Klim McPherson and Alison Metcalfe ('Inadmissible evidence', pages 2627, 30 March) are right to suggest that the clinical criteria for admission to intensive care are undefined and that refusal rates are a poor index of unmet need.

Many patients who would benefit from intensive care are either never referred (because of inexperience or temerity, or because the intensive care unit is always full), or referred so late that admission is then pointless.

Many patients are not referred because their operation is postponed, again because there is no intensive care bed.And others are not referred because informal enquiries have suggested that there are no beds. So there is an unknown but probably significant unmet need that is not reflected in refusal rates.

Further, there are the patients whose admission to intensive care is at the expense of another patient's premature discharge. Recent work has shown that premature discharges occurred twice as frequently in 199598 as in 1988-90, and that these patients were less likely to survive their hospital stay than those who were discharged appropriately.

1McPherson and Metcalfe are quite wrong to state that 'there is no indication of systematic underprovision of intensive care unit beds' because there is in fact no indicator of under-provision. Their mathematical legerdemain purporting to show an increase in refusal rates with an increase in intensive care unit beds is a disingenuous and unhelpful contribution to a problem which is costing people's lives.

Iain Mackenzie Anaesthesia and intensive care John Radcliffe Hospital Oxford

REFERENCE 1Goldfrad C, Rown K. Consequences of discharge from intensive care at night. Lancet 2000; 355:1138-1142.