Nuevo Javier, Tafalla Mónica, Cordero Luis, Ruiz Montserrat, Calvo Eva, Román Javier, Mayo Jesús
Departamento Médico, AstraZeneca Farmacéutica Spain SA, Madrid, España.
Gac Sanit. 2011 Mar-Apr;25(2):170-2. doi: 10.1016/j.gaceta.2010.12.007. Epub 2011 Mar 3.
Observational studies are subject to biases that may lead to misinterpretation of the results. This study aimed to determine the influence of omeprazole treatment on the duration of sick leave in patients with ankle sprains treated with non-steroidal anti-inflammatory drugs. We used the Ibermutuamur database. Contrary to our expectations, sick leave was longer in patients who received omeprazole than in those who did not. These findings were probably due to the influence of a bias due to confounding by severity, given that patients who received omeprazole had a worse kind of ankle sprain; however, a bias due to confounding by indication cannot be excluded. To avoid the influence of these systematic errors, biases should be monitored from the design stage to the data analysis stage.
观察性研究容易受到各种偏差的影响，这些偏差可能导致对结果的错误解读。本研究旨在确定奥美拉唑治疗对接受非甾体类抗炎药治疗的踝关节扭伤患者病假时长的影响。我们使用了Ibermutuamur数据库。与我们的预期相反，接受奥美拉唑治疗的患者病假时间比未接受治疗的患者更长。这些发现可能是由于严重程度混杂导致的偏差影响，因为接受奥美拉唑治疗的患者踝关节扭伤情况更严重；然而，也不能排除适应证混杂导致的偏差。为避免这些系统误差的影响，应在从设计阶段到数据分析阶段全程监测偏差。
