Cohort Studies, Quality Control
Conditions
Brief summary
This study is based on the Nicheng Cohort study. This study intends to analyze whether AI assistance can effectively improve the efficiency and accuracy of quality control of data collected in large-scale epidemiological surveys based on traditional quality control processes.
Detailed description
This study randomly divided quality control personnel into an experimental group and a control group. The experimental group adopts AI assisted quality control: the AI system automatically transcribes the recorded text of the questionnaire, extracts keywords, analyzes the consistency of the Q&A logic, and generates quality control prompts. The quality control personnel will verify the question fragments according to the prompts and determine the qualification of the questionnaire; The control group relies entirely on manual quality control: the quality control personnel listen to the recording word by word, manually record the content, independently identify keyword omissions, logical contradictions, or terminology deviations, and ultimately determine whether the questionnaire is qualified. The core difference lies in the fact that the experimental group uses AI technology to accurately locate risk issues, reducing the burden of manual comprehensive screening, while the control group requires full manual review without targeted support.
Interventions
OTHERAI-assisted quality control tool
In this study, quality control personnel will be randomly divided into an experimental group and a control group, and will use artificial intelligence-assisted quality control and manual quality control to conduct quality control on the data collected from the epidemiological survey. Experimental group: Quality control personnel will use the AI system to perform quality control on the questionnaire recordings. Control group: The questionnaire content was checked by manually listening back to the recordings and errors were recorded manually.
Sponsors
Shanghai 6th People's Hospital
Study design
Observational model
OTHER
Time perspective
CROSS_SECTIONAL
Eligibility
Sex/Gender
ALL
Healthy volunteers
No
Inclusion criteria
1. Be proficient in using computers; 2. The person responsible for questionnaire quality control needs to have good dialect recognition ability; 3. Have a basic understanding or high acceptance of AI-assisted tools, and be able to adapt to the learning and application of new technologies 4. Be able to participate in the research throughout the process, abide by the research process, receive training, and be willing to complete quality control tasks as required.
Exclusion criteria
1. The person responsible for questionnaire quality control cannot understand or recognize Shanghai Nanhui dialect proficiently; 2. Unfamiliar with AI-assisted tools and difficult to accept technical operations; 3. Unable to participate in the research, receive training or complete the specified tasks due to other work or academic reasons;
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Accuracy | From enrollment to the end of questionnaire quality control at 8 weeks | changes in the efficiency and accuracy of quality control of data collected in large-scale epidemiological surveys with AI-assisted tool |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Completeness and consistency | From enrollment to the end of questionnaire quality control at 8 weeks | AI-assisted quality control in questionnaire quality control improves the integrity and consistency of questionnaire data through automated keyword extraction and logical consistency checking. |
Outcome results
None listed