1 / 112

Research Design

Research Design. Dr Salah AbuRuz. HomeWork1: Research Title (Due 1/4). Research Title Introduction including: Statement of problem. Aims and objective Research questions and Providing a few hypotheses that come from the research problem. hierarchy of major study designs.

normanb
Download Presentation

Research Design

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Research Design Dr SalahAbuRuz

  2. HomeWork1:Research Title (Due 1/4) • Research Title • Introduction including: • Statement of problem. • Aims and objective • Research questions and Providing a few hypotheses that come from the research problem.

  3. hierarchy of major study designs systematicreview of RCTs RCT interventional cohort observational case control Cross sectional study Case report validity

  4. The reason that we use less reliable designs for some questions: • One reason for using a nonexperimental design is that a vast number of human characteristics are inherently not subject to experimental manipulation (e.g., blood type, personality, health beliefs, medical diagnosis); the effects of these characteristics on other phenomena cannot be studied experimentally. • Ethics - RCTs cannot usually be applied to questions about potential harm, because it is unethical to expose people actively to potential harms • Cost – More reliable designs are more expensive

  5. Practicality – Constraints might involve insufficient time, lack of administrative approval, excessive inconvenience to patients or staff, or lack of adequate funds. • Not all questions require RCTs or systematic reviews. For example, questions about prevalence can be quite satisfactorily answered by cross-sectional studies, rather than by an RCT. • Finally, nonexperimental research is usually needed before an experimental study can be planned. Experimental interventions are developed on the basis of nonexperimental research documenting the scope of a problem and describing critical relationships between relevant variables.

  6. Nonexperimental research

  7. Nonexperimental research (observational) • Correlational (ex post facto) research • Prospective designs • Retrospective designs • Cross-sectional correlation • Descriptiveresearch (survey/prevalence) • Purely descriptive

  8. Ex Post Facto/Correlational Research • This means that the study has been conducted after variations in the independent variable have occurred. • Ex post facto research attempts to understand relationships among phenomena as they naturally occur, without any intervention.

  9. correlation is an interrelationship or association between two variables, that is, a tendency for variation in one variable to be related to variation in another. • For example, in human adults, height and weight are correlated because there is a tendency for taller people to weigh more than shorter people. • Age and wealth • Age and height

  10. Correlation does not prove causation • The mere existence of a relationship— even a strong one—between variables is not enough to warrant the conclusion that one variable caused the other. • Although correlational studies are inherently weaker than experimental studies in elucidating cause-and-effect relationships, different designs offer different degrees of supportive evidence.

  11. NotOutcome Outcome Exposed a b a+b a b c d NotExposed c d c+d a+b+c+d b+d a+c Cohort Study

  12. Cohort studies • Cohort Study is a study in which patients who presently have a certain condition and/or receive a particular treatment are followed over time and compared with another group who are not affected by the condition under investigation. • Studies are generally concerned with what causes a disease. Problems include the time a study can take and the influence of other lifestyle variables. • For instance, since a randomized controlled study to test the effect of smoking on health would be unethical, a reasonable alternative would be a study that identifies two groups, a group of people who smoke and a group of people who do not, and follows them forward through time to see what health problems they develop.

  13. Characteristics 1. Begin by identifying individuals for study and control groups before the investigator is aware of whether they have or will develop the disease. 2. A cohort is a group of individuals who share a common experience. 3. Compare disease frequency over time between one cohort that possesses the characteristics under study (e.g., drug exposure) to a second cohort that does not possess these particular characteristics (e.g., no drug exposure).

  14. Characteristics • Since they identify exposure and then proceed to outcome (disease) they are less subject to bias than other observational studies. • They are especially useful when it is not possible for ethical reasons to use a RCT design (e.g. randomising to cigarette/no cigarette use). • Because cohort studies may run over many years and may need large numbers of participants, they are often very expensive.

  15. Ethically safe Cheaper than RCT?? No randomization Subject to confounders effect Long duration Cost?? Rare disease difficult to study Advantages: Disadvantages: Cohort studies

  16. NOTE • Outcomes are RR, NNT • Discuss more examples from the literature and experience

  17. Self study • Retrospective cohort study

  18. NotOutcome Outcome Exposed a b a+b a b c d NotExposed c d c+d a+b+c+d b+d a+c Case-Control Study

  19. Case control studies Case control studies are concerned with what causes a disease. • a group of patients who have experienced an outcome of interest (e.g., breast cancer) are retrospectively matched to a similar group of patients who lack the outcome of interest (e.g., women without breast cancer). (matched usually by age and gender) • The investigators then look for a difference in past exposures between the two groups (e.g., exogenous estrogen use) to determine the cause

  20. Characteristics 1. Begin after individuals already have developed or failed to develop the disease being investigated. 2. Go back in time to determine the characteristics of individuals before the onset of disease. 3. Cases: Individuals who have developed the disease already. 4. Controls: Individuals who have not developed the disease. 5. Retrospective studies are often cross-sectional, with data on both the dependent and independent variables collected once, simultaneously.

  21. the researchers must find controls without the disease or condition who are as similar as possible to the cases with regard to key extraneous variables (e.g., age, gender) and also obtain historical information about the presumed cause. • If controls are well chosen, the only difference between them and the cases is exposure to the presumed cause. • Researchers sometimes use matching to control for extraneous variables. • To the degree that researchers can demonstrate comparability between cases and controls with regard to extraneous traits, inferences regarding the presumed cause of the disease are enhanced.

  22. Matching ensures confounding is taken into consideration. , but rather than following the subjects into the future, data on past exposure to possible causal agents are collected (eg, by searching through medical records or interviewing subjects). • This type of trial may be the only feasible way to study rare disorders or those with a long lag time between exposure and outcome. However, the results can be subject to bias, such as when patients are asked to recall long-ago events, such as their prior use of certain drugs.

  23. Quick and cheap Excellent for rare diseases Potential recall bias Confounding of exposure variable Matching controls to cases is a “challenge” Advantages: Disadvantages: Case-control studies

  24. Notes • More examples • Outcomes: odds ratio

  25. Cross sectional Correlation Studies • Unlike other types of correlational research—such as the smoking and lung cancer investigations— the aim of descriptive correlational research is to describe the relationship among variables rather than to infer cause-and-effect relationships. • Descriptive correlation studies are usually cross-sectional (lung function and asthma control, age and weight)

  26. NotOutcome Outcome Exposed a b a+b a b c d NotExposed c d c+d a+b+c+d b+d a+c Cross-sectional Study

  27. Example of a descriptive correlational study: Morin, Brogan, and Flavin (2002) described therelationship between body image perceptions of postpartum African-American women on the one hand, and their weight (based on the body mass index) on the other. Irrespective of body masscategory, women usually considered themselves larger than they were.

  28. Descriptive Research • The purpose of descriptive studies is to observe, describe, and document aspects of a situation as it naturally occurs • sometimes to serve as a starting point for hypothesis generation or theory development.

  29. Cross sectional surveys A cross sectional survey is a measure of the frequency of a disease or risk factor in a defined population at a given time. • For example, the records of all patients in a number of general practices could be reviewed to determine the number of patients with heart failure (and therefore the prevalence of heart failure in that population) • Cross-sectional studies are useful for defining the prevalence of a disease or lifestyle patterns (e.g. Survey of Lifestyles, Attitudes and Nutrition - SLAN) but they cannot provide information on possible cause(s) of a disease.

  30. Some descriptive studies are undertaken to describe the frequency of occurrence of a behavior or condition rather than to study relationships. • For example, an investigator may wish to describe the health care and nutritional practices of pregnant teenagers.

  31. Cross-sectional Studies 1. prevalence studies 2. case control studies 3. Static pre-experminemntal studies 4. Descriptive Correlation Studies 5. Cross sectional surveys

  32. Cheap and simple Ethically safe No causality (only association) Unequal confounder distributions Recall bias Advantages: Disadvantages: Observational studies

  33. Cross sectional vs longitudinal • Comparing children health promotion activities 7 year old vs 10n year old is cross sectional study • Examining health promotion in 7 years old and then again when they are 10 year old is longitudinal • Which has a stronger design?

  34. Repeated cross-sectional studies may be carried out at different time points to assess trends over time. However, as these studies involve different groups of individuals at each time point, it can be difficult to assess whether apparent changes over time simply reflect differences in the groups of individuals studied.

  35. Case reports and case series • A case report is simply a presentation of one or more cases. While case reports are valuable for reporting rare clinical events and raising awareness of the possible harm or benefit associated with an agent, they are not an evidence-based tool for making clinical decisions. • Because there is no control group, case reports and series are not valid statistically.

  36. See full cohort study

  37. See full case control study

  38. Seen full cross sectional correlational study

  39. Experimental studies

  40. Characteristics of True Experiments 1. Manipulation—researcher does something to some subjects (introduces an intervention or treatment) 2. Control—researcher introduces controls, including a control group

More Related