Fundamental principles of validation, and reliability: rigorous science for the assessment of surgical education and training
Fundamental principles of validation, and reliability: rigorous science for the assessment of surgical education and training
复制标题
DOI：
10.1007/s00464-003-0035-4
复制
发表时间：
2003-10-01
影响因子：
3.1
通讯作者：
Satava, RM
中科院分区：
文献类型：
--
作者：
Gallagher, AG;Ritter, EM;Satava, RM
外科学一直是一门充满活力的学科。它有着悠久的历史，突破性的创新，大大改善了病人的治疗和手术实践的方式。然而，没有什么能让外科界为过去20年的革命性进步做好准备。几乎在一夜之间，传统的开放手术方法被微创手术（MIS）方法所取代[2]。实践者很快意识到，这种新方法需要新的培训类型。突然间，外科住院医师项目的负责人不得不与那些质疑世纪来一直为外科服务的培训模式的教员打交道。传统的Halsteadian方法似乎不适合MIS所需的新型技能。更糟糕的是，分配给培训住院医生的时间正在缩短。2001年，欧洲联盟扩大了《欧洲工作时间指令》的范围，将受训医生也包括在内，今年，美利坚合众国实行了每周工作80小时的强制性限制，这两项指令都大大减少了住院医生的工作/培训时间。与此同时，当外科研究的价值和相关性成为一个有争议的问题时，外科手术的实践在另一个显然无关的领域受到了审查。在一次前所未有的攻击中，《柳叶刀》的编辑[6]对当前外科研究的质量提出了质疑。在一篇题为《外科研究还是喜剧？》的社论中，霍顿博士对它的效用、有效性和有用性表示严重保留。尽管这篇社论遭到了外科界的愤怒回应，但霍顿确实提出了一些有效的观点。不一定是对这篇社论的回应，但肯定与之相邻，外科研究似乎正在从依赖回顾性数据转向前瞻性，随机化，并在可能的情况下盲法数据。我们的论点是，这两个问题，即MIS技能和外科研究，现在已经集中在我们的外科遗产的本质-严格使用“科学方法”和客观评估。也许这是最好的例证，相对较新的编辑实践，优先考虑前瞻性，随机，盲法手术试验的出版。虽然盲法、随机、对照试验设计可以用来明确回答许多外科研究问题，但我们必须记住，这种设计只是众多研究方法中的一种，并不总是首选的实验设计。好的科学的功能是显示清楚的因果关系，这并不总是需要前瞻性的，随机的，双盲的实验设计。实际上，最佳方法甚至可能不是实验性的，而是统计性的，例如结构方程模型[3]。这种方法几乎肯定需要回答的问题，哪些基本的心理能力预测MIS技能/性能或什么因素预测肥胖或治疗结果的肥胖。在这两个例子中，答案都是多因素的，这就排除了传统实验设计的使用。传统的实验设计所施加的控制将被同样强大的（但在方法上更复杂的）“随机”控制所取代。我们在这里要说明的一点是，外科研究没有“灵丹妙药”的实验方法;相反，有一系列的实验设计，应该适当地应用于回答相关的研究问题。因此，研究设计应根据...
Surgery has always been a dynamic discipline. It has a long history of groundbreaking innovations that dramatically improved the way patients were treated and surgery was practiced. However, nothing could have prepared the surgical community for the revolutionary advances of the last 2 decades. Almost overnight, the traditional open approach to surgery was replaced by a minimally invasive surgery (MIS) approach [2]. Practitioners quickly realized that this novel approach would require new types of training. Suddenly, the directors of surgical residency programs had to deal with faculty members who were questioning the training paradigm that had served surgery well for a century. The traditional Halsteadian method did not seem to be suitable for the new types of skills required for MIS. To make matters worse, the length of time allocated to training residents was contracting. In the European Union in 2001 the European Working time Directive was extended to include physicians in training and this year in the United States of America the mandatory 80 hour work week restriction has been instituted, both dramatically reducing work/training hours for residents. At the same time, the practice of surgery came under scrutiny in another, apparently unrelated area when the value and relevance of surgical research became an issue of controversy. In an unprecedented attack, the editor of the Lancet [6] publically called into question the quality of current surgical research. In an editorial entitled ‘‘Surgical Research or Comic Opera?’’Dr. Horton expressed grave reservations about its utility, validity, and usefulness. Although the editorial was met by the surgical community with the predictable indignant response, Horton did make some valid points. Not necessarily in response to this editorial but certainly contiguous to it, surgical research appears to be moving away from reliance on retrospective data to prospective, randomized, and—where possible—blinded data. It is our contention that these two issues, ie, MIS skills and surgical research, have now converged around the very essence of our surgical heritage—the stringent use of the ‘‘scientific method’’and objective assessment. Perhaps this is best exemplified by the relatively new editorial practice of prioritizing prospective, randomized, blinded surgical trials for publication. Although blinded, randomized, controlled trial design can be employed to answer definitively many surgical research questions, we must remember that this design is only one of a quiver full of research methodologies, and may not always be the experimental design of choice. The function of good science is to show clear cause-and-effect relationships, and this does not always require a prospective, randomized, double-blind experimental design. Indeed the optimal method may not even be experimental, but statistical, such as Structural Equation Models [3]. This methodology will almost certainly be required to answer the question of which fundamental psychological abilities predict MIS skills/performance or what factors predict obesity or treatment outcomes in obesity. In both of these examples the answer will be multifactorial, which precludes the use of traditional experimental design. The control exerted with traditional experimental design will be replaced with the equally powerful (but methodologically more sophisticated)‘‘statistical’’control. The point we want to make here is that there is no ‘‘magic bullet’’experimental methodology for surgical research; instead, there is a range of experimental designs that should be applied appropriately to answer the pertinent research question. Research design should therefore vary depending on the …