Background: In recent years, it has become common to publish a level of evidence grading for orthopaedic journal publications. Our primary research question is: is there an improvement in levels of evidence of articles published in pediatric orthopaedic journals over time? In addition, what is the current status of levels of evidence in pediatric orthopaedic journals?
Methods: All articles in Journal of Pediatric Orthopaedics-A (JPO-A) and Journal of Pediatric Orthopaedics-B (JPO-B) for 2001, 2002, 2007, and 2008 and those in Journal of Children's Orthopaedics (JCO) for 2007 and 2008, were collected by an independent reviewer. Of the 1,039 articles identified, animal, cadaveric and basic science studies, expert opinion and review articles were excluded. Seven hundred fifty remaining articles were blinded and randomized with respect to journal, title, publication date, author, and institution. According to the currently accepted grading system, study type and level of evidence was assigned to each article. Interobserver and intraobserver reliability were investigated. Statistical analysis was carried out using SPSS software.
Results: There were no statistically significant differences in study type or levels of evidence in articles published before and after 2003. Of articles published during 2007/2008, 3.0% were graded as level I, 5.0% as level II, 24.1% as level III, and 58.0% as level IV. Analysis of the separate journals for all 4 years revealed that JPO-A published 2.6% (13 of 503) level I studies, whereas JPO-B published 4.3% (7 of 163) and JCO published 1.2% (1 of 84). The intraobserver reliability was high for study type (κ, 0.842) and substantial for level of evidence (κ, 0.613). The interobserver reliability for study type and level of evidence was high (κ 0.921 and 0.860, respectively).
Conclusions: Since the introduction of levels of evidence to orthopaedic journals in 2003, there has been minimal change in the quality of evidence in pediatric orthopaedic publications. We note a modest increase in level III articles and a corresponding decrease in level IV articles. Articles can be reliably graded by nonepidemiologically trained individuals.
Level of evidence: Not applicable.