Point of care ultrasound screening, performed by midwives, outperforms standard selective screening for hip dysplasia. A diagnostic accuracy study

Abstract

Background and purpose

To investigate the accuracy of pubo-femoral distance (PFD) ultrasound (US) performed by midwives as a referral criterion for diagnostic Graf hip US, compared to standard screening for developmental dysplasia of the hip (DDH) in newborns.

Methods

PFD US screening was performed along standard screening for DDH in a prospective cohort. The accuracy of PFD US was compared to the standard referral criteria for diagnostic hip US in terms of sensitivity, specificity and positive predictive value in detecting abnormal hips, as well as proportion of abnormal hips found and referral rates.

Results

We included 2735 newborns of which 488 received a diagnostic hip US. Mean age at diagnostic US was 35.5 days (±12.8). 317 newborns (11.7 %) were referred by Danish standard screening criteria and 206 newborns (7.6 %) were referred by the PFD criterion.

Sensitivities/specificities for detecting ≥Graf type IIc hips were: 50 %/68 % for risk factors, 30 %/94 % for clinical examination, 60 %/62 % for risk factors + clinical examination, and 80 %/89 % for PFD US using a cut-off of 6.2 mm. Difference in sensitivity between PFD and standard screening criteria did not reach statistical significance, but specificity and PPV did.

The number of detected ≥ Graf type IIc hips were 6 and 9 for standard screening and PFD screening respectively.

Interpretation

Early point-of-care PFD US screening for DDH has significantly higher accuracy in classifying abnormal hips, when compared to standard screening. PFD US increases the detection rate of dysplastic hips (≥ Graf IIc) by 60 % while reducing referral rates.