Determinants of Perinatal Mortality in Tercha General Hospital, Southern Ethiopia; Facility Based Case Control Study.

Background: Ethiopia meets the target millennium development goal 4 on child survival three years ahead of time. However, there were high perinatal deaths in the country and the reduction was not impressive. Identifying determinants and implement evidence based interventions is crucial to reduce perinatal death. However, there were no clear evidences on determinants of perinatal mortality in Tercha General Hospital. Objective: To assess determinants of perinatal mortality in Tercha general hospital, Southern Ethiopia, January 1, 2014 and December 30, 2017. Method: An unmatched case control study using secondary data as a source of information was conducted in Tercha general hospital. Cases were stillbirths and early neonatal deaths. Controls were those newborns live till 7th days. Randomly selected 366 (183 cases and 183 controls) study subjects were constituted for this study. The data were collected from March 1-20/2018. Epi-Data version 3.1 and SPSS Version 23 were used for data entry and analysis, respectively. Descriptive statistics were used to describe the study population in relation to study variables. Logistic regressions were employed to identify determinants of perinatal death. Results: In multivariable logistic analysis, rural in residence of the mother [AOR=1.82; 95%CI:(1.04-3.19)], ANC booking [AOR=0.47; 95%CI:(0.27,0.83)], prolonged labour [AOR=2.75; 95%CI: (1.58-4.78)], low birth weight [AOR=1.78; 95%CI (1.06-2.97)], presence of obstetrics complication [AOR=2.15; 95%CI: (1.28-3.62)], using partograph [AOR=0.5; 95%CI: 0.25-0.9]. Using safe child birth checklist [AOR=0.52; 95%CI: 0.30-0.91], and coming with referral [(AOR=2.69; 95% CI: (1.51-4.8)] were significantly associated with perinatal mortality. Conclusion and Recommendation: Being rural in residence,

elevated the risk of perinatal mortality, and antenatal care booking, using partograph and using safe childbirthchecklist were associated with reduced risk of perinatal mortality. We therefore, recommend strengthening maternal health and newborn care servicesby taking into account these factors to reduce perinatal death.

Introduction
Perinatal mortality, as defined by World Health Organization (WHO), is total number of deaths in the perinatal period. This includes still birth (death of fetus after 28 weeks of gestation or above 1000gram birth weight) and early neonatal death (END) i.e. death of live new-born before the age of 7 completed days (1,2). While the WHO, the International Stillbirth Alliance and some developed countries utilize 22 weeks as their age of viability and a loss at that gestational age reported as perinatal mortality (3). In countries like Ethiopia, viability age greater than 28 week is considered for the report (4).
Newborns are at most risk of dying in their first week of birth, and globally three-quarters of neonatal deaths occur in the first week of life (5). Stillbirths account for over half of all perinatal deaths. One third of stillbirths takes place during delivery, and is largely avoidable. Intra-partum deaths (i.e. those occurring during delivery) are closely linked to place of care at delivery (3,6).
Every year, nearly 7 million perinatal deaths occur across the globe (3.5 million stillbirths and 4 million neonatal deaths), which is higher than the combined annual all age level deaths due to Acquired Immune Deficiencies Syndrome(AIDS) (2.1 million), tuberculosis (1.6 million) and malaria (1.3 million) (3,7,8).
Perinatal mortality is at an unacceptably high level in low and middle income countries, about 99% of these perinatal deaths occur in this country (9). In addition to this, the true mortality rate is under estimated in low and middle income countries where vital registration is not available (10).In Sub Saharan Africa approximately 30 million women become pregnant in a year. Of those, about 1 million deliveries are still birth; at least 1million babies die in their first month of life and 0.5 million die on the first day of life (11).
Ethiopia like other sub-Saharan countries has a high perinatal mortality. According to WHO report, PMR of Ethiopia in 2004 was 128,000 total deaths (34,000 still births and 94,000 early neonatal deaths) (12). According to Ethiopian Demographic and Health survey (EDHS-2016) perinatal mortality rate was 33 deaths per 1,000 pregnancies (13,14). The overall perinatal mortality reported from ten hospital based studies in Ethiopia was in the range of 66 to 124 per 1000 births. The report of the large scale community based perinatal mortality was also in the range of 37 to 52 per 1000 births (7).
Studies conducted in Ethiopia showed that perinatal mortality was associated with maternal socio demographic status like maternal age, occupation, birth interval, occupation (13,15,16). Maternal obstetrics and medical factors history such as Antenatal Care (ANC), Human immune deficiency Virus (HIV) status, history of obstetrics complications, parity, mode of delivery and neonatal related factors like fetal presentation, birth weight, and sex were variables which associated with perinatal mortality (27-31).
Southern Nations Nationalities Peoples and Representative (SNNPR), region is one of the regional states in Ethiopia which had perinatal mortality rate of 26 deaths per 1,000 pregnancies in EDHS 2016 report (13). Dawro zone is one of the largest populated zones in the region, with low health services coverage, institutional delivery and antenatal coverage (19). A study conducted at Tercha General Hospital (TGH) showed, perinatal mortality among the group of mothers undergone a major obstetrics intervention (Cesarean section(C/S),laparotomy for repair of uterus,hysterectomy, and destructive operation) were 258 per 1,000 live births (20).
Identifying determinants and implement evidence based interventions is crucial to reduce perinatal death. However, up to the knowledge of principal investigator while searching different literatures, there was no study done to assess determinants of perinatal mortality in TGH. Therefore, this study helps to assess determinants of perinatal mortality among hospital deliveries that will help to improve all concerned bodies understanding on the factors associated to perinatal mortality and serves as an important tool for planning and resource allocation that aimed to improving newborns survival.

Study setting and Period
The study was conducted at Tercah General Hospital which is found Tarcha town, SNNPR.
Tercha town is located at south west of Ethiopia, 517Kms away from Addis Ababa and 285Kms away from Hawassa (the capital city of SNNPR). The Hospital have 120 beds.
There are 40 beds in the maternity ward, 4 labor beds and 3 delivery Koch. There is also neonatal intensive care unit with kangaroo mother care room (equipped with 4 beds). Population:Cases were r andomly selected perinatal deaths that were attended in TGH from January 1, 2014 and December 30, 2017 and controls were randomly selected live births that were attended in TGH and alive up to 7 days in the same year with cases.

Sample size determination and Sampling technique
The required sample size was calculated by the statistical program of Epi-info stat Calc tool by considering different factors strongly associated with perinatal mortality from previous study conducted in Addis Ababa:birth interval <2 years, congenital anomalies, C/S delivery, Hg level<11gm/dl and partograph use (12). Based on the assumption of case to control ratio of 1:1, 95% confidence level, Power of 80% percentage of controls exposed (Hemoglobin level<11gm/dl) 8% and odds ratio of 2.6, the total sample size for this study became 366 (183 case and 183 controls).All perinatal cases and controls identification number (card number) in between January 1, 2014 and December 30, 2017 were taken from delivery room and neonatal intensive care unit registration log book then the required sample size were selected by using simple random sampling technique with computer generated random number. During data collection, delivery registry books were reviewed and selection of cases and controls was done from the registration books then the cards of both the selected cases and controls were traced from the archive (card room) using card numbers found in the registration book. For stillbirths and live births maternal cards were reviewed; and for early neonates death neonates card were assessed, if missing data in the early neonates card happened, their maternal cards were trace from the card room and complete it.

Data quality Assurance
Prior to data collection careful modification of the data collection tool were made. The tool was pre-tested (using 5% of sample size = 18) at TGH by using the source population (November-December 2013). Data collectors and supervisors were trained for 3 days. The tool also commented by two Epidemiology experts. Daily supervision of the data collection procedure was made.

Data processing and analysis
Before data entry data were cheeked for completeness, then data were coded and entered into Epi-Data version 3.1 then data were exported to Statistical package for social science (SPSS) Version 23 for checking the missing values, outliers, and analysis. Descriptive analysis was made to describe the study populations in relation to study variables.
First, bivariate logistic regression was done to select candidate variables for multi variable logistic regression. All variables having P-value ≤0.25 during bivariate analysis were considered as candidates for the multi variable logistic regression. After the multi variable logistic regression analysis variables having p-values <0.05 was considered as having statistical significant association with prenatal mortality. Model fitness was checked by Hosmer & Lemeshow goodness of test (p-value = 0.542). The degree of association between independent and dependent variables was assessed by using Adjusted Odds Ratio (AOR) with 95% CI. Finally, the data were presented by tables and frequencies.
Ethical Clearance Ethical clearance was obtained from Institutional Review Board (IRB) of Jimma University, Faculty of Public Health. Permission was obtained from Tercha General Hospital and consent was taken from the manager. Names and other personal information which can violate the confidentiality of the study subjects were not taken. Any information have been kept confidentially.

Socio-demographic characteristics of the respondents
The total of 366 study subjects (183 cases and 183 controls) were included in this study with 100% response rate. The mean age of mothers for cases and controls was 27.7± 6.2 and 26.4±5.9 years, respectively. The highest proportion of mothers of cases 135(73.7%) were in the age group 20-35 years. Majority of cases 153 (83.6%) were born from married mother. One third of mother of controls 66(36.07%) were at primary school level while 21(11.48%) of the controls were diploma and above. Most of mothers of cases (79.2%), were from rural area while 106 (57.9%) of the controls were from rural areas (Table 1).

Maternal Obstetrics Characteristics
Primipara mothers were higher in the control group than the cases. The proportion of mothers had antenatal care follow up at least one times in the current pregnancy in cases  (Table 2).

Maternal medical and Neonatal related factors
Regarding to maternal medical factors, 345 screened for HIV of which 340 (92.8%) were non-reactive and 4 (1.09%) were reactive among reactive mothers four of them were mothers of cases. Ninety seven percent of mothers were non-reactive for VDRL test.
Majority of participants (84.6%) were none reactive and 0.8% were positive for hepatitis B surface antigen testing. Forty-three mothers of (23.5%) cases and 16(8.7%) controls hemoglobin level were less than 11gm/dl. Thirty one (16.9%) of cases and 14(7.65%) of controls of were born from mothers who were diagnosed at least one type of medical illnesses in the current pregnancy. Concerning to neonatal related factors, One hundred seven (58.47%) cases and 136(74.3%) controls mode of delivery were vertex in fetal presentation. The proportion of congenital anomaly in was higher in cases (3.83%) than controls (1.69%). The proportion of low birth weight (Wt <2500gm) cases and controls were 78(42.62%) and 55(30.05%), respectively (Table 3).

Organizational related factors
The proportion of partograph use was high in controls (88.5%) than cases (69.9%). Safe childhood checklist was used in 138(75.4%) controls and 102(55.7%) cases. Among those 51(58%) of mothers had at least one obstetrics complication. Seventy two (39.3%) of mothers of cases were referred from other health institution with referral paper (Table 4) Determinants of perinatal mortality. were found to be independent determinants of perinatal mortality (Table 5).

Discussion
Mothers who were rural in residence were two times more likely to have perinatal mortality compared to those mother who were urban in residence [AOR = 1.82; 95%CI: (1.04-3.19), p = 0.035]. This finding is consistent with the study conducted in Jimma university referral hospital and showed that mothers who live outside Jimma is more likely to have perinatal death [AOR = 2.861; 95%CI:1.99-3.33] (21). This might be due to the fact that mothers live in rural area have exposure to lack of awareness, inaccessible healthy facility, poor transport access, or probably due to their big delay in health care seeking behavior; this all lead to obstetrics complications and increase the risk of perinatal loss.
In the present study, among socio demographic variables maternal level of education, age, occupational status, and marital status were not associated with perinatal mortality. The odds of perinatal mortality were 50% less likely among mothers whose labor was followed using partograph compared to their counter parts [AOR = 0.5; 95%CI: 0.25-0.9, P-value = 0.025]. This finding is consistence with the study conducted in public hospitals of Addis Ababa [AOR = 0.35; 95%CI: (0.18-0.66)] (12). This implies that using WHO recommended partograph is a protective factor for perinatal mortality. This might be due to use of partograph can help alert health care providers to pick any abnormalities during the course of labor. Therefore, it can prevent perinatal loss with early diagnosis and management of labor complications.
The odds of perinatal mortality were 48% less likely among mothers whose labor was followed using safe childhood checklist compared to their counter parts [AOR = 0.52;

Limitation of the Study
Since this study was conducted based on secondary data that was gathered for other purpose, it was difficult to gather all necessary variables. The confounding effect of unmeasured variables could not be controlled.

Conclusion
In conclusion, being rural in mother's place of residence, presence of obstetrics complication during delivery, duration of labour, low birth weight and coming with referral were positively associated with perinatal mortality; whereas antenatal care booking, using partograph, and using safe childbirth checklist were negatively associated with perinatal mortality. Therefore, managers should strictly monitor ambulance service utilization for minimizing delay during referral. Health care providers should give special attention for early recognition of abnormalities and manage accordingly while doing ANC and labor follow up. Similarly, they should do immediate newborn care with special attention to newborns with low birth weight. It will better if researchers conduct further longitudinal community based study to get other unmeasured risk factors.

Acknowledgements
We would like to thank Jimma University Institute of Health for giving us this chance. We want to sincerely acknowledge Tercha general hospital staffs, data collectors, and supervisors for their cooperation.

Funding
This study was sponsored by Jimma University, Institute of Health. The funder of the study had no role in study design, data collection, data analysis, data interpretation, writing of the report and in writing the manuscript.

Availability of data and materials
The data supporting our findings are found at, kept in confidential and stored at the correspondent author both in hard and soft copies. If someone wants our data, we are voluntary to share it and the correspondent author should be contacted through the email address under the author's information Authors' Contributions FTD, developed the proposal, carried out data collection, conducted the analysis, involved in reviewing the manuscript and had full access to all the data in the study and had final responsibility for the decision to submit for publication. CH and MB provided general guidance in overall study progress and participated in reviewing the proposal, reviewing the analysis and participated in final study document development. AA participate in reviewing the whole document and guide the preparation of manuscript.All authors read and approved the final manuscript and accountable for all aspects of the work.

Ethics approval and consent to participate
Ethical clearance was obtained from Institutional Review Board (IRB) of Jimma University, Faculty of Public Health. Following the endorsement by the IRB, a written permission obtained from Tercha general hospital. Furthermore, confidentiality was assured throughout the process.

Consent for publication
Not applicable.