Professional Documents
Culture Documents
Research Designs
Research Designs
RAMOS, RHONA
RESEARCH DESIGNS
CRUZ, CHARLIEMAIGN STANLEY
NI TRIALS
● is often considered when superiority of a test drug over a
reference drug (for instance standard of care) is not
anticipated
○ the objective is showing the test drug to be statistically
and clinically not inferior to the reference drug
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
NI TRIALS
● considered when the use of a superiority trial would be
considered unethical
○ Superiority trial of Drug A vs. placebo
○ it is unethical to use a placebo when there is available
effective treatment that possesses an important benefit
such as preventing death or irreversible injury to the
patient
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
NI TRIALS - DESIGN
● focuses on the mean difference in efficacy measures between
the test drug versus reference drug
● NI margin - can show how much better the reference drug's
effect can be and still allow the test drug to be considered
non-inferior
○ correctly determining this margin is considered the
greatest challenge in the design and interpretation of NI
trials
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
N-of-1 Trial
● attempts to apply the principles of clinical trials, such as
randomization and blinding, to individual patients
● useful when the beneficial effects of a particular treatment in
an individual patient are in doubt
● advantageous if the treatment has a short half-life (allowing
multiple crossover periods without carryover effects) and is
being used for symptomatic relief of a chronic condition
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
N-of-1 Trial
● attempts to apply the principles of clinical trials, such as
randomization and blinding, to individual patients
● useful when the beneficial effects of a particular treatment in
an individual patient are in doubt
● advantageous if the treatment has a short half-life (allowing
multiple crossover periods without carryover effects) and is
being used for symptomatic relief of a chronic condition
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
N-of-1 Trial Case Study
Retrospective
Design Prospective
(most often)
Predefined methods Yes No
Clearly defined outcome
Yes No
measures
Randomization Yes No
Blinding Yes No
Multiple treatment
Yes Not usually
periods
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Bioequivalence Trials
● establish that the quality, safety, and efficacy of generic drugs are the same as the
brand name product
● FDA regulations require that bioequivalence between the generic product and the
brand name product be demonstrated
● conducted under standardized conditions in normal healthy adult volunteers
Bioequivalent products - products that are equivalent in rate and extent of absorption
Programmatic Research
● focused on the impact and economic value of programs and services provided by
pharmacists in community and institutional settings
○ limited resources and budget constraints demand that only those services that improve patient care
in a cost-effective manner should be implemented
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
1. Cohort Studies
2. Case-control Studies
3. Cross-sectional Studies
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Exposure Outcome Study Study
Prospective Retrospective
Observational Known at Known at Determines Determines
Data Data
Study Design Beginning Beginning of Exposure Outcome
Collection Collection
of Study Study Status Occurrence
Cohort X X X
Case-control
X X X
(trohoc)
Cross-sectional X X X
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cohort Studies
● Cohort - describe a group of individuals with a common characteristic or
experience
○ a specific exposure to a particular agent such as a vaccine, medication, procedure, or
environmental toxin
● follow-up, incidence, and longitudinal studies
● participants are grouped by their exposure and followed over time to determine
the incidence of symptoms, disease, or death
● Two groups are usually compared
○ exposed
○ non-exposed
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cohort Studies
Categories:
Cohort Studies
Categories:
Cohort Studies
Other terms describing cohort studies:
Cohort Studies
Other terms describing cohort studies:
Cohort Studies
● used to investigate the cause of a disease or the benefits and safety risks of both
medication and procedures
● simplest approach for studying disease incidence
● disadvantages - expense and time consumption
● attempts to establish a relationship between an exposure or risk factor and a
subsequent outcome - relative risk
○ risk of developing a disease or adverse event in those participants exposed to a specific variable
compared to those not exposed to that variable
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cohort Studies
● no randomization process that occurs to ensure that each participant has an equal
opportunity to be in either the exposed or non-exposed group
● a cohort study has no assurance of this similarity
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cohort Studies
● two concerns arise:
○ selection bias - is a potential whenever the investigator is allowed to decide who is
brought into the study and who is not selected to participate
■ an investigator either knowingly or unknowingly selects from the general
population only the healthiest individuals to be assigned to the non-exposure
group while at the same time a mixture of healthy and unhealthy participants are
selected for the exposure group
○ confounder - is a variable related to one or more variables defined in the study
■ is common since they are the product of not using a randomization schedule
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cohort Studies
● accurate measurement of the outcome is essential
○ surveillance bias is a potential problem when one group, generally the
exposed group is more intensely monitored for changes in the outcome
measure than the comparison group
■ blinding the investigator to reduce bias
○ information bias can occur if the same efforts to measure outcomes are not
made for both the exposed and non-exposed groups
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Case-Controlled Studies
● case-referent / case history / retrospective studies /trohoc
studies
● seek to retrospectively identify potential risk factors of
diseases or outcomes
● subjects (cases) with a particular characteristic or outcome of
interest (e.g., disease) are recruited, matched with, and
compared to a similar group of subjects (controls) who have
not experienced the characteristic or outcome
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Case-Controlled Studies
● data are collected retrospectively via patient interviews or by
reviewing subject data records, and the two groups are
compared to identify possible risk factors or contributors for
development of the disease or outcome of interest
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Case-Controlled Studies
● more useful when diseases have a rare prevalence or when
many years of exposure to the risk factor is required
● predisposition to the disease of interest should be similar in
both cases and controls, except for exposure to the risk factor
under investigation; however, this is extremely difficult to
ensure - matching
● both should undergo the same diagnostic evaluation to
determine presence or absence of the disease under
investigation
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Case-Controlled Studies
Advantages:
Case-Controlled Studies
Disadvantages:
Case-Controlled Studies
Disadvantages:
Case-Controlled Studies
● Odds ratios - used to help interpret the results from
case-control studies
○ an estimate of relative risk
○ greater than one denotes increased risk, equal to one indicates no effect, and
less than one indicates a protective effect
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cross-Sectional Studies
● prevalence surveys
● can be thought of as a snapshot (or slice) of time because
data are collected and evaluated at a single point in time
● hypothesis generating as opposed to hypothesis testing
● not suited for testing the effectiveness of interventions
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cross-Sectional Studies
● only an association can be drawn from the results, not a
cause-effect relationship
● quick and easy to perform
● useful for measuring current health status or setting priorities for
disease control
● Ex. a survey of smokers is cross-sectional when the questionnaire is
administered once, even though the questions contained in the
survey may focus on smoking habits over the past 10 years
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
Cross-Sectional Studies
● problems are errors in data collection and transient effects that
may influence observations
● inaccuracies in data collection may go unnoticed because there are
no prior data for comparison
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● is information gathered from an identified group with
conclusions drawn and applied to a larger population
● considered either descriptive (such as opinions and attitudes)
or explanatory (such as explaining a cause and effect) in
nature
● validity of the results depends on the quality of the study’s
internal rigor
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● used to study the incidence, distribution, and relationships of
sociologic and psychological variables
● used to collect information from a sample and generalize the
findings to a larger, target population
● to identify, assess, and compare respondents’ ideas, feelings,
plans, beliefs, and demographics
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● used to determine how health care programs should be
implemented by utilizing the opinions of experts with
experience in a particular area
● to study effectiveness of a program by surveying individuals
who have used its services,
● to understand attitudes and behaviors of patients or members
of the profession the program affects
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Types:
1. descriptive surveys
■ attempt to identify psychosocial variables such as
attitudes, opinions, knowledge, and behaviors in a
population
2. explanatory surveys
■ attempt to explain causal relationships between
variables
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Types of data collected:
1. Incidence
2. Attitudinal
3. Knowledge
4. Behavior measurements
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Types of data collected:
1. Incidence
■ try to determine the occurrence of events without
drawing any relationships between variables
■ morbidity or mortality data, manpower data
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Types of data collected:
2. Attitudinal
■ try to compare this dependent variable with
independent variables such as age, education, or salary
■ job satisfaction surveys
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Types of data collected:
3. Knowledge
■ attempt to document a person’s knowledge or level of
understanding about a specific topic
■ physician’s knowledge of retail prices of medications
or pharmacist’s knowledge of state pharmacy laws
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Types of data collected:
4. Behavior
■ documents what a person actually does in a particular
situation rather than asking him or her in a survey,
which may reflect an attitude, rather than the actual
observed behavior
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Data collection:
○ questionnaires, examination of historical records,
telephone interviews, face-to-face interviews, Web-based
questionnaires, focus groups, or panel interviews
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Characteristics of well-conducted survey:
○ objective and carefully planned, data are quantifiable, and
subjects surveyed are representative of the target
population
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Errors affecting Precision and Accuracy
1. Coverage Error
■ a bias in a statistic that occurs when the target
population you want to survey does not coincide with
the sample population that is actually surveyed
■ can compromise the ability to generalize study results
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Errors affecting Precision and Accuracy
2. Sampling Error
■ refers to the difference between the estimate derived
from a sample survey and the true value that would
result if a census of the entire target population were
taken under the same conditions
■ occurs when the researcher surveys only a subset
(sample) of all possible subjects within the population
of interest
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Errors affecting Precision and Accuracy
3. Measurement error
■ occurs when the collection of data is influenced by the
interviewer or when the survey item itself is unclear
from the respondent’s point of view
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Errors affecting Precision and Accuracy
4. Nonresponse Error
■ occurs when a significant number of subjects in a
sample do not respond to the survey
■ nonresponse bias can result in surveys, where the
answers of respondents differ from the potential
answers of those who did not answer
■ strive for response rates in the 80% to 90% range
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
1. Factual Data and Internal Validity
■ There must be attempts to assess validity and
reliability of the survey and efforts made to validate
factual data
■ asking more than one question about a concept can
increase the internal validity of a survey
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
1. Factual Data and Internal Validity
■ correlation analysis such as the Cronbach alpha or
similar statistical test that measures correlation
between items should be calculated and the
coefficient factor(s) reported in the article
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
1. Factual Data and Internal Validity
■ coefficient alpha is interpreted in the same fashion
that coefficients of reliability are interpreted (i.e., 0
indicates no consistency between responses while
increasing consistency is seen as you approach 0.8 to
1)
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
2. Sample Size
■ methods section should report sample size, along with
a description of how it was determined
■ will give sufficient statistical power to demonstrate a
difference between two groups
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
2. Sample Size
■ investigators must then determine the minimum
number of subjects that must be sampled for the
sample to be representative of the entire population
● made by consulting references that describe
variability in sampling or consulting a statistician
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
3. Sample Frame
■ describes the population that will actually be drawn
from to make up the survey sample
■ a sample is comprehensive if all members of a
population had a chance to be chosen and no one was
systematically excluded
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
4. Sampling Strategy and Response Rates
■ methods section should supply the reader with
enough information to ensure that nonresponse error
was assessed and measures were taken to control the
error
■ authors should relate as much information about
nonresponders as possible
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
4. Sampling Strategy and Response Rates
■ methods section should supply the reader with
enough information to ensure that nonresponse error
was assessed and measures were taken to control the
error
■ authors should relate as much information about
nonresponders as possible
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
4. Survey Instrument Reliability
■ assess the reliability (i.e., can the results of the survey
be repeated by another investigator) of the survey
instrument and present the results of reliability
estimates
■ a statistical test that measures correlation between
items should be used as a reliability estimate
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
SURVEY RESEARCH
● Evaluation of Survey
4. Survey Instrument Reliability
■ a pretest or pilot test is an assessment of a
questionnaire made before full-scale implementation
to identify and correct problems such as faulty
questions, flawed response options, or interviewer
training deficiencies
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
1. narrative (nonsystematic) review (qualitative review)
2. systematic review (qualitative review)
3. meta-analysis (quantitative review)
○ the only type of review providing new quantitative data
derived from combining the results of each study in the
meta-analysis and performing a statistical analysis on that
data set
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
● classified as tertiary literature
○ except for meta-analyses which are classified as primary
literature since they create new data
● may be subject to author biases or inaccuracies in the
literature search
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
1. narrative (nonsystematic) review (qualitative review)
○ summary of research that lacks a description of
systematic methods
○ may pertain to one specific clinical question or disease
state, or to topics related to pharmacy administration
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
2. systematic review (qualitative review)
○ scientific investigations with predefined methods and
original studies as their subjects
○ summary of results of primary studies where the results
are not statistically combined
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
2. systematic review (qualitative review)
○ publication bias - when published studies are given more
consideration because they are positive in nature than
unpublished studies
■ drawback of unpublished studies - have likely not
undergone a peer-review and revision process; errors
and unclearly stated conclusions may be present
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
2. systematic review (qualitative review)
○ language bias - only articles published in the author’s
primary language are used
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
2. systematic review (qualitative review)
○ To reduce selection bias, authors must be blinded to:
1. names of the study authors (to avoid political or
personal issues)
2. institution of publication
3. results of the studies
○ For the initial choice of study inclusion, only the methods
section should be reviewed
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
○ widely used to provide supporting evidence for clinical
decision making
○ results of previously conducted clinical trials are
combined and statistically analyzed
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Types of Meta-Analyses
1. pair-wise meta-analysis
○ traditional method of synthesizing the results of
different trials to obtain an overall estimate of the
treatment effect of one intervention relative to the
control
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Types of Meta-Analyses
2. network meta-analysis
○ multiple treatment comparison or the mixed
treatment meta-analysis
○ a network of randomized controlled trials is
developed where all these trials have one
intervention in common
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Uses
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Uses
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Uses
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Issues
1. Publication bias
○ studies that show lack of efficacy are less likely to
be located than those that demonstrate beneficial
effects of a drug
○ identified by funnel plots
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Issues
1. Publication bias
○ funnel plot is a scatterplot of treatment effect
versus sample size of the studies included in the
meta-analysis
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Issues
1. Publication bias
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Issues
2. Presence of Divergences
○ heterogeneity (differences in the way the studies
being included in the meta-analysis were
conducted)
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Issues
2. Presence of Divergences
○ publication bias (tendency of investigators to
preferentially submit studies with positive results
for publication)
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
● sensitivity analysis
○ done on any articles that seemed to not be as
similar as the others combined
○ once identified, analyses are repeated without each
of those suspicious articles
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
Evaluation
● sensitivity analysis
○ to determine if this changes the overall results or
conclusions
○ to confirm the accuracy of the results produced by
the meta-analysis
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
REVIEW ARTICLES
3. meta-analysis (quantitative review)
● should be interpreted with caution
● remember that conclusions depend on the quality and
similarity of the studies included
● oversimplification may lead to inappropriate conclusions
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
PRACTICE GUIDELINES
Types:
1. evidence-based
2. formal consensus-based
3. mixture of EBM and consensus-based
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
PRACTICE GUIDELINES
Types:
1. evidence-based
● utilize a rigorous systematic process involving review and
critical evaluation of the medical literature to develop
final recommendations
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
PRACTICE GUIDELINES
Types:
2. formal consensus-based
● utilize experience of experts in their practice area to draw
conclusions and develop recommendations
● useful for those instances where the evidence does not
exist, is not complete, or not conclusive enough to allow
the development of a final recommendation
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
PRACTICE GUIDELINES
Types:
PRACTICE GUIDELINES
● created primarily for:
○ facilitating clinical decision making
○ improving the quality of health care
○ providing consistent treatment across environments
○ decreasing costs
○ diminishing professional liability
○ identifying individualized alternative treatment
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA
PRACTICE GUIDELINES
● provide information regarding therapeutic options and most
appropriate choices for a specific disease and patient
● research has shown that within 2 years of development, a
practice guideline may become outdated
CRUZ, CHARLIEMAIGN STANLEY
RAMOS, RHONA