Generating Evidence Through Quantitative and Qualitative Research

Research Is…

Research = external evidence

Quantitative research—the investigation of phenomena using manipulation of numeric data with statistical analysis.

Emphasizes precise measurement of variables

Use the most rigorous methodology possible

Types of quantitative research include:

Randomized controlled trials (RCTs)

Controlled trials without randomization

Case-controlled studies

Cohort studies

Descriptive studies

Getting Started: From Idea to Reality

Identify who will be the principal investigator (PI)—person responsible and accountable for overseeing all elements of the research study.

Create the team and bring them together for a research design meeting.

Team will address issues around:

Study design

Feasibility

Outcomes (“So What” outcomes)

Inclusion/exclusion criteria

How the integrity of the intervention will be maintained

Question #1

Is the following statement true or false?

The primary consideration when deciding if a clinical question merits research is whether the question has been addressed by earlier research.

Answer to Question #1

False

Rationale: When deciding whether research is merited, a researcher should prioritize the question of whether the answer would improve patient outcomes. The existing knowledge base must be assessed, but the potential impact on patients is central.

Steps in Designing a Quantitative Study #1

Step 1: Formulate the study question

Step 2: Establish the significance of the problem

Step 3: Search and appraise aggregated evidence

Step 4: Determine a theoretical or conceptual framework

Step 5: Generate hypotheses when appropriate

Step 6: Select the appropriate research design

Step 7: Identify the sample to enhance external validity

Steps in Designing a Quantitative Study #2

Step 8: Determine measures

Step 9: Outline the data collection plan

Step 10: Apply for human subjects approval

Step 11: Implement the study

Step 12: Prepare and analyze the data

Step 13: Interpret the results

Step 14: Disseminate the findings

Step 15: Incorporate findings into evidence-based practice and evaluate outcomes

Factors to Consider in Determining a Quantitative Study Design

Factors to Consider When Designing a Quantitative Study #1

Significance of the question

What is the “So What” Factor? (NOTE: Funding is more likely to be awarded if the problem being addressed is viewed as being important and impactful: For example

Does it affect a large population

Does it affect cost of care

Does it affect important outcomes (i.e. morbidity, mortality, nurse-sensitive indicators such as infection rates, length of stay, etc.).

Factors to Consider When Designing a Quantitative Study #2

To determine feasibility, ask yourself:

Can the study be conducted in a reasonable period of time?

Are there an adequate number of potential subjects to recruit into the study?

Have the settings for recruitment been identified and is accessibility a concern?

Does the lead person (PI) have sufficient time and expertise to spearhead the effort

Are there major legal and/or ethical constraints to undertaking this study?

Are there adequate resources available to conduct the study?

Progression of Quantitative Research

Figure 21.1: Progression of quantitative research.

Question #2

You are interested in studying the effects early term delivery (birth between 37 weeks 0 days and 38 weeks, 6 days gestation) has on breastfeeding success during the first two weeks of life. You search for and appraise the literature and discover a lot of descriptive and predictive studies. What would your next step be?

Conduct more descriptive and predictive research studies to build the body of evidence.

Conduct an experimental research study; testing an intervention to support breastfeeding success.

Conduct a qualitative research study to explore the maternal stress around breastfeeding her early term infant.

You don’t need to continue to conduct research as you’ve already discovered a robust body of evidence.

Answer to Question #2

A. Conduct an experimental research study; testing an intervention to support breastfeeding success.

Rationale: Conducting an experimental research study to test an intervention to support breastfeeding success in the early term neonate would progress the body of evidence by moving down the progression of quantitative research. Conducting more descriptive and predictive research studies would not further the depth of understanding when there is already an abundance of them in the literature. When resources are limited, it is more beneficial to allocate them to conduct experimental research to try to identify interventions that may improve the situation, rather than utilize resources to conduct qualitative research to explore a less than optimal state of affairs.

Why a Theoretical/Conceptual Framework Is Important

A theoretical or conceptual framework is comprised of interrelated statements that attempt to describe, explain, and/or predict a phenomenon. The framework can be used to guide the:

Selection of important study variables

Construction of interventions (where appropriate)

Development of or search for appropriate measurement instruments

Type of data analyses performed

Explanation of the findings

Question #3

A research team has identified the theoretical framework that they will use for their quantitative study. A theoretical framework is important because it:

Determines how many subjects will be needed in the study

Guides selection of variables that are important to control and/or measure

Explains factors that influence the choice of the study design

Determines the clinical significance of the research question

Answer to Question #3

B. Guides selection of variables that are important to control and/or measure

Rationale: A theoretical or conceptual framework guides investigators in knowing what variables to measure or control. It does not determine the size of the sample (this is done through a power analysis). Study design is selected after consideration of many factors, including feasibility of random selection, random assignment to groups, and the ability to control the independent variable. Finally, factors such as the frequency and severity of the clinical issue determine the clinical significance of the research question.

Factors in Designing a Research Study

A good quantitative design is one that:

Appropriately tests the hypotheses or answers the research questions

Lacks bias

Controls extraneous or confounding variables

Has sufficient ability to detect statistically significant findings

Nonexperimental vs. Experimental Designs

Nonexperimental studies are:

Used to describe, explain, or predict a phenomenon

Often used when it is undesirable or unethical to manipulate the independent variable

Nonexperimental designs include:

Descriptive studies—Includes survey research (cross-sectional and cohort), correlational studies, correlational-descriptive research and correlational-predictive research

Case–control studies

Cohort studies

Experimental Research Designs

Randomized controlled trials (RCTs) are a true experimental design. They provide the strongest evidence to support cause-and-effect. Three criteria must be met to support causality:

The independent variable (i.e., the intervention or treatment) must precede the dependent variable (i.e., the outcome) in terms of time sequence.

There must be a strong relationship between the independent and dependent variables.

The relationship between the independent and dependent variables cannot be explained as being due to the influence of other variables (i.e., other explanations of the relationship must be eliminated.)

The Five Phases of Development of Intervention studies or Clinical Trials

Phase I: Basic Research—usually exploratory and descriptive in nature; establishes variables that may be amenable to intervention and outcome measures

Phase II: Pilot Research—small-scale studies to test the intervention with a small number of subjects; feasibility determined.

Phase III: Efficacy Trials—evaluation of the intervention happens in an ideal setting; looks at efficacy (i.e., establish cost-effectiveness and internal validity of study)

Phase IV: Effectiveness Trials—analysis of intervention effect in clinical practice; clinical effectiveness determined

Phase V: Effects on public health—wide-scale implementation of intervention conducted to determine impact on public health.

Characteristics of Randomized Controlled Trials or True Experimental Design

RCTs are the “Gold Standard” for evaluating the effects of a treatment or intervention. They are the strongest design for determining cause-and-effect relationships.

Three characteristics of RCTs are:

An experimental group receives the treatment or intervention

A control, or comparison group, receives standard care or a comparison intervention that is different than the experimental intervention

Randomization or random assignment, whereby subjects are randomly assigned to either the experimental group or control group, has occurred. Methods to randomly assign include tossing a coin, pulling a colored card out of a box, etc.

Variations of Randomized Controlled Trials or True Experimental Design

Two-group RCT with pretest/posttest design and structurally equivalent comparison group

Two group RCT with posttest only design

Two-group RCT with long-term repeated measures follow-up

Two group RCT with true control group that receives no intervention

Three group RCT

Solomon four-group design

Factorial design

Factorial Design

Figure 21.10: 2 × 2 factorial experiment that generates four study groups. (A) A group that receives both education and exercise. (B) A group that receives education only. (C) A group that receives exercise only. (D) A control group that receives neither education nor exercise.

Question #4

Is the following statement true or false?

In experimental designs, the purpose of a posttest is to identify differences between the control and experimental groups after the intervention has taken place.

Answer to Question #4

True

Rationale: A posttest allows for comparison of groups in terms of the outcome variable after the application of an intervention.

Quasi-Experimental Designs

Designs in which the independent variable is manipulated or introduced, BUT there is:

No random assignment to groups and/or

No presence of a control group

Quasi-experiments are often more feasible, but less ability to establish cause-and-effect relationships

Examples of Quasi Experimental designs include:

Quasi-experiment with pretest and posttest design

Interrupted time series design

Other Types of Experimental Designs