Data collection
The registry data from DCCR were supplemented with data from medical records from all pediatric departments in Denmark. To assess the impact of a misdiagnosis on patient outcomes, we identified patients with a prior musculoskeletal misdiagnosis and compared their characteristics to the remaining patients without any musculoskeletal misdiagnosis. The following data were collected: demographic information such as age and gender, the clinical presentation (symptom, objective signs), the diagnostic intervals, tumor type and grade, metastases, treatment, comorbidities. We also collected data of cause and date of death, last day of follow-up, and presence of sequelae from tumor or treatment. Sequelae was categorized according to organ and severity, severe sequelae included plegia, incontinens or neurocognitive defects.
The musculoskeletal diagnosis was recorded as a misdiagnosis if the symptoms were later found to be due to the tumor. It was recorded as musculoskeletal comorbidity if it was a coexisting musculoskeletal diagnosis and the symptoms were not later explained by the tumor. The diagnoses included were from hospitalizations, emergency room, and outpatient visits but not from general practitioners.
We used a standardized definition of the diagnostic intervals, including the period from the first symptom until the start of treatment. Further, three additional time intervals were added: 1) a pre-diagnostic symptomatic interval, i.e., the time from the first symptom until diagnosis. 2) a first hospital doctor interval, i.e., the time from the first hospital contact until a specialist was involved; and 3) a specialist interval, i.e., the time from a specialist was involved until the final diagnosis was made. The specialist was defined as either a pediatric oncologist or neurosurgeon, (Fig. 1).