typesofresearchdesignsrsmehtod.pdf

TYPES OF RESEARCH & RESEARCH DESIGNS Dr. R S Mehta, MSND, BPKIHS 1  Types of Study Design: There is no best type of study design The context, assumptions, paradigms and perspectives decide the type of research methodology Dr. R S Mehta, MSND, BPKIHS 2 How t...

TYPES OF RESEARCH & RESEARCH DESIGNS Dr. R S Mehta, MSND, BPKIHS 1  Types of Study Design: There is no best type of study design The context, assumptions, paradigms and perspectives decide the type of research methodology Dr. R S Mehta, MSND, BPKIHS 2 How to Choose a Research Design Does it adequately test the hypothesis? Does it identify & control extraneous factors? Are results generalizable? Can the hypothesis be rejected or retained via statistical means? Is the design efficient in using available resources? Dr. R S Mehta, MSND, BPKIHS 3 Selecting a Research Design 1. Level of knowledge 2. Nature of the research phenomenon 3. Nature of the research purpose 4. Ethical considerations 5. Feasibility 6. Validity and availability of data 7. Precision 8. Cost Dr. R S Mehta, MSND, BPKIHS 4  Anatomy of Research 1. Define the problem ( Characteristics) 2. Specify the objectives (Hypothesis) 3. Select design or type of study 4. Select study population 5. Collect data 6. Analyze data 7. Determine conclusions Dr. R S Mehta, MSND, BPKIHS 5 Select design or type of study Dr. R S Mehta, MSND, BPKIHS 6  Types of Research From the view point of Type of Information Application Objectives Sought Exploratory Quantitative Pure Research Research Research Descriptive Qualitative Applied Research Research Research Correlation Research Explanatory Dr. RResearch S Mehta, MSND, BPKIHS 7 TYPE OF STUDIES Observational 1. Correlational study 2. Case reports and case series 3. Cross sectional survey 4. Case-control study 5. Cohort study Experimental 1. Community trials 2. Clinical trials –Dr. individuals R S Mehta, MSND, BPKIHS 8  Study Designs 1. Descriptive Studies 2. Cross-Sectional Studies 3. Cohort Study 4. Case Control 5. Randomized Controlled Trials 6. Survey Research Dr. R S Mehta, MSND, BPKIHS 9 Critical Thinking Decision Path: Non-experimental Design Choice Dr. R S Mehta, MSND, BPKIHS 10  Health Sciences and Nursing Research Non-interventional Interventional Pre-experimental True-Experiment experimental Explorative Descriptive Analytical Quasi- - Case study - Cross- - CRD - Cross-sectional sectional - RBD - Longitudinal - Case control - FD - Etc. - Cohort - etc - Etc Note: CRD-complete randomDr. design, RBD-random R S Mehta, block design, FD- factorial design MSND, BPKIHS 11  4 Types of Research Basic research Applied research Action research Evaluation research Dr. R S Mehta, MSND, BPKIHS 12  Basic Research Also known as fundamental research (sometimes pure research) is research carried out to increase understanding of fundamental principles. Many times the end results have no direct or immediate commercial benefits Basic research can be thought of as arising out of curiosity. However, in the long term it is the basis for many commercial products and applied research. Basic research is mainly carried out by universities Dr. R S Mehta, MSND, BPKIHS 13  Applied Research Concern with addressing problem of the world as they are perceived by participants, organization or group of people Action oriented and aims to assess, describe, document or inform people concerned about the phenomenon under investigation Findings are intended to have immediate and practical value In the field of education, policy, evaluation and contract are all examples of applied research Dr. R S Mehta, MSND, BPKIHS 14  Action Research Action Research is simply a form of self-reflective enquiry undertaken by participants in social situations in order to improve the rationality and justice of their own practices, their understanding of these practices, and the situations in which these practices are carried out. Wilf Carr and Stephen Kemmis (1986) Dr. R S Mehta, MSND, BPKIHS 15  Evaluation research Major concern is practical application Tends to be viewed as an isolated case study though the methodologies may be transferable Rooted in values and politics Is immediately prescriptive based upon logic and experience Reports are written for implementers, users and other interested people The extent of dissemination is controlled by sponsor Dr. R S Mehta, MSND, BPKIHS 16  RESEARCH DESIGNS QUANTITATIVE QUALITATIVE Experimental study Ethnography Quasi-experimental Case study Survey study Historical study Correlational study Dr. R S Mehta, MSND, BPKIHS 17 Types of Study Design: Details Dr. R S Mehta, MSND, BPKIHS 18 1. Descriptive Studies: Person, Place and Time Dr. R S Mehta, MSND, BPKIHS 19 Descriptive Epidemiology Includes activities related to characterizing the distribution of diseases within a population Concerns activities related to identifying possible causes for the occurrence of diseases Dr. R S Mehta, MSND, BPKIHS 20 Descriptive Epidemiology PLACE TIME PERSON Think of this as the standard dimensions used to track the occurrence of a disease. Dr. R S Mehta, MSND, BPKIHS 21 Descriptive Research Design: – Describe facts – Discover new facts – Not invent new theory and methods – Largest effort given on data collection – It answers questions: satisfy curiosity – Solve problems Dr. R S Mehta, MSND, BPKIHS 22 2. Cross-Sectional Studies Dr. R S Mehta, MSND, BPKIHS 23 Features of C-S Studies Snapshot in time –e.g. - cholesterol measurement and ECG measured at same time Determines prevalence at a point in time Therefore, C-S is a prevalence study Dr. R S Mehta, MSND, BPKIHS 24  Advantages of C-S Studies Short term Fewer resources required Less statistical analysis More easily controlled Design less complex Dr. R S Mehta, MSND, BPKIHS 25 Advantages of C-S Studies (Cont.) Provide relationship between attributes of disease and characteristics of various groups, e.g. elderly group Data is useful for planning of health services and medical programs Dr. R S Mehta, MSND, BPKIHS 26 Disadvantages of C-S Studies Represent only those who are surveyed Identify prevalence, not incidence necessarily – excludes cases that died before study was done Show association with survival - not risk of development Dr. R S Mehta, MSND, BPKIHS 27  Disadvantages of C-S Studies (cont.) People who are ill may not show up for survey -*Healthy Person Effect Often, not possible to establish temporal relationship between exposure and onset – e.g. does high cholesterol precede CHD? Not too effective if disease levels are low, as difficult to establish a causal relationship Dr. R S Mehta, MSND, BPKIHS 28 Design of a C-S Study Dr. R S Mehta, MSND, BPKIHS 29 Design of a C-S study (Cont.) Dr. R S Mehta, MSND, BPKIHS 30  Cross-Sectional Study Farm ers ineligible eligible participation no participation physically active physically active physically inactive physically inactive & & & & CHD no CHD CHD no CHD Dr. R S Mehta, MSND, BPKIHS 31  Cross-Sectional Study Disease Exposure yes no total yes a b a+b no c d c+d Dr. R S Mehta, MSND, BPKIHS 32 3. Cohort Study Dr. R S Mehta, MSND, BPKIHS 33  Cohort Studies Group by common characteristics Start with a group of subjects who lack a positive history of the outcome of interest yet are at risk for it (cohort). Think of going from cause to effect. The exposure of interest is determined for each member of the cohort and the group is followed to document incidence in the exposed and non-exposed members. Dr. R S Mehta, MSND, BPKIHS 34 When is a cohort study warranted? When good evidence suggests an association of a disease with a certain exposure or exposures. Dr. R S Mehta, MSND, BPKIHS 35  Cohort Effect Changes and variation in the disease or health status of a study population as the study group moves through time. “Generation effect” Dr. R S Mehta, MSND, BPKIHS 36  Types of Cohort Studies Prospective (concurrent) Retrospective (historical) Restricted (restricted exposures) Dr. R S Mehta, MSND, BPKIHS 37  Types of Cohort Studies Prospective – cohort characterized by determination of exposure levels (exposed vs. not exposed) at baseline (present) and followed for occurrence of disease in future Groups move through time as they age Retrospective - makes use of historical data to determine exposure level at some baseline in the past and then determine subsequent disease status in the present. Restricted - limited exposure, narrow behavior (e.g. military) Dr. R S Mehta, MSND, BPKIHS 38  Prospective Studies Also called – longitudinal – concurrent – incidence studies Looking into the future Example: Study of coronary heart disease (CHD) Dr. R S Mehta, MSND, BPKIHS 39  Design of a Cohort Experiment The essential characteristic in the design of cohort studies is the comparison of outcome in an exposed group and a nonexposed group (or a group with a certain characteristic and a group w/o that characteristic). A study population can be chosen by selecting groups for inclusion in the study on the basis of whether or not they were exposed Dr. R S Mehta, MSND, BPKIHS 40  Selection of Cohort Groups There are two basic ways to generate cohort groups. Select a cohort (defined population) BEFORE any of its members become exposed or before the exposures are identified. Select a cohort on the basis of some factor (e.g., where they live) and take histories (e.g., blood tests) on the entire population to separate into exposed and non- exposed groups. Regardless of which selection approach is used, we are comparing exposed and non-exposed persons. Dr. R S Mehta, MSND, BPKIHS 41 Design of a Cohort Experiment Dr. R S Mehta, MSND, BPKIHS 42  Design of a Prospective Cohort Experiment Major problem with a prospective cohort design is that the Dr. R S Mehta, MSND, BPKIHS 43 cohort must be followed up for a long period of time.  Data Gathering Person - to - person Drop off questionnaire Mailed to people Telephone interview Newsletter or magazine Dr. R S Mehta, MSND, BPKIHS 44  Potential Biases in Cohort Studies Information bias Bias in estimation of the outcome Bias from non-response Bias from losses to follow-up Analytic bias Dr. R S Mehta, MSND, BPKIHS 45  Advantages of Prospective Cohort Studies Large sample sizes Certain diseases or risk factors targeted Can be used to prove cause-effect Assess magnitude of risk Baseline of rates Number and proportion of cases that can be prevented Dr. R S Mehta, MSND, BPKIHS 46  Advantages of Prospective Studies (cont’d) Completeness and accuracy Opportunity to avoid condition being studied Quality of data is high Considers seasonal and other variations over a long period Tracks effects of aging process Dr. R S Mehta, MSND, BPKIHS 47  Disadvantages of Prospective Cohort Studies Large study populations required – not easy to find subjects Expensive Unpredictable variables Results not extrapolated to general population Study results are limited Time consuming/results are delayed Requires rigid design and conditions Dr. R S Mehta, MSND, BPKIHS 48  Disadvantages of Prospective Studies (cont’d) Subjects lost over time (dropouts) Costs are high Logistically demanding Maintaining quality, validity, accuracy and reliability can be a problem Dr. R S Mehta, MSND, BPKIHS 49  Cohort/ Longitudinal Studies Design High Exposure Outcome Medium Exposure Outcome Sample Of Population Low Exposure Outcome No Exposure Outcome Dr. R S Mehta, MSND, BPKIHS 50 Prospective Cohort Design Dr. R S Mehta, MSND, BPKIHS 51 Retrospective Cohort Design Dr. R S Mehta, MSND, BPKIHS 52  COHORT STUDY Source Population Cases (= Exposed, =Unexposed) (□= Exposed, =Unexposed)    □  □   □   □□  □□  □   Dr. R S Mehta, MSND, BPKIHS 53 4. Case Control Study Dr. R S Mehta, MSND, BPKIHS 54  CASE-CONTROL STUDIES SOME KEY POINTS Frequently used study design Participants selected on the basis of whether or not they are DISEASED (remember in a cohort study participants are selected based on exposure status) Those who are diseased are called CASES. Those who are not diseased are called CONTROLS. Dr. R S Mehta, MSND, BPKIHS 55  Case-Control Study S ou rce P opu lation in eligible eligible participation n o participation bad ou tcom e good ou tcom e (cases) (con trols) exposed u n exposed exposed u n exposed Dr. R S Mehta, MSND, BPKIHS 56  Case-Control Design Dr. R S Mehta, MSND, BPKIHS 57  Case-Control Design Design Subjects With Outcome of Interest Measure Compare factors factors Appropriate Control Group Without Outcome Of Interest Dr. R S Mehta, MSND, BPKIHS 58  Case-Control Studies Results Outcome Direction Of Present Absent Sampling Exposure to Yes a b intervention or causal factor No c d Dr. R S Mehta, MSND, BPKIHS 59  Case- Control Design: Advantages 1. Valuable for studying rare conditions. 2. Short duration 3. Relatively inexpensive 4. Relatively smaller sample needed 5. Yields odd ratio (usually a good approximation of relative risk) Dr. R S Mehta, MSND, BPKIHS 60  Case- Control Studies: Disadvantages 1. Limited to one outcome variable 2. Potential bias from selection of cases and controls 3. Does not establish sequence of events 4. Potential bias in measuring exposure 5. Potential survivor bias 6. Does not yield absolute risk estimates. Dr. R S Mehta, MSND, BPKIHS 61 Because participants are selected on the basis of disease, exposures for ALL PARTICIPANTS are obtained RETROSPECTIVELY………….. PAST PRESENT Exposure recall Cases & Controls Selected Example: lung cancer cases and non-cancerous controls recall past exposure to cigarette smoke Dr. R S Mehta, MSND, BPKIHS 62  SELECTION OF CASES Decide on a specific case definition based on a medically diagnosed condition.  When diagnosis relies on subjective assessment case definition will be less precise. Must consider what criteria will confirm the case definition:  Lung cancer confirmed by biopsy  Osteoporosis confirmed by bone density measurements  Studying mild forms of a disease results in largest possible case group but may include non-cases (misclassification)  Studying severe forms of a disease decrease the probability of misclassification Dr. R S Mehta, MSND, BPKIHS 63  SELECTION OF CONTROLS Controls should be representative of the referent population from which cases are selected (i.e. comparable) – Controls should have the potential to become cases; Controls should also be candidates for having the disease of interest Dr. R S Mehta, MSND, BPKIHS 64  SELECTION OF CONTROLS (2) Different Types of Controls……… –Population controls Randomly selected individuals from the population like RDD (random digit dialing) –Neighborhood controls Individuals that live in the same neighborhoods as MSND, Dr. R S Mehta, cases BPKIHS 65 SELECTION OF CONTROLS (3) – Friends controls best friends of cases spouses or siblings of cases – Hospital controls Individuals at the same hospital with cases Dr. R S Mehta, MSND, BPKIHS 66 SELECTION OF CONTROLS (4) The investigator can elect to use more than one TYPE of control for each case……. When there is no ONE group similar enough to cases. EXAMPLE: A particular leukemia case may have both a neighborhood control (similar to case in terms of environment) and a sibling control (similar to case in terms of genetic background). Dr. R S Mehta, MSND, BPKIHS 67  Cases & Controls For each CASE in the study, a control is selected How many controls should be selected per case? – 1:1 is usual – Increasing the ratio of controls to cases increases the precision and efficiency of the analysis – It also increases the cost to undertake the study Dr. R S Mehta, MSND, BPKIHS 68  MATCHING CHARACTERISTICS OFTEN USED –age –gender –body mass index (weight / height2) –smoking status –marital status Dr. R S Mehta, MSND, BPKIHS 69  MATCHING (2) GROUP MATCHING Based on proportions Idea is to select a control group with a certain characteristic identical to cases in the same proportion as it appeared in cases. Example: If 25% of cases in your study smoke you would select a control population that included 25% smokers. Dr. R S Mehta, MSND, BPKIHS 70  GROUP MATCHING EXAMPLE CASE POPULATION CONTROL POPULATION Smokers Smokers Non-Smokers Non-Smokers Dr. R S Mehta, MSND, BPKIHS 71  MATCHING (3) 2) INDIVIDUAL MATCHING (matched pairs) For every individual case a control is selected who is identical to the case on certain characteristics. Example: If your first case is a 25 year-old women who smokes then you would find a control who is 25, female and a smoker. So you are matching on age, gender, and smoking status. Dr. R S Mehta, MSND, BPKIHS 72 MATCHED PAIRS EXAMPLE CONTROL CONTROL CASE CASE Dr. R S Mehta, MSND, BPKIHS 73  POTENTIAL PROBLEMS WITH MATCHING It will be difficult to find controls if too many variables are selected for matching. Variables used for matching can not be studied as exposures or confounders. OVERMATCHING – when variables related to disease are inadvertently matched upon. Dr. R S Mehta, MSND, BPKIHS 74 Classic 2 x 2 Table for a Case-Control Study if in the POPULATION Disease No Disease Exposure A B No Exposure C D Odds Ratio = A/C = AD B/D BC Dr. R S Mehta, MSND, BPKIHS 75  Example: Hypothetical data Cases Controls Exposed 141 133 Unexposed 1250 4867 Total 1391 5000 ODDS RATIO = 141 * 4867 = 4.13 Dr. R S Mehta, MSND, BPKIHS 76 133 * 1250  Interpretation of the Odds Ratio… If: OR = 1 then exposure is NOT related to disease OR>1 then exposure is POSITIVELY related to disease OR