Introduction

Developmental dysplasia of the hip (DDH) is a condition that encompasses a range of abnormalities of the immature hip. In its mildest form, it consists of a shallow acetabulum with a stable well-located hip, and at its most severe it manifests as a dislocated hip joint. DDH is a dynamic condition, and hips with different degrees of dysplasia can go on to develop normally without intervention.1 When treatment is required, the condition is treatable in a relatively non-invasive manner by use of an abduction splint or harness if the condition is recognised ‘early’. Those treated successfully often require no further treatment; however, the classification of what constitutes DDH requiring treatment varies depending on anatomy, hip stability, age and the judgement of the treating physician.

Patients who present beyond 12–14 weeks of age are widely considered to be ‘late’ presentations and usually have dysplasia that is not amenable to harness or brace treatment. These patients invariably require a general anaesthetic to attempt reduction of the subluxed or dislocated hip. The need for surgical intervention to allow reduction of the hip into the acetabulum is also on a spectrum. This ranges from an arthrogram and closed reduction, with or without the release of soft tissues, to a formal open reduction with or without acetabular and femoral osteotomies. Those who require surgery may have problems with their hips in later life and are at an increased risk of osteoarthritis.2 There is no consensus on the ideal model of surveillance that should be used to detect cases of DDH3 4; however, most centres in the UK employ a combination of clinical examination and selective ultrasound (US) examination to identify DDH.

The Hall 4 report5 published in 2003 was adopted across the UK and recommended examination of the hips at 24 hours and a second examination between 6 and 8 weeks. These examinations are widely referred to as the ‘baby check’ and ‘GP check’, respectively. The recommendation also forms part of the recommendation from the National Screening Committee.6 In England practice has been standardised through the Newborn and Infant Physical Examination (NIPE) screening programme.7 All children are examined within 72 hours of birth and examined for symmetry of leg length, range of abduction and hip stability using the Ortolani8 and Barlow9 tests. If any abnormality is noted, then the infant is referred for a US examination to confirm or rule out the presence of DDH necessitating early treatment in a harness or splint. In addition, risk factors such as first-degree family history or breech presentation after 36 weeks’ gestation necessitate a US examination even in those infants with a normal examination. The children with a normal initial examination and no risk factors are examined again at 6–8 weeks, commonly in general practice.

It is often quoted that 40% of presentations of DDH have a recognised risk factor.10 However, Paton et al found that only 31% of the patients in their series had a recognised risk factor.11 Therefore, there is a reliance on effective clinical examination to identify the majority of cases of DDH in the UK. It has been reported in the literature that the effectiveness of the examination in the first days after birth is improved when performed by experienced staff.12–14 This is an area of concern for many clinicians as there has historically been no guidance in the UK on the need for examiner experience in this part of the examination process. Arrangements vary among hospitals, but anecdotally in many units the ‘baby examination’ has been traditionally allocated to junior staff with arguably the least experience of the examination.

While the NIPE screening programme is not established out with England, the criteria necessitating referral for US due to risk factors are similar to those undertaken in many other regions of the UK. Within our institution we follow the guidance on selective screening previously described with the addition of several other risk factors which necessitate US examination; these are torticollis, plagiocephaly, non-passively correctable foot deformity or other skeletal deformity.

Examination at 6–8 weeks, often referred to as the ‘GP examination’, forms an important part of the current model of screening in England and surveillance in the rest of the UK. It is often looked on by clinicians as a safety net in which to catch those patients with DDH that were not identified in the neonatal period.

Aims and methods

We aimed to assess the sensitivity, specificity and positive predictive values of examination at  6–8 weeks as part of the DDH surveillance system in our region from 2006 to 2010.

For the purposes of this study we define late presentation as a patient with a diagnosis of DDH, which required treatment, that was diagnosed after 14 weeks old. It was felt that the majority of children with a previously undetected DDH born during this period should have presented by December 2016 as they would all be of walking age. Data on late presentations are prospectively collected by the two senior authors who treat all DDH in the region. The number of live births between 1 January 2006 and 31 December 2010 within NHS Tayside was collated from the NHS Tayside Maternity and Neonatal Clinical Information Systems.

To identify all patients referred following examination at  6–8 weeks, all requests for hip US received by the radiology department during the study period were collated. In our region examination at 6–8 weeks is undertaken in general practice, and therefore the ‘Referring Location’ was initially used as a filter to identify only requests made by general practitioners (GPs). However, it was apparent that the majority of GP referrals generated following examination at 6–8 weeks were to orthopaedics or paediatrics rather than directly to radiology for US. Therefore, all outpatient referrals for US made between 4 and 14 weeks of age were presumed to be originating from assessments at 6–8 weeks. To ensure only the original referral for each child was analysed, only requests documented as being the first attendance for a hip US were included. The result of the scan for each patient was classified as a positive finding if the child required treatment for DDH, or a negative finding if they did not.

Data were also gathered on those patients with dates of birth during the study period who had presented with a late DDH. Information on attendance in the physical examination at 6–8 weeks for these patients was held by Information Services Division Scotland, which is a division of National Services Scotland and part of NHS Scotland.

In addition to the above data, the numbers treated for early DDH using a Pavlik harness were known from the existing prospective collection of data.

Results

There were 23 112 live births within Tayside throughout the 5-year study period and a total of 141 confirmed cases of DDH, which include both early and late presentations. The incidence of DDH at the time of the study was 6.1 per 1000 live births, which is in keeping with the expected incidence of DDH.15 16 Of the 141 cases of DDH diagnosed, 114 patients presented before 14 weeks and were classified as early presentations, and 27 patients presented after 14 weeks of age and were classified as late presentations. No patients referred before 14 weeks but scanned after were found to have DDH.

There were 2595 referrals for hip US in all children born during the study period (referred between 0 and 98 days of age). Using the above methodology, 400 US scans were attributed to being directly requested or prompted by the patient’s GP following concerns during examination at 6–8 weeks. Of these 400 referrals for US, 6 were found to confirm early DDH requiring treatment. All were treated with Pavlik harness, 4 successfully and 2 unsuccessfully.

Of the 27 late presentations, 2 did not attend examination at 6–8 weeks. However, the remaining 25 had all been examined at 6–8 weeks and considered normal at the time of examination. There were 19 917 patients who were not referred for US and did not present as a late presentation within the study period. We do not have data on the non-attendance rate of this part of the cohort, but the national attendance rate is known to be very good at around 94.5%.

These results were used to calculate the sensitivity and specificity of the examination detailed in table 1. The two patients with late presentations who did not attend examination at 6–8 weeks were excluded from the analysis. The sensitivity of the examination during this period was 19.4%, it had a specificity of 98% and a positive predictive value of 1.5%.

Discussion

For many years, examination at 6–8 weeks had been a psychological safety net for clinicians, but this study calls the reliability of this assumption into question. The results from this study show 1.7% of all live births were referred for US because of clinical concern at the time of examination at 6–8 weeks and that assessment is well attended with only a 7.4% DNA rate in our group of late presentations. However, the sensitivity of examination at 6–8 weeks for DDH was low at only 19.4%, which resulted in only one in five children with DDH being identified. Therefore, unfortunately despite the good attendance in examination at 6–8 weeks, 25 patients went on to present with a late DDH.

We would like to be very clear that this finding of poor sensitivity should not be seen as a reflection on the capability or experience of the GP team.

In our opinion, by this age the child and their soft tissues are both less compliant and examination using Ortolani and Barlow’s tests is less reliable in the hands of any examiner.

The merits and disadvantages of universal and selective US examination are well documented in the literature,17–21 and there remains no clear consensus on the best model. What is agreed on however is the necessity of examination experience12–14 22 when electing for a programme of selective screening; centres which have reported low rates of late presentation all have a system of experienced examiners performing the initial neonatal examination.

One of the issues is that, at the time of this study, there was no clear guidance in our institution on who should conduct the neonatal examination and there was no quality assurance of how the examination was being performed. This is something we have identified as an issue requiring attention and we will continue to monitor the presentations of late DDH as we strive to improve our systems in an effort to reduce late presentations.

We would note that there are some limitations in our methodology. The true number of referrals from GPs following the GP examination may be lower than our methodology suggests. However, should the number of GP referrals be lower than we have estimated, the sensitivity of the examination would be even poorer than we have reported. In addition, it is possible that children were seen for the GP assessment later than 12 weeks; between 14 and 20 weeks, there were a further 60 patients referred for US, and none of these patients was found to have DDH. We also note that there is a possibility that children born in our region may move and present as a late DDH in another region. We are not aware of this occurring in other regions in Scotland, but it is possible if children moved to another country. Again, if this has occurred, then the true sensitivity would be lower.

We would reiterate that that the results above should not be viewed as a reflection on the expertise or experience of the GP or other healthcare professionals examining a child’s hips during examination at 6–8 weeks. The conclusions in this paper should not be viewed in isolation and should be considered as part of a review of the selective examination process overall. Our results highlight that attention must be given to both improving the effectiveness of neonatal examination component of the system and understanding the limitations of examination at 6–8 weeks.

It has been suggested that the current timing of examination at 6–8 weeks may be too early. If the infant is examined after 8 weeks, Choudry et al 23 report that unilateral limitation of hip abduction may be a more appropriate test, with a positive predictive value of 54.7% and a sensitivity of 78.3%. It may therefore be that a modification to the existing examination arrangements could improve outcomes, although identification at this later time reduces the window for Pavlik harness treatment.

Unfortunately, we conclude that examination at 6–8 weeks in its current form is not reliable. Therefore, when electing to employ a model of selective examination, the importance of the initial neonatal examination to be performed by trained and experienced examiner is pivotal. It is essential efforts are made to get it right for every child as the long-term consequences of late presentation can be life-changing.