Evaluation of the Oxford and Sheffield SIDS risk prediction scores.

STUDY OBJECTIVE

To evaluate the clinical usefulness (sensitivity and specificity) of the Oxford and Sheffield birth scores for prospective identification of infants at high risk of SIDS.

DESIGN

Retrospective medical record reviews of prospectively identified, autopsy-validated SIDS and living control infants.

STUDY SUBJECTS

Consecutive sample of 140 infants, born between 1/1/83 and 12/31/87, who died suddenly and unexpectedly in the Avon Area Health Authority in southwest England between 1/1/84 and 12/31/88. Seventeen of the cases were excluded: 6 because they lacked adequate clinical records, 11 because they were not SIDS. The 637 control infants were comprised of every 80th delivery between 1/1/83 and 12/31/87 in the three major hospitals in the area.

RESULTS

SIDS incidence was 2.85/1,000 live births. Using standard cut scores to define high SIDS risk (2.0 for Oxford and 500 for Sheffield), sensitivities were 0.55 and 0.35 and specificities were 0.78 and 0.89 for the Oxford and Sheffield scores, respectively. SIDS risk for infants in the high risk group was 7.3/1,000 (Oxford) and 9.3/1,000 (Sheffield).

CONCLUSIONS

Since there is no intervention with proven efficacy for SIDS prevention, and since approximately one half of SIDS cases occur in low risk groups, clinical use of these scoring systems for allocation of health care resources or personnel for the sole purpose of SIDS prevention is not justified.