Diagnostic accuracy of clinical examination for detection of non-cephalic presentation in late pregnancy: cross sectional analytic study

Patients, setting, and data collection

We carried out the study at an antenatal clinic in a tertiary obstetric hospital between September 2003 and December 2004. Women with a singleton pregnancy at 35-37 weeks' gestation were eligible. A midwife, resident, registrar, or obstetrician, all of whom were aware of the study, provided routine antenatal care. All eligible women underwent clinical examination to assess fetal presentation. Subsequently, those who consented to participate underwent ultrasonography to confirm the diagnosis. The ultrasound examination was conducted with a portable hand held machine following a standard protocol. The operators were blinded to the result of the clinical examination until after the ultrasonography.

We collected data from the antenatal record and recorded it on a standard data abstraction form. We assessed the accuracy of clinical examination in diagnosing fetal presentation by calculating sensitivity, specificity, and positive and negative predictive values.

Sample size and analysis

To determine a sensitivity of 75% (with a 95% confidence interval plus or minus 10%) we required 100 women with a breech presentation. As 6-8% of singleton pregnancies are breech at 35-37 weeks' gestation,7 we needed between 1250 and 1700 women to gain a sample of 100 with a breech presentation. We investigated predictive factors associated with correct diagnosis of fetal presentation using contingency tests and used sensitivity analyses to examine specific accuracy rates by maternal characteristics. P < 0.05 was considered significant and analyses were conducted with SAS version 8.2 (SAS Institute, Cary, NC).

Strengths and limitations of the study

We included a large unselected sample and used appropriate timing of the clinical examination relevant for management of non-cephalic presentation in late pregnancy. Previous reports of the sensitivity of clinical examination for detecting non-cephalic presentation have ranged from 28-88%. These studies were small, underpowered, and included selected high risk pregnancies and low gestational ages (range 20-42 weeks).^3–6 Our observed prevalence of non-cephalic presentation was consistent with rates found in longitudinal studies of fetal presentation,9 suggesting that our findings may be applied in other obstetric settings.

We did not collect information on individual clinicians and were unable to ascertain whether particular individuals may have biased results. As all examiners were aware of the study and assessments were recorded and verified, we assumed that assessors would be vigilant. Nevertheless, it is possible that some clinicians may not have been as attentive because diagnoses were going to be checked with ultrasonography.

Room for improvement

Introduction of routine ultrasonography to assess fetal presentation in late pregnancy would improve diagnostic accuracy. However, costs, resource availability, and feasibility need to be considered, as well as the potential deskilling of care providers in performing clinical examination. A cost effectiveness analysis would be necessary before implementation and change in clinical obstetric practice. However, lower rates of accuracy found among overweight or obese women suggest that formal ultrasonography in late pregnancy for these women is required.

Clinical examination to assess fetal presentation is a relatively simple procedure and, with ongoing diligence and regular audit and feedback, accuracy may be increased. Variability in accuracy rates by examiner and level of experience also suggest there is room for improvement by all pregnancy care providers.³⁵¹⁰