Prediction model developed on the basis of meta-analysis in the field of medicine: a systematic survey and methodological summaries
Abstract Background We performed a systematic survey to collate studies on prediction models developed on the basis of meta-analysis in medicine and summarize key steps involved in model development. Methods We systematically searched Web of Science, PubMed and Embase until April 2023 and included t...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-08-01
|
| Series: | BMC Medical Research Methodology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12874-025-02639-6 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Background We performed a systematic survey to collate studies on prediction models developed on the basis of meta-analysis in medicine and summarize key steps involved in model development. Methods We systematically searched Web of Science, PubMed and Embase until April 2023 and included those that developed prediction models on the basis of meta-analysis. The data were summarized via narrative synthesis. Results The search strategy identified 23 studies with 25 prediction models that met the eligibility criteria. The predicted outcomes focused on complications of diabetes, mortality, cognitive impairment, gestational diabetes, bronchopulmonary dysplasia, bacterial sinusitis, seizures, and psychosis. Twenty-three prediction models reported AUCs, with a median value of 0.77 (range: 0.59–0.91). Ten prediction models were developed with sample sizes exceeding 10,000 participants. The key steps in developing prediction models via meta-analysis include (1) confirming that the prediction model meets clinical needs before starting, (2) collecting data through meta-analysis to select predictors, (3) developing prediction models, (4) validating prediction model performance, (5) presenting and interpreting models and (6) reporting models. Conclusion The development of prediction models based on meta-analysis may represent a promising approach, which demonstrates good discrimination. Guidance is necessary to support the design, implementation and reporting of future prediction model development via meta-analysis. |
|---|---|
| ISSN: | 1471-2288 |