Time to recovery of neonatal sepsis and determinant factors among neonates admitted in Public Hospitals of Central Gondar Zone, Northwest Ethiopia, 2021

Background Neonatal sepsis is a leading cause of neonatal morbidity and mortality, particularly in developing countries. Time to recovery is an indicator of the severity of sepsis, and risk factors varied significantly according to study population and settings. Moreover, published literature regarding the time to recovery of neonatal sepsis is scarce. Objective The aim of this study was to assess the time to recovery of neonatal sepsis and determinant factors among neonates admitted in the Public Hospitals of Central Gondar Zone, Northwest Ethiopia. Methods An institution-based prospective follow-up study design was conducted among 631 neonates with sepsis. A structured, pre-tested, interviewer-administered questionnaire was used. The median time to recovery, life-table, the Kaplan Meier curve, and the log-rank test were computed. Both bi-variable and multivariable Cox regression models were applied to analyze the data. Results Of all septic neonates, 511 successfully recovered. They were followed for a total of 4,740- neonate day’s observation and the median time to recovery was 7 days (IQR = 5-10 days). After adjusting for covariates, intrapartum fever (AHR = 0.69, 95%CI: 0.49, 0.99), induced onset of labor (AHR = 0.68, 95%CI: 0.49, 0.94), chest indrawing (AHR = 0.67, 95%CI: 0.46, 0.99), late onset sepsis (AHR = 0.55, 95%CI: 0.40, 0.75), non-oral enteral feeding (AHR = 0.38, 95%CI: 0.29, 0.50), assisted with bag and mask (AHR = 0.72, 95%CI: 0.56, 0.93), normal birth weight (AHR = 1.42, 95%CI: 1.03, 1.94), gestational age of 37-42 weeks (AHR = 1.93, 95%CI: 1.32, 2.84), septic shock (AHR = 0.08, 95%CI: 0.02, 0.39), infectious complications (AHR = 0.42, 95%CI: 0.29, 0.61), being in critical conditions (AHR = 0.68, 95%CI: 0.52, 0.89), and early recognition of illness (AHR = 1.83, 95%CI: 1.27, 2.63) were independently associated with the time to recovery of neonatal sepsis. Conclusions and recommendations The time to recovery of this study was moderately acceptable as compared to the previous studies. The above-mentioned factors could be used for the early identification of neonates with sepsis at risk for protracted illness and it could guide prompt referral to higher centers in primary health sectors. This also will provide prognostic information to clinicians and families as longer recovery time has economic and social implications in our country. This study was conducted at NICU, Neonatology Ward, in Public Hospitals (randomly selected) of Central Gondar Zone, Gondar, Northwest Ethiopia. The Central Gondar Zone is one of the largest administrative zones in Gondar Province. It includes Gondar City and the surrounding areas, such as Lay-Armachiho, Tach-Armachiho, Gondar Zuria, Chiliga, Tegedea, East Dembiya, West Dembiya, Alefa, Takusa, Wogera, West Belessa, East Belessa, and Kinfaz-Begela Districts. Hospitals found in this zone are Sanja (serving 121, 321 populations), Aykel (158, 587), Shawra (233, 917), Koladiba (211,790), Deligi (181, 603), Tegedea (96, 035), Gohala (146, 599), the University of Gondar Comprehensive Specialized Hospital (UoGCSH), Arbaya (168, 491), and Wogera (249, 412) Hospital. The number of delivery services in Tegedea, Arbaya, Gohala, Wogera, Sanja, Deligi, Shawra, Koladiba, and Aykel Hospitals were 127, 490, 595, 659, 763, 770, 850, 1303, and 1432, respectively. According to the UoGCSH Information Center, around 410,000 people visit the hospital every year. Total delivery reaches up to 8,000 each year on average (845 births per month) (the list of hospitals, districts, and services were obtained from the Central Gondar Zone Health Office). The study was conducted from 15/04/2021 to 29/09/2021. The multicenter institution-based prospective follow-up study design was undertaken to determine the time to recovery of NS. All neonates admitted with sepsis in the Public Hospitals of Central Gondar Zone were a source population. All neonates admitted with sepsis in selected Public Hospitals of Central Gondar Zone who were available during the data collection period were a study population. All neonates admitted with the diagnosis of NS in Public Hospitals of Central Gondar Zone during the study follow-up period were included in the study. Neonates who died before taking the treatment were excluded from the study. The sample size was calculated using STATA Version 16 Statistical Software, a sample size for time to event data; by considering alpha (0.05), the hazard ratio for mentioned factors (Respiratory distress and meconium aspiration), percent of survival, power 0.80, ratio (1:1), and withdrawal 10% for a sample size of Log-rank test and the sample size for the two variables was 154 and 278. Furthermore, we considered alpha 0.05, the hazard ratio for mentioned factors, power 0.80, SD 0.5, and withdrawal 10% for the sample size of Cox PH regression, and the sample size for the two variables was 20 and 14. The sample size for incidence of recovery was also calculated using a precision approach formula (n = (Zα/2)2 * P(1-P)/d2 = 574); by considering the proportion value of 0.84, 95% of the confidence interval (CI), 3% margin of error, and 10% of non-response rate (57.0). Accordingly, the sample size was 631. The above information, to estimate the sample size of this study, was taken from the study conducted in the Felege Hiwot Referral Hospital [1]. By comparing the sample size obtained, the highest sample size was selected among the three. Therefore, the final sample size was 631 mother-newborn pairs. Among ten hospitals found in Central Gondar Zone, the five of them, 50%, (Shawra Hospital, Sanja Hospital, Aykel Hospital, UoGCSH, and Koladiba Hospital) were selected randomly using the lottery method. Then, all neonates who met the inclusion criteria during the study period were included in the study in each proportionally allocated hospital (S1 File). The data collection was started in the five sites at the same time. Time to recovery of neonatal sepsis was a dependent variable. Socio-demographic variables. Maternal age, place of residence, religion, marital status, educational status, educational status of the husband, occupational status, monthly income, and family size. Maternal-related variables. Parity, gravidity, the onset of labor, duration of labor, mode of delivery, place of delivery, delivery attendant, number of ANC visits, twin pregnancy, obstructed labor, foul-smelling liquor, UTI/STD during pregnancy, Pregnancy-Induced Hypertension (PIH), antepartum hemorrhage, intrapartum fever, diagnosed chorioamnionitis, duration after the ROM, maternal infection history, and presence of chronic illness. Clinical and medical care-related variables. Have fever, apnea, respiratory distress, tachycardia, poor feeding, dehydration, vomiting, lethargy, convulsion/seizure, irritability, drowsiness, hypothermia, CRT, pallor, cyanosis, severe jaundice, chest indrawing, bulging fontanel, blood culture, complete blood count (WBC, platelet count, etc.), radiological finding, sepsis type, the onset of infection, bacterial isolates, major co-morbidities, non-oral enteral feeding, assisted with bag and mask, medications, supportive care, duration of treatment, respiratory failure, septic shock, hypoxemia, meningitis, neurological sequelae, organ dysfunction, DIC, acute kidney injury, infectious complications, being in critical conditions, and discharge and outcome status variables. Health care service-related variables. Satisfied with services, appropriately trained health workers, early care seeking at the household level, quality status of NICU, early recognition of illness, early initiation of treatment, the distance to the nearest health facility, fast and adequate transport access, the cost of transportation, and time of visiting health facility after the neonate get sick. Neonate-related variables. Age of neonate at admission, sex of neonate, Birth Weight (BW), GA at birth, admission weight, vital signs, EBF initiated within one hour, the first minute APGAR score, fifth minute APGAR score, resuscitated at birth, RDS, MAS, and kept in KMC within one hour. If a neonate was recovered from the infection after completing the treatment according to physician diagnosis. Refers to neonate left (or stops treatment) the treatment unit against medical advice or the treatment. A neonate died by NS during the treatment or at the treatment unit. It refers to a neonate defaulted from the treatment, referred, died, or transferred. A time from the admission date by NS to the discharge date while the neonate is recovered. It was measured by subtracting the date of admission from the discharge date (time in days until recovery/discharge). If sepsis occurred from birth up to seven days of age. If sepsis occurred between eight and twenty-eight days of age. Neonates with possible serious bacterial infections were considered as sepsis based on the physician’s diagnosis. Data were collected using an interviewer-administered questionnaire with direct face-to-face interviews with the mothers. Document reviews were also considered. The main questions that are included in the questionnaire were socio-demographic variables, maternal-related factors, neonatal-related factors, health care service-related characteristics, and clinical and medical care-related factors (clinical feature, diagnostic/laboratory test, management, complication, and outcome status characteristics) (S2 File). A well-developed checklist was used to collect additional data, such as data on general information, from the follow-up, or recorded data in a chart. The questionnaire was constructed after the review of relevant literature in order to maintain the standards of the questionnaire [1, 3, 13–16, 22–41, 48–53]. Then, the validity was established by doing expert discussions (Pediatricians and Public Health experts) and pre-test study. As a result, changes were made based on both a pre-test and expert opinion to make the questionnaire measure what is intended to measure. After data were collected using a pre-test study, the questionnaire was tested for reliability (Alpha/reliability coefficient = 0.7622, acceptable reliability) and it was assessed for suitability of the content, clarity, sequence, and flow of the questionnaire. To ensure accuracy and consistency of meaning, the data collecting questionnaire was first written in English, then translated into Amharic, and then back to English (S3 File). Two neonatal nurse data collectors, with one immediate supervisor (physician) in each hospital in addition to the investigator, collected the data in each respective NICU of the hospital. Information about the conditions during delivery, neonatal factors, maternal factors, and socio-demographic characteristics were obtained from the mother and attending physician. The GA of the neonate was determined by the first date of the last normal menstrual period (nine months of amenorrhea) as reported by the mother and new Ballard score assessment [54]. The mothers were assessed for the regular cycle of menstruation and history without contraception. Neonates were considered appropriate for GA if their BW and head circumference were between the 10th and 90th percentile using the Lubchenco chart [55]. Anthropometric measurements and physical examination were considered to collect data from study participants. At admission, the data collectors assessed the condition of the neonate (All assessments were made and data were collected). During every follow-up visit, the neonates were examined and the necessary data were collected (Neonatal measurements, clinical features, and diagnostic/laboratory test results, for example). Besides, during medication time, all essential treatments, medications, or procedures prescribed were recorded, and the outcome status of the neonates was assessed. To diagnose NS, the World Health Organization Integrated Management of Neonatal and Childhood Illness (IMNCI) guideline was considered, and NS was suggested with the presence of any one of the seven clinical signs and two or more hematologic criteria. These include the presence of difficulty of feeding, convulsions, the movement only when stimulated, severe chest retractions, change in the level of activity, respiratory rate ≥ 60 breaths per minute, and oral temperature ≥ 37.5˚C or < 35.5˚C. Furthermore, other signs like tachycardia, bradycardia, irritability, oxygen requirement, increased frequency of apnea, poor CRT, and ≥ 2 hematological criteria (total leukocyte count <5,000 or >12,000 cells/μl, absolute neutrophil count <1,500 cells/μl or >7,500 cells/μl, erythrocyte sedimentation rate >15/1h, platelet count <150x103 or >450×103 cells/μl, elevated C-reactive protein>1mg/dl, and glucose intolerance confirmed at least two times: hyperglycemia (blood glucose >180 mg/dL) or hypoglycemia (glycaemia <45 mg/dl) when receiving age-specific normal range glucose amounts) were considered [6, 56–58]. Notably, the diagnosis included history taking, clinical manifestations (physical examination), and laboratory tests. All neonates were observed for clinical events and managed according to the hospitals’ standard protocol, and followed up to the outcome of interest. All infection prevention precaution standards were used during the time of measurement. Following the measurement of each neonate, a handwashing procedure was performed. Standard precautions were also applied for measuring equipment. Materials like a balance beam neonate scale, calibrated non-elastic plastic tape, etc. were used to measure parameters. All measurements were recorded on the questionnaire and checklist designed for this study. The mothers of each neonate were orientated verbally about the purpose and usefulness of the study. The collected data were also checked on each day of activity for consistency and completeness by the immediate supervisors. Besides, the data collectors (and supervisors) were trained and closely supervised. Furthermore, the data collection questionnaire and all data collection processes were ensured, checked, and supervised for content and completeness. More importantly, the questionnaire was pretested in a similar setting by the research investigators prior to the data collection on five percent of the total sample size at two of the hospitals (Arbaya Hospital and Wogera Hospital) that were not part of the main study. Revisions and adjustments were performed after the pre-test. Health education on the outcome of interest was provided to each participant during the follow-up and at the time of discharge. The collected data were checked for completeness, accuracy, and clarity. The collected data were entered into Epi-Info version 7.2.2 and exported to Stata Version 16 Statistical Software for further analysis. The information that needs coding was coded and missing values were considered before analysis. As result, findings were presented in the form of text, tables, and figures using frequencies and summary statistics. Descriptive analyses (percentages, median, IQR, mean, and SD) were done to describe the frequency and percentage of the dependent and independent variables. Mean ± SD were presented for normally distributed continuous covariates while median with IQR was presented for skewed covariates. Meanwhile, numbers (percentage) were presented for categorical variables. The median time to recovery, life-table, Kaplan Meier curve, and log-rank test were computed. Both graphically and through Schoenfeld residual global tests, the proportional hazard assumption was verified. Both the bi-variable and multivariable Cox regression models were applied to describe the association between the dependent and independent variables and independent predictors of the time to recovery. To control the possible confounding covariates simultaneously, the covariates that showed a P-value ≤ of 0.05 in bivariate analysis were entered into a multivariable regression analysis. The Cox Snell residual test was used to assess the model goodness of fit. The Crude Hazard Ratio (CHR) and Adjusted Hazard Ratio (AHR) were used to test the strength of association between the independent and dependent variables. In all, a P-value ≤ of 0.05 was considered statistically significant (or AHR with their respective 95% CI). Ethical clearance was obtained from the University of Gondar, Institute of Public Health Ethical Review Committee (Ref No/IPH/1543/2013 E.C.). The objective of the study was described to the mothers of all neonates, including the reasons for assessment of the time to recovery of NS (S2 File). In addition to this, we informed the mothers that all information obtained from them will be secured and kept confidential (S2 File). To ensure confidentiality, the names were avoided in the questionnaire and reporting the results of the study. All data involving measurements were gathered without any harm to the neonates. During data collection, a copy of a written informed consent form approved by the Ethical Review Committee of Institute of Public Health, College of Medicine and Health Science, the University of Gondar, was given to each participant. It was read aloud in Amharic to the mothers who could not read. Written informed consent was taken from the neonate’s mother or father (S2 File).