NHS consultant report on needless hip surgeries

Staff at Children’s Health Ireland (CHI) have alleged that Dublin hospitals are conducting needless hip surgeries on children. An external report by an NHS consultant has concluded in one Dublin hospital 80 percent of osteotomies on children between the ages of one and seven were unnecessary. This draft report was delivered to CHI management in January. Months later affected families have neither seen the report nor been contacted by CHI. Minister for health Jennifer Carroll MacNeill has said she hasn’t seen the report. The Ditch is now publishing the report in full.

Independent audit of indications for pelvic osteotomy in DDH

February 2025

Independent Draft External Medical Audit for Children’s Health Ireland on Indications for Pelvic Osteotomy in Children with Developmental Hip Dysplasia

PRIVATE & CONFIDENTIAL NOT FOR DISCLOSURE

1. AUTHOR DETAILS AND DISCLOSURE

1.1. I am not aware of any actual or potential conflicts of interest in conducting this audit. All surgeon and patient identifiable data have been removed from the records and X rays.

1.4. I have been asked by Children’s Health Ireland (CHI) to audit decision making for pelvic osteotomy, in young children with DDH, against published consensus principles. The audited period was from January 2021 to December 2023. Children were included over the age of one year but less than 7 years.

2. METHODOLOGY AND BACKGROUND DATA

2.1. Redacted orthopaedic clinic records and X rays for children with DDH undergoing pelvic osteotomy at Temple Street Hospital (TSH), Crumlin Hospital (CRH) and National Orthopaedic Hospital Cappagh (NOHC) were made available to me in batches for each of the surgeons performing the procedure. Surgeons were anonymised “A-N”. A random selection of 10 cases from each surgeon’s series at each institution was made. This audit is based on reviews of those records and radiographs as provided to me. It is not a complete review of practice but rather an audit specifically to look at the indications for pelvic osteotomy that have been advised and performed for children aged 1-7 years between January 2021 and December 2023.

Additional information has been recorded in some cases to assist with understanding the decision making.

2.2. At the outset the expectation was that most of the children undergoing treatment by pelvic osteotomy would have had persistent hip dysplasia after failure of bracing, or delayed diagnosis, requiring closed hip reduction, with muscle releases and prolonged spica casting, or open hip reduction and spica casting. There are good quality data on radiological indices that predict risks for persisting dysplasia in these settings, based in the main on age at time of reduction and acetabular index (AI). Persisting dysplasia, after successful hip reduction, would warrant pelvic osteotomy in those children with a reduced likelihood of spontaneous improvement after a period of observation (1). An audit form set out the planned methodology based on these assumptions.

2.3. In many of the TSH and NOHC cases sampled, there was no record of closed or open hip reduction after delayed diagnosis, no record of either hip ever having been dislocated and often no history of treatment soon after birth by splint or brace for hip instability. Pelvic X rays have often been arranged because of a family history of hip dysplasia, report of a “clicky” hip or asymmetric thigh creases identified in the community. Some children have been treated in an abduction brace around the age of one year following referral with pelvic X rays showing acetabular immaturity that was sometimes termed “stable acetabular dysplasia” (2). Hip instability, subluxation or dislocation were not described in these cases.

2.3.1. It would be unusual for corrective pelvic osteotomy in childhood to be indicated in large numbers of essentially untreated hips. There are no reliable data predicting which of these, if any, might be at particular risk of idiopathic adolescent dysplasia (3,4) and/or adult hip osteoarthritis (2).

2.3.2. The peer reviewed literature examining the relationship between acetabular dysplasia and future risk of arthritis, for hips with silent or stable dysplasia identified serendipitously on pelvic X rays, does not reach consistent conclusions. It is generally accepted, however, that a centre edge angle (CEA) below 20-25 degrees at skeletal maturity probably predicts a significant future risk of hip arthritis. CEA is not a reliable measure under the age of 5 years, which is the age group studied in this audit. An AI during this period that is within a reasonable range of normal, typically 2 standard deviations from the mean, based on contemporary normative population data (5), would generally be considered reassuring. Children with an AI between one and two standard deviations from the mean would reasonably be considered to have borderline dysplasia requiring ongoing surveillance. A high rate of spontaneous resolution, to an AI within one standard deviation of the mean, would be anticipated in this group.

2.3.3. Intervention, by pelvic osteotomy in childhood, for persistent elevation of AI outside two standard deviations of the mean would be considered reasonable with the goal of preventing persistent acetabular dysplasia at skeletal maturity. However this is neither actively screened for, in any country to my knowledge, nor often required for children with no prior personal history of treated DDH.

2.3.4. The normative data for AI used in this audit are presented with percentiles and Z scores (5). The latter are the number of standard deviations of a value above or below the mean. With a normal distribution of AI about the age-specific mean, 68% and 95% of the data will fall within one and two standard deviations of the mean respectively. At the upper range, the 84th and 98th percentiles are equivalent to one and two standard deviations, or Z scores, from the mean. In the actual data for AI that is used for this audit, the 90th and 99th percentiles are provided and progressively darker shaded above the 50th percentile. They equate to Z scores of 1.28 and 2.33 respectively.

2.4. For children undergoing follow-up after successful closed or open hip reduction and spica casting, predictive data from Iowa (USA) use AI measurements at different time points after successful reduction to predict future risk of persistent dysplasia (1).

2.4.1.These data indicate that, while a hip joint without prior treatment for DDH and an AI within a reasonable range of the age-specific mean should have a good long term outcome, a hip with the same AI at the same age that has been treated for DDH by closed or open reduction may have residual acetabular dysplasia at skeletal maturity.

2.4.2.A recent study from Hong Kong confirmed the same in a cohort of children undergoing Pavlik harness treatment or closed hip reduction and casting for DDH (open reduction cases were excluded). In hips treated thus, an AI within two standard deviations of the age-specific mean at the age of 2-3 years could still be associated with a poor outcome at skeletal maturity (6). This is not to say that all DDH cases treated by Pavlik harness or closed reduction, with an AI at the upper limit of two standard deviations at 2-3 years of age, require immediate pelvic osteotomy; rather that a further period of observation is indicated for the small number within this subgroup that may go on to have persisting dysplasia.

2.4.3. It is critically important therefore to apply the appropriate predictions, based on different AI datasets, to the correct groups of DDH treated and untreated children.

2.5. Centre edge angle (CEA) is potentially unreliable in younger children under the age of 5 years who have not ossified sufficient of the acetabular anlage (immature cartilage model) for this to be measured on plain X rays. It was not therefore used in this audit as a primary measure for quantifying risk of persisting dysplasia. However it was considered reassuring when it measured above 20-25 degrees in older cases. It is important, nonetheless, to follow up children undergoing closed or open reduction and casting, and especially those requiring pelvic osteotomy, to skeletal maturity to confirm that an adequate CEA has been achieved at an age when this measurement is reliable and can be used as a predictor of outcome.

2.6. Some of the surgeons cited a “short sourcil” as a concerning feature to justify pelvic osteotomy. This description is not established as a valid qualitative means to predict persisting dysplasia, to my knowledge, and was not therefore used in this audit. In cases where the lateral edge of the sourcil could be identified it was used, in this audit, to measure AI in cases of “stable acetabular dysplasia” diagnosed by serendipitous pelvis X rays. This is the measure used in the referenced dataset (5) and is equally as reliable as the classic lateral edge method (7).

2.7. A lateral acetabular notch was also cited by some surgeons as a potential justification for pelvic osteotomy. This has been studied in the peer reviewed literature as potentially indicative of prior neonatal hip instability when seen on hip X rays in the first few months of life (8). It is not a valid predictor for persisting dysplasia and probably represents a biological failure of ossification rather than a true deficiency (9). Hip arthrography is typically reassuring in confirming a stable hip joint with adequate coverage by an immature cartilage anlage.

2.8. Acetabular floor thickness was not cited by surgeons to justify pelvic osteotomy in the records that I reviewed. This was not measured in this audit.

2.9. Reviewing the redacted patient records and making accurate radiological measurements was time consuming. The audit form planned to analyse a “snapshot” of 20-50 cases from each surgeon.

2.9.1. A sample of 10 cases was taken at random for every surgeon operating at each of the centres. For those surgeons performing less than 10 pelvic osteotomies at an institution, all the cases were reviewed.

2.9.2. Only 10 cases were reviewed for each of the CRH (H-N) surgeons. Most of the TSH surgeons A-G had also done pelvic osteotomies at NOHC during the audit period. More cases of the TSH surgeons were therefore sampled but a higher percentage of all pelvic osteotomies at CRH were sampled because of the smaller total done there.

2.9.3. I was satisfied that the analysis of 10 cases for the CRH surgeons was sufficient for me to understand their decision making. They were mostly operating for children with persisting acetabular dysplasia after closed or open hip reduction. This simplified the task of analysing the indications because the pre-planned methodology in the audit form could be applied.

2.10. Some surgeons’ samples indicated a very high rate of bilateral, rather than unilateral, pelvic osteotomies.

2.10.1. This audit references the epidemiological data of Loder et al that predict: one third of DDH cases will be bilateral, two thirds of unilateral cases will affect the left hip and three quarters of all DDH cases will be female (10).

2.10.2. There is good anecdotal and some peer reviewed evidence (11) that bilateral DDH has a higher risk of persistent dysplasia. Some excess of bilateral pelvic osteotomies, above the predicted incidence of bilateral DDH, would therefore reasonably be anticipated.

2.11. For hips undergoing Pavlik harness (or equivalent abduction brace) treatment after early identification of instability on ultrasound, successful reduction and stabilisation are likely in 95% of all hips based on benchmark European and North American data (12-14).

2.11.1. A 2-5% rate of late dysplasia (actual and borderline) at 5 years after successful reduction in the harness is typical with an anticipated requirement for pelvic osteotomy in around 2% (12,14).

2.11.2. A 1991 study reported a 17% rate of acetabular changes on radiographs at skeletal maturity in a group of 74 hips treated in a Pavlik harness (15). The changes reported were of uncertain significance. A more recent systematic review (16), of late acetabular dysplasia after successful DDH treatment in a Pavlik harness, included 6029 hips. There was late dysplasia in 280 (4.6%) and a requirement for additional surgery in 109 (1.8%).

3. SUMMARISED RESULTS BY INSTITUTION

3.1. Totals

3.1.1. 147 cases audited of 14 surgeons across 3 institutions.

3.1.1.1. 563 cases* at TSH & NOHC; records & X rays of 241 cases provided for audit.

3.1.1.2. 164 cases* at CRH; records & X rays of 101 cases provided for audit.

3.2. TSH

3.2.1. 49 cases sampled of a total 127 cases (39%).

3.2.2. 85 pelvic osteotomies; 34 indicated (40%).

3.3.NOHC

3.3.1. 51 cases sampled of a total 114 cases (44%).

3.3.2. 70 pelvic osteotomies; 15 indicated (21%).

3.4. CRH

3.4.1. 47 cases sampled of a total 101 cases (47%).

3.4.2. 63 pelvic osteotomies; 62 indicated (98%).

* Estimated numbers provided by CHI of children aged 1-7y undergoing pelvic osteotomy between 01/2021-12/2023.

4. CONCLUSIONS AND RECOMMENDATIONS

4.1.Summary

4.1.1. CRH surgeons perform pelvic osteotomies, mostly for children with previously treated DDH, that are reasonably indicated, with one exception (one side of a bilateral case), according to consensus published radiological and clinical criteria that have been retrospectively applied for the purposes of this audit.

4.1.1.1.In many cases the radiological criteria are recorded and tracked by CRH surgeons over time. There is sometimes documented discussion with colleagues +/- arthrography for cases with borderline indications for surgery.

4.1.1.2. Children appear to be under ongoing follow-up to skeletal maturity after pelvic osteotomy.

4.1.1.3.Record keeping in some of the CRH clinic correspondence that I have seen is poor. In the event that complications were to arise it would not be possible, in some cases, to establish the risk/benefit analysis behind the decisions to operate or whether the families had been involved in an exploration of their attitudes to and tolerance of the risks.

4.1.2. Some of the TSH and NOHC surgeons have been performing significant numbers of pelvic osteotomies for children that are not indicated based on the consensus criteria that have been retrospectively applied for the purposes of this audit.

4.1.2.1.

4.1.2.2.m There will inevitably be variation between observers when recording radiological measurements. Diﬀerences also arise between the classic lateral and sourcil methods of AI measurement (7). The variance identified in this audit cannot be accounted for by measurement error or observer variability alone.

4.1.2.3. For two surgeons at NOHC there were missing records and X rays that excluded some cases.

4.1.2.4. It is likely that the TSH/NOHC surgeons would benefit from peer review across all 3 sites for decision making to agree and confirm reasonable clinical and radiological indications and timings for pelvic osteotomy. This should be a routine part of surgical planning going forward.

4.1.2.5. Qualitative assessments of the radiographs, including terms such as “short sourcil” and “acetabular notching”, have been used to justify intervention with no evidence linking these descriptions to persistent adolescent hip dysplasia or adult hip arthritis. Some surgeons at TSH/NOHC have set out percentage risks of persistent dysplasia and hip arthritis, for children with no prior history of DDH and/or normal neonatal hip ultrasound, that are not an accurate or balanced reflection of the peer reviewed orthopaedic literature.

4.1.2.6. Children appear often to be discharged from TSH/ NOHC at around one year after pelvic osteotomy. Planned review of ongoing acetabular development and assessment of the outcome at skeletal maturity appear to be lacking.

4.1.2.7. The high rates of bilateral pelvic osteotomy exceed the predicted incidence of bilateral DDH (10,11). In one case, there was an adverse outcome from the osteotomy in a bilateral case; on the side for which surgery was not indicated according to the consensus criteria used for this audit.

4.2. Recommendations for follow-up

4.2.1. All 561 cases that have undergone pelvic osteotomy at NOHC/ TSH require ongoing follow-up to skeletal maturity. These families should all be recalled and informed that this is a matter of routine good practice. It is critically important to confirm satisfactory acetabular development now and at skeletal maturity with measurements that should include (lateral) centre edge and Tonnis angles on pelvic X rays (19).

4.2.2. This audit raises concerns about the indications for pelvic osteotomy in many of the cases at TSH/NOHC but some of those pelvic osteotomies were reasonably indicated. The surgeons there will be able to review and articulate their reasons for recommending additional surgery in all the cases sampled. The first step should be to recall all the 561 cases, that underwent pelvic osteotomy at TSH/NOHC between 2021-2023, for ongoing follow-up that should be routine in any event.

4.2.3. Clinical and radiological assessment of these cases by independent and appropriately experienced paediatric orthopaedic surgeons is recommended. Intermittent follow-up should continue until skeletal maturity at 14-16 years of age. The majority is likely to be functioning well but the outcome after pelvic osteotomy, performed for children with acetabular indexes within acceptable limits of a normal range, is not known. There is a risk, for example, of symptomatic hip impingement arising from the altered orientation of the socket. Over-coverage or an overly “deep” socket have been identified as significant risk factors for the development of osteoarthritis in adults (21).

4.2.4. Any complications that have arisen from pelvic osteotomy must be acknowledged and discussed with the families. There should be an assessment of the original decision making, in cases of concern, based on the appropriate radiological indices and clinical history informing the decision to operate. The appropriate datasets to use, for the other cases not sampled in this audit, depend upon whether there was any previous history of treated DDH and, if so, whether this was management in a harness or open/closed reduction. The appropriate methodology is set out in the preamble to this audit and in the further recommendations below.

4.2.5. The majority at TSH/NOHC has undergone pelvic osteotomy using a novel technique. Published follow-up for this less invasive osteotomy is only to an average 24 months for a series in which all but one was unilateral (17). The longer term outcome to skeletal maturity is critically important but as yet unknown. The outcome of the less invasive osteotomy is not characterised for the significant numbers of children sampled that underwent bilateral surgery with radiological indices that were within an acceptable range of normal.

4.2.6

4.3. Recommendations for ongoing practice at all institutions

4.3.1. Children with “stable acetabular dysplasia” (acetabular immaturity) on pelvic X ray in the first 6-18 months of life, after serendipitous referral in from the community for concerns such as a hip click or thigh crease asymmetry:

4.3.1.1. Children in this age group will have a high rate of apparent acetabular immaturity on plain X ray related to physiological ossification of the cartilage anlage of the acetabulum (9).

4.3.1.2. If the hip joints are concentrically located and there is no prior treatment for DDH then these cases require a further X ray at 2 years and reassurance once the acetabular index (AI) is within one standard deviation of the age-specific mean (Z score of 1). This corresponds to most of the light grey area in the figures below (5).

4.3.1.3. The rate of pelvic osteotomy in this group should be very low. Cases deemed to require it, because of an AI persistently in the dark grey ranges below, should be peer reviewed with other DDH surgeons from all 3 institutions. Families should also be involved in a thorough discussion of decision making.

4.3.2. Children undergoing successful treatment early in life in a Pavlik harness for DDH identified on neonatal hip ultrasound:

4.3.2.1. This group requires pelvic X rays at walking age. Reassurance and discharge is appropriate once the acetabular index is within one standard deviation of the age and sex-specific mean (Z score of one). On that basis, the majority can be discharged at 2 years of age but some will require ongoing follow-up to 5 years (12-14).

4.3.2.2. For the minority with residual significant dysplasia on X rays at 2 years, most will have resolved by 5 years. Pelvic osteotomy should be required in only around 2% of this group (13-16). Audit of this rate, based on the denominator of children treated in a Pavlik harness, would be straightforward and is recommended.

4.3.2.3. Cases deemed to require pelvic osteotomy should undergo peer review as part of surgical planning, with surgeons from all 3 institutions, and families should be involved in decision making.

4.3.2.4. Based on recently published data, children with acetabular indices at or above 24 degrees at 2-3 years of age may benefit from ongoing follow-up but pelvic osteotomy should not be recommended unless the AI exceeds 25 degrees at 4 years or 24 degrees at 5 years (6).

4.3.3. Children with delayed diagnosis of DDH or failure of harness treatment undergoing successful closed or open hip reduction:

4.3.3.1. This group requires close follow-up for the first 2 years after successful reduction to assess the acetabular index. The evidence base of Albinana et al should be used to assess the risk of persisting acetabular dysplasia at skeletal maturity according to time from successful reduction (1):

4.3.3.2. At 2 years from successful reduction, the risk of persisting dysplasia should be approximated from the table and graph above, discussed with the family and peer reviewed with DDH surgeons from all 3 institutions if pelvic osteotomy is advised. AI values should be recorded in the medical notes with an estimated risk of residual dysplasia that is communicated to families as part of the process of decision making and subsequent consent.

4.3.3.3. In cases where pelvic osteotomy is advised by the surgeon, agreed by the peer review process and accepted by the family after discussion of the potential risks and benefits, a further appointment should be arranged for a proper process of consent. The minimum standards to document are:

4.3.3.3.1. An explanation of the condition of the hip and the percentage likelihood of a poor outcome (persistent acetabular dysplasia) at skeletal maturity with and without any intervention.

4.3.3.3.2. Description of the proposed intervention and the anticipated benefit. If the modified osteotomy is recommended, it must be disclosed that it is a novel procedure with limited published follow-up currently. The advantages and disadvantages in comparison to the standard open procedure should be set out.

4.3.3.3.3. A discussion of the postoperative course including immobilisation and any anticipated requirements for rehabilitation.

4.3.3.3.4. Disclosure and explanation of all important risks.

4.3.4. Less invasive pelvic osteotomy

4.3.4.1. The procedure described as a modified Salter osteotomy at TSH and NOHC is substantially different to the open procedure originally described by Salter (18). It would be better described as a less invasive pelvic osteotomy with either no eponym or a different name attached to it.

4.3.4.1.1.The original open procedure provides a predictable correction based on direct identification of anatomical landmarks (20).

4.3.4.1.2. The modified procedure is performed through a smaller incision using radiological guidance and bioresorbable pins. 50 procedures in 49 cases have been reported with average 2 year follow-up (14). Bioresorbable pins avoid the need for surgical removal but can cause osteolysis with the formation of bone cysts or a sterile abscess.

4.3.4.2. Parents should be informed that the modified osteotomy is novel, has specific risks and currently lacks longer term follow-up. The option of an alternative procedure with longer follow-up should be provided. Most are likely to accept a shorter procedure through a smaller scar but time should be set aside for a balanced, documented discussion of the potential risks, benefits and current uncertainties of the less invasive osteotomy.

4.3.4.3. There should be a prospective registry to collect data on outcomes and adverse events with the less invasive osteotomy. These data should be published at appropriate time intervals. A retrospective clinical and radiological review of cases already performed is also recommended when the TSH/NOHC cohorts are recalled.

PRIVATE & CONFIDENTIAL NOT FOR DISCLOSURE