This section has been addressed under the five case identification quandary lessons which had appeared worthy to share as part of the findings by the principal investigator.
All these experiences have been discussed under account of the case identification as presented underneath.
The need for examining maternal mortality and newborn death in Ethiopia is increasingly become a gleaming concern. This is largely because of (a) the unacceptably high maternal mortality and newborn death in the country; and, (b) the Ethiopian government's commitment to its health sector development partners, and/or global communities interest that pledge to reduce or halt the unfair and high maternal mortality and newborn deaths. Nonetheless, one may conclude that in Ethiopia, information about maternal mortality and newborn death is needed, not only for what it tells about but also for what it implies about the women's and her baby’s health, social and economic status in general. in effect, WHO and UNICEF described maternal mortality and newborn deaths as litmus test of the status of women and her baby, their access to health care and the adequacy of the health care system in responding adequately to their health care needs (UN 2010a:3-4; UN 2010b:13-25). The positive effects beginning to surface in maternal mortality reviews currently in operation in very few selected parts of the country has also added impetus to such thinking (MoH 2013:16-23).
Generally, maternal mortality and newborn death is a reportable incident in the country;
however, in reality lots of them occur neither recorded nor reported. Putting in place a system of routine reviews of maternal mortalities and newborn deaths in Ethiopia has to overcome certain practical challenges. During the course of fieldwork in the present study, some key bottlenecks and organisational issues were identified, and the lessons are presented, which may also poses a big challenge to any review of maternal mortalities and newborn deaths in Ethiopia. These includes:
4.2.1 Identification and reporting of case
It is generally believed that deaths that have occurred in health facilities are easy to identify and report. This may be true for maternal mortality and newborn deaths that occurred in a maternity ward and neonatal unit but not the case for those taking place in other units. Quite a lot of women and significant number of newborn babies die in the outpatient units of the selected hospitals. Some of these are maternal mortalities and newborn deaths, but go unrecorded or misclassified. During the course of this study, a lot of female deaths within reproductive age were recorded at the medical and other wards of the hospital. However, none of them could be classified or suspected as a maternal death because their pregnancy status was not known during the time of admission. Currently, in Ethiopia's health information system there is nowhere in the admission records where the pregnancy status of women admitted is stated. This is also true for the newborn cases assorted as children. Outpatient deaths observed to have occurred at the hospital during the study were generally not documented within the hospital recording system. In effect, none had case notes so consequently were nowhere in the hospital’s information system.
Likewise, identification and reporting of abortion related deaths are generally difficult particularly in places where abortion is top secret. In Ethiopia, such deaths stand a high chance of not being reported for various reasons including illegality. Furthermore, in a society such as Ethiopia where out of marriage pregnancies are socially unacceptable for religious and other social reasons such deaths will be hard if not impossible to identify. In the present study, no abortion related death was reported nor identified. This does not signify that such deaths did not happen in the study area. Considering the high fertility rate in the area, documented early sexual debut among young girls and related abortion cases coupled with low contraceptive utilisation (CSA 2011:59-80), theoretical there may be significant pregnancies which might have been aborted, and some of these are most likely unsafe that result in maternal death.
What's more, reporting of maternal mortalities and newborn deaths in the selected zones and regional states was not satisfactory even with the existence of professional skilled birth attendants in the health facilities, health extension workers (HEW) and/or traditional birth attendant (TBA) in the communities of the selected districts or zones and regional states. The situation seems similar in the other parts of the country, too.
Despite the size of maternal mortality and newborn death cases they may have, almost all hospitals appears to be willing to communicate only limited number of cases (usually about 10 maternal death cases) for a twelve month time. Interestingly, even this is possible only if the researcher have official government letter. In one of the hospitals, the principal investigator has actually come to a clear discrepancy of what is registered on log-book and the deceased case cards retrieved for review during the study period.
Following the introduction of maternal death surveillance and response (MDSR) a couple of months prior to this study by the Ethiopian government in a very few pilot-health-facilities, maternal mortality and newborn death appears to be highly sensitive and politicised issue. As it could be seen in this study, 118 out of the 133 maternal deaths identified were women permanently resident in the selected study sites but only 72 of them were actually reported.
4.2.2 Participants indecisiveness
Families or relatives of the deceased women or the newborn babies were generally open and willing to be interviewed on events surrounding the death only if they trust you and are confident that the information being generated will not be used to punish or warning anybody. A lot of deaths including maternal deaths occur in Ethiopian communities but almost none has been audited or deeply reviewed. Therefore, auditing maternal deaths now will naturally cause undue hesitancy; more precisely, if the interviewer is not familiar to interviewees. In most instances the interviewer may even be viewed as an intelligence agent. Ethiopians are socially interconnected, very interactive and generally people know each other. But they very much mind “who said what” and “who did what”. They generally don’t like to offend each other or to be known to hurt or implicate another person particularly the one you know. Paradoxically, the culture of silence is deep rooted among Ethiopians. People don’t like to be quoted especially on issues viewed as negative. During the course of data collection in the present study, there were instances in which families felt reluctant to be interviewed.
Health care providers also seem to be concerned by the investigative approach to deaths they may be involved in providing care or at least knew about the case. While at the hospital, trying to interview health care providers who actually provided care to a case, some of the staff pretends not to remember that particular case just to avoid the interview. During informal discussions with some of the staff it was realised that their
behaviour were understandable and justified. They feared being reprimanded by the hospital management for revealing what happened within the hospital as what happened in the hospital is generally believed to be kept confidential. They also feared being punished by the higher authorities at ministry level if deficiency in care was highlighted. In effect they were confronted with 'double fear'. This illustrates that in many instances, 'supervision' is not really an enabling activity, rather a control mechanism to ensure that things don’t go wrong.
4.2.3 The health records
Incomplete health data, missing records or lack of clarity in such records remains a big issue. Tracing and retrieving of health facility kept records were labour intensive and time consuming. In situations where they exist, most of the time they are chaotically stored. Generally, record keeping relatively appears better at the referral hospitals. At the peripheral hospitals, health centres and health posts, the situation was totally different. The recording system in these facilities was not uniform and very inconsistent, with key features of the recording (such as column headings) varying in different facilities and even at different times or with different staff in the same facility. Chronic lack of reliable stationery supply at peripheral health care facilities also makes matters worse. What contributed to the deplorable situation is the virtually non-existent supportive supervision in these health facilities. Past monthly returns may be available but in none of the health facilities were service statistics in the form of picture diagrams displayed. This implies that peripheral health staffs do not make simple analysis of the data collected to observe trends.
4.2.4 The data collector and the route to data collection
The data collector either male or female; health worker or a lay person does not matter.
What is essential is that the data collector must win the confidence of the informants, should exercise high degree of confidentiality and patience. The interviewer must be one who has detective skills, non-dominant and considerate. Above all he/she should be aware of local customs and traditions and be sensitive to them. On several occasions, families were visited for an interview but seem not prepared as the head of the house hold was not in. They are usually men, who must authorise family members to be interviewed. Some families refuse interviews of some specific days, linked with
their beliefs. In such situations a revisit was arranged and done just to show respect to local customs, beliefs and traditions. Revisits could be both time consuming and expensive.
The maternal mortality and newborn deaths examining process predominantly starts at what happened at the health facilities, where care was assumed to be sought; and, radiate to the community cases where bulk of deceased cases believed to dwell. It is generally important to assess both levels in a comprehensive way to have a holistic picture of the 'path-to-death'. This approach is, however, labour intensive and could be very expensive. It means spending longer time in examining a case because different places and often many people need to be interviewed. It may also require long distance walking often on dirt terrain roads. The poor road network and conditions in Ethiopia, worse in the rainy season, may make it more difficult. During data collection, sometimes the research team had to walk for over 25 kilometres just to follow-up a case.
Sometimes the key informant for that case is away in the farm field. Data collectors end up walking to meet him/her at the farms as the road was too muddy for a vehicle. On the other hand, these difficult access issues are also relevant in the reverse direction – when a client needs to seek health care urgently.
4.2.5 Standards of practice
Any form of health review or an audit is only effective if care provided is compared with explicit standards. This implies that service standards, protocols or guidelines must first be in place before instituting an audit or a review system. In Ethiopia, one of the biggest challenges is having effective maternal and newborn health care standards with which care will be compared. During the course of data collection, certain aspects of substandard care were noted but there were no written procedures in the maternal and newborn health services in Ethiopia addressing most of those issues. In such instances, internationally agreed standards were used to compare with current care practices. This may not be appropriate as standards and protocols should be adapted to meet local situations.
To conclude, more could have been said under this section - account of the case identification − which one may expect as lessons or challenges. But these five case
identification quandary lessons appear worthy to share as part of the findings by the principal investigator.
4.3 QUANTITATIVE RESEARCH FINDINGS, INTERPRETATIONS AND