Output guide

What to Expect

A breakdown of what each tool produces and how to use the outputs in your manuscript.

Study Designer

AI-Powered Study Protocol

Enter a clinical research question in natural language. The Study Designer parses it into PICO components, searches PubMed for relevant literature, and generates a complete TriNetX study protocol grounded in published evidence.

What you get

PICO Framework

Population, Exposure/Intervention, Comparator, and Outcome parsed automatically from your question.

Cohort Construction Steps

Numbered steps with ICD-10 and CPT codes, plus TriNetX-specific instructions for building each cohort.

Outcome Definitions

Primary and secondary outcomes with ICD-10 codes, measurement windows, and literature citations.

Covariates for Matching

Suggested covariates with rationale and codes — ready for PSM setup in TriNetX.

Statistical Analysis Plan

Recommended methods (Cox regression, logistic regression, KM, etc.) with evidence-based rationale.

Suggested Figures

Recommended visualizations (forest plots, KM curves, flowcharts) for your manuscript.

Supporting Literature

Full citations from PubMed with direct links — every recommendation is traceable to published evidence.

PDF Export

Download the complete study design as a formatted PDF for sharing with collaborators or IRB review.

Analyzer

Upload your TriNetX CSV exports exactly as downloaded — no preprocessing, renaming, or reformatting required. Each analysis produces a ZIP file containing publication-ready outputs.

Compare Outcomes

Uses CSVs from the TriNetX Compare Outcomes analysis. Each CSV is one outcome comparison between two cohorts, containing cohort statistics, risk difference, risk ratio, and odds ratio blocks.

.csvSTACKED_OUTCOME_TABLE.csv — raw stacked data with all timepoints

.docxSTACKED_OUTCOME_TABLE.docx — formatted Word table with bold section headers, right-aligned counts, bold significant p-values (Times New Roman, 10pt)

.pdfforest_plot.pdf — publication forest plot with odds ratios, 95% CI whiskers, and timepoint-prefixed labels (300 DPI, embedded fonts)

Table columns

Outcome | Cohort 1 Events | Cohort 2 Events | Odds Ratio (95% CI) | P-value

Section headers

Each folder you upload becomes a bold section header row showing the folder name and cohort N values. Upload folders in the order you want them stacked (e.g. 90 Days, 1 Year, 3 Years).

P-value formatting

<0.001 shown as <.001**, <0.05 as <.05*, others shown to 3 decimal places. Significant values are bold in the Word doc.

PSM Baseline

Before/after matching tables with standardized mean differences and p-values from PSM baseline CSV exports.

.csv*_PSM_STACKED.csv — long-format stacked table (Before + After matching)

.xlsx*_PSM_JOURNAL.xlsx — journal-formatted hierarchical table

.docx*_PSM_JOURNAL.docx — publication Word table with merged variable cells, italic variable names, bold p-values, proper column widths

Table columns

Layout

"Before matching" and "After matching" appear as bold full-width divider rows. Each demographic variable shows Cohort 1 and Cohort 2 data in consecutive rows with the variable name vertically merged.

Cohort names

If TriNetX embeds cohort names in the CSV headers (e.g. "Depression", "No Depression"), they're extracted automatically. Otherwise, the names you enter in the form are used.

Kaplan-Meier

Beta — Free

Upload pairwise KM comparison CSVs from TriNetX. Each file compares one exposure group against a reference. The first number in the filename identifies the group.

.csvresults_summary.csv — group, PSM N, events, survival probability, HR, CI, p-value

.pdfpublication_summary_table.pdf — color-coded results table (green = reference, red = baseline)

.pdfforest_plot.pdf — hazard ratio forest plot with 95% CI

.pdfsurvival_probability_chart.pdf — bar chart of end-of-window survival by group

.pdfdose_response.pdf — dose-response curve with 95% CI band

Input format

Upload pairwise KM comparison CSVs (e.g. "2_v_1_TKA_2_years.csv"). The first number in the filename identifies the exposure group.

Reference group

The reference group is auto-detected from the comparison files (typically the second-smallest group). It appears as HR = 1.0 in all plots. Group 0 is highlighted as the baseline.

Customization

You provide the Outcome Name (used in titles and y-axis labels) and Exposure Label (used as x-axis label) on the upload form.

Absolute Risk

Beta — Free

Upload Measures of Association CSVs from TriNetX (one per exposure group). The first number in each filename identifies the group. Files with "no_" or "none" in the name are mapped to group 0.

.csvrisk_summary.csv — absolute risk, risk difference, risk ratio, NNH per group

.pdfabsolute_risk_bar.pdf — color-coded bar chart with % labels and sample sizes

.pdfrisk_difference_bar.pdf — absolute risk difference vs baseline (percentage points)

.pdfnnh_curve.pdf — Number Needed to Harm curve

Baseline

Group 0 is used as the baseline for risk difference and NNH calculations by default.

Customization

You provide the Outcome Name and Exposure Label on the upload form. These appear on all chart titles and axis labels.

Cox Trend

Beta — Free

Upload Cox Proportional Hazards text exports from TriNetX (one per exposure level). The first number in each filename identifies the exposure group.

.csvresults.csv — comparison, HR, coefficient, SE, z-stat, 95% CI, p-value, interpretation

.pdfforest.pdf — hazard ratio forest plot with error bars

.pdfdose_response_styled.pdf — publication-quality dose-response with CI band, color-coded markers, reference annotation

Trend test

A linear trend of log(HR) vs exposure is computed automatically. The R-squared and slope are included in the output.

Why These File Types?

DOCX

Word Documents

Tables are delivered as native .docx files — not images or PDFs. This means you can directly edit cell values, reword headers, adjust column widths, change fonts, or copy the table into your manuscript. No retyping, no screenshot cropping. The formatting (bold p-values, merged cells, borders) is already set to journal standards, but every element is fully editable in Word or Google Docs.

PDF

Vector PDF Figures

All plots are saved as vector PDFs, not rasterized PNGs. They scale to any size without losing quality — perfect for journal figures that get resized during typesetting. Fonts are embedded as Type 42 (TrueType) at 300 DPI so they render identically on any system. Open them in Adobe Illustrator, Inkscape, or Affinity Designer to tweak before submission.

CSV

Raw CSV Data

Every analysis includes the underlying data as a CSV file alongside the formatted outputs. Import into Excel, R, SPSS, or any tool for custom analyses, supplementary tables, or to verify the computed statistics. The CSV is what the Word table and figures are built from — a complete audit trail.

XLSX

Excel Workbooks (PSM)

PSM baseline tables include an Excel fileas an intermediate format. Re-sort variables, add calculated columns, or use Excel's built-in formatting tools before final publication. The companion Word doc is generated from this Excel file automatically.

General Notes

Delivery

All analyzer outputs are delivered as a single .zip download. Unzip and you have everything — data, tables, and figures ready for your manuscript.

Figure Quality

300 DPI with embedded fonts. Figures meet the submission requirements of all major medical journals (JAMA, NEJM, Lancet, etc.) without further processing.

Word Formatting

Times New Roman 10pt, publication-standard borders (heavy top/bottom, light internal), bold significant p-values. Matches what reviewers expect to see.

Input Compatibility

Accepts standard TriNetX CSV exports exactly as downloaded. No preprocessing, renaming, or reformatting required.

Processing Speed

Analyzer results are generated in under 30 seconds. Study Designer protocols take 30-60 seconds depending on PubMed search depth.

Complete Workflow

Use the Study Designer to plan your research, run the study in TriNetX using the generated protocol, then upload the results to the Analyzer for publication-ready outputs.

Ready to get started?

Design a study or upload your TriNetX exports now.

Study Designer Analyzer