Saturday, September 21, 2013

Research: Study Designs

Evidence-based medicine is based on research which comes into account through several different study designs. Different study designs give different outcomes. In other words, in different circumstances, the researcher requires to use different but the best design to carry out a study. At the same time, each study design has strengths and limitations. However, the researcher tries to use the design where advantages overweigh the disadvantages. Study designs are usually based on:
  • The prevalence of health problems
  • Health problems
  • Evaluation of therapy, including treatment and prevention
Qualitative research: In qualitative research, the researcher determines the relationship between one thing (an independent variable) and another (dependent or outcome variable) in a population that is selected for the study purpose.
Goals: To explore the depth and complexity of a phenomenon

Quantitative research: It is a formal, objective, systemic way of obtaining information about the world. Here, a method used to describe test and examine cause-effect relationship. Quantitative studies are then divided into two experimental and observational.
Goals: To test relationships, describe and examine cause-effect relations   

Types of study designs:

Observational studies

The researcher just observes and collects information, but does not try to change the people (or animals or reagents) being observed or the outcome.

Examples

  • A survey of smoking habits among the factory workers
  • A researcher joins to study life style of a gang (the researcher dose not try to change his style or behavior, this is observational study)
  • You take blood samples to measure blood alcohol levels during specific days (yes, you are intervene to take the blood, but you do not alter the outcome).

Observational Designs

Cross-sectional surveys: 

Cross-sectional surveys determine the frequency of a particular attribute, for instance, a specific exposure or disease or any other health-related event, in the study population at a particular point in time.

Cohort, or longitudinal, or prospective studies: 

It is a forward study that may run over a long time. It is type of clinical or observational study in which the researchers analyze a risk factor or cause in a particular groups of un-affected people e.g. people exposed to pollution, taking same type of vaccination or drugs etc. In other words, comparison between two groups is done. Subgroups in a cohort may be compared with each other.


Case-Control study:

It is the type of observational study where the researcher selects two groups from the same population; one subgroup having specific disease and the other subgroup does not. Both subgroups are exposed to same circumstances. It leads to the conclusion where the researcher can recommend and control the outcome of the specific diseases in the population.
Example:
A group is taken from American whites who are equally exposed to HIV some people belong to this group are affected by HIV and some are not. So it will give the way to researcher why some are affected and some are not.

Experimental studies

 The researcher intervenes to change something (e.g. gives some patients a drug) and then observes what happens.  In an observational study there is non-intervention. 

Examples

  • Offer a beer to law students and see whether they argue better or not when they are drunk.
  • Encourage one group of bikers to stop smoking the funny-looking cigarettes to check whether they get less or more belligerent.
  • Warn one group of students who drink that they would raise their blood alcohol levels next Monday and then compare level of blood alcohol levels of this group with those who were not warned.
  • A school teacher gave two different types of pens to two groups of the students. He informed the one group that it will improve your writing skills and he did not inform the other group. Next week he observed that there was no improvement in the handwriting of the students in either group. It means that the pens had no effect on the writing skills.

Experimental Designs

Randomized controlled trial (RCT): 
The RCT is one of the simplest but most powerful tools in the research. Kane and Radosevich state as “Clinical research worships at the shrine of RCT”. It refers to a study where the elements are randomly selected to receive one of several clinical interventions. The term “intervention” used in RCTs usually refers to treatment; however, it can be used in much wider sense referring to clinical maneuver/s that may have an effect on the health status of the participants. The clinical maneuvers may include:
·       Prevention strategies
·       Screening programs
·       Diagnostic tests
·       Interventional procedures
·       The setting in which health care is provided
·       Educational models
In radiology, RCT can play an important role in the assessment of screening programs, diagnostic tests and procedures.
Example:
RCT has studied effect of raloxifene on fracture risk in postmenopausal women and reported that the women who used raloxifene drug over the same five year period of time as the women who did not take the drug reduced their risk of clinical vertebrate fracture.

Quasi-experimental designs: 
Quasi study design estimates the causal impact of an intervention on its target population or study population. It has many similarities with traditional experimental design or RCTs, but they lack randomization. Quasi-experimental design allows the researcher to control the assignment to the treatment condition without random assignment (e.g. an eligibility cutoff mark). In some cases, the researcher may have no control over assignment to treatment condition.

Advantages and disadvantaged of study designs:




Advantages and Disadvantages of study designs

Study Design
Advantages
Disadvantages
Cross sectional
Quick, covers whole population and offers representativeness
Causes biases, diagnostic inaccuracy and no causal sequence
Cohort
Prospective, establish casual sequence and estimates incidence
Costly and time-consuming
Case control
Economical to focus casual effect
Need recall of past events and cases may not be equal to controls
RCT
Removes biases, offers etiologic and evaluative research
Ethical issues and generalizability
Quasi experimental
Can use natural experiments and may be more effective than RCT
Significant bias

References

Kane RL, Radosevich DM. Conducting Health Outcomes Research. USA: Michael Brown 2011. 

Study designs [Internet]. [Cited 2013 Aug 15]. Available from: http://www.med.uottawa.ca/sim/data/Study_Designs_e.htm.

 

 Next Post

Keywords: Study designs, observational study designs, Experimental study designs,  Randomized controlled trial (RCT), Quasi-experimental designs



Visit Below

Disqus Shortname

Comments system