We thank Diarmuid Quinlan for the interest in our study, and he correctly points out that the patients included in the study have a substantially higher risk compared with the average adult patient assessed by a GP. We do not advocate to use the new derived model to screen for sepsis in all adult patients with a suspected infection presenting in primary care, as this would unnecessarily increase the workload for GPs and result in many false-positive cases. In our view, a predictive model should be used only in cases where a GP perceives a clinical problem. Therefore, we decided to include only study subjects with relevant signs and/or symptoms and those in whom the GP suspected a serious infection. This however did not lead to a study population with ‘only the sickest of patients’. The proportion of patients meeting the Sepsis-31 criteria was high (42%), but this does not mean all patients were critically ill at the time of inclusion. The rate of referral to the hospital after the index consultation was 56%, and 3% of the patients were admitted to the ICU.
We do not feel the phrasing ‘early recognition of sepsis’ in the title is misleading. The population included in the study is a high-risk population for sepsis, but we included patients directly after the first contact with the out-of-hours GP cooperative and almost half of the patients were not referred to the hospital after inclusion. We expect the model to be valid for all patients suspected of sepsis or otherwise having signs of a serious infection in the primary care setting, but, as we stated in the article, additional research is needed before widely adopting the model in practice.
Footnotes
on behalf of the co-authors
- © British Journal of General Practice 2022