clinical-2x2-study-analyzer

description: Extracts binary outcome data from raw clinical study text, reconstructs 2x2 contingency tables, and produces publication-ready epidemiological risk analysis including RR, OR, ARR, NNT, confidence intervals, and statistical tests. name: clinical-2x2-study-analyzer tools: python

Clinical Study 2×2 Outcome Analyzer

Purpose

Analyze clinical research text describing binary outcomes and automatically:

1. Extract treatment and control group outcome data
2. Construct a 2×2 contingency table
3. Perform epidemiological risk calculations
4. Conduct statistical significance testing
5. Produce publication-ready results suitable for manuscripts

Capabilities

✔ Parse raw study text
✔ Detect treatment and control arms
✔ Extract event and non-event counts
✔ Construct contingency tables
✔ Calculate clinical effect measures
✔ Compute confidence intervals
✔ Run hypothesis tests
✔ Produce publication-quality summaries

Input

Accept raw clinical study text.

Example:

In a randomized trial of 400 patients, 200 received Drug A and 200 received placebo. In the Drug A group, 45 patients experienced the primary outcome while 155 did not. In the placebo group, 60 patients experienced the outcome and 140 did not.

Alternate formats:

Drug A: 45/200 events
Placebo: 60/200 events

Treatment group had 12 deaths out of 100 patients compared to 20 deaths among 100 controls.

Step 1 --- Information Extraction

Detect treatment and control arms using keywords.

Treatment indicators: treatment, intervention, drug, exposed, experimental

Control indicators: control, placebo, standard care, unexposed

Extract:

events_treatment
total_treatment
events_control
total_control

Step 2 --- Reconstruct 2×2 Table

          Event   No Event

Treatment a b Control c d

Where:

a = events_treatment
b = total_treatment − events_treatment
c = events_control
d = total_control − events_control

Apply Haldane--Anscombe correction if any cell is zero.

Step 3 --- Event Rates

Experimental Event Rate

EER = a / (a + b)

Control Event Rate

CER = c / (c + d)

Step 4 --- Effect Measures

Relative Risk

RR = EER / CER

Odds Ratio

OR = (a * d) / (b * c)

Risk Difference

RD = EER − CER

Absolute Risk Reduction

ARR = CER − EER

Relative Risk Reduction

RRR = ARR / CER

Step 5 --- Clinical Impact

Number Needed to Treat

NNT = 1 / ARR

Round upward.

Number Needed to Harm

NNH = 1 / ARI

Step 6 --- Confidence Intervals

Relative Risk

SE = sqrt((1/a) − (1/(a+b)) + (1/c) − (1/(c+d)))

lower = exp( ln(RR) − 1.96 * SE ) upper = exp( ln(RR) + 1.96 * SE )

Odds Ratio

SE = sqrt(1/a + 1/b + 1/c + 1/d)

lower = exp( ln(OR) − 1.96 * SE ) upper = exp( ln(OR) + 1.96 * SE )

Step 7 --- Statistical Tests

Chi-square test

χ² = Σ (observed − expected)² / expected

Fisher exact test for small samples.

Step 8 --- Publication Table

Group Event No Event Total Event Rate

Treatment a b a+b EER Control c d c+d CER

Step 9 --- Manuscript Results

Report:

Relative Risk (95% CI)
Odds Ratio (95% CI)
Absolute Risk Reduction
Relative Risk Reduction
Number Needed to Treat

Step 10 --- Interpretation

Provide neutral academic interpretation describing effect size, direction, and statistical significance.

Step 11 --- Meta-analysis Output

treatment_events
treatment_total
control_events
control_total
RR
logRR
SE_logRR

Output Order

1. Extracted data
2. Contingency table
3. Event rates
4. Effect measures
5. Confidence intervals
6. Hypothesis tests
7. Publication-ready table
8. Manuscript-ready summary

Quality Guidelines

Use precise calculations internally.
Round to 2--3 decimals for reporting.
Prefer Fisher test when expected counts are small.
Always present absolute and relative risks together.