Prognostic accuracy of antenatal neonatology consultation

Abstract

Objective:

Neonatologists provide antenatal counseling to support shared decision-making for complicated pregnancies. Poor or ambiguous prognostication can lead to inappropriate treatment and parental distress. We sought to evaluate the accuracy of antenatal prognosticaltion.

Study Design:

A retrospective cohort was assembled from a prospectively populated database of all outpatient neonatology consultations. On the basis of the written consultation, fetuses were characterized by diagnosis groups (multiple anomalies or genetic disorders, single major anomaly and obstetric complications), assigned to five prognostic categories (I=survivable, IIA=uncertain but likely survivable, II=uncertain, IIB=uncertain but likely non-survivable, III non-survivable) and two final outcome categories (fetal demise/in-hospital neonatal death or survival to hospital discharge). When possible, status at last follow-up was recorded for those discharged from the hospital. Prognostic accuracy was assessed using unweighted, multi-level likelihood ratios (LRs).

Results:

The final cohort included 143 fetuses/infants distributed nearly evenly among the three diagnosis groups. Over half (64%) were assigned an uncertain prognosis, but most of these could be divided into ‘likely survivable’ or ‘likely non-survivable’ subgroups. Overall survival for the entire cohort was 62% (89/143). All but one of the fetuses assigned a non-survivable prognosis suffered fetal demise or died before hospital discharge. The neonatologist’s antenatal prognosis accurately predicted the probability of survival by prognosis group (LR I=4.56, LR IIA=10.53, LR II=4.71, LR IIB=0.099, LR III=0.040). The LRs clearly differentiated between fetuses with high and low probability of survival. Eleven fetuses (7.7%) had misalignment between the predicted prognosis and outcome. Five died before discharge despite being given category I or IIA prognoses, whereas six infants with category IIB or III prognoses survived to discharge, though some of these were discharged to hospice care.

Conclusions:

The neonatologist’s antenatal prognosis accurately predicted fetal–neonatal outcome. Infants with non-survivable or uncertain but likely poor prognoses had a very low probability of survival, whereas those with good or uncertain prognoses had a high probability of survival. There were few cases of prognostic failure with most occurring in fetuses with one major or multiple anomalies. The few cases of prognostic failure suggest a need for caution. Honest disclosure of prognostic uncertainty and shared decision-making with families utilizing their personal values is critical in the antenatal encounter.