Iain Chalmers’s research while affiliated with The Cochrane Collaboration and other places

In an influential book published more than twenty years ago, Archie Cochrane drew attention to our great collective ignorance about the effects of health care, and explained how evidence from randomized controlled trials (RCTs) could help us to use resources more rationally. He recognized that people who want to take more informed decisions about health care do not have ready access to reliable reviews of the available evidence. In 1979, he wrote: "It is surely a great criticism of our profession that we have not organised a critical summary, by specialty or subspecialty, adapted periodically, of all relevant randomised controlled trials." The Cochrane Collaboration has evolved in response to this challenge and will eventually cover all areas of health care. Contributors in many countries and specialties are preparing and maintaining systematic reviews of RCTs, and reviews of other evidence when appropriate. These reviews will be disseminated using electronic media through the Cochrane Database of Systematic Reviews.

People providing health care tend to turn to review articles rather than reports of primary research when seeking information to guide their practices. For this reason, the process of reviewing the results of primary research must respect scientific principles. Reviews must be kept up to date and disseminated in appropriate ways to the people who make decisions about health care, including policy makers, practitioners, and users of the health services. This article describes the methods developed and used in an attempt to address these challenges for care during pregnancy and childbirth.

Continuing differences of opinion among obstetricians and neonatologists about the place of corticosteroid administration before preterm delivery have prompted us to carry out a systematic review of the relevant controlled trials, using methods designed to minimize systematic and random error. Data from 12 controlled trials, involving over 3000 participants, show that corticosteroids reduce the occurrence of respiratory distress syndrome overall and in all the subgroups of trial participants that we examined. This reduction in respiratory morbidity was associated with reductions in the risk of intraventricular haemorrhage, necrotizing enterocolitis and neonatal death. There is no strong evidence suggesting adverse effects of corticosteroids. The risks of fetal and neonatal infection may be raised if they are administered after prolonged rupture of the membranes, but this possibility is not substantiated by the results of the available trials. The available data on long-term follow-up suggest that the short-term beneficial effects of corticosteroids may be reflected in reduced neurological morbidity in the longer term.

Investigations in which statistically significant differences between treatment groups have not been observed are less likely than others to be reported in scientific journals. In clinical research, this selective suppression of "negative" results may lead to the adoption of ineffective or hazardous treatments. In an attempt to obtain information about unpublished trials in perinatal medicine, letters were sent to 42,000 obstetricians and pediatricians in 18 countries. As a result, we were notified of 395 unpublished randomized trials. Only 18 of the trials had been completed more than 2 years before the survey, a period during which at least 2300 reports of perinatal trials had been published. Of the 395 unpublished trials, 125 had ceased recruitment within the 2 years prior to the survey, 193 were actively recruiting at the time of the survey, and 59 were about to begin recruitment. It was concluded that publication bias will not be addressed successfully by attempts to obtain information about unpublished trials retrospectively. However, since the response rate to our request for details about ongoing and planned trials was good, prospective registration of trials at inception appears to be a feasible approach to reducing publication bias and its adverse consequences. An additional merit of prospective registration of clinical trials is that it should reduce unnecessary duplication (as opposed to necessary replication) in research and promote more effective collaboration.

Controversy continues about the use of beta-mimetic drugs in preterm labour. One reason for this is that the adequately controlled trials of these drugs have all been small and have thus provided very imprecise estimates of their effects. We have therefore conducted a‘meta-analysis’ using data relating to 890 women who participated in the 16 methodologically acceptable controlled trials of these agents in the treatment of preterm labour. This analysis demonstrates an unequivocal effect of beta-mimetic tocolytic administration in delaying delivery, and this is reflected in a reduction in the frequency of preterm birth and low birthweight. However, no beneficial effect of this treatment on perinatal mortality or severe neonatal respiratory disorders could be detected.

Recent claims that routine active management of the third stage of labour increases rather than decreases maternal and neonatal morbidity have prompted us to conduct a systematic review of the relevant controlled trials. In this paper we have analysed data derived from a total of nine published reports of controlled trials in which an oxytocic drug was compared with either a placebo or no routine prophylactic. Oxytocic drugs used routinely appear to reduce the risk of postpartum haemorrhage by about 40% (typical odds ratio 0.57, 95% confidence interval 0.44-0.73) implying that for every 22 women given such an oxytocic, one postpartum haemorrhage could be prevented. The available data are insufficient to assess the possible effects of this policy on the incidence of retained placenta, hypertension and other possible adverse effects.

Recent experimentally derived evidence has confirmed earlier suggestions that seizures which occur within 48 h of birth in babies born at or later than 37 completed weeks gestation are particularly likely to reflect intrapartum asphyxia. We have compared 54 cases of such seizures with 41,090 controls in a geographically defined population. Nulliparity, hydramnios, post-term pregnancy, oxytocin augmentation of labour, abnormalities of fetal heart rate and/or meconium-stained amniotic fluid, prolonged second stage of labour, emergency caesarean section, assisted vaginal delivery, low Apgar score and resuscitation at delivery and subsequent ventilatory support were all statistically significantly more common among cases than among controls. Five of the 54 babies who developed seizures died within 28 days of birth and 11 of the 49 survivors had an impairment diagnosed by 3 years of age which was usually associated with some degree of cerebral palsy. Comparison of the frequency of antecedent perinatal risk factors in the seizure babies who died, those who survived with disabilities and normal survivors failed to reveal any clear pattern.

A database of perinatal trials is currently being established to provide a resource for reviews of the safety and efficacy of interventions used in perinatal care and to foster cooperative and coordinated research efforts in the perinatal field. The database will ultimately comprise four main elements: a register of published reports of trials; a register of unpublished trials; a register of ongoing and planned trials; and data derived from pooled overviews (meta-analyses) of trials. This article describes the development of the first of these four elements.