The following is the text of a confidential letter sent to the new health secretary, Alan Milburn.

Dear Alan, Congratulations on your appointment and all good wishes for the task ahead: we need an efficient health secretary and are confident you may be him.

A range of 'elephant traps' need careful negotiation by you and Sir Humphrey. First, the National Institute for Clinical Excellence. You will have noted the 'drug pedlar' response to the Relenza decision. At least three lessons follow.

Ensure NICE publishes with its decisions on particular technologies its reasons, putting the data and judgements on the web.

Prepare for more difficult decisions.

The classic will be beta interferon for multiple sclerosis, where the evidence suggests it is clinically effective but not cost-effective. The mandarins and NICE have fudged the issue of clinical versus cost-effectiveness. But our goal is NHS efficiency and this means the choice rule is cost-effectiveness.

Handle NICE communications better. For example, in the Relenza case, the department should have simultaneously increased influenza vaccination of chronic and elderly groups where cost-effectiveness is well documented. This would have informed the media and shifted much needed pressure onto primary care.

Primary care is the second elephant trap. You and your political colleagues have gone on about a 'primary careled' NHS without a clue what it means.

Such wheezes are dangerous.

Furthermore, the NHS Centre for Reviews and Dissemination at York has demonstrated the existence of cost-effective technologies which many hard-pressed GPs ignore.

For example, pharmaceutical treatment of menorrhagia reduces hospital referrals, appropriate treatment of leg ulcers avoids amputations, 'quit smoking' clinics can be cost-effective, and appropriate use of aspirins and beta blockers can reduce cardiac mortality. To be brutal , you are permitting GPs to let patients die and live in distress when it could be avoided.

Why don't GPs practise appropriate medicine in these areas, and why do you tolerate influenza vaccination rates of high-risk groups at a level of less than 50 per cent? Some argue that fees per item of service would alter behaviour as they did with the new contract after 1991. Wow! Didn't vaccination, immunisation, cervical cytology and minor surgery rates zoom then, when there was loot available? Perhaps you should give practice nurses fees for services and let them make hay and improve population health. Go on, be radical.

The management of practice is a task for the Commission for Health Improvement. It seems CHI will initially focus on adherence to cancer guidelines. OK, but don't let it ignore the primary care issues above and ensure the audit is comprehensive.

A comprehensive compliance mechanism, like that used by the Audit Commission, will be costly, but the cost of black spots surviving and rupturing later will be even greater.

One of the nicest challenges for CHI and local managers is variations in clinician activity levels. John Yates of Birmingham University has argued that, instead of employing more doctors, you could pursue waiting-list and waiting-time targets by raising the activity of the laggards. He has shown great variations in NHS activity rates, especially in his analyses of the West Midlands data. Why are you reluctant to pursue these issues?

One explanation is the cowardly way you deal with private practice. In 1997 I wrote to Frank Dobson, then health secretary, asking him to require private hospitals to report activity and case-mix activity data for all NHS consultants. You sent me a classic 'yes, minister' letter which I cherish and which makes me doubt your seriousness about dealing with consultant (and GP) productivity levels.

Is Dr Yates right to argue that low productivity NHS consultants are spending undue amounts of time in the private sector? My guess is that the evidence would be uneven as some of these guys are workaholics who love to be exploring someone's guts with a knife.

After decades of 'redisorganising' the NHS, what effects have there been on GP and consultant productivity?

Maybe such reforms merely shift the deck chairs on the Titanic and have little real effect on the behaviour of many. Did Mrs Thatcher's reforms increase NHS productivity? Why do you expect Labour reforms to increase productivity, or is it all hot air and airy-fairy talk? You have been poacher (minister of health) and then gamekeeper (Treasury minister) and are now back to poaching. The Treasury will soon ask about productivity, so get your story organised.

The final elephant trap is the revision of GP and consultant contracts. What is its purpose? To increase NHS productivity? To raise NHS quality? If so, how will you measure this? Be clear about your goals and recognise that successive health ministers have avoided this wasps' nest.

There is a risk that the NHS will be stung by a higher wage bill for little increased productivity and that doctors, like teachers, will be alienated.

GPs and consultants often work long hours of unpaid overtime. If their contracts are crudely redefined, payment for hours worked may absorb a lot of growth cash when you have to deliver on the prime minister's waiting list targets as well as cancer and heart service improvements. Uncontrolled pay rises, the reforms and all these essential targets have to be met by a service where deficits - sorry, 'gaps' in Blair-speak - are ubiquitous.

Recognise, dear thing, that the 'slack' is increasingly difficult to manage out of the service and that misdirected vigour risks alienating the provider tribes who will determine your fate.

Good luck and go well in these interesting times.

Yours sincerely, Na la Dranyam Alan Maynard is professor of health economics at York University.