted and new themes added as they emerged from the data. As analysis was to be performed by two teams (CISM and UBC), the coding structured was in English. Theme structure The two Mozambican social scientists coded all transcriptions in Portuguese, by reading the text in Portuguese and labeling the concepts using the codes which were written in English. Three IDI and two FGD transcriptions were translated from Portuguese to English and coded by a social scientist based at UBC for three purposes: first, to support the discussions on the development of the coding structure; second, for the UBC collaborator to be familiar with the raw data, so as to assist interpretation; and finally, for quality control of the coding. To allow the two teams to work independently, the data was split into two Nvivo projects, but the same coding structure was used for both teams. Coding consensus meetings to discussing data analysis strategy and findings were held via Skype™. Coding agreement between the coders was very high. When the coding was completed the analysed data was merged into single project managed by the Mozambican team, form which the final queries were run. Ethical approval for this study was granted by the CISM Institutional Review Board in Mozambique (CIBS_CISM/08/2013), as well as by the UBC C&W Research Ethics Board in Canada (H12-00132). Written informed consent was sought from each participant before data collection. For the illiterate participants a literate witness was involved in the consent process whereby they were asked to read and explain to the participant the contents of the participant information sheet. The consent form was signed by the witness and the field worker, after the participant finger print was taken. All identifiable data of participants were codified through attribution of unique identification numbers or pseudonymous to guarantee anonymity. When needed the respondent was identified by stating the administrative post or the province in the illustrative quotes.

• boene, helena
• vidler, marianne
• augusto, orvalho
• sidat, mohsin
• macete, eusebio
• menéndez, clara
• sawchuck, diane
• qureshi, rahat
• von dadelszen, peter
• munguambe, khátia
• sevene, esperança