System Overview

🧠Theory and Research
1

Concept

Structured questions, clear literature retrieval strategies (collected from around the world with certain inclusion criteria), quality assessment of eligible literature, data extraction, drawing conclusions, and timely updating of data afterward.

Conventional Meta-Analysis Production Steps

Whether intending to publish a Cochrane systematic review or a non-Cochrane systematic review, the core steps of production are the eight steps just mentioned. The only difference is that publishing a Cochrane systematic review requires regular updates and demands more detailed specifications.

  1. Topic selection: choose research content that is controversial, has large differences in results among original studies, and is clinically meaningful.
  2. Determine inclusion/exclusion criteria: based on the topic, set inclusion and exclusion criteria from the perspectives of study subjects, interventions, outcome indicators, etc.
  3. Literature retrieval: formulate search strategies (PICOS principle, determine keywords) + search databases (CNKI/CBM/PubMed, etc.)
  4. Study selection: screen and select valid studies based on inclusion and exclusion criteria (EndNote/NoteExpress software)
  5. Quality assessment of literature: Cochrane tool (for RCTs), NOS scale (for cohort and case-control studies)
  6. Extract bibliographic and data information: extract sample sources, basic sample information, outcome variable data, etc.
  7. Data synthesis and statistical interpretation: heterogeneity, random/fixed effects, sensitivity, etc. (STATA/RevMan software)
  8. Result interpretation: clinical significance of meta-analysis results, study limitations, and related explanations

Steps for Writing a Meta-Analysis Paper

  1. Title
  2. Abstract
    1. Objective
    2. Methods
    3. Results
  3. Keywords
  4. Research background
    1. Disease status, harm, burden
    2. Current research status on the disease
    3. The value of this study and the problems it solves
  5. Materials and methods
    1. Inclusion and exclusion criteria
    2. Literature retrieval strategy
    3. Literature screening and data extraction
    4. Risk of bias assessment of included studies
    5. Statistical analysis
  6. Results
    1. Literature screening results
    2. Basic characteristics of included studies
    3. Literature quality assessment results
    4. Meta-analysis results
  7. Discussion
    1. Summary of meta-analysis results
    2. Study limitations and weaknesses
    3. Clinical significance of meta-analysis conclusions
    4. Prospects and implications
  8. References
  9. Appendices

Effect Size

The essence of meta-analysis is combining results (effect sizes) from multiple similar original studies. Clinical research has various types, such as case-control studies, cohort studies, cross-sectional studies, randomized controlled trials, before-and-after studies, etc.

Effect size in case-control studies: OR/HR/median survival time
Effect size in cohort studies: RR/HR/median survival time
Effect size in randomized and non-randomized controlled studies: MD/RR/OR
Effect size in cross-sectional studies: incidence, mortality, prevalence

Basic Clinical Research Concept Classification

graph TD

A[Whether the study has an intervention factor] -->|Yes| B[Experimental Study]
A -->|No| C[Observational Study]

%% ---- Experimental Studies ----
B --> D{Is it randomly assigned?}
D -->|Yes| E[Randomized Controlled Trial]
D -->|No| F[Non-randomized Controlled Trial]

E --> G{Is the intervention cross-over?}
G -->|Yes| H[Cross-over Trial]
G -->|No| I[Parallel Trial]

F --> J[Concurrent Controlled Trial (similar to parallel trial but without randomization)]
F --> K[Non-concurrent Controlled Trial]
K --> L[Self before-and-after controlled]
K --> M[Different cases before-and-after controlled]

%% ---- Observational Studies ----
C --> N{Is there a control group?}
N -->|No| O[Descriptive Study (can only propose hypotheses, cannot confirm)]
N -->|Yes| P[Analytical Study (can verify hypotheses)]

O --> Q[Cross-sectional Study]
O --> R[Case/Case Series Study]

P --> S[Case-control Study (outcome → cause, can analyze possible causes of existing cases)]
P --> T[Cohort Study (cause → outcome, observe from now whether specific causes lead to disease)]

Principles for Topic Selection

  • Important
    • ① Provide important evidence for clinical practice (solve practical problems)
    • ② Research conclusions have clinical guidance significance
    • ③ Research conclusions benefit the subjects involved
  • Controversial
    • ① Contradictions exist between different research results (literature)
    • ② Clinical practice conclusions differ greatly
    • ③ Single independent studies have low credibility (small sample studies, studies with poor trial design)
  • Feasible
    • ① There is a certain number of original studies based on the topic
    • ② There is a certain number of high-quality original studies based on the topic
    • ③ Original studies have clear research methods and clear effect size values
    • ④ The topic should be neither too large nor too small (must consider workload of conducting research)
  • Innovative
    • ① Choose content and directions that others have not done
    • ② For content already done by others, analyze from other perspectives or update
    • ③ For content currently in progress by others, analyze from other perspectives (search databases for already published Meta, and check ongoing Meta on PROSPERO, Inplasy registration platforms)
  • Scientific — Redundant
    • The topic must have theoretical and practical basis and must conform to scientific principles and natural laws