ObjectiveTo review individual treatment effect (ITE) models developed from randomized controlled trials, with the aim of systematically summarizing the current state of model development and assessing the risk of bias. MethodsPubMed and Embase databases were searched for studies published between 1990 and 14 June 2024. Data were extracted using the CHARMS inventory, and the PROBAST risk of bias tool was used to assess model quality. ResultsA total of 11 publications were included, containing 19 ITE models. The ITE modelling methods were regression models with interaction terms (n=8, 42.1%), dual-range models (n=5, 26.3%) and machine learning (n=6, 31.6%). The ITE models had a reporting rate of 78.9%, 73.2% and 10.5% for differentiation, calibration and clinical validity, respectively. Fourteen models were assessed as having a high risk of bias (73.7%), particularly in the area of statistical analysis, due to inappropriate handling of missing data (n=15, 78.9%), inappropriate consideration of model fit issues (n=5, 26.3%), etc. ConclusionCommon approaches to ITE model development include constructing interaction terms, dual procedure theory, and machine learning, but suffer from a low number of model developments, more complex modeling methods, and non-standardized reporting. In the future, emphasis should be placed on further exploration of ITE models, promoting diversified modeling methods and standardized reporting to improve the clinical promotion and practical application value of the models.
ObjectiveTo estimate the individual treatment effect (ITE) of Shenlingcao oral liquid (SOL) combined with adjuvant chemotherapy in patients with stage Ⅱ-ⅢA non-small cell lung cancer (NSCLC) after complete resection using the Rboost method, and to identify characteristics of the potential high-benefit population, thereby providing a decision-making basis for clinical precision medicine. MethodsBased on data from a previously conducted nationwide, multicenter, pragmatic, open-label randomized controlled trial (n=516), the change in the overall quality of life score from baseline to the 4th chemotherapy cycle, as measured by the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30), was used as the primary outcome. The Rboost method was employed to establish an ITE estimation model. Internal validation was performed using 500 bootstrap resampling iterations. The model's ability to discriminate between beneficiaries was evaluated using the C-for-benefit. Patients were divided into high- and low-benefit groups based on the median ITE. A linear regression model incorporating an interaction term between treatment allocation and benefit group was constructed to verify treatment effect heterogeneity. Differences in baseline characteristics between the two groups were compared. ResultsThe established Rboost model yielded a C-for-benefit of 0.52 (95%CI 0.38 to 0.66). The median individual treatment effect was 5.84 (range: –0.86 to 12.50). The estimated individual treatment effect was 3.79 (95%CI –0.25 to 5.74) in the low-benefit group and 7.89 (95%CI 5.98 to 11.35) in the high-benefit group. After adjusting for covariates including sex, age, and BMI, the heterogeneity of treatment effect was statistically significant (P=0.01). Comparisons between benefit groups revealed that patients who were male, had a smoking history, had higher BMI, had squamous cell carcinoma, were at stage ⅢA, or had higher levels of white blood cells, neutrophils, hemoglobin, total bilirubin, aspartate aminotransferase, or alanine aminotransferase derived greater benefit from treatment with SOL. ConclusionThe Rboost model can effectively identify the advantageous population for SOL. The findings provide evidence-based support for the precise adjuvant use of traditional Chinese medicine during chemotherapy.