Standardization of retinopathy of prematurity clinical examinations in clinical trials: risks, benefits and alternatives.
Standardization of retinopathy of prematurity clinical examinations in clinical trials: risks, benefits and alternatives.
复制标题
临床试验中早产儿临床检查视网膜病变的标准化：风险、益处和替代方案。
DOI：
10.1111/j.1442-9071.2008.01673.x
复制
发表时间：
2008
影响因子：
4
通讯作者：
Good,WilliamV
中科院分区：
文献类型：
--
作者：
Good,WilliamV
早产儿视网膜病变 (ROP) 临床试验带来的一个不为人知的好处是改善了患有这种复杂疾病的婴儿的临床护理。 CRYO-ROP 研究小组、STOP-ROP 小组、Light-ROP 小组、ETROP 研究小组和许多其他研究小组年复一年地致力于标准化 ROP 的诊断和治疗标准。 Antaean 参与该标准化的工作包括定期技术小组会议，眼科医生在会上审查 ROP 的诊断结果，现场访问直接观察 ROP 婴儿并比较结果，根据配对、现场检查对研究眼科医生进行认证，并在研究方案的关键点确认检查。尽管有互联网和远程医疗，这种高水平的护理标准化仍然需要研究人员的承诺和牺牲。因此，这些临床试验的结果可以被认为是高度可靠的，参与这些研究的眼科医生提高了他们的 ROP 水平。与此同时，ROP 的物理发现现已得到数千名未直接参与任何试验的眼科医生的认可。这是因为眼科医生在这些研究中所取得的临床智慧扩展到中心之外，从而提高了国内外 ROP 护理的质量。 ROP 治疗决策基于研究结果，而这些结果又基于熟练的检查。此外，鉴于严重 ROP 的发病率较低，新的 ROP 诊断医生需要相当长的时间才能了解该疾病的细微差别。通过临床试验，ROP诊断和治疗的学习曲线已经缩短。因此值得注意的是达洛等人。本版《临床和实验眼科》1 描述了澳大利亚和新西兰新生儿网络 (ANZNN) 各中心急性 ROP 发病率的显着差异。该网络维护有关 ROP 发生率和严重程度的数据，但没有针对眼科医生的严格认证流程。经过仔细的统计分析，作者得出结论，观察者偏差可能导致其网络内 ROP 发生率的变化。他们的结论似乎无可辩驳：应该对以 ROP 为结果的研究制定认证流程。当然，无论疾病或试验，中心之间的差异是不可避免的。即使采用高度标准化的诊断和治疗程序，这种差异也会发生。单个或小型中心之间的差异是协作工作的理由。必须获得足够的样本量来检验假设，对于 ROP 等低发病率疾病，大规模合作是必要的。但 ANZN 网络了解到，例如，第 3 阶段的治疗范围在 15% 到 120% 之间，因此一些眼睛必须接受低于第 3 阶段的治疗水平。统计分析控制了其他变量，例如出生体重和体重。
An unsung benefit stemming from clinical trials on retinopathy of prematurity (ROP) has been the advancement of clinical care for infants with this complicated disease. Year after year while ongoing, the CRYO-ROP Investigative Group, STOP-ROP Group, Light-ROP Group, ETROP Investigative Group and many others strived to standardize diagnostic and treatment criteria for ROP. The Antaean effort involved in this standardization included regular technical group meetings, where ophthalmologists reviewed diagnostic findings for ROP, site visits for direct observation of infants with ROP and comparison of findings, certification of study ophthalmologists based on paired, live examinations and confirming examinations at crucial points in study protocols. The Internet and telemedicine notwithstanding, this standardization of care at such a high level has required commitment and sacrifice by study investigators.As a consequence, findings from these clinical trials can be considered highly reliable, and ophthalmologists involved in these studies have elevated their ROP game. Meanwhile, physical findings in ROP are now recognized by thousands of ophthalmologists who were not directly involved in any trial. This occurred as clinical acumen achieved by ophthalmologists within these studies spread beyond centres, improving the quality of ROP care nationally and internationally. ROP treatment decisions are based on study findings, and these, in turn, are based on skilled examinations. Furthermore, given the low incidence of severe ROP, it would otherwise take a new ROP diagnostician considerably longer to learn the nuances of the disease. Thanks to clinical trials, the learning curve for ROP diagnosis and treatment has been shortened. So it is noteworthy that Darlow et al. in this edition of Clinical and Experimental Ophthalmology1 describe considerable variation in the incidence of acute ROP at centres involved in the Australia and New Zealand Neonatal Network (ANZNN). This Network maintains data on ROP incidence and severity, but does not have a rigorous certification process for ophthalmologists. After careful statistical analysis, the authors conclude that observer bias likely contributed to the variation in ROP incidence within their Network. Their conclusion seems irrefutable: that a certification process for studies in which ROP is an outcome should be in place. Of course, variation among centres, no matter the disease or trial, is inevitable. Such variation will occur even with highly standardized diagnostic and treatment procedures. Variation among single or small centres is the rationale for collaborative work. An adequate sample size must be achieved to test a hypothesis, and in low incidence diseases such as ROP, large-scale collaboration is necessary. But the ANZN Network learned that treatment for stage 3, for example, varied between 15% and 120%, and thus some eyes had to receive treatment at levels less than stage 3. Statistical analysis controlling for other variables such as birthweight and