Bernard-Bonnin Anne-Claude, Pless I Barry, Robitaille Yvonne, Leblanc John, King W James, Tenenbein Milton, Klassen Terry P
Department of Pediatrics, University of Montreal, Montreal, Quebec.
Paediatr Child Health. 2003 Sep;8(7):433-7. doi: 10.1093/pch/8.7.433.
Many intervention studies typically require data from several centres to ensure adequate power. The usual intention is to pool data after testing for heterogeneity. Sites that differ in sample characteristics may, on the one hand, complicate the assessment of the intervention, but on the other hand, they may add important insights through analysis of site-specific findings.
The aims of the present paper were to compare the distribution of injuries and risk factors among children participating in a five-centre study of a home-based injury prevention program, and to contrast parental injury awareness and knowledge with home safety measures.
Five children's hospitals in Canada agreed to participate in a case-control study combined with a randomized controlled trial. Patients were children zero to seven years of age presenting to a hospital emergency department with a fall, burn, ingestion or choking. Two controls were matched to each case, one with another injury and another with a minor illness. A home visitor completed a home hazard assessment based on observed safety measures. To determine whether data could be pooled, comparisons across sites were made with respect to types of injuries seen, sociodemographic characteristics, observed hazards and the parents' reported beliefs about severity of injuries, safety measures, preventability of injuries and susceptibility to injuries.
There were few differences between the five hospitals. The mean age was 2.2 years (range 1.4 to 3.3). There were 219 falls (56%), 80 burns (20.4%), 54 poisonings (13.8%), and 38 chokings (9.7%), all distributed in a proportionately similar manner, except for poisoning, at each site. There were significantly more well-educated fathers at one hospital and younger parents with less education at another. Homes were generally lacking five recommended safety measures. However, most parents at all sites perceived their home as being very safe for any of the specific injuries, and their child as being at low risk of sustaining any of these injuries.
The similarity across sites supports the pooling of these data regarding hospital-treated injuries in young children in urban Canada. Most parents at all sites perceived their home as being very safe in spite of their homes lacking one-quarter of the recommended safety measures. This discrepancy between parental perception and home safety highlights the needs for further education and prevention efforts.
许多干预研究通常需要来自多个中心的数据以确保有足够的效力。通常的意图是在检验异质性后合并数据。样本特征不同的研究地点一方面可能会使干预评估复杂化，但另一方面，通过对特定地点的研究结果进行分析，它们可能会提供重要的见解。
本文的目的是比较参与一项基于家庭的伤害预防计划的五中心研究的儿童中伤害和危险因素的分布情况，并将父母的伤害意识和知识与家庭安全措施进行对比。
加拿大的五家儿童医院同意参与一项病例对照研究并结合随机对照试验。患者为零至七岁因跌倒、烧伤、中毒或窒息到医院急诊科就诊的儿童。为每个病例匹配两名对照，一名有其他伤害，另一名有轻微疾病。一名家访员根据观察到的安全措施完成家庭危险因素评估。为了确定数据是否可以合并，对各研究地点在所见伤害类型、社会人口学特征、观察到的危险因素以及父母报告的关于伤害严重程度、安全措施、伤害可预防性和易受伤害性的信念方面进行了比较。
五家医院之间差异很少。平均年龄为2.2岁（范围1.4至3.3岁）。有219例跌倒（56%）、80例烧伤（20.4%）、54例中毒（13.8%）和38例窒息（9.7%），除中毒外，各研究地点的分布比例相似。一家医院受过良好教育的父亲明显更多，而另一家医院的父母年龄较小且受教育程度较低。家庭普遍缺乏五项推荐的安全措施。然而，所有研究地点的大多数父母都认为他们的家对任何特定伤害来说都非常安全，并且他们的孩子遭受任何此类伤害的风险很低。
各研究地点之间的相似性支持合并这些关于加拿大城市地区幼儿医院治疗伤害的数据。尽管家庭缺乏四分之一的推荐安全措施，但所有研究地点的大多数父母都认为他们的家非常安全。父母认知与家庭安全之间的这种差异凸显了进一步开展教育和预防工作的必要性。
