Distinguishing the signal in the noise

BMJ 2000; 321 doi: (Published 23 September 2000) Cite this as: BMJ 2000;321:0

As Nicholas Peters says in his editorial this week, the word “cure” is not widely applicable in medicine—and that's exemplified by most of this week's BMJ, where few outcomes are as clear as a cure and where it's often hard to distinguish the signal from the noise.

The exception seems to be catheter ablation for cardiac arrhythmias. Peters (p 716) cites a prospective study of 1050 patients undergoing catheter ablation which produced a 95% cure rate. He claims the procedure offers an alternative for younger patients to long term drug treatment.

In another editorial, however, David Noble and Henrik Kehlet unearth what must be an old problem that lurks around surgery—that of interrupting drug treatment (p 719). The consequences, they argue, may be as serious as those arising from drug errors. A recent study showed that patients who had to interrupt their treatment had three times the normal rate of postoperative complications.

“Chaotic” is the word used by Jill Parmanum and colleagues about the current arrangements for high risk perinatal care in Britain. Their census (p 727) of all transfers from tertiary perinatal centres because of lack of space or staff showed large regional variations. They suggest these reflect the lack of a strategy for the provision of high risk perinatal care.

Two papers studied whether what routine data are telling us is true. Such data tell us, for example, that in north east England perinatal mortality is five times higher and congenital malformation four times higher in pregnancies in diabetic women than in women without diabetes, yet this difference does not occur in Norway. On p 730 G Hawthorne and colleagues report a prospective study with standardised data collection in both places to see whether these differences were real: sadly, they were. Babatunde Olowokure and colleagues, however, found that their data let them down (p 731). Routine surveillance of Haemophilus influenzae disease suggests that its incidence fell 15-fold after the introduction of a vaccine in 1992. But when they compared it with active surveillance the fall was less dramatic.

Public inquiries send out a strong signal, but they too may get it wrong. A powerful article by Edmund Hey and Iain Chalmers criticises the way the inquiry into the research framework at North Staffordshire Hospital NHS Trust went about its task (p 752) and prompts Richard Smith in an editorial to call for clarity about the purpose and process of inquiries (p 715). In particular, he argues, they should “publish their material and methods, check oral allegations against documentary evidence [and] be held in public—otherwise there will always be suspicions of bias, corruption, or incompetence.”


View Abstract

Log in

Log in through your institution


* For online subscription