Letters Hip arthroplasty

Authors’ reply to National Joint Registry

BMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e7246 (Published 31 October 2012) Cite this as: BMJ 2012;345:e7246
  1. D J W McMinn, consultant orthopaedic surgeon1,
  2. J Daniel, director of research1,
  3. R B C Treacy, consultant orthopaedic surgeon2,
  4. R D Riley, reader in biostatistics3
  1. 1McMinn Centre, Edgbaston, Birmingham B15 3DP, UK
  2. 2Royal Orthopaedic Hospital, Northfield, Birmingham
  3. 3School of Health and Population Sciences, University of Birmingham, Birmingham
  1. r.d.riley{at}bham.ac.uk

The National Joint Registry (NJR) was aware that the data supplied would be used by clinical teams associated with Smith & Nephew (including first author McMinn).1 2 The data were originally requested because the NJR’s editorial chairman invited McMinn “to criticise registry data . . . to support the value of resurfacing” at the British and Irish Orthopaedic Congress 2011. McMinn therefore asked Smith & Nephew to apply, and the application stated that the data would be used by clinical teams to investigate survivorship. After data release, the NJR was made publicly aware of early findings at the above congress. Subsequent emails and discussions between the NJR and the author(s) document the NJR’s awareness of our possession of the data and our intent to publish revision and mortality analyses. The NJR was invited to be involved in the analyses but did not respond.

The fact that the data were potentially incomplete for patients after a revision does not change the conclusions. We were interested in death associated with the first procedure alone, and because a different procedure may be used on revision, our analyses censored patients at revision. Hence, we needed to use the revision time only for these patients.

Finally, we corrected statements about episode statistics linkage in our paper because we did not have dates of death. However, we used the provided date and age at primary operation with the age at death or revision (available to multiple decimal places) to calculate the time from surgery to death, revision, or censoring for each patient.

Our paper aimed to compare the survival of implant classes and patients, and it identified potential differences in both revision and mortality rates between different classes of implant. We believe this is in keeping with the NJR’s mission statement: “to collect high quality data . . . to provide an early warning of issues relating to patient safety.”


Cite this as: BMJ 2012;345:e7246