Predictors of Pregnancy-Related Anemia: A Logistic Regression Study at a Maternity Facility in Ghana
- 1Tarleton State University
Abstract
BackgroundAnemia during pregnancy remains a major public health concern, particularly in low- and middle-income countries, where it contributes substantially to maternal and neonatal morbidity and mortality. Identifying women at increased risk is essential for timely intervention and improved pregnancy outcomes.
ObjectiveThis study aimed to identify significant predictors of anemia among pregnant women using logistic regression and to evaluate the association between selected clinical characteristics and anemia status.
MethodsA retrospective study was conducted using secondary data obtained from Ayiwa Memorial Maternity Home in the Suame Municipality of the Kumasi Metropolis, Ashanti Region, Ghana, covering the period 2018 to 2023. A sample of 396 pregnant women with complete information on hemoglobin status and candidate predictor variables was analyzed. The outcome variable was whether a woman had ever had anemia (hemoglobin concentration below 11 g/dL). Candidate predictors included maternal age, diastolic and systolic blood pressure, parity, height, weight, gestational age, sickle cell status, employment, and education. Multivariable logistic regression was fitted, with multicollinearity among predictors assessed using variance inflation factors (VIF) and tolerance statistics. Model performance was evaluated using the Hosmer-Lemeshow test, Nagelkerke R2, and the area under the receiver operating characteristic curve (AUC).
ResultsOf the 396 pregnant women studied, 69.44% were classified as anemic. In the multivariable logistic regression model, gestational age, maternal weight, and diastolic blood pressure were significantly associated with anemia status (p < 0.05). Higher gestational age was associated with increased odds of anemia (AOR = 1.124, 95% CI: 1.073-1.180), while higher weight (AOR = 0.975, 95% CI: 0.955-0.995) and higher diastolic blood pressure (AOR = 0.964, 95% CI: 0.934-0.993) were associated with reduced odds of anemia. Systolic blood pressure, height, and education were not significantly associated with anemia in the final model. The model showed acceptable discriminatory ability (AUC = 0.752, 95% CI: 0.694-0.810) and adequate calibration (Hosmer-Lemeshow{chi} 2 = 13.323, p = 0.101), explaining approximately 23.9% of the variation in anemia status (Nagelkerke R2 = 0.239), with high specificity (91.8%) but limited sensitivity (36.5%).
ConclusionGestational age, maternal weight, and diastolic blood pressure were identified as significant predictors of anemia during pregnancy in this population. These findings support intensifying anemia screening as pregnancy advances and incorporating nutritional assessment into routine antenatal care. However, given the models limited sensitivity, clinical risk models cannot replace routine laboratory hemoglobin testing. Further prospective, multi-center studies incorporating dietary, infectious, and socioeconomic determinants are recommended to validate these findings.
Related
- From survival to positive childbirth experiences: An integrative review of facility-based maternity care in South-East Asia.
- Maternity versus non‐maternity healthcare settings for the diagnosis of gestational diabetes or diabetes in pregnancy
- Blood biomarkers and penalized logistic regression for the prediction of complicated appendicitis in children: a retrospective cohort study
- Prevalence and risk factors of postmenopausal osteoporosis: a systematic review and meta-analysis based on multivariate logistic regression
- Hepatitis B vaccination coverage and factors associated with higher uptake among children aged 5–35 months in eight selected sub-Saharan African countries: a multilevel ordinal logistic regression analysis
BibTeX
@misc{kusi2026predictors,
title = {{Predictors of Pregnancy-Related Anemia: A Logistic Regression Study at a Maternity Facility in Ghana}},
author = {R. Y. Kusi and F. Y. Anyan and G. O. Agyekum},
howpublished = {medRxiv},
year = {2026},
month = jul,
doi = {10.64898/2026.07.16.26358280},
url = {https://www.medrxiv.org/content/10.64898/2026.07.16.26358280v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - Predictors of Pregnancy-Related Anemia: A Logistic Regression Study at a Maternity Facility in Ghana
AU - R. Y. Kusi
AU - F. Y. Anyan
AU - G. O. Agyekum
PY - 2026
DA - 2026/07/16
PB - medRxiv
DO - 10.64898/2026.07.16.26358280
UR - https://www.medrxiv.org/content/10.64898/2026.07.16.26358280v2
AB - BackgroundAnemia during pregnancy remains a major public health concern, particularly in low- and middle-income countries, where it contributes substantially to maternal and neonatal morbidity and mortality. Identifying women at increased risk is essential for timely intervention and improved pregnancy outcomes. ObjectiveThis study aimed to identify significant predictors of anemia among pregnant women using logistic regression and to evaluate the association between selected clinical characteristics and anemia status. MethodsA retrospective study was conducted using secondary data obtained from Ayiwa Memorial Maternity Home in the Suame Municipality of the Kumasi Metropolis, Ashanti Region, Ghana, covering the period 2018 to 2023. A sample of 396 pregnant women with complete information on hemoglobin status and candidate predictor variables was analyzed. The outcome variable was whether a woman had ever had anemia (hemoglobin concentration below 11 g/dL). Candidate predictors included maternal age, diastolic and systolic blood pressure, parity, height, weight, gestational age, sickle cell status, employment, and education. Multivariable logistic regression was fitted, with multicollinearity among predictors assessed using variance inflation factors (VIF) and tolerance statistics. Model performance was evaluated using the Hosmer-Lemeshow test, Nagelkerke R2, and the area under the receiver operating characteristic curve (AUC). ResultsOf the 396 pregnant women studied, 69.44% were classified as anemic. In the multivariable logistic regression model, gestational age, maternal weight, and diastolic blood pressure were significantly associated with anemia status (p < 0.05). Higher gestational age was associated with increased odds of anemia (AOR = 1.124, 95% CI: 1.073-1.180), while higher weight (AOR = 0.975, 95% CI: 0.955-0.995) and higher diastolic blood pressure (AOR = 0.964, 95% CI: 0.934-0.993) were associated with reduced odds of anemia. Systolic blood pressure, height, and education were not significantly associated with anemia in the final model. The model showed acceptable discriminatory ability (AUC = 0.752, 95% CI: 0.694-0.810) and adequate calibration (Hosmer-Lemeshow{chi} 2 = 13.323, p = 0.101), explaining approximately 23.9% of the variation in anemia status (Nagelkerke R2 = 0.239), with high specificity (91.8%) but limited sensitivity (36.5%). ConclusionGestational age, maternal weight, and diastolic blood pressure were identified as significant predictors of anemia during pregnancy in this population. These findings support intensifying anemia screening as pregnancy advances and incorporating nutritional assessment into routine antenatal care. However, given the models limited sensitivity, clinical risk models cannot replace routine laboratory hemoglobin testing. Further prospective, multi-center studies incorporating dietary, infectious, and socioeconomic determinants are recommended to validate these findings.
ER -