Baker R, Tata LJ, Kendrick D, et al. Identification of incident poisoning, fracture and burn events using linked primary care, secondary care and mortality data from England: implications for research and surveillance. Inj Prev 2016;22:59–67.

The authors missed a small number of deaths from injuries for the period 1^st April 1997–31^st December 2000. During this period, deaths in England were initially recorded using ICD-9, with transfer to ICD-10 over this period. They have identified that their dataset only included deaths recorded using ICD-10 for these 3–4 years. The impact of this error is minimal, as this would only affect deaths not also recorded in the primary care or hospital admission records of the patient. Using data from the subsequent years of the study, the authors have estimated that they may have missed up to 41 deaths from poisonings, 23 deaths from fractures and 4 deaths from burns. As their dataset contains 42,985 poisonings 185,517 fractures and 36,719 burns identified using linked primary care, hospital admission and mortality data, this small number of deaths does not change their overall estimates of injury incidence using the three linked data sources. Estimated death rates from poisonings, fractures and burns included in Table 3 of the paper may be underestimates as a result of not identifying some deaths coded using ICD-9 for the period affected. This minor error does not change the key findings or conclusions of the paper.