8.4 Evidence-Based Decision Making, Research Design, Control Groups & Sampling
Key Takeaways
- The Hierarchy of Evidence ranks research designs from highest internal validity (Systematic Reviews & Meta-Analyses, Level 1) down to Randomized Controlled Trials (Level 2), Cohort Studies, Case-Control Studies, and Expert Opinion/In Vitro studies.
- Randomized Controlled Trials (RCTs) represent the gold standard for clinical intervention studies due to random allocation, control groups, and blinding techniques that minimize confounding bias.
- Prospective Cohort Studies track exposed and unexposed groups over time to determine relative risk and disease incidence, whereas Retrospective Case-Control Studies compare individuals with a disease (cases) to those without (controls) to identify past exposures (odds ratio).
- In experimental research, the Independent Variable is the intervention manipulated by the investigator (e.g., chemical agent or device), while the Dependent Variable is the measured outcome metric (e.g., plaque score or probing depth).
- Random sampling provides every individual in a population an equal and independent chance of selection, optimizing sample representativeness and generalizability, whereas convenience sampling carries the highest risk of selection bias.
8.4 Evidence-Based Decision Making, Research Design, Control Groups & Sampling
Evidence-Based Decision Making (EBDM) in dental hygiene practice is the formalized integration of three core elements: the best available scientific research evidence, clinical expertise, and the patient's unique values and preferences within specific clinical circumstances. Mastering research methodology allows dental hygienists to critically appraise scientific literature and translate clinical research into safe, effective patient care.
The Hierarchy of Evidence (The Evidence Pyramid)
Scientific study designs are structured hierarchically based on their ability to control bias and establish cause-and-effect relationships. NBDHE candidates must recognize the relative strength and positioning of research designs on the Evidence Pyramid.
+------------------------------------------------------------------------------------+
| THE EVIDENCE PYRAMID |
| |
| [LEVEL 1] Systematic Reviews & Meta-Analyses (Highest Scientific Evidence) |
| [LEVEL 2] Randomized Controlled Trials (RCTs - Gold Standard for Interventions) |
| [LEVEL 3] Cohort Studies (Prospective Observational / Relative Risk) |
| [LEVEL 4] Case-Control Studies (Retrospective Observational / Odds Ratio) |
| [LEVEL 5] Cross-Sectional Studies (Prevalence / Single Timepoint) |
| [LEVEL 6] Case Series & Case Reports (Descriptive / No Control Group) |
| [LEVEL 7] Expert Opinion, In Vitro & Animal Research (Lowest Evidence Level) |
+------------------------------------------------------------------------------------+
Level 1: Systematic Reviews & Meta-Analyses
- Systematic Review: A exhaustive, standardized synthesis of all existing high-quality clinical studies addressing a specific PICO question, using rigorous predetermined inclusion/exclusion criteria.
- Meta-Analysis: A quantitative statistical pooling technique that combines data from multiple independent studies into a single statistical calculation, providing maximum statistical power.
Level 2: Randomized Controlled Trials (RCTs)
- The gold standard for testing therapeutic interventions, preventive agents, or clinical devices.
- Key Features: Random allocation of participants, presence of concurrent control groups, and blinding (Single-blind: participants unaware of assignment; Double-blind: both participants and outcome evaluators unaware of assignment to eliminate examiner bias).
Level 3: Cohort Studies (Prospective Observational)
- Tracks a defined group of individuals (cohort) sharing a specific exposure (e.g., smokers vs. non-smokers) forward through time to monitor disease development.
- Used to calculate Incidence (rate of new cases) and Relative Risk (RR). The researcher observes without manipulating the intervention.
Level 4: Case-Control Studies (Retrospective Observational)
- Compares individuals who already have a specific disease (cases) against matched individuals without the disease (controls), looking backward in time to evaluate historical exposures.
- Used to calculate the Odds Ratio (OR). Highly efficient for studying rare diseases or long latency periods.
Level 5: Cross-Sectional Studies
- Observes exposure status and disease prevalence in a population at a single point in time (like a snapshot). Cannot establish temporal cause-and-effect relationships.
Level 6 & 7: Case Reports & Expert Opinion
- Case reports describe rare clinical conditions in 1 to 5 patients without control groups. Level 7 represents expert consensus, narrative reviews, and laboratory in vitro research.
Variables & Experimental Control Principles
Understanding variable definitions is critical for interpreting research literature on licensing exams.
| Variable Classification | Definition & Research Function | Clinical Study Example |
|---|---|---|
| Independent Variable (IV) | The experimental intervention, treatment, or condition manipulated by the researcher to test its effect. | Type of therapeutic mouthrinse (e.g., 0.12% Chlorhexidine vs. Placebo). |
| Dependent Variable (DV) | The measured outcome variable that depends on the independent variable. | Final Gingival Index score or percent plaque reduction. |
| Extraneous / Confounding Variable | Uncontrolled variables that may distort or obscure the true relationship between the IV and DV. | Patient smoking status, brushing frequency, or unrecorded systemic disease. |
Control Group Mechanisms
- Negative Control Group: Receives a placebo or no active treatment (establishes baseline change occurring without therapy).
- Positive Control Group: Receives an established standard-of-care treatment (compares whether the new experimental agent is superior or non-inferior to current market standards).
Sampling Methodologies & Selection Bias
Sampling refers to selecting a representative subset of individuals from a broader target population.
+------------------------------------------------------------------------------------+
| SAMPLING METHODOLOGIES |
| |
| PROBABILITY SAMPLING (Low Bias) NON-PROBABILITY SAMPLING (High Bias) |
| - Simple Random Sampling - Convenience Sampling (Highest Bias) |
| - Stratified Random Sampling - Purposive / Judgmental Sampling |
| - Systematic Sampling - Quota Sampling |
+------------------------------------------------------------------------------------+
Probability Sampling Methods (Minimizes Selection Bias)
- Simple Random Sampling: Every individual in the population has an equal and independent chance of being selected (e.g., computer random number generator). Delivers the highest internal and external validity (generalizability).
- Stratified Random Sampling: Population is divided into homogeneous subgroups (strata, e.g., age bracket, periodontitis stage), and random samples are drawn proportionally from each stratum. Ensures representation of sub-populations.
- Systematic Sampling: Selecting every $N^{th}$ individual from an ordered list following a random starting point.
- Cluster Sampling: Selecting entire naturally occurring geographic units or groups (e.g., all 3rd graders in randomly chosen school districts).
Non-Probability Sampling Methods (High Risk of Bias)
- Convenience Sampling: Selecting readily accessible subjects (e.g., recruiting dental hygiene students or patients sitting in a waiting room). Carries the highest risk of selection bias; limits external validity.
- Purposive / Judgmental Sampling: Researcher uses personal judgment to select subjects believed to be representative.
- Quota Sampling: Selects non-random subjects to meet predetermined subgroup proportions.
A team of researchers evaluates the clinical efficacy of a new essential oil mouthrinse compared to a sterile water placebo by measuring Plaque Index scores after 6 weeks of daily use. In this study design, what variable does the essential oil mouthrinse represent?
Which clinical research design occupies the highest level of scientific evidence on the Hierarchy of Evidence Pyramid for evaluating treatment interventions?
An epidemiologist selects 100 individuals diagnosed with severe aggressive periodontitis and matches them with 100 periodontally healthy individuals. The researcher reviews historical medical records to compare prior dietary habits and systemic risk factors over the preceding 10 years. What type of study design is being utilized?
Which sampling technique ensures that every member of the target population possesses an equal and independent probability of being selected, yielding the highest generalizability?