**9. Statistical analysis**

We reviewed 19 articles (15 case reports1,2,4,5,7,10-19, three studies pertaining to musculoskeletal infection in general3,8,9, and one study focused on an osteomyelitis population6). In estimating the heterogeneity of the three musculoskeletal infection papers (k = 3), we measured the I2 index as our test of heterogeneity and found the I2 index to be 7.7%. The Cochran's Q value was found to be 2.153 (p = 0.3407). This statistic signifies a degree of heterogeneity between the three papers which makes it difficult to derive meaningful conclusions from the summary data from these studies. However, this level of heterogeneity is expected considering several factors such as different study methodology, different timing of the onset of disease, different disease severity and different patient groups.

The incidence rate of DVT was calculated to compare the difference between an incidence rate specific to osteomyelitis and the incidence related to musculoskeletal infection in general. From our results, the number of DVTs in the three musculoskeletal infection papers was 23 with a sample size of 450.3,8,9 Therefore the incidence of DVT in children with musculoskeletal infection was 5.1%. However, in the one paper in which osteomyelitis alone was considered, the number of DVTs was 9 with a sample size of 116.6 Hence, the incidence of DVT with osteomyelitis was 7.8%. Although the differences in incidence could be due to different populations being studied, there does appear to be a higher incidence of DVT specifically with pediatric osteomyelitis as compared to pediatric musculoskeletal infection in general. Further research is required in this area.
