TRUST Research Center

07/05/2026

🍎🍊 “We can’t compare apples to oranges. Or May be we can!!!"

Propensity Score Matching (PSM)
---

⚠️ **The core problem: Selection Bias**

In real-world studies, treatment groups are often different *before treatment even starts.*

Example:

💊 New treatment group → younger, lower-risk patients
🏥 Standard treatment group → older, sicker patients

So if outcomes differ later…
❌ it may NOT be because of the treatment itself.

This is the classic:

🍎 **Apples vs Oranges problem**

---

🎯 **What PSM tries to do**

PSM creates *fairer comparisons* by matching patients with similar baseline characteristics.

Think of it as:

🧍‍♂️ Matching each treated patient
with
🧍‍♀️ a similar control patient

based on:
✔ age
✔ s*x
✔ comorbidities
✔ disease severity
✔ other baseline variables

---

📌 **The 4-step idea behind PSM**

1️⃣ Collect baseline characteristics
2️⃣ Calculate a “propensity score”
→ probability of receiving the treatment

3️⃣ Match patients with similar scores
4️⃣ Compare balanced groups

⚖️ After matching, groups become more comparable → less biased analysis

---

💡 Important insight:

PSM does **NOT** create perfect randomization.

It only balances:
✔ measured variables

But:
❌ hidden or unmeasured confounders may still exist

That’s why:
🧪 RCTs remain the gold standard
📊 But PSM greatly improves observational studies

---

📌 Bottom line:

Without adjustment:
🍎 ≠ 🍊

With proper matching:
🍎 🤝 🍊

---

To learn more about observational studies and PSM, join our upcoming course

👉 https://wa.me/201119678899?text=I%20want%20to%20join%20the%20observationa%20studies%20course

07/05/2026

📊 **Correlation vs Regression — one of the most misunderstood concepts in research**

Many treat them as interchangeable… they’re not.

This infographic breaks it down clearly 👇

---

🔗 **Correlation = Relationship (No direction)**
It answers: *Are two variables linked?*

✔ Measures **strength + direction** (r)
✔ **Symmetrical** → X with Y = Y with X
✔ No prediction, no causation

---

📈 **Regression = Prediction (Directional model)**
It answers: *Can X predict Y?*

✔ Builds an equation:
Y = b₀ + b₁X

✔ **Directional** → X predicts Y
✔ Quantifies **how much Y changes when X changes**

---

🏥 **Simple example:**

💰 Hospital expenses = fixed costs + (cost per patient × number of patients)

👉 **Intercept (b₀):** Running costs even with zero patients
👉 **Slope (b₁):** Added cost per patient

📊 If R² = 79% → most variation in cost is explained by patient numbers

---

⚠️ **The BIG trap: Correlation ≠ Causation**

🍦 A real-life analogy:

A man noticed:
• When he ordered **vanilla ice cream** → his car didn’t work ❌
• When he ordered **strawberry** → his car worked ✔️

At first glance:
👉 Ice cream flavor seems “linked” to car failure!

But the real reason was:

⏱️ **Preparation time**
Vanilla was served faster → he returned to the car sooner → engine still overheated
Strawberry took longer → engine had time to cool → car worked fine

💡 Hidden factor = **waiting time**, not the flavor

👉 So yes, there is **correlation**
❌ But no **causation**

---

📌 **What to report in research:**

✔ Correlation → r, p-value
✔ Regression → slope (B), CI, R²
✔ Always include descriptive statistics

---

💡 Bottom line:
Correlation tells you **“they move together”**
Regression tells you **“how one predicts the other”**
Neither proves causation on its own.

---

If you want to *actually understand statistics and apply it in clinical research*…

👉 Message us here to join **Statistics for Clinicians (S4C):**
https://wa.me/201119678899?text=I%20want%20to%20join%20the%20Statistics%20for%20Clinicians%20course

04/05/2026

Bias can silently undermine your research ⚠️

This infographic highlights the **main sources of bias in observational studies**:
🔹 Selection Bias (who gets included)
🔹 Information Bias (how data is collected)
🔹 Confounding & Interaction (hidden variables)
🔹 Detection & Time Biases (when and how outcomes are measured)

💡 It also shows practical strategies to **prevent and adjust for bias** during study design, data collection, and analysis.

Because strong research is not just about results — it is about how reliable those results are.

📩 Want to confidently identify and handle bias in your studies? join our upcoming course:

👉 https://wa.me/201119678899?text=I%20want%20to%20join%20the%20Observational%20Studies%20course

04/05/2026

Observational study designs made simple 👇

This infographic breaks down the core analytical designs:
🔹 Case-Control → looking from present to past
🔹 Cohort → following from present to future
🔹 Cross-Sectional → a snapshot at one point in time
🔹 Hybrid designs (Nested case-control & Case-crossover)

💡 Each design serves a different purpose — choosing the right one is critical for generating valid evidence.

If you work with real-world data or are planning a research project, understanding these designs is essential.

📩 Want to learn how to choose and apply the right design with confidence? click on the following link

👉 https://wa.me/201119678899?text=I%20want%20to%20join%20the%20Observational%20Studies%20course

03/05/2026

📊 Comparing 3 or more groups? This is where most people go wrong

Many still run multiple t-tests ❌
That’s a fast track to **false positives (Type I error inflation)

👉 The correct approach: **One-Way ANOVA**

This infographic simplifies it—but here’s the key idea using a simple analogy 👇

🏠 The “House Analogy” (ANOVA made intuitive)

Think of your data as a house:

🏠 = Total variability (everything happening in your data)
🚪 Rooms = Groups (e.g., different treatments)

Now inside the house:

👥 Inside each room → Within-group variability
(people differ even if they receive the same treatment)

🚶‍♂️ Space between rooms → Between-group variability
(differences caused by the treatment itself)

🔍 What ANOVA does:
It compares between-group variation vs within-group noise

👉 If rooms are very different from each other (big between-group)
👉 And people inside each room are similar (low within-group)

💥 Then you get a **significant result**

---

📌 Example: Different patient groups (cirrhosis with HCC, without HCC, normal) showed clear differences in fasting blood sugar → confirmed by a strong F-ratio and very small p-value.

---

⚠️ **But don’t stop at “p < 0.05”**

A proper ANOVA report MUST include:
✔ Means & SDs for each group
✔ F-statistic + degrees of freedom + p-value
✔ Effect size (η² or ω²)
✔ Post-hoc tests (to know *which* groups differ)

---

💡 Bottom line:
ANOVA tells you **“there is a difference”**
Post-hoc tests tell you **“where the difference is”**

---

If you want to confidently apply this in your research and *actually understand what you're doing*…

👉 Message us here to join **Statistics for Clinicians (S4C):**
https://wa.me/201119678899?text=I%20want%20to%20join%20the%20Statistics%20for%20Clinicians%20course

03/05/2026

📊 **Struggling to choose the right statistical test? Start here.**

The **Unpaired t-test** is one of the most essential tools in clinical research—but many people use it without fully understanding *when*, *why*, and *how*.

This infographic simplifies it for you:

🔹 **When to use it?**
To compare the means between two independent groups (e.g., Drug A vs Drug B)

🔹 **Key idea:**
You’re comparing **signal (mean difference)** to **noise (standard error)** → giving you the *t-value*

🔹 **Critical insight:**
Small samples = more uncertainty → higher critical values → heavier tails (that’s why we use the t-distribution, not Z)

🔹 **What most people miss:**
✔ Checking variance (equal vs unequal) changes the calculation
✔ Reporting is not just p-value — you need CI + effect size (Cohen’s d)

🔹 **Real-world takeaway:**
Statistical significance is not enough—interpret magnitude and clinical relevance.

---

💡 If you want to **master statistics in a practical, clinician-friendly way**, join our upcoming **Statistics for Clinicians (S4C)** course. click on the following link and click send:
https://wa.me/201119678899?text=I%20want%20to%20join%20the%20Statistics%20for%20Clinicians%20course

30/04/2026

🔬📊 THE CLINICIAN'S GUIDE TO CHI-SQUARE TESTING 📊🔬
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Ever wondered how researchers test whether occupation influences disease — or whether a drug truly changes outcomes? The answer lies in one powerful tool: the Chi-Square Test (χ²) ✅

🧩 WHAT DOES IT DO?
✔️ Goodness of Fit → Does your observed data match a theoretical model?
✔️ Test of Independence → Are two qualitative variables associated with each other?

🏥 REAL CLINICAL EXAMPLES:
👨‍⚕️ Are doctors more hypertensive than nurses?
💊 Does the type of anticoagulant affect thromboembolic complications?
👶 Is the gender ratio of newborns consistent with the expected 1:1 ratio?

🧮 THE FORMULA:
χ² = Σ [(O − E)² / E]
→ Sum the standardized squared differences between Observed (O) and Expected (E) frequencies

📏 KEY RULES TO REMEMBER:
📌 Reject H₀ when p < 0.05
📌 Always report Effect Size (Phi φ or Cramér's V) — not just p-value!
📌 Apply Cochran's Rule: 80% of cells must have expected count > 5
📌 If conditions are violated → use Yates' Correction or Fisher's Exact Test

🔍 WHEN YOU HAVE >2 GROUPS:
Don't stop at the overall chi-square!
➡️ Perform Post Hoc 2×2 comparisons
➡️ Apply Bonferroni correction to avoid inflated p-values

💡 PRO TIP:
A significant p-value tells you THAT a difference exists.
Effect Size (like Odds Ratio or Relative Risk) tells you HOW BIG that difference is.
Both matter! 🎯

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎓 Save this post for your next research project!
📲 Share with your medical colleagues & students
❤️ Like if you found this helpful!

23/04/2026

📊 Measure of Uncertainty — What does the p-value really tell us?

Let’s simplify it with a practical example 👇

Imagine 200 patients randomized into two groups:
🔹 Treatment A → 90% success
🔹 Treatment B → 80% success

That’s a 10% difference. Sounds convincing, right?

When we analyze this difference, we get a p-value = 0.047 (4.7%)

💡 So what does this actually mean?

The p-value is:
👉 The probability of observing this 10% difference if the two treatments were actually equal

Since 4.7% is less than the accepted 5% threshold:
✔️ We reject the idea that both treatments are equal
✔️ We conclude that Treatment A is likely superior

—but here’s the critical part many people miss 👇

⚠️ The p-value is NOT telling you how big or important the difference is**
👉 The 10% difference* shows the size (effect)
👉 The p-value* shows how uncertain we are about that conclusion

🧠 Think of it this way:

* Smaller p-value → more confidence in your conclusion
* Larger p-value → more uncertainty
* But NEVER a measure of effect size or clinical importance

🎥 Watch the full video “Measure of Uncertainty” to grasp this concept clearly and avoid one of the most common mistakes in research interpretation.
https://youtu.be/zMpYHcKUIJY?si=XrNf0_qylrwIrrME

3 likes. "TRUST Pills - Measure of uncertainty"

20/04/2026

🎯 What’s the real meaning of “statistically significant”?

In biological research, variability isn’t a flaw—it’s the rule. Even when patients receive the *same* treatment, their responses can vary widely: some improve dramatically, others moderately, and some may not improve at all.

So here’s the reality:
We can *never* be 100% certain in our conclusions.

👉 That’s why researchers made a “deal”:

✔️ Every study must calculate the probability of error (using statistical tests)
✔️ The scientific community agrees to take results seriously only if this error is ≤ 5%

This 5% threshold is what we call alpha (α) — the acceptable risk of being wrong.

And the value we calculate in each study?
That’s the p-value.

💡 So what does “statistically significant” actually mean?
It simply means:
➡️ The p-value should respect alpha (0.05)

📌 If p ≤ 0.05 → Statistically significant
📌 If p > 0.05 → Not statistically significant

No magic. No mystery. Just a practical agreement to manage uncertainty in science.

🎥 Watch the full video: https://www.youtube.com/watch?v=QTzrL7P1Uxo to understand this concept in a simple, intuitive way.

2 likes. "TRUST Pills - The Deal"

20/04/2026

🚀 Starting in 10 Days! Don’t Miss Out

Our course “Validity, Reliability & Diagnostic Accuracy Measures” is starting in just 10 days.

If you’re aiming to strengthen your research skills and confidently handle statistical concepts, this course will take you step by step through:
📊 Validity & reliability
📈 Questionnaire validation
📉 Agreement measures
🔍 Diagnostic accuracy analysis

Whether you're a beginner or looking to refine your understanding, this course is designed to make complex concepts clear and practical.

📩 Get full course details here:
https://wa.me/201119678899?text=Validity+Reliability+course+details

📢 Join our WhatsApp channel for updates and more educational content:
https://whatsapp.com/channel/0029VbCipww4yltWlq7IDY2J

17/04/2026

Ever wondered how we prove a new treatment is NOT WORSE THAN the standard one? 🤔
Welcome to the world of Non-Inferiority & Equivalence Studies 🔬✨

👉 Non-inferiority = proving the new treatment is not worse than the standard
🚫 But it does NOT mean they are equal!
👉 Equivalence = essentially two non-inferiority tests combined to show both treatments are similar

🧩 What You’ll Learn:
1️⃣ Distinguish between superiority, non-inferiority & equivalence
2️⃣ Calculate unilateral 95% & bilateral 90% confidence intervals 📊
3️⃣ Define non-inferiority margins & equivalence limits
4️⃣ Perform sample size calculations
5️⃣ Analyze using Intention-to-Treat & Per-Protocol methods
6️⃣ Interpret results & recognize limitations 🧠

💰 Research Design Package Offer:
📚 Observational Studies
📚 Clinical Trials
📚 Non-inferiority & Equivalence Studies

💡 1 Course: 1200 EGP
💡 2 Courses: 2100 EGP (instead of 2400)
💡 3 Courses: 2900 EGP (instead of 3600)

🎟️ Register now:
https://trustresearch.org/event/non-inferiority-equivalence-studies-6/

📲 Get details on WhatsApp:
https://wa.me/201119678899?text=Non-inferiority+Course+Details

🔥 Master one of the most misunderstood concepts in clinical research and boost your analytical edge!