Chapter 2: The Nature of Evidence

epidemiology according to WHO

the study of the distribution and determinants of health-related states or events in specified populations, and the application of this study to control of health problems

any health-related data (PT, OT, etc.)

epidemiological studies

•Mainly observational

•Allow nature to take its course and analyze relationships between indices of health status and other variables

lab based studies

•Mainly experimental

•Intervene to affect what happens to one or more of the groups

types of epidemiological studies

case reports, correlational, cross-sectional, case-control, cohort, RCT (gold standard, most difficult, most expensive)

correlational example

the higher your cholesterol, the higher your chance of dying of coronary heart disease

cross sectional survey

after a survey, you divide population into groups (ex: high, moderate, and low PA) and compare prevalence of health related outcomes between groups

cohort study design

a cross-sectional survey (establish a baseline, divide into groups), then follow up over several years (live your life, then we will ask about your choices and use that to compare outcomes across groups)

prevalence

the proportion of individuals in a population that exhibits the outcome of interest at a specified time. (ex: how many people have diabetes in 2025)

incidence

the number of new occurrences of an outcome that develop during a specified time interval (ex: how many people have been diagnosed with diabetes in the past 5 years)

measuring tools for comparisons of disease occurrence between exposed and unexposed groups

–Risk difference

–Relative risk

–Odds ratio

–Hazard ratio

Population-attributable risk

the incidence of a disease in a population that can be attributed to exposure. exposure to a certain germ for example. Could be health choices (sedentary behaviors, drink a bunch of alcohol, fast food diet)

randomized control trial (RCT)

recruit individuals from population. Random allocation into a PA intervention group and a control group (live your normal life). Compare outcomes between randomized groups

outcome measures

it can be inferred that there is a causal link between increased PA and lower risk of CVD. Because PA lowers blood pressure, and lowered blood pressure lowers the risk of CVD.

parallel RCT study design

while individuals are coming in for intervention, control is living their own life at same time

cross over RCT study design

every single person does both the control and the intervention. This reduces error of individual differences. One group does the control and then after 6 months do the intervention. And the other group foes the intervention during those first 6 months then does the control for the last 6 months

Individuals are compared to themselves in both the control condition and the intervention condition

before after RCT study design

least recommended. do pretesting, everyone does the intervention. Compare the pre and post

accuracy and precision

a good study needs both. Different tools have different accuracy and precision data (V O2max test is much more accurate than a survey. But it is more expensive and time consuming. So you have to decide what you can do and what you need to measure)

accuracy

data is close to the true values

precision

the same measurement, when repeated, consistently yields similar values

random error

due to chance:

biological variation (we are all different and respond to things differently)

sampling error (how are you recruiting? Are you recruiting at a YMCA for a study about sedentary behavior?)

measurement error: human (miscommunication, not explaining well) or measurement error (due to tools)

bias

results differ systematically from true values:

selection bias (how are you recruiting)

measurement bias (specific tool not validated for age group you are studying)

confounding

observed association is due to a third (confounding) factor related to the exposure that independently affects the risk of developing the disease

Sometimes can explain why there is a relationship/correlation

Ex: there is high rate of shark attacks and high amount of ice cream being consumed. Shark attacks are not caused by eating ice cream, but it is just summertime and those things increase in the summer (summertime is the confounding variable)

behavior and genetics

have an impact on physiological characteristics (ex: you may look muscular, but not live a very healthy life)->almost a 50/50 split

measuring physical activity/sedentary behavior

Job classification;

Leisure-time activity: (1) questionnaire (used the most), (2) wearable monitors (pedometers, accelerometers), (3) total energy expenditure by doubly labelled water (most accurate)

pedometer

measure steps. no idea what intensity is (walking, running?) ($20)

accelerometer

better than pedometer. Measure how fast you move (can’t measure any strength training, on a treadmill, or on a stationary bike). More for natural every day movement ($100-200)

doubly labelled water

looks at total energy expenditure (super expensive, invasive (drink water, urine sample)). Uses what physiologically going on in your body

measure cardiovascular fitness

–direct VO₂max (treadmill or cycle ergometer);

–predict VO₂max from sub-maximal heart rate;

functional measure, e.g. level in shuttle test, time to complete a set distance (wide range of accuracy)

direct VO2max test

gold standard. Hooked up to a mouth piece. Tube goes to computer. Run on a treadmill. Measures oxygen, CO2, how many breaths you are taking, volume of breaths, etc.

Calculate how efficiently you use oxygen

Expensive and finnicky (can’t use in humid areas). 8-12 minutes (go until you drop basically)

predict VO2max test

sub max test. Still have mouth piece and tube, but exercise to a certain heart rate or cut off

Use math to predict the max (so, it is an estimation)

Good for older adults or sedentary individuals because a max test may cause heart attacks

measuring strength

–isokinetic or isometric strength using lab-based dynamometer;

–grip strength using hand-grip dynamometer (important for older adults because we use our hands so much)

–functional measure, e.g. push-up capacity, sit-to-stand test, one max test

dynamometer

measures some type of force

one max test

what is the max weight can you safely do one and only one rep on a machine (ex: chest press)

IPAQ

international physical activity questionnaire (#1 in the world)

establishing causality: Appropriate sequenced

want to have some sort of pretesting (need a starting point to see progress)

establishing causality: Plausibility

use logic, does this make sense? (ex: ice cream and shark attack correlation doesn’t logically make sense. There must be another variable)

establishing causality: Consistency in different populations

compare 18-25 year olds who are active to 18-25 year olds who are sedentary will see very different results if they get the same intervention

establishing causality: strength of association

strength of relationship between the 2 factors

establishing causality: does response

ex is when someone is sedentary, they will see no benefits. If they start doing low levels, they will start to see benefits. If they do vigorous exercise, then benefits build up

establishing causality: reversibility

if you get injured, in 2 weeks, you can lose 1-2 months of fitness (how long you have trained helps to protect you if you get injured and are out for a period of time). To gain it back, it will probably take 2+ months

Genetics: you may take longer, you may be ok

establishing causality: strong study design

RCT is the only one that can truly examine cause and effect