Prediction models for remission of school-age asthma are limited, and many available tools focus on preschool diagnosis or require lung function measurements. The authors developed a pragmatic, history-based prognostic tool — the School-age Asthma Prognosis Score (SAPS) — intended to predict likelihood of clinical asthma remission in children and to support clinicians in tailoring follow-up intensity.
The development dataset was drawn from the Swiss Paediatric Airway Cohort (SPAC), a prospective cohort that included 1,860 children aged 5–16 years with physician-diagnosed asthma. The external validation cohort was the German All-Age Asthma Cohort (ALLIANCE). Both studies are registered clinical cohorts and include paediatric arms with ethics committee approvals as described by the authors.
The study defined clinical remission at 2–3 years after asthma diagnosis. Remission required both absence of wheeze and absence of inhaler use during the preceding 12 months. This outcome is based on parental questionnaire responses collected in the cohorts.
Candidate predictors were obtained exclusively from parental questionnaires and captured demographics, symptom patterns, triggers, and family history. The initial modelling considered 12 candidate variables. The approach prioritized readily available clinical history items so the resulting score would not depend on lung function testing or specialized investigations.
The prognostic model was developed using LASSO regression to select predictors and shrink coefficients. Multiple imputation was applied to address missing data. Model performance was evaluated by area under the receiver operating characteristic curve (AUC) for discrimination, the Hosmer–Lemeshow (HL) test, and calibration plots to examine agreement between predicted and observed remission probabilities.
From the 12 candidate variables, the final simplified score retained eight predictors: sex; wheeze frequency; night-time awakening; exercise-induced wheeze; pollen-triggered wheeze; animal-triggered wheeze; maternal asthma; and paternal asthma. The authors report a simplified score derived from the model coefficients to facilitate clinical use; details of the point allocation and exact scoring algorithm are provided in the study materials and supplementary files referenced by the authors.
In the SPAC development cohort the SAPS achieved moderate discrimination with an AUC of 0.71. Calibration was assessed with the Hosmer–Lemeshow test and calibration plots, though numerical HL results are reported in the full manuscript rather than the abstract. Overall, discrimination and calibration assessments supported the model's internal performance.
The SAPS was externally validated in the German ALLIANCE cohort. The score maintained similar discriminative ability on external validation, with an AUC of 0.71, indicating reproducible moderate performance across European paediatric cohorts.
Because the SAPS uses only clinical history items available from parental questionnaires and does not require lung function testing, it offers a practical tool for routine clinical settings. The authors propose that SAPS can help clinicians identify children at lower probability of remission who may benefit from closer monitoring and follow-up, supporting targeted resource allocation in paediatric asthma care.
Both SPAC and ALLIANCE are registered (SPAC: NCT03505216; ALLIANCE paediatric arm: NCT02496468; ALLIANCE adult arm: NCT02419274). SPAC received approval from the Bern Cantonal Ethics Committee; ALLIANCE received approval from relevant local ethics committees with the leading committee at University Luebeck. Participants or their legal guardians provided written informed consent.
Study protocols, consent forms, definitions, derivation of clinical characteristics and outcomes, and other study materials are available online (www.spac-study.ch and www.dzl.de). SPAC individual-level data are not deposited in an open repository because participants were not asked to consent to public deposition; requests for partial datasets and analysis documentation can be made to the corresponding author and Prof Kuehni under reasonable request conditions.
The manuscript is posted as a preprint and has not been certified by peer review. The authors provide detailed competing interest statements: several authors declare no conflicts of interest, while others disclose industry-related grants, consulting fees, honoraria, advisory roles, or leadership positions outside the submitted work. These declarations are reported by the authors and pertain to work outside the submitted study.
The SAPS offers a history-based, easily applicable prognostic score for predicting 2–3 year asthma remission in school-aged children. Performance was moderate and reproducible across two European cohorts (development in SPAC, external validation in ALLIANCE). Because this report is a preprint, the score should be interpreted with caution until peer-reviewed publication and further independent validation are available. For access to full methods, scoring details, and supplementary materials, the authors have provided links and indicate that additional data and code can be made available on reasonable request.