Public Health Nutrition Exam 2 Study Material

Why is research so important?

Stakeholders, including government and community agencies, hospitals, policymakers, and the general population, heavily depend on research outcomes to inform decisions, enhance medical practices, and formulate policies and procedures aimed at safe-guarding populations and optimizing health outcomes.

Quantitative research

Based on numeric data and techniques. (Including surveys, experiments, observational studies.) It quantifies the prevalence of health issues, identifies risk factors, and assesses the effectiveness of interventions.

Qualitative research

The collection and analysis of non-numeric data to gain deeper insights to complex social phenomena.

(Including interviews, focus groups, participant observations) Explores experiences, perceptions, and behaviors related to health.

Mixed Media Research

Uses both quantitative and qualitative data. Explores the breadth (trends and stats) and captures the depth (experiences, perspectives, context)

Research procedural steps

1. - Identify the problem/formulate research Q/hypothesis

2. - Review relevant literature

3. - Revise Research Question

4. - Select study design

5. - Plan data and analyze strategy

6. - Develop and submit project description to IRB

7. - Collect data

8. - Analyze data

9. - Present results

When formulating a research Question, it often falls into 3 categories:

Descriptive or exploratory, Relational, Causal

(Helps establish if purpose is to describe phenomena, explore relationships between variables, or establish causal connections within the studied context)

Descriptive Research

Provides a detailed account and summary of the characteristics, behaviors, and patterns within a given population.

The data describes the prevalence of disease, distribution of health behaviors, and demographics of a community. This provides the snapshot.

Exploratory Research

Investigating new and uncharted areas within public health. The goal is not to confirm pre-existing hypothesis but to discover and explore potential avenues for future, more tailored research.

Relational and Causal

Aim to elucidate the relationships between variables, investigating whether altering one variable, such as implementing or enhancement of a health program, leads to subsequent modifications in behavior, knowledge, and health outcomes. It helps better understand the underlying mechanisms.

What is an effective Research Question?

It is theoretically important and practically useful

It is answerable using data, realistically collected, and contributes to policy/program development.

Two commonly used techniques to formulate and organize research questions.

EOP and PICOT

EOP: Exposures and outcomes in a population

• - factors result from exposures which affect people of interests/communities

PICOT: Patient/population/problem, intervention, comparison/control, outcome, and timeframe

• - Determines the research’s purpose + well defined parameters. Evaluates participant adherence, exploring socioeconomic influences, and understanding the intervention’s long-term effects on overall health outcomes.

3 Pivotal Questions

• - Define the problem

• - Elucidate its significance

• - Outline the proposed solution

• 
  

Quantitative research 2 types and splitem

Determining how much, aka the magnitude of an occurence or association.

ex: epidemiologists/economists

Observational

• cross sectional

• case control

• cohort

• ecological

Interventional

• randomized control trials (single/double blind)

Quantitative studies - Cross sectional study

Tailored intervention program

ex: prevalence of vaping among teens in a specific region

Quantitative studies - Cohort study

A group of individuals are followed over time to observe development of health outcomes and identify potential risk factors. Helps establish causal relationships.

ex: Framingham Heart Study (more expansion in other cards trust)

Quantitative studies - Randomized control trials

Randomly assign participants to either an intervention or control group. Test effectiveness of new treatments, interventions, and policies.

• ex: random sample, single blind, double blind

Qualitative Studies

Explores the who, what, why, when, how

Explores phenomenology, uncovering the essence of a phenomena experienced by individuals. (in-depth interviews, and focus group discussions)

ex: social scientists/anthropologists

Methods: observation, interview, focus groups, surveys

Qualitative studies - Grounded theory

Develop theories grounded in data collection, allowing new concepts and insights to emerge. Helps develop evidence-based strategies to improve public health.

Qualitative studies - Purposeful sampling and Saturation

P.S.: Selecting people from diverse socioeconomic statuses who live in the community of interest.

Saturation: the researcher gathers a rich and diverse amount of information

What is the most imperative principal of ethical research?

Informed consent

ex: always remember and or reference Tuskeegee experiment

What’s the IRB

Institutional Review Board. Approves all research and makes sure all makes sense and is ethical.

Quantitative data analysis - Descriptive statistics

Summarizes key features, central tendencies, and the distribution of variables. A snap spot of the characteristics of a population.

Quantitative data analysis - Inferencial statistics

Allows researchers to make inferences and draw conclusions about a population based on a sample of data. (hypothesis testing)

Qualitative analysis - Spatial analysis

Helps researchers explore geographic patterns of health outcomes and explore potential environmental or social determinants.

Qualitative analysis - Immersion

Repeated readings or viewing. Coding, labeling, and categorizing segments of data based on recurring themes or concepts. (Creates tailored interventions, policies, and strategies that are contextually relevant and sensitive to the needs of diverse populations)

Statistics refers to both the…

Numbers and the science

Numbers that describe the health of populations and science that helps interpret those numbers. Sometimes data alone doesn’t give enough certainty or its too ambiguous to allow a valid conclusion.

What is the p-value? What is a confidence interval?

The p-value is the probability that the observed result could have occurred by chance.

A confidence interval is the range of values within which the true result is likely to fall.

What’s a false negative? What’s a false positive?

A false negative is when you find no effect when there is one.

A false positive is when you find an effect that is not real.

What is screening? What are some caveats of screening + an example?

Screening is the best public health measure. Screening at-risk populations helps detect the disease early, when it’s most treatable. It enables early detection and containment and helps mitigate the virus's spread. It helps healthcare systems effectively allocate resources and ultimately reduces the burden on healthcare infrastructure while safeguarding vulnerable populations.

Screening through sensitive testing can be helpful but can sometimes yield false positives, leading to unnecessary treatment and longer follow-up tests to confirm accuracy. (ex. mammography in women in their 40s)

Screening can also lead to overdiagnosis bias, which occurs when a tumor is detected by screening, but they are tumors unlikely to progress to the stage that causes symptoms and isn’t life-threatening.

What are rates?

Rates put raw numbers into perspective by relating them to the size of populations being considered.

How can you compare populations that are equivalent? What’s the most common figure used to compare the health of various populations?

BY converting crude rates to adjusted rates. (ex. death rates among males and females)

The most common figure to compare is the “life expectancy at birth”.

What is Years of Potential Life Lost? (YPLL)

It is a measure of premature mortality. It shows the weight of death of young people.

Dying before 75 is a premature death. Homicides rank high in YPLL

Public health workers monitor statistics to…

• Identify special risk groups

• Detect new health threats

• Plan public health programs and evaluate their success

• Prepare government budgets

National Center for Health Statistics (NCHS)

A federal repository for vital statistics. It’s an agency that collects, analyzes, and reports data on Americans' health.

• States transmit data they have compiled from local records, such as vital statistics like births and deaths.

• Periodic surveys of representative samples of the population (health status, lifestyle, health-related behavior, onset and diagnosis of illness and disability, and the use of medical care)

What are vital statistics?

Stats of populations. Common ones include birth and death certificates.

• Includes marriages, divorces, spontaneous fetal deaths, abortions

Note: The U.S. Census provides data on geographic distribution of the population and its sex, age, ethnic characteristics, and social and economic characteristics, including education, housing, and health insurance status.

American Community Survey

Useful for planning transportation systems, zoning, schools, healthcare facilities, housing, as well as estimate the need for social services.

What does NHANES stand for? What does it do?

It stands for the national health and nutrition examination survey.
Surveyors go into different neighborhoods at random and analyze whole families to create a generalized sample of the United States.

• more detailed and accurate

• doctors + nurses sent in vans to conduct

• data to see prevalence of chronic diseases

• smoking, physical fitness, diet (ex. 2019, seeing how teens are losing weight)

What does BRFSS stand for? What does it do?

BRFSS stands for the Behavioral Risk Factor Surveillance System.

This is the largest continuous health survey conducted in the world. It is self-reported and less reliable than NHANES.

• Health Q’s about risk factors: high b.p., high cholesterol, diabetes, weight, screening

What is the goal of an epidemiologic study?

• Observe a possible association between an exposure and an illness

• Develop a hypothesis of a cause-and-effect relationship

• Test the hypothesis through a formal epidemiologic study

What does RCT stand for? What is it?

RCT stands for randomized control trial. It is an intervention study. Caveats include trying to tell people what to do and they don’t follow protocol. Easier for mice or situations when you can control all the factors.

Cohort

This observational methodology usually follows a group of people over time and compares the results to whatever you’re trying to quantify. This doesn’t require people to change their behavior, but it can change how people interpret the results. Ex: Framingham Heart study with how smoking increases risk fo CVD, now less people smoke :)

Case-Control Study

Involves cases who have disease and controls who don’t. Ask about their diets and see the prevalence of a specific type of phenomena in relation to each group.

Confounding variables

“Heavy coffee consumption leads to cancer” confounding variable: the coffee drinkers smoke which leads to cancer

“Ice cream consumption leads to increased shark bites” confounding variable: It’s a hot sunny day at the beach

What’s selection bias? What’s recall bias?

S.B: When there’s a systematic difference between the groups you’re testing for.

R.B.: Subject’s report data unreliably.

Can epidemiologic studies prove cause or effect?

NOOOOO

What can strengthen an epidemiologic study?

• High volume of participants

• Stronger association between exposure and disease

• Dose-response relationship

• Known biological explanation

(Ex., bad example with Vioxx: lied about results that showed this drug increased the risk of heart attacks; funnily enough, it was sold to people with heart problems. Now things HAVE to be FDA-approved and tested first with RCT’s)

What are the 4 sampling techniques of probability sampling?

Random, stratified, cluster, and systematic

Random sampling?

Every member of a population has an equal chance of being selected.

Stratified?

Stratified sampling involves dividing a population into mutually exclusive, homogeneous subgroups called strata based on specific characteristics such as age, gender, or income level. A random sample is then drawn from each stratum, ensuring that all subgroups are represented in the final sample. This method reduces sampling error, improves representativeness, and allows for detailed analysis of subgroups. For example, a school principal might divide students by grade level and randomly select students from each grade to survey

Cluster?

Cluster sampling divides a population into naturally occurring, heterogeneous groups called clusters, such as neighborhoods, schools, or regions. Instead of sampling individuals from each cluster, entire clusters are randomly selected, and all members within those clusters are included in the sample. This method is cost-effective and practical for large or geographically dispersed populations. For instance, a researcher might randomly select several city districts and survey all residents within those districts

Systematic?

Assign a number to every population member, select a random number, and select sample members at random intervals. (Every 10th student would be selected)

What is the 4 non-probability sampling? (no random element involved)

Convenience, snowball, purposive, voluntary

Convenience?

Samples developed by what’s readily available. Often leads to biased results.

Snowball?

Refer a friend, message gets passed and spreads

Purposive?

Researchers intentionally choose participants based on their knowledge, experience, or relevance to the study topic. Focuses on depth and relevance.

Voluntary?

People self-select themselves.