====== Observational study ======

In this type of study, researchers observe and record data without intervening or manipulating variables. They study how variables naturally interact with each other in real-world settings.
----

===== Types of Observational Studies =====

Observational studies are research designs where investigators observe and analyze subjects without intervening or assigning specific treatments. They help identify associations between exposures and outcomes in real-world settings.

----

=== 1. Cohort Studies ===
A **cohort study** follows a group of people (cohort) over time to examine how exposures affect outcomes.

**Types:**
  * **Prospective Cohort Study**: The researcher follows participants ''forward in time'' from exposure to outcome.
  * **Retrospective Cohort Study**: The researcher uses past data to track exposure and outcome relationships.

**Example:** Following smokers and non-smokers for 10 years to observe lung cancer rates.

**Advantages:**
  * Establishes **temporal relationships**.
  * Can study **multiple outcomes**.
  * Less recall bias compared to case-control studies.

**Limitations:**
  * Expensive and time-consuming.
  * Loss to follow-up can affect validity.

----

=== 2. Case-Control Studies ===
A **case-control study** compares individuals with a specific outcome (**cases**) to those without it (**controls**) to identify past exposures.

**Example:** Studying patients with brain tumors (**cases**) and comparing them to a group without brain tumors (**controls**) to investigate mobile phone use as a risk factor.

**Advantages:**
  * Quick and inexpensive.
  * Good for **rare diseases**.
  * Requires fewer participants than cohort studies.

**Limitations:**
  * **Recall bias**: Patients may inaccurately remember past exposures.
  * **Cannot establish causality**, only associations.

----

=== 3. Cross-Sectional Studies ===
A **cross-sectional study** collects data at a **single point in time** to analyze associations between exposure and outcome.

**Example:** A survey measuring **obesity** and **physical activity** in a population at one point in time.

**Advantages:**
  * Quick and low-cost.
  * Useful for assessing **disease prevalence**.
  * Helps generate hypotheses for further studies.

**Limitations:**
  * **Cannot determine causality**.
  * Susceptible to **survivor bias**.

----

=== 4. Ecological Studies ===
An **ecological study** analyzes data at the **population level**, rather than individuals, to identify trends and associations.

**Example:** Comparing **air pollution** levels and **asthma rates** across different cities.

**Advantages:**
  * Useful for generating public health policies.
  * Relatively simple and inexpensive.

**Limitations:**
  * **Ecological fallacy**: Associations at the population level may not apply to individuals.
  * Limited ability to control for confounders.

----

=== 5. Registry-Based Studies ===
A **registry-based study** uses **pre-existing data** from patient registries to study outcomes, trends, and treatment effectiveness.

**Example:** Using a **stroke registry** to analyze the impact of **thrombolysis** on patient survival.

**Advantages:**
  * Large sample size with real-world data.
  * Cost-effective compared to prospective studies.

**Limitations:**
  * Limited by data quality and completeness.
  * Risk of confounding variables.

----

=== Comparison Table ===

^ Study Type ^ Timeframe ^ Data Collection ^ Best for ^ Weaknesses ^
| **Cohort** | Longitudinal (past or future) | Exposure → Outcome | Rare exposures, multiple outcomes | Expensive, long duration |
| **Case-Control** | Retrospective | Outcome → Exposure | Rare diseases, quick studies | Recall and selection bias |
| **Cross-Sectional** | Single point in time | Exposure & Outcome | Disease prevalence, correlation studies | No causality, survivor bias |
| **Ecological** | Aggregate data | Population-level exposure | Public health trends | Ecological fallacy |
| **Registry-Based** | Retrospective or prospective | Pre-existing registry data | Treatment effectiveness, real-world data | Data quality limitations |

----

=== Choosing the Right Study Type ===
  * **If studying rare diseases → Case-control.**
  * **If assessing exposure before outcome → Cohort.**
  * **If needing quick prevalence estimates → Cross-sectional.**
  * **If analyzing population trends → Ecological.**
  * **If using healthcare databases → Registry-based.**

----