What Evidence Underlies Clinical Practice in Paediatric Surgery? A Systematic Review Assessing Choice of Study Design

OBJECTIVE: Identify every paediatric surgical article published in 1998 and every paediatric surgical article published in 2013, and determine which study designs were used and whether they were appropriate for robustly assessing interventions in surgical conditions. METHODS: A systematic review was conducted according to a pre-specified protocol (CRD42014007629), using EMBASE and Medline. Non-English language studies were excluded. Studies were included if meeting population criteria and either condition or intervention criteria. Population: Children under the age of 18, or adults who underwent intervention for a condition managed by paediatric surgeons when they were under 18 years of age. Condition: One managed by general paediatric surgeons. Intervention: Used for treatment of a condition managed by general paediatric surgeons. MAIN OUTCOME MEASURE: Studies were classified according to whether the IDEAL collaboration recommended their design for assessing surgical interventions or not. Change in proportions between 1998 and 2013 was calculated. RESULTS: 1581 paediatric surgical articles were published in 1998, and 3453 in 2013. The most commonly used design, accounting for 45% of studies in 1998 and 46.8% in 2013, was the retrospective case series. Only 1.8% of studies were RCTs in 1998, and 1.9% in 2013. Overall, in 1998, 9.8% of studies used a recommended design. In 2013, 11.9% used a recommended design (proportion increase 2.3%, 95% confidence interval 0.5% increase to 4% increase, p = 0.017). CONCLUSIONS AND RELEVANCE: A low proportion of published paediatric surgical manuscripts utilise a design that is recommended for assessing surgical interventions. RCTs represent fewer than 1 in 50 studies. In 2013, 88.1% of studies used a less robust design, suggesting the need for a new way of approaching paediatric surgical research.