Berkson's Bias is a type of selection bias that occurs in case-control studies, particularly when the cases and controls are selected from a hospital or clinical setting. It arises when the selection of cases (individuals with the disease) and controls (individuals without the disease) is influenced by the presence of other conditions or factors, leading to a distortion in the association between exposure and outcome.

Key Features of Berkson's Bias

1. Hospital-Based Selection: Berkson's Bias typically occurs in studies where both cases and controls are drawn from the same hospital or clinical setting. This can lead to a situation where the controls are not representative of the general population.

2. Association with Other Conditions: Individuals who are hospitalized may have multiple health issues or risk factors that are not present in the general population. This can create a misleading association between the exposure being studied and the disease outcome.

3. Underestimation or Overestimation of Risk: Because the controls may have different health profiles compared to the general population, the odds ratio calculated in the study may be biased. This can lead to either an overestimation or underestimation of the true association between the exposure and the disease.

Example of Berkson's Bias

Consider a study investigating the relationship between smoking and lung cancer, where both cases (lung cancer patients) and controls (patients without lung cancer) are selected from a hospital. If the controls are patients with other diseases that are also related to smoking (e.g., chronic obstructive pulmonary disease), this could lead to Berkson's Bias. The controls may have a higher prevalence of smoking than the general population, which could distort the perceived association between smoking and lung cancer.

Implications of Berkson's Bias

• Misleading Conclusions: Berkson's Bias can lead researchers to draw incorrect conclusions about the relationship between exposures and outcomes, which can affect public health recommendations and clinical practices.
• Generalizability Issues: Findings from studies affected by Berkson's Bias may not be generalizable to the broader population, limiting the applicability of the results.

Mitigating Berkson's Bias

To reduce the risk of Berkson's Bias in research, researchers can:

1. Select Controls from the General Population: Instead of selecting controls from a hospital, researchers can use population-based controls to ensure a more representative sample.

2. Use Multiple Control Groups: Employing different control groups can help identify and account for potential biases.

3. Stratify Analyses: Stratifying analyses based on relevant characteristics (e.g., age, sex, comorbidities) can help to control for confounding factors.

4. Conduct Sensitivity Analyses: Performing sensitivity analyses can help assess how robust the findings are to different assumptions about the data.

EPIDEMIOLOGY

Epidemiology is the study of the Distribution and determinants of disease frequency in Humans.

Epidemiology— study of health and disease in human populations and how these states are influenced by the environment and ways of living; concerned with factors and conditions that determine the occurrence and distribution of health. disease, defects. disability and deaths among individuals

Epidemiology, in conjunction with the statistical and research methods used, focuses on comparison between groups or defined populations

Characteristics of epidemiology:

1. Groups rather than individuals are studied

2. Disease is multifactorial; host-agent-environment relationship becomes critical

3. A disease state depends on exposure to a specific agent, strength of the agent.  susceptibility of the host, and environmental conditions

4. Factors

• Host: age, race, ethnic background, physiologic state, gender, culture
• Agent: chemical, microbial, physical or mechanical irritants, parasitic, viral or bacterial
• Environment: climate or physical environment, food sources, socioeconomic conditions

5. Interaction among factors affects disease or health status

Uses of epidemiology

I. Study of patterns among groups

2. Collecting data to describe normal biologic processes

3. Understanding the natural history of disease

4. Testing hypotheses for prevention and control of disease through special studies in populations

5. Planning and evaluating health care services

6. Studying of non disease entities such as suicide or accidents

7. Measuring the distribution of diseases in populations

8. Identifying risk factors and determinants of disease

1. Disease is multifactorial in nature; difficult to identify one particular cause

 a. Host factors

(1) Immunity to disease/natural resistance

(2) Heredity

(3) Age, gender, race

(4) Physical or morphologic factors

b. Agent factors

(1) Biologic—microbiologic

(2) Chemical—poisons, dosage levels

(3) Physical—environmental exposure

c. Environment factors

(1) Physical—geography and climate

(2) Biologic—animal hosts and vectors

(3) Social —socioeconomic, education, nutrition

2. All factors must be present to be sufficient cause for disease

3. Interplay of these factors is ongoing: to affect the disease, attack at the weakest link

Some Terms

1. Epidemic—a disease of significantly greater prevalence than normal; more than the expected number of cases; a disease that spreads rapidly through a demographic segment of a population

2. Endemic—continuing problem involving normal disease prevalence; the expected number of cases; indigenous to a population or geographic area

3. Pandemic—occurring throughout the population of a country, people, or the world

4. Mortality—death

5. Morbidity—disease

6. Rate—a numerical ratio in which the number of actual occurrences appears as the numerator and number of possible occurrences appears as the denominator, often used in compilation of data concerning the prevalence and incidence of events; measure of time is an intrinsic part of the denominator.

Common tests in dental biostatics and applications

Dental biostatistics involves the application of statistical methods to the study of dental medicine and oral health. It is used to analyze data, make inferences, and support decision-making in various dental fields such as epidemiology, clinical research, public health, and education. Some common tests and their applications in dental biostatistics include:

1. T-test: This test is used to compare the means of two independent groups. For example, it can be used to compare the pain levels experienced by patients who receive two different types of local anesthetics during dental procedures.

2. ANOVA (Analysis of Variance): This test is used to compare the means of more than two independent groups. It is often used in dental studies to evaluate the effectiveness of multiple treatments or to compare the success rates of different dental materials.

3. Chi-Square Test: This is a non-parametric test used to assess the relationship between categorical variables. In dental research, it might be used to determine if there is an association between tooth decay and socioeconomic status, or between the type of dental restoration and the frequency of post-operative complications.

4. McNemar's Test: This is a statistical test used to analyze paired nominal data, such as the change in the presence or absence of a condition over time. In dentistry, it can be applied to assess the effectiveness of a treatment by comparing the presence of dental caries in the same patients before and after the treatment.

5. Kruskal-Wallis Test: This is another non-parametric test for comparing more than two independent groups. It's useful when the data is not normally distributed. For instance, it can be used to compare the effectiveness of three different types of toothpaste in reducing plaque and gingivitis.

6. Mann-Whitney U Test: This test is used to compare the medians of two independent groups when the data is not normally distributed. It is often used in dental studies to compare the effectiveness of different interventions, such as comparing the effectiveness of two mouthwashes in reducing plaque and gingivitis.

7. Regression Analysis: This statistical method is used to analyze the relationship between one dependent variable (e.g., tooth loss) and one or more independent variables (e.g., age, oral hygiene habits, smoking status). It helps to identify risk factors and predict outcomes.

8. Logistic Regression: This is used to model the relationship between a binary outcome (e.g., presence or absence of dental caries) and one or more independent variables. It is commonly used in dental epidemiology to assess the risk factors for various oral diseases.

9. Cox Proportional Hazards Model: This is a survival analysis technique used to estimate the time until an event occurs. In dentistry, it might be used to determine the factors that influence the time until a dental implant fails.

10. Kaplan-Meier Survival Analysis: This method is used to estimate the probability of survival over time. It's commonly applied in dental studies to evaluate the success rates of dental restorations or implants.

11. Fisher's Exact Test: This is used to test the significance of a relationship between two categorical variables, especially when the sample size is small. It might be used in a study examining the association between a specific genetic mutation and the occurrence of oral cancer.

12. Spearman's Rank Correlation Coefficient: This is a non-parametric measure of the correlation between two continuous or ordinal variables. It could be used to assess the relationship between the severity of periodontal disease and the patient's self-reported oral hygiene habits.

13. Cohen's kappa coefficient: This measures the agreement between two or more raters who are categorizing items into ordered categories. It is useful in calibration studies among dental professionals to assess the consistency of their diagnostic or clinical evaluations.

14. Sample Size Calculation: Determining the appropriate sample size is crucial for ensuring that dental studies are adequately powered to detect significant differences. This is done using statistical formulas that take into account the expected effect size, significance level, and power of the study.

15. Confidence Intervals (CIs): CIs provide a range within which the true population parameter is likely to lie, given the sample data. They are commonly reported in dental studies to indicate the precision of the results, for instance, the estimated difference in treatment efficacy between two groups.

16. Statistical Significance vs. Clinical Significance: Dental biostatistics helps differentiate between results that are statistically significant (unlikely to have occurred by chance) and clinically significant (large enough to have practical implications for patient care).

17. Meta-Analysis: This technique combines the results of multiple studies to obtain a more precise estimate of the effectiveness of a treatment or intervention. It is frequently used in dental research to summarize the evidence for various treatments and to guide clinical practice.

These tests and applications are essential for designing, conducting, and interpreting dental research studies. They help ensure that the results are valid and reliable, and can be applied to improve the quality of oral health care.

Plaque index (PlI)    

    0 = No plaque in the gingival area.
    1 = A thin film of plaque adhering to the free gingival margin and adjacent to the area of the tooth. The plaque is not readily visible, but is recognized by running a periodontal probe across the tooth surface.
    2 = Moderate accumulation of plaque on the gingival margin, within the gingival pocket, and/or adjacent to the tooth surface, which can be observed visually.
    3 = Abundance of soft matter within the gingival pocket and/or adjacent to the tooth surface.

Gingival index (GI)    

    0 = Healthy gingiva.
    1= Mild inflammation: characterized by a slight change in color, edema. No bleeding observed on gentle probing.
    2 = Moderate inflammation: characterized by redness, edema, and glazing. Bleeding on probing observed.
    3 = Severe inflammation: characterized by marked redness and edema. Ulceration with a tendency toward spontaneous bleeding.

Modified gingival index (MGI)    

    0 = Absence of inflammation.
    1 = Mild inflammation: characterized by a slight change in texture of any portion of, but not the entire marginal or papillary gingival unit.
    2 = Mild inflammation: criteria as above, but involving the entire marginal or papillary gingival unit.
    3 = Moderate inflammation: characterized by glazing, redness, edema, and/or hypertrophy of the marginal or papillary gingival unit.
    4 = Severe inflammation: marked redness, edema, and/or hypertrophy of the marginal or papillary gingival unit, spontaneous bleeding, or ulceration.
    
Community periodontal index (CPI)    

    0 = Healthy gingiva.
    1 = Bleeding observed after gentle probing or by visualization.
    2 = Calculus felt during probing, but all of the black area of the probe remains visible (3.5-5.5 mm from ball tip).
    3 = Pocket 4 or 5 mm (gingival margin situated on black area of probe, approximately 3.5-5.5 mm from the probe tip).
    4 = Pocket > 6 mm (black area of probe is not visible).
    
Periodontal screening and recording (PSR)    

    0 = Healthy gingiva. Colored area of the probe remains visible, and no evidence of calculus or defective margins is detected.
    1 = Colored area of the probe remains visible and no evidence of calculus or defective margins is detected, but bleeding on probing is noted.
    2 = Colored area of the probe remains visible and calculus or defective margins is detected.
    3 = Colored area of the probe remains partly visible (probe depth between 3.5-5.5 mm).
    4 = Colored area of the probe completely disappears (probe depth > 5.5 mm).
 

When testing a null hypothesis, two types of errors can occur:

1. Type I Error (False Positive):

   • Definition: This error occurs when the null hypothesis is rejected when it is actually true. In other words, the researcher concludes that there is an effect or difference when none exists.
   • Consequences in Dentistry: For example, a study might conclude that a new dental treatment is effective when it is not, leading to the adoption of an ineffective treatment.

2. Type II Error (False Negative):

   • Definition: This error occurs when the null hypothesis is not rejected when it is actually false. In this case, the researcher fails to detect an effect or difference that is present.
   • Consequences in Dentistry: For instance, a study might conclude that a new dental material is not superior to an existing one when, in reality, it is more effective, potentially preventing the adoption of a beneficial treatment.