Knowledge of Neonatal Danger Signs and Associated Factors Among Mothers Who Gave Birth in the Past Six Months in Chole District, Arsi Zone, South East Ethiopia: Cross sectional Study

Background: Early detection of neonatal illness by mothers is an important step towards improving newborn survival. Even though mother’s knowledge of neonatal danger signs plays a critical role in reducing neonatal morbidity and mortality, studies on the area are limited and most of them are institution-based study which do not include rural mothers and mothers with home delivery. Therefore, the study aimed to assess knowledge of neonatal danger signs and associated factors among mother’s who gave birth in the past 6 months. Methods: Community-based cross-sectional study was conducted on 520 postnatal mothers by using multistage sampling method from March 1-15, 2019. The data were entered into Epi data version 3.1 and then exported into SPSS version 20 for analysis. Bivariate and multivariate analysis were used by using binary logistic regression to identify factors associated with mother’s knowledge of neonatal danger signs. Statistical signicance was declared at p-value less than 0.05. Results: Mother’s level of knowledge on neonatal danger signs was found to be 50.2% (95%CI: 46.3%, 54.3%). Mother’s and husband’s secondary and above educational level [AOR= 2.15 95%CI: (1.11, 4.17)], and [AOR: 2.05, 95%CI: (1.07, 3.94) respectively, being in Urban [AOR=5.83, 95%CI: (2.77, 12.24)], had four or more antenatal visits [AOR: 2.10, 95%CI: (1.13, 3.90)], counseled during antenatal care [AOR= 4.33 95%CI: (1.88, 9.98)] and knowledge about essential newborn care [AOR=3.91 95%CI: (2.05, 7.48)] were signicantly associated with mother’s knowledge of neonatal danger signs. Conclusion: The study revealed that mother’s level of knowledge towards neonatal danger signs was low. Mother’s educational status, husband’s educational status, place of residence, number of antenatal visits, counseled during antenatal visit and knowledge about essential newborn care were signicantly

Knowledge about the severity of an apparent neonatal illness (i.e., knowing when to act) and knowledge about the appropriate lifesaving action is very important to prevent or avoid baby's health problem. Thus, improving maternal knowledge towards signs of neonatal illness will signi cantly decrease neonatal morbidity and mortality.(Robert Kuganab-Lem and Adadow Yidana, 2014; Mekdes ,et al., 2018).
Lack of speci city of the clinical manifestation of various neonatal morbidities, resulting in di culty in making early diagnosis, and delay in seeking care and result high mortality are some of the plausible explanation for new born health problems (Ekwochi ,et al., 2015).
Study conducted in Ethiopia showed that there was poor knowledge of mothers towards neonatal danger signs. Mothers practice for neonatal danger sign was unsafe; most mothers take their sick neonate to traditional healers and give home remedies. Most neonatal death take place at home, this indicating that lack of early recognition of the danger signs and low treatment seeking practice of mothers (caretaker) towards modern health care service (Walellign ,et al., 2017;Melkamu ,et al., 2016).
Even though, Ethiopian Government increases the provision of quality community-based newborn care services including management of newborn sepsis; and strengthens the supportive systems with a focus of woreda capacity building, the change in neonatal mortality is not as signi cant as the change in postneonatal and child mortality(CSA, 2016). Despite the fact that mother's knowledge on neonatal danger signs play a critical role in reducing neonatal morbidity and mortality, studies on the area seem to be limited, as far as researcher has investigated various sources of information. Since the available studies were conducted in the urban area or at the institution level (Awoke, 2011), they do not address the rural community in which knowledge of neonatal danger signs is relatively low and may not consider mothers with home delivery. Previous studies assess knowledge from postnatal mothers who delivered two years back (Tesfaye, 2018), this may result in recall bias and failure to differentiate between the neonatal and post neonatal period. So, the level of maternal knowledge might be miss-reported. Therefore, study aimed to assess the knowledge of neonatal danger signs and associated factors among mothers who gave birth in the past six months in Chole district South Eastern Ethiopia from March, 1-15/2019.

Study setting and period
The study was conducted from March 1-15, 2019 at Chole District, one of the districts in Arsi zone, which is found in the South Eastern part of Ethiopia. The district is about 291 km from Addis Ababa, the capital city of Ethiopia. Chole district has a total of 20 kebeles (4 urban and 16 rural kebeles). The district has a total population of 120,764 of which 61,568 are males while 59,196 of them are females and 4191 annual live births. Estimated women with child bearing age of 21,738(7386 urban and 14,352 rural).
There were 2095 (712 Urban and 1383 Rural) mothers who gave live birth in the past six months at the district. The district has 4 health centers, 18 health posts, 10 private clinics and 8 drug stores. According to 2017 report of Chole health bureau, the health coverage of the district reached 64.56% in 2016 (Chole district health o ce, 2017).

Study design and population
Community based cross-sectional study design using quantitative method was applied.
All mothers who gave birth within the past six months prior to the study period and resident at least for 6 months in the chole district were included in this study. All Mothers who gave birth in the past six months but unable to communicate because of serious illness or impaired cognition during data collection period were excluded from the study.

Sample size determination and Sampling procedure
The sample size was calculated by using a single population proportion formula with assumptions of con dence level at 95% = 1.96, a margin of error (d) = 0.05, Design effect = 1.5 and a reasonable proportion of mothers knowledge of neonatal danger signs (P = 0.313) from a previous study conducted at Wolkite Town, Gurage Zone, SNNPR, Ethiopia, 2017 (Walellign ,et al., 2017) and adding 5% nonresponse rate, the nal sample size became 520.
Multi stage sampling method was employed in selecting the study subjects. Chole District has 16 Rural and 4 Urban kebeles (kebele is the smallest administrative unit in Ethiopia), 5 kebeles from the Rural and two kebeles from the urban were selected randomly. The sample size was distributed to the seven kebeles proportionate to the size of their population. Finally, the study subjects that had been included in the study from each kebele were identi ed by using simple random sampling technique (computer based) based on the sampling frame obtained from kebele health extension workers registration books (N = 2095). The selected households were located with the help of kebele health extension workers and administrators of the given kebeles. For household with more than one mother who gave birth in the past 6 months, one of the mothers was selected using lottery method.

Data collection methods
Data collection was undertaken using an interviewer administered structured questionnaire that was adopted from the Safe Motherhood questionnaire developed by the Maternal and Neonatal Health Program of Johns Hopkins Program for International Education in Gynecology and Obstetrics (JPHIEGO) (Del Barco, 2004). The data were collected by trained 7 diploma nurses and supervised by three BSc holder nurses who were uent in local language "Afan Oromo and Amharic". The reason why nurses were chosen is that, at the end of data collection health education was given for the respondents with poor knowledge of neonatal danger signs. A brief introductory orientation was given for the study participants by data collectors about the purposes of study. Explanation was given on the importance of their involvement, then mothers who were volunteer were interviewing face to face using structured and pretested questionnaires by going into household level. Study participants were contacted at their home both at morning and afternoon.

Operational de nition
Key danger signs: Are those signs which warrant survival chance of a neonate and demands immediate medical care. Thus, 1) poor/not sucking, 2), fever is high body temperature or being hot body of the selected child as perceived and reported by mothers or care givers/hyperthermia. 3) hypothermia is decreased body temperature or cold to touch of the baby 4) Convulsion(twisting of body), 5) increased respiratory rate/fast breathing (more than 60 per minute when counted for at least one minute) and, 6) Vomiting, 7) chest retractions/in drawing, 8) Jaundice (yellow soles, palms and sclera) 9) Lethargy-only moves when stimulated or not at all and 10) umbilical redness or draining pus/sign of infection.(UNICEF, 2014; WHO, 2017b) Knowledge was measured by mother's capability of mentioning the 10 WHO listed neonatal danger signs without prompt by interviewer. Then knowledge is categorized into good knowledge and poor knowledgebased on ability to mentioning three out of ten WHO listed neonatal danger signs.
Knowledge: State of awareness of mothers on WHO listed key neonatal danger signs (Solomon ,et al., 2015; WHO, 2017b).
Good knowledge: Mothers who were capable of mentioning three or more key WHO identi ed danger signs for neonate (Lancet, 2008).
Poor knowledge-Those mothers who were able to mention two or less key WHO identi ed neonatal danger signs (Lancet, 2008).
Neonatal danger signs: Are symptoms that complicate the lives of the neonate and happen during the neonatal period (during the 1st 28 days) (WHO, 2017a).

Data quality control
The questionnaire was rst prepared in English language then it was translated to Amharic and Afan Oromo languages, which are used for communication in the local community and back to English by different language experts to check consistency of the data. The questionnaire was pre-tested on postnatal women in the nearby district (Guna district) before the beginning of the actual data collection on 5% of total sample size. Findings and experiences from the pre-test were utilized in modifying and reshaping the research data collection tools. Training was given for data collectors and supervisors about the objective of the study, con dentiality of information, respondent's right, privacy and techniques of interview prior to data collection. Completeness of questionnaire was checked by principal investigator and supervisors on daily basis. Double data entry was done by two data clerks and consistency of entered data were cross checked by compering two separately entered data into Epi-Data.

Data processing and analysis
The data were rst coded, entered and cleaned by Epi data statistical software version 3.1 and then were exported to SPSS window version 20 for analysis. A descriptive statistical analysis was employed to describe the characteristics of participants. For analysis of the outcome variable, good knowledge was coded as 1 and poor knowledge was coded as 0. The information was presented using frequencies, tables and gures. Multi collinearity was checked using VIF (Variance in ation factors) and standard error (SE) and variables with SE of > 2 or VIF > 10 were dropped. The goodness of t was tested by Hosmer-Lemeshow statistic and Omnibus tests. The model was considered good t since it is found to be insigni cant for Hosmer-Lemeshow statistic (p = 0.648) and signi cant for Omnibus tests (p = 0.000). Bi-variate and multivariate analysis were used to observe the association between each independent variable and the outcome variable by using binary logistic regression. All variables with P ≤ 0.25 in the bivariate analysis were included in the nal model of multivariate analysis in order to control all possible confounders. In addition, variables which were signi cant in previous studies and from context point of view were included in the nal model even if the above criteria were not met. The direction and strength of statistical association was measured by odds ratio with 95% CI. Adjusted odds ratio along with 95% CI was estimated to identify associated factors with knowledge about neonatal danger signs by using multivariate analysis in the binary logistic regression. Finally, statistical signi cance was declared at pvalue < 0.05.

Ethical considerations
Ethical clearance to conduct this study was obtained from Haramaya University, College of Health and Medical Sciences, Institutional Health Research Ethics Review Committee (HU-IHRERC) before starting data collection process. O cial letter to make a study was also obtained from the Woreda and Kebele Government o cials as needed.
Informed, voluntary, written and signed consent was obtained from each study participants prior to the interview after explaining about the purpose of the study, their right to refuse or discontinue the interview at any time if they did not want it. They were also informed as information obtained from them were treated with complete con dentiality (respondents name and other identi cation were not written in the questionnaire) and in case it might have minimum risk on them (may consume time).

Socio-demographic Characteristics
A total of 510 out of 520 mothers of babies aged up to six months were recruited in the study yielding a response rate of 98.1%. The mean age of the participants was 28.48(SD ± 4.68) years. Majority of the mothers were married 467(91.6%) and rural resident 333(65.3%). About 337 (72.7%) of them are Orthodox Christian followers and 255(50%) belong to Amhara ethnic group.    Mother's knowledge of neonatal danger signs The majority (93.5%) of the participants were knowledgeable about at least one of ten WHO recognized neonatal danger signs, but only 50.2% (95%CI: 46.3%, 54.3%) of interviewed mothers were able to mention at least three neonatal danger signs (had a good knowledge). Figure 2 Total knowledge of neonatal danger signs among mothers who gave birth in the past six months at Chole District, Arsi Zone, South East Ethiopia, 2019 (n = 510) The most frequently recognized neonatal danger sign by 423(82.9%) mothers was hotness of the body (fever) followed by vomiting 306 (60%) and lethargy 260 (51%). Fig. 3

Discussion
The knowledge of mothers on neonatal danger signs was found to be 50.2% (95%CI: 46.3, 54.3) (able to mention three or more WHO identi ed neonatal danger signs). Factors that signi cantly associated to women knowledge of neonatal danger signs were mother's educational level, husband's educational level, place of residence, number ANC visits, counseling during ANC and knowledge about essential newborn care.
The knowledge of neonatal danger signs in this study was consistent with studies done in Southern Ethiopia (Abera ,et al., 2017) which was 50.3% and Tigray region, Ethiopia (Nuredin ,et al., 2017) which was 50.6%. But higher than studies conducted in Kenya (Kibaru and Otara, 2016), Eastern Ethiopia (Tesfaye, 2018) and Gondar, Ethiopia (Solomon ,et al., 2015) which was (15.5%), (9.38%) and (18.2%) respectively. This difference might be due to study period difference, a slight difference in the data collection tools, number of neonatal danger signs included in this study were higher than those studies and in this study exposure to ANC and percentage of institutional delivery (mothers who delivered at health institutions had better exposure to post-natal counseling) were found to be higher. In another ways, previous studies included mothers who gave birth two years back which might lead to mothers fail to recall and loose caring responsibility. Meanwhile, this study is lower than studies conducted in ). This might be due to involvement of rural women in this study and study area difference that, this study was community-based study. In this study, a low level of mothers' knowledge of the neonatal danger signs was observed even though the majority of the women had attended more than 3 visits of antenatal care; this led to an idea that the antenatal care providers may not have proper resources and facilities to educate mothers about the neonatal danger signs. Poor knowledge of mothers on neonatal danger signs will have negative impact on Integrated Management of Childhood Illnesses (IMCI) Program in Ethiopia, because the program is based on early identi cation of newborn danger Signs by caregivers and appropriate referral aiming at reduction in neonatal mortality.
The most frequently mentioned danger signs were fever, vomiting and lethargy which is consistent with study conducted in Kenya (Kibaru and Otara, 2016) and Ethiopia (Abera ,et al., 2017). This might be due to these signs were commonly affect the health of neonates and were relatively easily detected by caregivers. And incongruent with studies conducted in Nigeria (Ekwochi ,et al., 2015) and Ethiopia (Mekdes ,et al., 2018;Solomon ,et al., 2015). This might be due to the difference in extensive HEWs counseling in the community and socio-cultural variation between the study participants. Overall, this implies Poor recognition of danger signs is the main barrier in order not to seek care even if they are manifesting those signs and also caregivers do not seek medical care if they do not recognize it as severe (because only 61.2% of mothers in this study perceived the recognized signs as severe).
In this study, mothers having secondary and above educational level were about two times more likely to know about neonatal danger signs as compared to those mothers with primary or below educational level. This is nearly consistent with study conducted in Gondar town (Solomon ,et al., 2015). The possible justi cation could be educated mothers acquire knowledge about disease and human health through their academic life and education increase to get health service and increase tendency to read and understand materials related to newborn health.
Similarly, husband educational level was signi cantly associated with mothers' good knowledge about neonatal danger signs. The odds of being having knowledge about neonatal danger signs was two times among mothers whose husbands achieved secondary and above educational level. This is consistent with the study conducted in Wolkite and Gondar, Ethiopia (Walellign ,et al., 2017;Solomon ,et al., 2015). This could be explained as educated husbands are more informed and help the mother in recognition of neonatal danger signs and this might positively affect the knowledge of the mothers.
The study showed that mothers who live in urban were 5.83 times more likely knowledgeable as compared to mothers who lived in rural area. This is congruent with the study conducted in Southern Ethiopia (Abera ,et al., 2017) and Woldia, Ethiopia (Mekdes ,et al., 2018) where living in rural increase the odds of good knowledge of neonatal danger signs. This might be due to mothers who live in urban were more likely to seek health care and health information from different sources as compared to mothers who living in rural parts. This study showed that 82% of home delivery was conducted at rural area and this leads to less chance to contact with health professionals. This leads to an idea that, mothers who live in rural area had tendency to deliver at home and had less chance to get immediate PNC counseling on neonatal danger signs and remain uninformed.
ANC visit was signi cantly associated with mother's knowledge about neonatal danger signs. In line with that, mother who got counseling during ANC follow up were 5.71 times more knowledgeable as compared to mothers who did not get counseled during ANC. This is consistent with study conducted in Ghana (Okawa ,et al., 2015), Eastern Ethiopia (Tesfaye, 2018), Jimma, Ethiopia (Melkamu ,et al., 2016) and in Gondar (Solomon ,et al., 2015). This might be due to exposure to ANC follow up repeatedly might increase the chance to get more information related to neonatal danger signs from health professionals.
These might indicate the need to improve the counseling given to mothers during ANC and PNC giving due emphasis to these signs and symptoms of serious newborn illness, given the fact that the majority of the mothers were having ANC/PNC follow-up and delivered at health facilities.
Mothers who had knowledge about essential newborn care had signi cant association with knowledge about neonatal danger signs. This is consistent with study conducted in Ethiopia (Abera ,et al., 2017). This might be due to the case that recognition of neonatal danger signs is one of the components of essential newborn care. So, those mothers who have knowledge on essential newborn care practices are more likely knowledgeable about neonatal signs of illness. Meanwhile, there was high institutional delivery in this study (73.7%), which probably allow the chance to get contact with health professionals at health facility and get counseled on neonatal danger signs on appropriate time.
The study might not show cause and effect relationship because of the nature of the study design (crosssectional). It might be affected by recall bias because it allowed mothers who gave birth in the past 6 months, while neonatal danger signs occurred during the rst one month of life.

Conclusion And Recommendations
The study revealed that the overall level of mother's knowledge about neonatal danger signs was low. They were also informed as information obtained from them were treated with complete con dentiality Consent to publication

Non applicable
Availability of data and materials Pertinent data were presented in this manuscript. Additional data can be requested from the corresponding author upon reasonable request.

Con ict of interest
There is no con ict of interest Funding Haramaya university Authors' contributions TG, the corresponding author, worked on designing the study, trained, and supervised the data collectors, checked the completeness of collected data, entered, analyzed, and interpreted the result, and prepared the manuscript. The co-authors namely TA, MD, AE, and TG played their role in re-analyzing and writing the nal draft of the results. Moreover, the co-authors wrote the manuscript. All authors were involved in reading and approving the nal manuscript.  Total knowledge of neonatal danger signs among mothers who gave birth in the past six months at Chole District, Arsi Zone, South East Ethiopia, 2019 (n=510) Figure 2 Total knowledge of neonatal danger signs among mothers who gave birth in the past six months at Chole District, Arsi Zone, South East Ethiopia, 2019 (n=510) Figure 3 Maternal knowledge on speci c neonatal danger signs among mothers who gave birth in the past six months at Chole District, Arsi Zone, South East Ethiopia, 2019 (n=510)