Research methods Final
Research protocol
A detailed written description of all the processes and procedures that will be used for data collection and analysis
uses of a cohort study
- Used to study incidence, causes, and prognosis - effects of a suspected risk factor - a birth cohort study is looking at a group of people born at the same time
establishing a case study
- case definitions are important in describing and determining who fits within the population of interest (study population) based on the inclusion and exclusion criteria * inclusion criteria: the criteria a participant needs in order to be included in a particular study * exclusion criteria: the criteria that will determine whether a participants needs to be excluded from the study
key statistical measure of a correlational study
- correlation - the line of best fit shows the correlation between variables - r to the second demonstrates the strength of association - r demonstrates the direction of the association (positive or negative)
cross sectional surveys are used to
- describe communities - assess population needs - support program planning - monitor and evaluate programs - establish a baseline data prior to the initiation of longitudinal studies - most common study design used in epidemiology
case series uses
- describing the characteristics of and similarities among a group of individuals with the same signs and or symptoms of a disease - identifying new syndromes and refining case definitions - clarifying typical disease progression -describing atypical presentations of a disease or unusual complications from a treatment - developing hypothesis for future research
coding
- develop codes, group codes into categories, identify themes - inductive: codes created based on the data - deductive: codes created because of the literature - invivo: based on the actual spoken words of participants - draws on relevant frameworks such as narrative analysis
correlational
- experimental study in which researchers measure two variables and explore the statistical relationship between the two - uses population-level data to look for associations - nearly all are secondary analyses (uses existing data sources)
cohort
- follows participants over a long period of time with common characteristics to follow a certain disease or outcome - investigate the causes of disease and establish links between the risk factors and health outcomes Key statistical measure: relative risk
recruiting participants for a qualitative study
- goal is to maximize participation - recruitment methods often paired with the data collection method - provide multiple opportunities to participate - consider appropriate incentives
case-control study
- goal: compare exposure histories in people with disease (cases) and people without diseases (controls)
experimental study
- goal: compare outcomes in participants assigned to an intervention or control group.
cohort study
- goal: compare rates of new (incident) disease in people with different exposure histories or follow a population forward in time to look for incident diseases
correlation (ecological) study
- goal: compare the average levels of exposure and disease in several populations.
case series
- goal: describe a group of individuals with a disease
cross sectional survey
- goal: describe exposure and/or disease status in a population
qualitative study
- goal: seek to understand how individuals and communities perceive and make sense of the world and their experiences
review/meta-analysis
- goal: synthesize existing knowledge
methods/techniques for a qualitative study
- in person or telephone interview - may promote accuracy and completeness but may be expensive because of personnel costs - semi-structured interview guide: includes general questions you're going to ask and probes - in-depth interview guides: a list of very specific questions that must be followed closely - record and transcribe interviews and code/analysis with a software that helps map out common ideas/themes - training interviews are important because we want uniformity in data collection
A well thought out and well written protocol can be judged according to the three main criteria
- is it adequate? - is it feasible? - does it provide enough detail
cross sectional study characteristics
- made over a short period of time - use of representative samples
key statistical analysis for case study
- most case study reports do not require any numbers beyond simple counts and percentages - some case study reports may benefit from the use of well-defined measures pertaining to morbidity and mortality (case fatality rate)
three options of matching
- no matching: assume that similar inclusion and exclusion criteria for cases and controls will result in case and control populations that have similar distributions according to sex, age group, socioeconomic status, and other characteristics that may be confounders of the association between the key exposure and the disease - frequency (group matching): a sampling design that is used to assure that cases and controls in a case control study have similar demographic characteristics -matched pairs (individual) matching: a study in which each case is personally linked to a particular individual control, such as genetic sibling
aggregate data
- one characteristic of a population is exposed, one is the outcome/disease - data must be first entered into a spreadsheet before conducting statistical analysis--each population assigned their own column
special considerations case control
- overmatching: recruiting challenges and possible statistical bias resulting from matching cases and controls, or exposed and unexposed participants, on too many characteristics - misclassification bias: bias that occurs when participants are not correctly categorized , such as when some controls in a case-control study are incorrectly classified as cases - recall bias: bias that occurs when cases and controls in a case control study systematically have different memories of the past
case series steps
- select a disease of interest - determine what new information or new ideas will be analyzed using the study - identify a source of cases
prospective
- the exposed and unexposed are followed into the future over time to see the development of disease
ecological fallacy
- the incorrect attribution of population-level associations to individuals - ex: falsely assumes every individuals IQ is high since the class average is high - how to avoid: compare groups rather than individuals - no individual-level data in the analysis
ecological study
- type of correlational study - explores an environmental exposure
retrospective
- uses baseline information collected from the past - information is collected from past clinical records and the outcome of interest is investigated ex: researchers used previously collected data to investigate whether there was an association between birth experience and maternal caregiving attitudes and behavior over a 12 month period.
PubMed
service of the US National Library of Medicine. These supplemental searches are important to help narrow down the yield of hits.
testability
the ability of a research question to be measured and examined
prevalence rate
the percentage of the population with a given trait at the time of the survey * prevalence rate ratios compare the prevalence of a characteristic in the two population groups by taking the ratio of their prevalence rates - CANNOT SHOW THAT AN EXPOSURE CAUSED A DISEASE
internal validity
to see how well a study was designed , conducted, interpreted, and reported
Writing a research protocol primary study
Main goal/ objective, desired sample size/ steps taken to acquire participants, how participants recruited/ contracted, exact questions participants asked, plans for data analysis
Researcher responsibilities
Make sure clear guidelines are set and responsibilities are clear, have a plan to resolve conflicts
Key statistical analysis of case control
Odds ration (OR) OR= a/c / b/d= ad/bc analysis: OR=1: the odds of exposure were the same for cases and controls OR>1: cases had a higher odds of exposure than controls, implying that the exposure was risky OR<1: cases had lower odds of exposure than controls implying that the exposure was protective
case fatality rate (case study)
The proportion of persons with a particular disease who die as a result of that condition * case fatality rate= (total number of deaths from a specified disease or condition) / (total number of cases of a specified disease or condition)
cross sectional study population
These studies are done by selecting participants to represent a population, so representativeness is very important, making sure the participants match the population that is being studied - often uses a questionnaire - KAP survey- asks participants about their knowledge, attitude, and practices - uses a prevalence rate - prevalence in exposures, histories, diseases, states, and demographic characteristics
Gantt chart
Very helpful for visually displaying the research timeline
blinding (experimental)
When participants (and sometimes providers) are unaware is in the intervention or control group
Keyword
Word, mesh term, or short phrase used in a database search
photo voice
a problem-posing participatory methodology that uses photography to record and reflect a community's assets and concerns , discuss issues of importance to promote dialogue and produce new knowledge, and reach influencers to generate strategies to effect change
brainstorming
a process of gathering long lists of spontaneous ideas about possible research questions
population of a cohort study
a relatively large-sized group of people with the same characteristics followed over a long period of time (years) to find certain risk factors for a disease.
case-control
a study that compares the exposure histories of people with disease (cases) and people without the disease (controls) * both cases and controls are asked the same set of questions about past exposures
MESH
acronym for medical subject headings, a dictionary used for searches in MEDLINE
PICOT
acronym for patient/population, intervention, comparison, outcome, timeframe, a framework that is helpful for developing clinical research questions and designing intervention studies, points towards selection of key indicators that would provide evidence for the success of the intervention.
Primary study
collecting new data from individuals
experimental study
compare outcomes in participants assigned to an intervention or control group -study question: does exposure cause outcome
key statistical analysis: incidence rate ratios
The number of new cases in a population during a period of time divided by the total number of persons in the population who were at risk during that period RR=1: the incidence rate was the same in the exposed and in the unexposed RR>1: the incidence rate was higher in the exposed than in the unexposed, suggesting that the exposure was risky RR<1: the incidence rate was lower in the exposed than in the unexposed suggesting that the exposure was protective
qualitative survey
may yield a higher sample size, be more cost effective, and be better for sensitive questions
gaps in literature
missing pieces of information that a new study could fill
matching
protocols for recruiting one or more controls who are demographically similar to each other in a case-control study or recruiting one.
qualitative study
seeks to understand how individuals and communities perceive, process, and navigate their experiences
types of cohort studies
-prospective - retrospective -longitudinal - is exposure associated with an increased incidence of disease?
Literature searches
-start the literature search to inform your topic choice -Where to get the literature: a. international and national organizations have factsheets, brochures, and websites that have basic information to use. b. you have to be cautious with sources that contradict others.
approaches to experimental studies
-treatment received approach: limits analysis to the participants that are fully compliant - treatment assigned approach: includes all participants even if they weren't fully compliant
PICOT questions
-what is the patient/population group that will be studied? -what is the intervention that will be tested? -what will the intervention be compared to? -what is the outcome of interest? -what is the timeframe for follow-up?
C.R.E.A.T.E
1. Consider the relationships between the key concepts presented in the introduction 2. Read the methods and results sections 3. Elucidate the hypothesis based on the results 4. analyze and interpret the data 5. think of the next experiment * this process is used when reading and analyzing existing works
The health research process
1. identify a study question 2. Select a study approach 3. design study and collect data 4. analyze data 5. report findings
reasons to initiate a health research study
1. needs assessment (what is the health status of this population? What are the major health concerns of members of this population?) 2. risk assessment (what are the threats to health in this population? What are the risk factors for morbidity, mortality, disability, and other health issues?) 3. applied practice (how well are we preventing, diagnosing, and treating health concerns in the populations we serve?) 4. outcome evaluation (was this intervention successful at improving health status in this population?
Research Questions
1. what is the goal/purpose 2. what is the population/data source 3. finalizing the study question - use E,D,P -exposure, disease, population -forms basis of research question ex: is [exposure] related to [disease/outcome] in [population]?
analysis of qualitative data
3 main steps: - read transcript - code - develop themes
qualitative study focus groups
8-10 participants purposively sampled participants in a 1 to 2 hour conversation moderated by researchers (a moderator and a note taker)
abstract databases
Allows researchers to find other abstracts for keywords or other terms.
Primary investigator
Responsible for guarenteeing that the protocol is followed, budget is properly managed and any adverse outcomes are immediately reported to institutions research ethics committee
annotated bibliography
At minimum a full reference for the document being reviewed and a brief summary of the article or report
Target population
Broadest group to which the results of a study should be applicable
Research timelines
Create list of steps, create calendar, identify fixed deadlines, regular meetings
key statistical measure of an experimental study
Efficacy: the proportion of individuals in the control group who experience an unfavorable outcome when they could have been expected to have a favorable outcome if they had been assigned to the active group instead of the control
Purpose statement
How + What + Where + When 1. How: study approach (primary or secondary) 2. What: exposures being investigated 3. Who: Population 4. Where: study location 5. When: study time period
Collection of new data
Identify appropriate way to sample / recruit participants, develop questionnaire/ other data collection tools, select methods for gathering/recording responses from participants, prepare an application for research ethics review committee
Originality
Research needs to have only one substantive difference between previous work. Important to synthesize everything that has already been published
Research protocol
Should describe the exact procedures that will be used for every step of the research process
vital statistics
birth and death rates and other demographic information
recruiting process for case control study
controls must be similar to cases except for their disease status , so the inclusion and exclusion criteria for the cases that do not specifically relate to the disease must also apply for the controls ex: if cases must be males between 25-39 years of age controls must also be men in this age group
topic development
create a question from observation or interest that can be tested, brainstorm and topic map to identify a topic, compile a list of keywords, refine each of them in the outcome disease population formula (exposure) and (disease/outcome) in (population)
Case series
describes a group of individuals with the same disease or who have undergone the same procedure
codebook
describes each variable and specifies how the collected information will be entered into a computer database
tertiary study
existing literature is reviewed
longitudinally
follows a large group of people over a long period of time. ex: female nurses who smoke and female nurses who do not smoke
PPT (case series)
generally more is needed than just an ICD code so PPT (person, place, and time) characteristics are often used as a supplement to the inclusion criteria, as they specify the desired study population
cross sectional
goal- measure the proportion of a population with a particular exposure or disease (also called a prevalence study)
ICD codes (case series)
international classification of disease, they are helpful in establishing study participants as they are contained in a central database and describe the various criteria of each disease
external validity/generalizing
likelihood that the results of a study with internal validity can be generalized to other populations, places, and times
abstract
paragraph summary of an article that includes a brief description (glimpse into the content of the article of interest)
experimental study population
randomly assign participants to intervention or control - how to randomize: - simple random sample: using coin toss or a random number generator to assign individuals - cluster: all individuals in strata receive the same assignment - stratified: individuals are grouped into strata and then a simple random sample is done within the strata to assign individuals
Efficacy equations
rate of intervention (ri)=a+ (a+b) rate of control (rc)= c/(c+b) efficacy= (rc-ri)/rc
repeated cross sectional survey
re-samples and resurveys representatives from the same source population at two or more different time points - does not resample the exact same people - uses a new group of people from the population - if the same person is selected twice their answers will not be linked and they will be considered separate answers - can reveal trends in population-level metrics over an amount of time - do not allow the individual change to be studied
secondary study
use existing data
concept mapping
visual method for listing ideas and then grouping them to reveal relationships; this technique can be useful when identifying a study question and as part of narrative analysis of qualitative data.
When to use a case-control approach
when a disease is relatively uncommon but a source of cases is available - best study approach for identifying possible risk factors for disease (especially for uncommon diseases)
