The probability of spontaneous delivery -- possibility and limits of obstetric risk scores

Abstract

For a sound consultation of pregnant women e. g. regarding the choice of delivery place, a simple, reliable prepartal risk estimation system would be desirable. After exclusion of all "hospital required" risk cases, out of 203,111 records from the Berlin Clinical Perinatal Data 1993 - 1999 176,734 births remained, for which all risks already discernible before delivery (catalogues of the German Maternity Guidelines) were retrospectively valued and weighted by means of a logistic regression analysis on their real importance for the prediction of a spontaneous birth. On the basis of logarithms of the odds ratio for the individual risk factors a point system was developed. There is a clearly higher risk for a surgical delivery in primiparae compared with multiparae. For multiparae 27, for primiparae 26 statistically significant (p < 0.01) risk factors could be identified. After conversion of the odds ratio-values into the point system of the score, for the group of primiparae the cut off value was > 1 point (with a specificity of 0.5 and a sensitivity of 0.5), in the multiparae > 3 points (sensitivity 0.7, specificity 0.7). As for primiparae the resulting prognosis of birth mode (spontaneous vs. surgical birth) using the described score is false at 46 %, in multiparae at 33 %, the prognosis power of the risk score can therefore only be classified as relatively poor. The risk assessment system is therefore not suitable e. g. as aid to decision-making for or against an extra-clinical delivery place where no obstetrical operations are possible.