Serum Surfactant Protein D Predicts The Outcome Of Patients With Idiopathic Pulmonary Fibrosis Treated With Pirfenidone

Jie Weng,Mengyun Tu,Bo Yang,Zhiyi Wang

DOI: https://doi.org/10.1016/j.rmed.2017.11.014

IF: 4.3

2018-01-01

Respiratory Medicine

Abstract:•The factors with P values <0.2 in the univariate analysis should be proceeded to the multivariate analysis.•The proportional hazard assumption should been evaluated.•The prediction models need to be validated through using some appropriate methods. We took an interest to read the above-published article conducted by Ikeda K and colleagues that was published in Respiratory Medicine recently [[1]Ikeda K. Shiratori M. Chiba H. Nishikiori H. Yokoo K. Saito A. et al.Serum surfactant protein D predicts the outcome of patients with idiopathic pulmonary fibrosis treated with pirfenidone.Respir. Med. 2017; 131: 184-191https://doi.org/10.1016/j.rmed.2017.08.021Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar]. This was a well-written article. To make the results more accuracy and convincing, some statistical methods should be reconsidered seriously. Firstly, logistic regression analysis was performed in this study by authors. The factors showing p < 0.05 in the univariate analysis proceeded to the multivariate analysis. However, this strategy can overestimate the effect of factors and induced the Testimation bias [2Steyerberg E. Clinical Prediction Models: a Practical Approach to Development, Validation, and Updating. Springer Science & Business Media, 2008Google Scholar, 3Weng J. Wu H. Xu Z. Xi H. Chen C. Chen D. et al.The role of propionic acid at diagnosis predicts mortality in patients with septic shock.J. Crit. Care. 2017; 43: 95-101https://doi.org/10.1016/j.jcrc.2017.08.009Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar]. The factors with p value less than 0.2 in the univariate analysis proceeded to the multivariate analysis may be more appropriate. Secondly, the Cox proportional hazard model was used in this study. Cox proportional hazard model means that the predictors are assumed to have proportional effects during follow-up on the outcome. However, the proportional hazards assumption was forgot by the authors. The wrong conclusions will be made when the proportional hazards assumption is violated [[4]Kleinbaum D.G. Klein M. Survival Analysis: a Self-learning Text. Springer Science & Business Media, 2006Google Scholar]. We suggest the authors to examine the proportional hazard assumption in their study. Finally, in order to draw a more scientific conclusion, especially in small sample size study, the prediction models need to be tested by some appropriate methods such as cross-validation and bootstrapping. Otherwise, these conclusions would be optimistic interpretation [[5]Smith G.C. Seaman S.R. Wood A.M. Royston P. White I.R. Correcting for optimistic prediction in small data sets.Am. J. Epidemiol. 2014; 180: 318-324https://doi.org/10.1093/aje/kwu140Crossref PubMed Scopus (221) Google Scholar]. None. None.