Answers are grounded in this paper and cite their sources.
BACKGROUND: In the Republic of Tajikistan, the prevalence of preterm birth has shown no downward trend in recent years. Early prediction and timely prevention therefore remain relevant strategies for reducing the incidence of preterm birth. Existing nomograms for short-term prediction of preterm birth rely on costly laboratory and instrumental methods and offer limited practical value in low-resource settings. Taking into account the regional characteristics of Tajikistan, the search for tools enabling the long-term prediction of preterm birth based on data collected at a pregnant woman’s first antenatal care appointment to primary health care facilities represents a relevant line of research. AIM: The aim of this study was to identify independent clinical, social, and anamnestic predictors of preterm birth in a population of pregnant women residing in Tajikistan and, based on these, to develop a prognostic model with a clinical risk calculator. METHODS: This retrospective case–control study included delivery records and prenatal care charts of 180 women with singleton pregnancies registered in the first trimester: 120 with preterm birth ( 37+0 weeks) and 60 with term birth (≥37+0 weeks). We studied clinical, social, and anamnestic factors recorded at the initial visit to a primary health care facility. Independent predictors were identified using multivariable logistic regression, with model performance assessed by ROC analysis and internal validation performed using the bootstrap resampling method (1,000 iterations). RESULTS: Multivariable logistic regression identified the following independent predictors of preterm birth: history of induced abortion [odds ratio (OR) 4.65; 95% confidence interval (CI) 1.47–14.74; p = 0.009], history of preterm birth (OR 3.73; 95% CI 1.53–9.08; p = 0.004), physical labor (OR 3.22; 95% CI 1.53–6.76; p = 0.002), anemia (OR 2.90; 95% CI 1.28–6.57; p = 0.011), and obesity (OR 2.65; 95% CI 1.15–6.11; p = 0.022); all these factors were incorporated into the model. Bacteriuria was included in the model due its clinical relevance (OR 2.90; 95% CI 0.95–8.84; p = 0.062). The resulting prognostic model demonstrated the following discriminatory performance: area under the ROC curve was 0.776 (95% CI 0.708–0.845), and overall accuracy was 74.4%. The optimal probability threshold, determined by the Youden index, was p = 0.615 (sensitivity 72.5%, specificity 73.3%). Internal validation using the bootstrap resampling method (1,000 iterations) confirmed the model robustness (95% CI for the area under the ROC curve 0.708–0.845). An online calculator derived from the model is designed to stratify the risk of preterm birth at the stage of the woman’s first prenatal appointment. CONCLUSION: Adaptation of the model to the regional conditions of Tajikistan makes it applicable to primary obstetric care settings in Central Asian countries with comparable socioeconomic profiles.
No takes yet. Share an insight, caveat, or question.
Ishan-Khodzhaeva et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: