Ch. 14 Biostatistics

Continuous Data

Data with a logical order and value

Types of continuous data

- Interval : No meaningful zero (ex: Celsius temp)

- Ratio : Meaningful zero (0 = nothing, lack of)

Discrete Data

Data that is classified into categories, and do not take on value

Types of Discrete Data

- Nominal : Categories w/o order (ex: male/female or yes/no)

- Ordinal : Categories with order (ex: pain scale

What are the measures of central tendency?

Mean, Median, Mode

What is Mean and with what type of data is it utilized?

Mean is the average of all data points and is preferred when working with continuous data with a normal distribution (no outliers)

What is Median and with what type of data is it utilized?

Median is the midpoint of data points within a data set and is preferred when working with ordinal or continuous data that is skewed

What is Mode and with what type of data is it utilized?

Mode is the most frequently-occurring data point in a data set and is preferred when working with nominal data

Another name for a Bell-curve is ____________________ distribution

Gaussian

Gaussian distribution (Bell-curve)

68% of values will fall within ______ SDs of the mean

1

Gaussian distribution (Bell-curve)

95% of values will fall within ______ SDs of the mean

2

How to determine is Gaussian distribution is skewed?

If 68% of the values do not fall within 1 SD of the mean

A _______________________ variable can be manipulated by the researcher

independent

A ________________________ variable is what is being examined to see the effect that a change in the independent variable may cause

dependent

Null hypothesis

No difference

Alternative hypothesis

Difference

Alpha level

Maximum error margin

if Alpha is set at 0.05 and the p-value is 0.023, should the null hypothesis be (accepted/rejected)?

Rejected (There is a difference)

if Alpha is set at 0.05 and the p-value is 0.085, should the null hypothesis be (accepted/rejected)?

Accepted (No difference shown)

How to calculate Confidence interval

CI = 1 - alpha

If alpha is set at 0.05, what is the confidence interval?

Confidence interval is 95%

(meaning that we are 95% confident that the conclusion is correct)

When comparing difference data (means) that confidence interval must not include ______ to be considered significant

0

When comparing ratio data (relative risk, odds ratio, hazard ratio) the confidence must not include ______ to be considered significant

1

Type I Error

False Positive (Mistakenly rejecting the null hypothesis)

Type II Error

False negative (Mistakenly accepting the null hypothesis)

What is Study Power?

Probability that a test will reject the null hypothesis correctly (avoid a type II error)

Relative Risk (RR)

RR = Risk in treatment group / Risk in control group

Relative Risk Reduction (RRR)

RRR = 1 - RR

Absolute Risk Reduction (ARR)

ARR = % risk of control - % risk of treatment

Number Needed to Treat (NNT)

NNT = 1 / ARR (Round up)

Number Needed to Harm (NNH)

NNH = 1 / ARR (Round down)

Odds Ratio

Y Exposure Y Outcome = A

Y Exposure N Outcome = B

N Exposure Y Outcome = C

N Exposure N Outcome = D

OR = AD / BC

Hazard Ratio (HR)

HR = HR of treatment / HR of control

T-tests

Continuous data with 2 groups

Paired t-test

when a single sample group is used for a pre- / post- measurement

Student t-test

when the study has two independent samples

ANOVA

Continuous data with 3 or more groups

Chi-Square test

Nominal/Ordinal data with 1 or 2 groups

Fisher's exact test

Nominal/Ordinal data with 2 groups

True/False Correlation proves Causality

False

Sensitivity

How effectively a test identifies patients with the condition

[A / (A + C)] x 100

Specificity

How effectively a test identifies patients without the condition

[D / (B + D)] x 100

Intention to treat analysis

includes data for all patients even if they did not complete the trial according to protocol

Per protocol analysis

Includes only data for patients that completed the trail as intended by protocol

Equivalence study

As good as the standard of care

Non-inferiority study

No worse than standard of care

Forest plots

Used for meta-analysis

• Provides CIs for difference data or ratio data

Case-control study

Retrospective observational comparisons of cases & controls

Cohort study

Retrospective or Prospective comparison of patients with an exposure to those without an exposure

Randomized controlled trial

Prospective comparison of patients who were randomly assigned to groups

Meta-analysis

Analyzes the results of multiple studies

ICER

Change in costs and outcomes when two treatment alternatives are compared

(C2-C1)/(E2-E1)

Cost-minimization analysis

when 2 or more interventions have demonstrated equivalence in outcomes, and the costs of each intervention are being compared

Cost-benefit analysis

comparing benefits and costs of an intervention in terms of monetary units

Cost-effectiveness analysis

compare the clinical effects of two or more interventions to the respective costs

* Most common pharmacoeconomic methodology seen in biomedical literature

Cost-utility analysis

quality-adjusted life years (QALYs) and disability-adjusted life years (DALYs)