Corticosteroids in acute traumatic brain injury: systematic review of randomised controlled trials.

OBJECTIVE: To quantify the effectiveness and safety of corticosteroids in the treatment of acute traumatic brain injury. DESIGN: Systematic review of randomised controlled trials of corticosteroids in acute traumatic brain injury. Summary odds ratios were estimated as an inverse variance weighted average of the odds ratios for each study. SETTING: Randomised trials available by March 1996. SUBJECTS: The included trials with outcome data comprised 2073 randomised participants. RESULTS: The effect of corticosteroids on the risk of death was reported in 13 included trials. The pooled odds ratio for the 13 trials was 0.91 (95% confidence interval 0.74 to 1.12). Pooled absolute risk reduction was 1.8% (-2.5% to 5.7%). For the 10 trials that reported death or disability the pooled odds ratio was 0.90 (0.72 to 1.11). For infections of any type the pooled odds ratio was 0.92 (0.69 to 1.23) and for the seven trials reporting gastrointestinal bleeding it was 1.05 (0.44 to 2.52). With only those trials with the best quality of concealment of allocation, the pooled odds ratio estimates for death and death or disability became closer to unity. CONCLUSIONS: This systematic review of randomised controlled trials of corticosteroids in acute traumatic brain injury shows that there remains considerable uncertainty over their effects. Neither moderate benefits nor moderate harmful effects can be excluded. The widely practicable nature of the drugs and the importance of the health problem suggest that large simple trials are feasible and worth while to establish whether there are any benefits from use of corticosteroids in this setting.     

Does home visiting prevent childhood injury? A systematic review of randomised controlled trials.

OBJECTIVE--To quantify the effectiveness of home visiting programmes in the prevention of child injury and child abuse. DESIGN--Systematic review of 11 randomised controlled trials of home visiting programmes. Pooled odds ratios were estimated as an inverse variance weighted average of the study specific odds ratios. SETTING--Randomised trials that were available by April 1995. SUBJECTS--The trials comprised 3433 participants. RESULTS--Eight trials examined the effectiveness of home visiting in the prevention of childhood injury. The pooled odds ratio for the eight trials was 0.74 (95% confidence interval 0.60 to 0.92). Four studies examined the effect of home visiting on injury in the first year of life. The pooled odds ratio was 0.98 (0.62 to 1.53). Nine trials examined the effect of home visiting on the occurrence of suspected abuse, reported abuse, or out of home placement for child abuse. Because of the potential for bias in outcome reporting in these studies, pooled effect estimates were not calculated. CONCLUSIONS--Home visiting programmes have the potential to reduce significantly the rates of childhood injury. The problem of differential surveillance for child abuse between intervention and control groups precludes the use of reported abuse as a valid outcome measure in controlled trials of home visiting.     

Commentary: educational initiatives deserve randomised controlled trials.

The media have been full of reports of crisis in the NHS. Although national analyses suggest that the NHS should be able to cope within the increases in spending it has been given, local pressures can leave parts of the service struggling. Firstly, the change to allocation of funds on the basis of population needs has meant that some authorities and trusts have had effective cuts in their budgets, requiring them to trim services. Secondly, the government''s insistence on an annual 3% increase in efficiency may have resulted in authorities taking short term measures that actually decrease efficiency in the long term. Thirdly, health authorities have had to bear the costs of national targets such as reducing waiting lists and junior doctors'' hours as well as local problems such as higher numbers of mentally disordered offenders. However, all these factors can be controlled by national or local management and so their impact is not inevitable.     

Developing a register of randomised controlled trials in primary care.

OBJECTIVE--To determine the number, nature, site of publication, and feasibility of identifying randomised controlled trials relevant to primary care. DESIGN--Review of literature using three strategies: approaching journal editors, Medline search, and manual search of individual journals. SETTING--Journals containing publications of studies based in primary care. MAIN OUTCOME MEASURES--The number, site of publication, and subject of trials identified. RESULTS--No journal had a system which enabled identification of all the randomised controlled trials it published. 266 trials relevant to primary care were identified from 110 different journals during 1987-91 by Medline. Of these, only 62 trials were published in primary care journals. Hand searching of seven major primary care research journals showed that between 13% and 38% of the trials had been missed by the Medline search. Of the trials identified, 47 (18%) were concerned with mental disease (including neuroses, tobacco misuse and alcohol misuse) and 43 (16%) were concerned with hypertension. CONCLUSION--Given the diversity of publication sources and topics, this supports the need for a centrally based register of randomised controlled trials that may be relevant to primary care overviews in the future.