Page 135 - 2018_11-Haematologica-web

Pediatric CT scans and childhood leukemia
year latency period was used, in part to deal with reverse causa- tion due to confounding by indication.31 Also, multiple predis- posing factors (Online Supplementary Table S1) were accounted for with outpatient register data. The methods are described in more detail in the Online Supplementary Material.
We obtained data on all CT scans performed on pediatric patients (<15 years) from all five university hospitals and the five
largest central hospitals in Finland (Table 1, Figure 2). The period of data availability varied between hospitals, because radiologi- cal databases with information on each CT scan for individual patients were introduced at different times. We estimated that the data from the study hospitals covered 87% of all pediatric CT scans performed in Finland during 1975–2011 (see the Online Supplementary Material for details). For each CT scan, we
Table 1. The collection and availability of electronically stored computed tomography scans.
Hospital
Helsinki University Hospital
Tampere University Hospital Oulu University Hospital
Turku University Hospital
Kuopio University Hospital Central Finland Central Hospital Satakunta Central Hospital Seinäjoki Central Hospital Päijänne Tavastia Central Hospital North Karelia Central Hospital TOTAL
City
Helsinki
Tampere Oulu Turku Kuopio Jyväskylä Pori Seinäjoki Lahti Joensuu
Data availability
1990–2011
1978–2011 1993–2011 1996–2011 1996–2011 2002–2011 1995–2011 1999–2011 2000–2011 1993–2011
Number of CT scans
31,825
9,236 7,513 7,360 5,408 2,571 1,948 1,759 1,597 1,191 80,783
All Finnish university hospitals are listed first followed by the five chosen central hospitals.
Figure 2. Flow chart linking the collected computed tomography scans to the study subjects The flow of the CT scans is represented on a white back- ground and the CT scans to different body parts on a light gray background. The necessary exclusions are shown in red boxes.
haematologica | 2018; 103(11)
1875