Abstract
Background:
Low-birth-weight (LBW) remains a leading cause of neonatal mortality in Ethiopia. This study aimed to identify individual and institutional determinants of mortality among LBW infants in Southern Ethiopia.
Methods:
A hospital-based prospective cohort study was conducted among 768 LBW neonates admitted to 8 public hospitals in South Ethiopia from September 2023 to February 2024. Data were collected using a structured questionnaire and analyzed using STATA 16. Multilevel logistic regression was applied to identify predictors of mortality.
Results:
Overall mortality was 28.3%. Key individual-level predictors included lower birth weight, respiratory distress syndrome, necrotizing enterocolitis, and perinatal asphyxia. Institutionally, absence of isolated preterm rooms, inadequate physician staffing and distant NICU location significantly increased mortality risk.
Conclusions:
LBW mortality in Southern Ethiopia is alarmingly high. Both clinical comorbidities and modifiable institutional factors, especially infrastructure and staffing, are critical determinants. Urgent improvements in neonatal care facilities and human resources are needed.
Keywords: individual level, institutional level, neonatal intensive care unit, random effect
Introduction
Low-birth-weight (LBW), defined as the weight of a newborn less than 2500 g, is a significant contributor to neonatal mortality worldwide, particularly in low-and middle-income countries (LMICs).1,2 LBW infants are vulnerable to feeding problems, hypoglycemia, hypothermia, respiratory problems, infections, and fluid and electrolyte imbalances, which significantly increase the risk of death within the first 28 days of life.3,4 Approximately 2.3 million neonatal deaths occurred globally in 2022, equivalent to 6300 deaths per day. The leading causes include prematurity and /or LBW, birth asphyxia, congenital anomalies, and sepsis. 5
Low- and middle-income countries (LMICs) experience a higher percentage of LBW neonates, resulting in a greater burden of neonatal morbidity and mortality during the neonatal period. 6 Sub-Saharan Africa has also seen only a 34% reduction in neonatal mortality rates since 2000, underscoring the urgent need for targeted interventions to address this pressing issue. 5 LBW infants account for a substantial proportion of neonatal deaths in this region, with 80% of neonatal deaths occurring in this group globally. 7 In Ethiopia, the proportion of mortality attributable to LBW varies from 3.63% to 4.53% of all deaths.1,8,9
Although improvements have been made in neonatal care, mortality among LBW infants remains high. 10 Besides, newborn survivors are more likely to have stunted growth and a reduced IQ in the early years of development.11,12 Beyond the immediate medical consequences, it imposes significant emotional and financial stress on families and long-term economic burdens on health systems.13 -15 This all highlights that identifying factors related to LBW mortality is critical to avert this burden. However, existing evidence conducted in Southern Ethiopia mainly focused on individual-level factors, overlooking the influence of the healthcare system and facility-level determinants.16 -18 This presents a critical gap, as most LBW infants are admitted to neonatal intensive care units (NICUs), where institutional factors such as staffing ratios, infrastructure, and proximity to labor wards may significantly influence outcomes.
Recognizing this gap, the current study aims to identify the factors associated with mortality in neonates with low-birth-weight in Southern Ethiopia. By incorporating a multilevel analytical approach, this study seeks to provide comprehensive evidence to inform policymakers, guide program implementers, and promote timely interventions to improve neonatal survival in the study area and other similar settings.
Methods
Study Design, Study Area, and Period
A facility-based cohort study design was employed between September 2023 and February 2024 in 8 public hospitals across the Gedeo Zone and Sidama Region in Southern Ethiopia. These hospitals (Adarye General Hospital, Yirgalem General Hospital, Kebado Primary Hospital, Aleta Wondo Primary Hospital, Leku Primary Hospital, Dilla University Referral Hospitals, Hawassa University Referral Hospitals, and Yirgacheffe Hospital) represent key referral and primary care centers serving large and diverse catchment populations. The study sites were selected to capture both urban and rural healthcare settings, ensure adequate patient flow for follow-up, and reflect the regional variations in neonatal care practices. All participating facilities provide routine maternal and newborn services, including delivery care, neonatal observation units, and essential newborn care, enabling consistent data collection throughout the study period.
Study Population
The study included all LBW neonates (birth weight [BW] <2500 g, gestational age [GA] ≥28 weeks) who were admitted to the NICUs of the selected hospitals. Neonates with major congenital anomalies or GA <28 weeks were excluded, as infants below 28 weeks have extremely low viability in this setting due to limited advanced respiratory support, and their inclusion would markedly inflate mortality estimates.
Sample Size and Sampling Procedure
Sample size was calculated using 2 approaches. First, using the single proportion formula with a 37.8% mortality rate, 19 95% confidence interval, and a 5% margin of error, we obtained n = 362. After adjusting for 10% non-response and a design effect of 2, this yielded 797. Second, using the double population proportion formula and considering different predictors significantly associated with low-birth-weight mortality from previous studies (prematurity, young maternal age, unimmunized status, and infant birth weight). 20 The largest required sample size was 384; adding 10% for non-response and applying a conservative design effect of 2 (due to the absence of ICC estimates in prior studies) yielded a final sample size of 845. This higher figure (845) was adopted. Participants were proportionally allocated across the 8 hospitals. Consecutive recruitment continued until the target was approached, resulting in 768 neonates with complete data included in the analysis (Figure 1).
Figure 1.
Schematic representation of sampling procedure at public hospitals of Gedeo zone and Sidama region, South Ethiopia, 2024.
Variables of the Study
Dependent variables : Neonatal mortality (coded 1 = dead, 0 = alive).
Independent variables were categorized into individual and institutional-level factors.
Individual-level factors: Maternal age, antenatal follow-up, types of gestation, parity, gestational hypertension, steroid use, place of delivery, mode of delivery, neonatal sex, GA, BW, and comorbidities (Sepsis, Respiratory Distress Syndrome, Necrotizing Enterocolitis, Perinatal Asphyxia, Hypoglycemia, and Hypothermia).
Institutional-level factors : Availability of isolated preterm rooms, infection prevention quality, patient-physician ratio, patient-nurse ratio, and NICU distance from labor ward (categorized as near, far, very far).
Operational Definition
Level of NICU: Classified according to the facility’s capacity to provide neonatal care, adapted from the Ethiopian national neonatal management guidelines: Level I (Basic Neonatal Care), Level II (Special Care Nursery), and Level III (Intensive Neonatal Care). 21
Isolated room: A dedicated room with incubators and individual care equipment for preterm/LBW. 21
Infection prevention: “Good” if facilities avoided overcrowding, had clean equipment, hand washing supplies, and weekly cleaning; otherwise “poor.”
Patient-physician ratio: “Adequate” if at least one certified pediatrician and one resident were always available; otherwise “inadequate..” 22
Nurse-to-patient ratio: “Adequate” in Level III NICUs: One nurse assigned to care for no more than 1 critically ill neonate (1:1 ratio), in Level II NICUs: One nurse assigned to care for no more than 3 moderately ill neonates (1:3 ratio) and in Level I units: One nurse assigned to care for up to 5 stable neonates (1:5 ratio); otherwise “inadequate.” 22
NICU distance: refers distance between the Labor Ward and the NICU. Categorized in “Near” = adjacent to labor ward; “Far” = separate building, same compound (50-200 m); “Very far” = off-site or external facility (>200 m).
Data Collection Tool and Procedure
A structured questionnaire was developed based on extensive literature review8,19 -23 and pre-tested on 5% of the sample for validity and reliability, with subsequent refinements (Supplemental file 1). Twelve trained data collectors, who received 2 days of standardized training, used the pre-tested tool to abstract data at both admission and discharge.
All neonates with BW ≤2500 g and GA ≥28 weeks admitted to the NICUs of the 8 selected hospitals between September 2023 and February 2024 were consecutively screened. After obtaining informed written consent from parents/guardians, maternal, neonatal, and facility-related variables were collected within 24 hours of admission. Neonatal comorbidities and final outcomes (survival or death) were recorded before the medical records were archived. Each participant was assigned a unique identification code to ensure confidentiality and accurate follow-up.
Data Processing and Analysis
Data were entered and cleaned using EpiData version 4.6 and analyzed with STATA version 16.0. Descriptive statistics were used to summarize the data. Multilevel logistic regression with random intercept was applied to account for clustering at the hospital level. Variables with P < .2 in bi-variable analysis were entered into the multivariable model. Four sequential models were constructed: null model (intercept only), Model I (individual-level factors), Model II (institutional-level factors), and Model III (both individual and institutional factors).
Adjusted odds ratios (AOR) with 95% confidence intervals were reported; statistical significance was declared at P < .05. Model comparison and selection were based on log-likelihood, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Multicollinearity was assessed using variance inflation factor (mean VIF = 1.99; all <10). The intra-cluster correlation coefficient (ICC) was 10.3%, confirming the need for multilevel modeling. Median odds ratio (MOR) was calculated to quantify between-hospital variation.
Ethical Approval and Informed Consent
Ethical approval for this study was obtained from the appropriate Institutional Review Board. All study procedures were conducted in accordance with the principles of the Declaration of Helsinki and national ethical guidelines for health research. Written informed consent was obtained from the parents or legal guardians of all participating neonates after providing clear information about the study objectives, procedures, voluntary participation, and the right to withdraw at any time without affecting clinical care. Participant confidentiality was strictly maintained through the use of anonymized unique identifiers, and no personal identifiers were recorded. Electronic data were stored in password-protected files accessible only to the research team.
Results
Individual and Institution-Related Characteristics of Study Participants
The initial sample size was 845. During data cleaning, participants with incomplete or invalid questionnaires, non-response, or other exclusion criteria were removed, resulting in 768 (91%) participants included in the final analysis. Of these, 217 (28.3%) died during their hospital stay. About one-fourth of the dead low-birth-weight babies were born to mothers who had not received antenatal follow-up, and 76.2% of the babies were singleton pregnancies.
One hundred twenty-one (55.8%) of the neonates were male. Respiratory distress syndrome and hypothermia were the most common diagnoses at admission. The average gestational age for the entire group was 34.3 weeks (standard deviation, 2.7 weeks), and the average birth weight was 1825.6 g (standard deviation, 390.8 g). For those who died, the average birth weight was 1663.7 g (with a standard deviation of 431.7 g), and the average gestational age was 33.1 weeks (with a standard deviation of 3.1 weeks; Table 1).
Table 1.
Individual Characteristics of Dead Low-Birth Weight Infants Admitted to Public Hospitals’ NICUs in Southern Ethiopia, 2023 (n = 768).
| Covariates | Discharge alive | Died | ||||
|---|---|---|---|---|---|---|
| N (%) | Total # of deaths, N (%) | Level 1 NICUs, N (%) | Level 2 NICUs, N (%) | Level 3 NICUs, N (%) | ||
| Age of the mother in years mean (Sd) | 27 (+0.21) | 26.7 (±0.34) | ||||
| Antenatal follow-up | No | 81 (14.7) | 55 (25.4) | 3 (5.45) | 23 (41.82) | 29 (52.73) |
| Yes | 470 (85.3) | 162 (74.6) | 11 (6.79) | 51 (31.48) | 100 (61.73) | |
| Types of gestation | Single | 420 (76.2) | 156 (71.9) | 12 (7.69) | 57 (36.54) | 87 (55.77) |
| Multiple | 131 (23.8) | 61 (28.1) | 2 (3.28) | 17 (27.87) | 42 (68.85) | |
| Parity | Primipara | 219 (37.7) | 106 (48.9) | 10 (9.43) | 34 (32.08) | 62 (58.49) |
| Multipara | 304 (55.2) | 102 (47) | 4 (3.92) | 37 (36.27) | 61 (59.8) | |
| Grand multipara | 28 (5.1) | 9 (4.1) | 0 | 3 (33.33) | 6 (66.67) | |
| Gestational hypertension | No | 505 (91.6) | 181 (83.4) | 11 (6.08) | 63 (34.81) | 107 (59.12) |
| Yes | 46 (8.4) | 36 (16.4) | 3 (8.33) | 11 (30.56) | 22 (61.11) | |
| Taking steroid | No | 511 (92.7) | 193 (88.9) | 11 (5.7) | 65 (33.68) | 117 (60.62) |
| Yes | 40 (7.3) | 24 (11.1) | 3 (12.5) | 9 (37.5) | 12 (50) | |
| Place of delivery | Home | 25 (4.5) | 16 (7.4) | 2 (12.5) | 6 (37.5) | 8 (50) |
| Health facility | 526 (95.5) | 201 (92.6) | 12 (5.97) | 68 (33.83) | 121 (60.2) | |
| Mode of delivery | Vaginal delivery | 444 (80.6) | 160 (73.7) | 9 (5.63) | 58 (36.25) | 93 (58.13) |
| Cesarean delivery | 107 (19.4) | 57 (26.3) | 5 (8.77) | 16 (28.07) | 36 (63.16) | |
| Sex of baby | Male | 292 (52.9) | 121 (55.7) | 7 (5.79) | 39 (32.23) | 75 (61.98) |
| female | 259 (47.1) | 96 (44.3) | 7 (7.29) | 35 (36.46) | 54 (56.25) | |
| Age of newborn in days mean (Sd) | 1.73 (±0.11) | 1.85 (±0.19) | ||||
| Gestational age in weeks mean (Sd) | 34.6 (±0.11) | 33.1 (±0.21) | ||||
| Birth weight in grams mean (Sd) | 1889.4 (±15.1) | 1663.7 (±29.3) | ||||
| Sepsis | No | 237 (43.1) | 94 (43.3) | 9 (9.57) | 27 (28.72) | 58 (61.7) |
| Yes | 314 (56.9) | 123 (56.7) | 5 (4.07) | 47 (38.21) | 71 (57.72) | |
| Respiratory distress syndrome | No | 405 (73.5) | 108 (49.7) | 9 (8.33) | 42 (38.89) | 57 (52.78) |
| Yes | 146 (26.5) | 109 (50.3) | 5 (4.59) | 32 (29.36) | 72 (66.06) | |
| Necrotizing enterocolitis | No | 528 (96) | 184 (84.8) | 12 (6.52) | 60 (33.61) | 112 (60.8) |
| Yes | 22 (4) | 33 (15.2) | 2 (6.06) | 14 (42.42) | 17 (51.52) | |
| Perinatal asphyxia | No | 481 (87.5) | 155 (71.4) | 11 (7.1) | 56 (36.13) | 88 (56.77) |
| Yes | 69 (12.5) | 62 (28.6) | 3 (4.84) | 18 (29.03) | 41 (66.13) | |
| Hypoglycemia | No | 443 (80.4) | 159 (73.3) | 9 (5.66) | 59 (37.11) | 91 (57.23) |
| Yes | 108 (19.6) | 58 (16.7) | 5 (8.62) | 15 (25.86) | 38 (65.52) | |
| Hypothermia | No | 358 (64.9) | 121 (55.8) | 10 (8.26) | 47 (38.84) | 64 (52.89) |
| Yes | 193 (35.1) | 96 (44.2) | 4 (4.17) | 27 (28.13) | 65 (67.71) | |
Regarding hospital factors, over a third (72 LBW babies) of the deceased low-birth-weight infants were admitted to the same room as normal-weight or term babies, meaning they shared the same ward. Additionally, 27 LBW babies experienced poor infection prevention practices, such as inadequate hygiene protocols. Furthermore, nearly 30% of the low-birth-weight babies died in hospitals with a low neonate-to-physician ratio (Table 2).
Table 2.
Health Facilities Related Characteristics of Died Low-Birth Weight Babies Admitted in Public Hospitals NICUs in Southern Ethiopia, 2023 (n = 768).
| Covariates | Total, (N) | Died, N (%) | Level 1 NICUs, N (%) | Level 2 NICUs, N (%) | Level 3 NICUs, N (%) | |
|---|---|---|---|---|---|---|
| Isolated preterm room | Available | 562 | 133 (23.7) | 3 (2.26) | 63 (47.37) | 67 (50.38) |
| Not Available | 206 | 84 (40.8) | 11 (13.1) | 11 (13.1) | 62 (73.81) | |
| Infection prevention | Good | 654 | 187 (28.6) | 8 (4.28) | 58 (31.2) | 121 (64.71) |
| Poor | 114 | 30 (26.3) | 6 (20) | 16 (53.33) | 8 (26.67) | |
| Patient - physician ratio | Adequate | 582 | 152 (26.1) | 12 (7.89) | 57 (37.5) | 83 (54.61) |
| Inadequate | 186 | 65 (34.9) | 2 (3.08) | 17 (26.15) | 46 (70.77) | |
| Patient - nurse ratio | Adequate | 247 | 74 (29.9) | 7 (9.46) | 17 (22.97) | 50 (67.57) |
| Inadequate | 521 | 143 (27.4) | 7 (4.9) | 57 (39.86) | 79 (55.24) | |
| NICU distance from labor ward | Near | 139 | 25 (17.9) | 6 (24) | 9 (36) | 10 (40) |
| Far | 412 | 113 (27.4) | 8 (7.08) | 35 (30.97) | 70 (61.95) | |
| Too far | 217 | 79 (36.4) | 0 | 30 (37.97) | 49 (62.03) | |
Model Comparison and Random Effect Analysis
The null model revealed an intraclass correlation coefficient (ICC) of 10.3%, indicating that 10.3% of the total variation in LBW mortality was attributable to differences between the 8 NICUs. The median odds ratio (MOR) was 1.79, meaning that, if 2 identical neonates were randomly placed in different NICUs, the one in the higher-risk facility had 1.79 times higher odds of death. Model comparison using log-likelihood, AIC, and BIC identified the full model containing both individual and institutional factors (Model III) as the best-fitting model (AIC 728.74, BIC 835.51, log-likelihood −341.37; Table 3).
Table 3.
Measures of Variation (Random Intercept Models) and Model Fit Statistics.
| Model measures | Model-0 | Model-I | Model-II | Model-III |
|---|---|---|---|---|
| Random effect | ||||
| ICC (%) | 10.3 | 8.5 | 6.6 | 6.3 |
| Variance (95% CI), (SE) | 0.37 (0.06-2.32), (0.35) | 0.31 (0.04-2.01), (0.29) | 0.23 (0.04-1.36), (0.21) | 0.22 (0.04-1.39), (0.2) |
| MOR (%) | 1.79 | 1.69 | 1.57 | 1.56 |
| Model comparison | ||||
| –2Log likelihood | −445.64 | −368.07 | −406.42 | −341.37 |
| LR test | χ2 = 23.17*** | χ2 = 16.5 *** | χ2 = 22.9*** | χ 2 = 15.9*** |
| AIC | 895.28 | 770.14 | 828.83 | 728.74 |
| BIC | 904.57 | 849.06 | 865.98 | 835.51 |
Abbreviations: LR, likelihood ratio test; SE, standard error; ICC, intraclass correlation coefficient; MOR, median odds ratio; AIC, Akaike’s information criterion; BIC, Bayesian information criterion.
P < 0.001.
The final model demonstrated excellent performance: bootstrap-corrected AUC 0.835 (95% CI 0.793-0.901), near-perfect calibration (slope 0.98, intercept −0.05; Supplemental figure 1), and robustness in sensitivity analyses. Residual diagnostics and leave-one-cluster-out analyses confirmed no undue influence of outliers or individual hospitals. Residual between-NICU heterogeneity remained modest (random-effect SD 0.42; MOR ≈ 1.5). These results indicate that the model fits the data well and reliably identifies determinants of neonatal mortality.
Multilevel Logistic Regression Analysis
The fixed-effect multilevel logistic regression model revealed significant individual- and group-level predictors of mortality among low-birth-weight (LBW) neonates (P < .05).
Individual-Level Predictors
Several clinical and demographic factors at the individual level were significantly associated with neonatal mortality. For each 100 g increment in birth weight, the odds of mortality decreased by 16% (AOR 0.84, 95% CI 0.79-0.89). Low-birth-weight babies diagnosed with respiratory distress syndrome had nearly twice the odds of dying compared to those without the illness (AOR = 1.94, 95% CI: 1.27, 2.94). Similarly, the presence of necrotizing enterocolitis and perinatal asphyxia increased the risk of mortality threefold (AOR = 3.02, 95% CI: 1.54-5.87) and 3.2 times (AOR = 3.18, 95% CI: 1.94-5.22), respectively.
Institutional-Level Predictors
Different institutional factors also contributed significantly to mortality outcomes. Babies admitted to the same room as other neonates were 4.6 times more likely to die compared to those LBW admitted to the isolated rooms (AOR = 4.61, 95% CI: 2.83, 7.49). Additionally, LBW infants admitted to NICUs with an inadequate neonate-physician ratio had 3.7 times higher odds of mortality (AOR = 3.68, 95% CI: 1.53, 8.88). Regarding NICU distance, neonates delivered in units farther from the labor ward faced increased risks, with those in NICUs far from the labor ward being 2.9 times more likely to die (AOR = 2.94, 95% CI: 1.34, 6.45), and those in units considered “too far” having an even greater risk, at 8.5 times more likely to die (AOR = 8.51, 95% CI: 3.51, 20.6; Table 4).
Table 4.
Multivariable Multilevel Logistic Regression Analysis Results of Both Individual and Health Institution-Level Factors Associated With Mortality of LBW babies in South Ethiopia 2023 (n = 768).
| Category | Adjusted odds ratio (95% CI) | |||
|---|---|---|---|---|
| Model-0 | Model-I | Model-II | Model-III | |
| Individual related factors | ||||
| Gestational age | 0.91 (0.85, 0.98) * | 0.95 (0.88, 1.03) | ||
| Birth weight | 0.87 (0.82, 0.92) *** | 0.84 (0.79, 0.89) *** | ||
| ANC visit (yes) | ||||
| No | 1.70 (1.06, 2.74)* | 1.51 (0.91, 2.47) | ||
| Types of gestation (single) | ||||
| Multiple | 1.37 (0.89, 2.11) | 1.28 (0.82, 2.03) | ||
| Taking steroid (no) | ||||
| Yes | 1.28 (0.67, 2.45) | 1.27 (0.65, 2.50) | ||
| Parity (grand multi) | ||||
| Primi-para | 1.77 (0.68, 4.58) | 1.67 (0.63, 4.43) | ||
| Multi-para | 1.09 (0.42, 2.81) | 0.98 (0.37, 2.61) | ||
| Gestational hypertension (no) | ||||
| Yes | 1.58 (0.87, 2.86) | 1.43 (0.76, 2.69) | ||
| Place of delivery (home) | ||||
| Health facility | 0.52 (0.24, 1.12) | 0.60 (0.26,1.35) | ||
| Mode of delivery (vaginal delivery) | ||||
| Cesarean section | 1.75 (1.09, 2.83)* | 1.42 (0.86, 2.35) | ||
| Respiratory distress syndrome (no) | ||||
| Yes | 1.97 (1.34, 2.89)** | 1.94 (1.27, 2.94)** | ||
| Necrotizing enterocolitis (no) | ||||
| Yes | 3.54 (1.84, 6.84)*** | 3.02 (1.54, 5.87)** | ||
| Perinatal asphyxia (no) | ||||
| Yes | 3.31 (2.05, 5.29)*** | 3.18 (1.94, 5.22)*** | ||
| Hypoglycemia (no) | ||||
| Yes | 0.98 (0.61, 1.58) | 0.87 (0.53, 1.43) | ||
| Hypothermia (no) | ||||
| Yes | 1.03 (0.68, 1.57) | 0.85 (0.54, 1.35) | ||
| Institutional-related factors | ||||
| Isolated preterm room (available) | ||||
| Not Available | 4.06 (2.66, 6.19)*** | 4.61 (2.83,7.49)*** | ||
| Infection prevention (good) | ||||
| Poor | 0.92 (0.56, 1.49) | 1.16 (0.67, 1.99) | ||
| Patient physician ratio (adequate) | ||||
| Inadequate | 5.45 (2.44, 12.2)*** | 3.68 (1.53, 8.88)** | ||
| Patient nurse ratio (adequate) | ||||
| Inadequate | 2.08 (1.01, 4.32)* | 1.79 (0.81, 3.94) | ||
| NICU distance (near) | ||||
| Far | 3.47 (1.76, 6.81)*** | 2.94 (1.34, 6.45)** | ||
| Too far | 10.78 (5.1, 22.8)*** | 8.51 (3.51, 20.6)*** | ||
Abbreviation: NICU, neonatal intensive care unit.
P < 0.05. **P < 0.01. ***P < 0.001.
Discussion
This study found that 1 out of 4 low-birth-weight (LBW) infants died (28.3%) among those admitted to public hospitals in Southern Ethiopia. Also this study revealed that both individual and institution-related factors associated with LBW neonatal outcomes. Key predictors included birth weight, respiratory distress syndrome, necrotizing enterocolitis, perinatal asphyxia, isolated preterm room, patient-physician ratio, and NICU distance from the labor ward, all linked to death of LBW neonates.
The results indicated an inverse association between birth weight and neonatal mortality, consistent with earlier research.1,24,25 This finding suggests that LBW infants, who are more physiologically vulnerable, may require closer monitoring and timely supportive care. Conditions such as hypoglycemia, hypothermia, hyperbilirubinemia, and coagulation abnormalities are more common among LBW neonates and have been associated with poorer outcomes when not promptly recognized and managed.2,26 These patterns underscore the need for heightened clinical attention in the care of LBW infants, particularly during the early neonatal period.
Neonatal illnesses such as respiratory distress syndrome (RDS), necrotizing enterocolitis (NEC), and perinatal asphyxia (PNA) were key factors associated with mortality in our study, consistent with previous findings.8,23,27 -29 RDS is more common in premature babies and requires prompt treatment like surfactant therapy and breathing support to prevent serious complications such as hypoxic-ischemic encephalopathy8,23 Likewise, NEC and PNA pose serious management challenges for LBW infants. NEC complications, including extended restrictions on enteral feeding, can be associated with higher mortality, highlighting the need for standardized feeding protocols and close monitoring.29 -31 PNA, caused by poor oxygen supply during or after birth, stresses the importance of quick resuscitation and stabilization efforts2,23,32 These findings highlight the need to address these health issues and improve access to specialized neonatal care through stronger healthcare policies.
In addition to clinical conditions, institutional-related factors were shown to play a key role in neonatal outcomes. This study found that LBW infants admitted to isolated preterm rooms had higher survival rates compared to those placed in shared rooms with term babies. This finding aligns with existing literature, which suggests that separate room settings have been associated with fewer healthcare-associated infections and lower mortality. Moreover, it promotes parental involvement and enhances both clinical and psychosocial outcomes30,31,33,34 Additionally, an inadequate neonate-to-physician ratio was linked to increased mortality among LBW infants, highlighting the importance of sufficient staffing for timely decision-making and continuous follow-up. In Southern Ethiopia, where healthcare resources are often limited, improving staffing levels and training specialized neonatal care providers should be prioritized.
Another striking finding was the strong association between NICU distance from the labor ward and mortality, with odds rising up to 8.5-fold when the NICU was very far. This likely reflects critical delays in initiating specialized care, increased risk of hypothermia, and unsafe transport conditions; factors repeatedly linked to adverse outcomes in resource-limited settings.10,35 Although kangaroo mother care during transport is recommended, its implementation remains inconsistent in the study region. The observed association may partly be explained by residual confounding from unmeasured variables, including referral patterns, illness severity at birth, or other hospital characteristics that co-determine both NICU location and mortality risk. Nonetheless, the findings strongly support urgent infrastructural redesign to co-locate delivery and neonatal intensive care services, alongside establishment of standardized, hypothermia-preventing neonatal transport protocols.
Limitations
This hospital-based study may not fully represent community-managed LBW infants. Important variables such as maternal socioeconomic status and care-seeking behavior were not collected, potentially leading to residual confounding. Due to unavailability of precise time-to-death records, survival analysis could not be performed, limiting adjustment for differing durations of exposure to risk. The observed intra-cluster correlation (ICC = 10.3%) exceeded the conservative design effect used in sample size calculation, which may have reduced statistical power. Finally, with only 8 hospitals (clusters), estimates of facility-level effects, although derived from multilevel modeling, may have limited precision and generalizability. These limitations warrant cautious interpretation with caution, particularly for institutional factors.
Conclusion
This multicenter study identifies critical individual- and facility-level determinants of mortality among low-birth-weight neonates in Southern Ethiopia. Key clinical predictors include lower birth weight, respiratory distress syndrome, necrotizing enterocolitis, and perinatal asphyxia. Equally important institutional factors were the absence of dedicated preterm rooms, inadequate physician staffing, and distant NICU location from labor wards. Reducing LBW mortality demands concurrent improvements in clinical management and health-system readiness. Priority actions include establishing isolated preterm care areas, optimizing staffing ratios, ensuring close NICU proximity to delivery units, strengthening referral systems, and enhancing infection prevention. Future research should explore community-level determinants, socioeconomic factors, and care-seeking behavior and evaluate the impact of scalable NICU organizational and infrastructural interventions across Ethiopia.
Supplemental Material
Supplemental material, sj-pdf-1-gph-10.1177_30502225251414623 for Neonatal Mortality in Low-Birth-Weight Infants in Southern Ethiopia: A Multicenter Cohort Study With Multilevel Analysis by Jemberu Nigussie, Fekadeselassie Belege Getaneh, Dires Birhanu Mihretie, Bekahegn Girma, Tewodros Mulugeta, Endalk Birrie Wondifraw, Alemu Gedefie, Yibeltal Asmamaw Yitayew and Getu Degu Alene in Sage Open Pediatrics
Supplemental material, sj-png-2-gph-10.1177_30502225251414623 for Neonatal Mortality in Low-Birth-Weight Infants in Southern Ethiopia: A Multicenter Cohort Study With Multilevel Analysis by Jemberu Nigussie, Fekadeselassie Belege Getaneh, Dires Birhanu Mihretie, Bekahegn Girma, Tewodros Mulugeta, Endalk Birrie Wondifraw, Alemu Gedefie, Yibeltal Asmamaw Yitayew and Getu Degu Alene in Sage Open Pediatrics
Acknowledgments
The authors would like to thank the parents or guardians of all study participants, the NICU healthcare professionals, and the hospital management teams at the participating facilities for their cooperation and support during the study.
Footnotes
ORCID iDs: Jemberu Nigussie 
https://orcid.org/0000-0002-2003-8933
Fekadeselassie Belege Getaneh
https://orcid.org/0000-0001-8729-1784
Dires Birhanu Mihretie
https://orcid.org/0000-0002-1492-3733
Yibeltal Asmamaw Yitayew
https://orcid.org/0000-0003-4161-3500
Ethical Considerations: The study was approved by the Institutional Review Board of the College of Medicine and Health Sciences, Dilla University (Ref: duchm/irb/059/2023). The study was conducted in accordance with the Declaration of Helsinki and national ethical guidelines for health research in Ethiopia.Written informed consent was obtained from the parents or legal guardians of all participating neonates after providing clear information about the study objectives, procedures, voluntary participation, and the right to withdraw at any time without affecting clinical care. Participant confidentiality was strictly maintained through the use of anonymized unique identifiers, and no personal identifiers were recorded. Electronic data were stored in password-protected files accessible only to the research team.
Consent to Participate: Written informed consent was obtained from the parents or legal guardians of all participating neonates prior to data collection.
Author Contributions: JN, FBG, BG, and DBM conceived and designed the study, participated in data collection, conducted data analysis, and drafted the manuscript. TM, YA, and FBG contributed to data analysis and manuscript preparation. EBW, AG, and GDA critically reviewed the manuscript and provided substantial intellectual input. All authors reviewed and approved the final version of the manuscript.
Funding: The authors received no financial support for the research, authorship, and/or publication of this article.
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement: The datasets generated and/or analyzed during the current study are not publicly available due to ethical and confidentiality considerations but are available from the corresponding author upon reasonable request.*
Patient and Public Involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Supplemental Material: Supplemental material for this article is available online.
References
- 1. Alemu AA, Aynalem BY. Low birth weight in Ethiopia: a multilevel analysis of 2016 Ethiopian demographic and health survey PAMJ - One Health. 2022;9:1-11. [Google Scholar]
- 2. Kliegman R, Geme JS, Stanton B, Schor N, eds. (eds) Nelson Textbook of Pediatrics. 21th ed. Elsevier; 2020. [Google Scholar]
- 3. Blencowe H, Krasevec J, Onis MD, et al. National, regional, and worldwide estimates of low birthweight in 2015, with trends from 2000: a systematic analysis. Lancet Glob Health. 2015;7(7):e849-e860. doi: 10.1016/S2214-109X(18)30565-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tesfaye N. Clinical Reference Manual for Advanced Neonatal Care in Ethiopia. Federal Democratic Republic of Ethiopia, Ministry of Health; 2021. [Google Scholar]
- 5. Hug L, Liu Y, Nie W, et al. Levels & Trends in Child Mortality. Report 2023. 2024. https://data.unicef.org/resources/levels-and-trends-in-child-mortality-2024/
- 6. Katz J, Lee AC, Kozuki N, et al. Mortality risk in preterm and small-for-gestational-age infants in low-income and middle-income countries: a pooled country analysis. Lancet. 2013;382(9890):417-425. doi: 10.1016/s0140-6736(13)60993-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Dhaded SM, Somannavar MS, Vernekar SS, et al. Neonatal mortality and coverage of essential newborn interventions 2010 - 2013: a prospective, population-based study from low-middle income countries. Reprod Health. 2015;12(2):S6. doi: 10.1186/1742-4755-12-S2-S6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kebede BF, Genie YD, Aregawi DH, Tadele BA. Survival status and predictors of mortality among low birthweight neonates admitted in Amhara region referral hospitals of Ethiopia: retrospective follow-up study. Health Serv Res Manag Epidemiol. 2022;9:1-9. doi: 10.1177/23333928221117364 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. World Health Organization. World Health Statistics. World Health Organization (WHO); 2017. [Google Scholar]
- 10. World Health Organization. WHO Recommendations for Care of the Preterm or Low-Birth-Weight Infant. World Health Organization (WHO); 2022. [PubMed] [Google Scholar]
- 11. Gu H, Wang L, Liu L, et al. A gradient relationship between low birth weight and IQ: a meta-analysis. Sci Rep. 2017;7(1):1-13. doi: 10.1038/s41598-017-18234-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Christian P, Lee SE, Donahue Angel M, et al. Risk of childhood undernutrition related to small-for-gestational age and preterm birth in low- and middle-income countries. Int J Epidemiol. 2013;42:1340-1355. doi: 10.1093/ije/dyt109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Petrou S, Sach T, Davidson L. The long-term costs of preterm birth and low birth weight: results of a systematic review. Child Care Health Dev. 2001;27:97-115. Published online 2000. [DOI] [PubMed] [Google Scholar]
- 14. Sicuri E, Bardají A, Sigauque B, et al. Costs associated with low birth weight in a rural area of southern Mozambique. PLoS One. 2011;6(12):e28744. doi: 10.1371/journal.pone.0028744 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Victor A, Wesley I, Aguiar O, et al. Social inequalities in child development: analysis of low-birth-weight trends in Brazil, 2010-2020. J Prev. 2022. [DOI] [PubMed] [Google Scholar]
- 16. Jemal B, Abebe T, Zemedkun A, et al. Determinants of neonatal mortality in the neonatal intensive care unit of Dilla University referral hospital, southern Ethiopia; 2019-2020; a matched, case-control study. Heliyon. 2022;8(9):e10389. doi: 10.1016/j.heliyon.2022.e10389 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Bogale A, Borko UD, Abreha S, Balta B. Determinants of mortality among preterm neonates admitted at Wolaita Sodo University comprehensive specialized hospital, southern Ethiopia: an unmatched case-control study. PLoS One. 2024;19(12):e0314632. doi: 10.1371/journal.pone.0314632 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wondie WT, Zeleke KA, Wubneh CA. Incidence and predictors of mortality among low birth weight neonates in the first week of life admitted to the neonatal intensive care unit in northwestern Ethiopia comprehensive specialized hospitals, 2022. Multi-center institution-based retrospective follow-up study. BMC Pediatr. 2023;23(1):13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Genie YD, Kebede BF, Silesh Zerihun M, Tilahun Beyene D. Morbidity and mortality patterns of preterm low birthweight neonates admitted to referral hospitals in the Amhara region of Ethiopia: retrospective follow-up study. BMJ Open. 2022;12(7):1-8. doi: 10.1136/bmjopen-2021-054574 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Coulibaly A, Baguiya A, Millogo T, Meda IB, Koueta F, Kouanda S. Predictors of mortality of low birth weight newborns during the neonatal period: a cohort study in two health districts of Burkina Faso. Int J Gynecol Obstet. 2016;135:S89-S92. doi: 10.1016/j.ijgo.2016.08.006 [DOI] [PubMed] [Google Scholar]
- 21. Fmoh E. Clinical Reference Manual for Advanced Neonatal Care in Ethiopia Ministry of Health 2021. Ministry of health; 2021. [Google Scholar]
- 22. Chuan PTC. Guidelines for Hospitals with Neonatal Intensive Care Service: Regulation 4 of the Private Hospitals and Medical Clinics Regulations. Minster of Health; 2001. [Google Scholar]
- 23. Necho W, Alebachew W, Munye T, Amera T. Impact of respiratory distress syndrome and birth asphyxia exposure on the survival of preterm neonates in East Africa continent: systematic review and meta analysis. Heliyon. 2021;7:1-10. doi: 10.1016/j.heliyon.2021.e07256 [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 24. Michaelis IA, Krägeloh-Mann I, Manyisane N, Mazinu MC, Jordaan ER. Prospective cohort study of mortality in very low birthweight infants in a single centre in the Eastern Cape province, South Africa. BMJ Paediatr Open. 2021;5:1-7. Published online 2021. doi: 10.1136/bmjpo-2020-000918 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bührer C. Mortality and major morbidity of very-low-birth-weight infants in Germany 2008 – 2012: a report based on administrative data. Front Pediatr. 2016;4:1-8. doi: 10.3389/fped.2016.00023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Behrman RE, Harris JE, Hobel CJ, Klein JO, Luella Klein MCM. Preventing Low Birthweight. National Academy Press; 1985. [Google Scholar]
- 27. Sutrisno AA, Hikmahrachim HG, Hardiyanti N. Risk factors of necrotizing enterocolitis-related mortality and patient survival: a preliminary prospective stud. Paediatr Indones p-ISSN. 2022;62(3):186-191. doi: 10.14238/pi62.3.2022.186-91 [DOI] [Google Scholar]
- 28. de Almeida MF, Moreira LM, Vaz dos Santos RM, Kawakami MD, Anchieta LM, Guinsburg R. Early neonatal deaths with perinatal asphyxia in very low birth weight Brazilian infants. J Perinatol. 2015;35(11):954-957. doi: 10.1038/jp.2015.114 [DOI] [PubMed] [Google Scholar]
- 29. Olack B, Santos N, Inziani M, et al. Causes of preterm and low birth weight neonatal mortality in a rural community in Kenya: evidence from verbal and social autopsy. BMC Pregnancy Childbirth. 2021;21:536-539. Published online 2021. doi: 10.1186/s12884-021-04012-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Jansen S, Berkhout RJM, te Pas AB, et al. Comparison of neonatal morbidity and mortality between single-room and open-bay care: a retrospective cohort study. Arch Dis Child - Fetal Neonatal Ed. 2022;107(6):611-616. doi: 10.1136/archdischild-2021-323310 [DOI] [PubMed] [Google Scholar]
- 31. Manzionna M, Di Mauro A. Rooming-in organization to prevent neonatal mortality and morbidity in late preterm infants. Ital J Pediatr. 2014;40(2):A4. doi: 10.1186/1824-7288-40-S2-A4 [DOI] [Google Scholar]
- 32. Horn AR, Swingler GH, Harrison MC, et al. Defining hypoxic ischemic encephalopathy in newborn infants: benchmarking in a South African population. J Perinatal Med. 2012;41(2):211-217. doi: 10.1515/jpm-2012-0107 [DOI] [PubMed] [Google Scholar]
- 33. van Veenendaal NR, van Kempen AAMW, Franck LS, et al. Hospitalising preterm infants in single family rooms versus open bay units: a systematic review and meta-analysis of impact on parents. ECM. 2020;23:100388. doi: 10.1016/j.eclinm.2020.100388 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Pammi M, Davis RJ, Gordon A, Starke J. Infant isolation and cohorting for preventing or reducing transmission of healthcare-associated infections in neonatal units. Cochrane. 2016;12. doi: 10.1002/14651858.cd012458 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Bull MJ, Engle WA, Care P. Safe transportation of preterm and low birth weight infants at hospital discharge. Pediatrics. 2009;123(5):1424-1429. doi: 10.1542/peds.2009-0559 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material, sj-pdf-1-gph-10.1177_30502225251414623 for Neonatal Mortality in Low-Birth-Weight Infants in Southern Ethiopia: A Multicenter Cohort Study With Multilevel Analysis by Jemberu Nigussie, Fekadeselassie Belege Getaneh, Dires Birhanu Mihretie, Bekahegn Girma, Tewodros Mulugeta, Endalk Birrie Wondifraw, Alemu Gedefie, Yibeltal Asmamaw Yitayew and Getu Degu Alene in Sage Open Pediatrics
Supplemental material, sj-png-2-gph-10.1177_30502225251414623 for Neonatal Mortality in Low-Birth-Weight Infants in Southern Ethiopia: A Multicenter Cohort Study With Multilevel Analysis by Jemberu Nigussie, Fekadeselassie Belege Getaneh, Dires Birhanu Mihretie, Bekahegn Girma, Tewodros Mulugeta, Endalk Birrie Wondifraw, Alemu Gedefie, Yibeltal Asmamaw Yitayew and Getu Degu Alene in Sage Open Pediatrics