How to Create a Systematic Review Data Extraction Table

A data extraction table collects what you need from each included study in a structured grid, with one row per study and columns for the variables you plan to analyse. A characteristics-of-included-studies table in the final paper is a cleaned version of it.

Two tables, two purposes

1. Working extraction sheet: large, detailed, used by the team. Includes all variables, source locations and notes.
2. Published summary table: concise, with key characteristics that help readers judge each study.

Suggested columns for the working sheet

Study ID | Title | Country/setting | Design | N | Population | Intervention | Comparator | Outcome definition and time | Result (effect and CI) | Funding/conflicts | Risk of bias | Notes | Extracted by | Verified by | Date.

For the published table

Include first author and year, country, design, sample size, participants (age, condition), intervention and comparator, main outcomes, follow-up and risk of bias. Keep entries short and consistent, and use footnotes for abbreviations.

Formatting tips

• Use data validation (dropdown lists) for fields with fixed options.
• Use "NR" for not reported and leave a note if calculated.
• Freeze the header row, and keep one study per row.
• Back up and version the file.

Illustrative example

Invented for teaching. Row: Smith 2019 | Kenya | RCT | 240 | Adults with hypertension | Community health worker follow-up | Usual care | Systolic BP at 6 months | MD −5.1 mm Hg (95% CI −8.0 to −2.2) | Government funding | Some concerns. (Invented values.)

Adding risk-of-bias and certainty information

Many reviews add a column or a linked table for domain-level risk-of-bias judgements, so readers can see why a study was rated as it was. Keep the judgement and the supporting quotation from the paper in the working sheet. For the published version, consider a separate figure or table rather than crowding the characteristics table, and place details in a supplement if the journal limits tables.

Common pitfalls

• Too many columns that no one uses.
• Inconsistent units across rows.

Key takeaways

• Separate the working sheet from the published table.
• Make fields consistent, sourced and verifiable.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.