Hip and Knee Arthroplasty Orthopedic Literature in Medical Journals-Is It Negatively Biased?

Abstract

Background: Healthcare policy is often determined by well-designed studies most often published in high-impact medical journals. However, concern about the presence of publication bias against lower-extremity arthroplasty-related studies has called into question some of the validity of certain reports. There are only a few studies investigating the presence of the bias in high-impact medical journals against lower-extremity arthroplasty intervention, particularly in the Journal of American Medical Association (JAMA), New England Journal of Medicine (NEJM), and the Lancet. Thus, the purpose of this study was to assess (1) the distribution of positive, neutral, and negative results; (2) the number of reports focused on lower-extremity arthroplasty complications among these 3 journals; and (3) difference in bias between 2 time periods (1975 to 1990 and 2000 to 2016).

Methods: A review of the literature from 3 major medical journals (NEJM, Lancet, and JAMA) was performed using PubMed electronic databases, which retrieved articles between January 1976 and December 2016. Studies were categorized as being positive, neutral, and negatively biased by 2 reviewers. Studies were categorized as reporting on lower-extremity arthroplasty-related complications if they were based on complications including deep vein thrombosis, infection, metal-related complication, fat embolism, readmission, or mortality. In addition, we have compared the journal bias between 2 different time points (1975 to 1990 and 2000 to 2016). Descriptive analyses were performed to assess frequencies. Chi-squared analysis was conducted for categorical variables, whereas a z-test was performed for dichotomous data.

Results: When assessing all 3 journals, there were 46 positive (30.3%), 46 negative (30.3%), and 60 neutral reports (39.5%). There was no statistically significant difference in classification proportions between the 3 groups (P = .905). There was a higher percentage of medical literature reporting on the complications of arthroplasty (55.9%); however, this was not statistically significant (z-score = 1.38; 95% confidence interval, 0.48-0.64; P = .17). There was no difference in overall journal reporting between 1975 to 1990 and 2000 to 2016 (P = .548).

Conclusion: There was no evidence of publication bias of lower-extremity arthroplasty reports in the major medical journals (JAMA, NEJM, and Lancet). However, there were more published studies reporting on complications of lower-extremity arthroplasty. This may be due to systematic bias among journal editors in these journals, or due to low journal submission reporting noncomplications after arthroplasty intervention. We did not find the time period to be a factor in bias reporting of orthopedic literature in major medical journals. More work is needed to verify the results of this study.

Keywords: arthroplasty; bias; healthcare; orthopedic literature; publication bias.