The London Health Observatory has studied 34 selected procedures for which the clinical case is sometimes debatable, and drawn up 'access criteria', which would sharply restrict their use.
Its report found 'significant and unexplained variation in planned surgical care in London for procedures where there is evidence that service access criteria should be in place'. Wide variation in admission rates for the top 10 procedures were found between primary care and acute trusts, with no association to deprivation at PCT level.
Commission for Equityconcludes that£28m-£93m could have been saved by contractual caveats stipulating the conditions under which acute trusts could perform certain procedures, and that this money should have been reinvested in more appropriate care.
Based on 2005-06 figures, it estimates that£9.3m could be saved on minor skin procedures;£6.6m on tonsillectomy;£4.5m on varicose vein surgery; and£7.6m on hysterectomy for heavy menstrual bleeding.
LHO director and report co-author Dr Bobbie Jacobson said: 'Most PCTs are using commissioning by access criteria but they are at different stages, and local decisions are tempered by providers and sometimes patients.'
The variation in rates by hospital was at least as much as that by PCT, suggesting that hospital practice is as important as referral practice. Dr Jacobson said: 'We have looked at procedures where access criteria are justified by the evidence.'
The list of procedures to which access criteria could be applied was taken from work carried out by the South West London public health network and includes grommets, hip, knee and joint replacements and cataract surgery.
The LHO examined the 10 most common procedures for London from the list; they were estimated to cost£149m in total for 2005-06. From those 10 alone,£17m-£63m could have been saved to reinvest, the report said.
The new guidance is being implemented in Croydon; Sutton and Merton; Wandsworth; Kingston; and Richmond and Twickenham.
NHS London interim chief executive Ruth Carnall welcomed the work, which she said NHS London's commissioning group would examine closely. 'Despite the major strides taken in recent years to ensure that healthcare is based on strong evidence of what works, we - and London PCTs - recognise that there is still a way to go.'
Croydon PCT acting chief executive Dominic Conlin said: 'The report is a really helpful academic model, which PCTs are looking at, but we must work with local practitioners and secondary care clinicians to underpin it.
'We have been working with GPs over the summer to develop some protocols on the procedures they are most concerned about, and putting mechanisms in place. If surgery is the right course of action for a patient, then we do have exceptional treatment arrangements but that is a decision for clinicians.'
Under review are hysterectomy for heavy periods; cosmetic surgery; surgery for varicose veins; grommets for glue ear; minor skin lesions that are benign and not causing symptoms; tonsillectomy and adeno-tonsillectomy; and cochlear implants.
Robert Creighton, chief executive of Ealing PCT and chair of the London PCT chief executives' group, said: 'Primary care trusts are working to strengthen their commissioning along the lines suggested, However, making these standards apply on the ground is challenging, as it involves changing the practice of many individual clinicians.'
The LHO report shows how the common commissioning approach to reducing hysterectomy rates across south west London might be used across the whole of London.
In London, hysterectomy is the tenth most common procedure on what has been dubbed the 'Croydon list', costing over£10m in 2005-06.
National Institute for Health and Clinical Excellence guidelines and a systematic review of the evidence concluded that treatment with intrauterine levonorgestrel improved the quality of life for women with heavy menstrual bleeding of benign origin as effectively as hysterectomy, and was more cost effective.
Dr Jacobson said: 'The PCT could say it would not fund hysterectomy for non-cancerous bleeding unless the alternatives defined by NICE had failed.
Read the report here