STUDY DESIGN IN SCIENTIFIC RESEARCH

Design the study

Aspects of study design: Experimental Units Type of study and sampling What to measure? How to measure? How to analyze?

Design the study

Observational Descriptive study: Here observations made of a single sample are reported. Observational Analytical study: Here observations of similarities/differences in two or more samples are reported Experimental/Interventional Study: Here the outcome of an intervention in one group is compared to another group where no intervention is made.

Observational Descriptive Studies These are simplest types of studies and include case reports and case series.
Case Reports: These are best looked upon as a means to generate hypotheses about disease processes. They usually report an unusual event in respect to a disease. Case Series: These report observations in a group of patients. They are good for studying rare disorders or reporting treatment outcomes of interventions like surgery.

Observational Analytical Studies Here observations in two or more groups are compared without researchers making an intervention. Two main types are casecontrol studies and cohort studies
Case-Control Studies Cohort Studies

Observational Analytical Studies Case-Control Studies Here the researcher is looking for the possible causes of an identified outcome.
A group of subjects with the outcome is compared with a matched control group without the outcome. The difference in exposures and risk factors between the two groups is then analysed retrospectively.

Observational Analytical Studies Case-Control Studies Example: TMD in orthodontic patients compared to non orthodontic patients case-control studies are usually retrospective, but not always

Observational Analytical Studies Case-Control Studies

Pros: Cheap to run, few subjects required thus quicker yield of results Cons: Difficult to recruit matching control group, relies on recall and reporting of risk factors, not good for rare exposures, difficult to establish temporality of exposure and outcome.

Case Control Studies

Time Direction of inquiry
Finish Exposed Not exposed Start Cases (People with the disease) Population Exposed Not exposed Controls (Similar people without the disease)

Observational Analytical Studies Cohort Studies Here the researcher is looking at whether an exposure or risk factor results in an outcome.
A group of subjects with the exposure is compared with a matched group without the exposure. Both groups are then observed or followed up to see the difference in development of the outcome between the two groups.

Cohort Studies
Time Direction of inquiry
Start People without the disease Exposed Not exposed Finish Disease Population No disease Disease No disease

Observational Analytical Studies Cohort Studies

Pros: Good for rare exposures, assessment of temporality, estimation of disease incidence. Cons: Expensive to set up, unsuitable for rare outcomes, prolonged follow-up, selection bias, drop outs, influenced by confounders, no randomisation.

Experimental/Interventional Studies Here the researcher intervenes in a group of subjects to assess the outcomes that the intervention produces. There are several types of experimental/interventional studies:
Open Trial Controlled Trial Randomised Controlled trial (RCT) Crossover Trial

Experimental/Interventional Studies Open Trial

Here all the test subjects receive the intervention. There is no control group. This is a cheap and easy method of assessing efficacy of an intervention.

Experimental/Interventional Studies Controlled Trial

Here there are two groups; one receives the intervention and the other a placebo or sham treatment. The outcomes in the two groups are then compared.

Experimental/Interventional Studies Randomised Controlled trial (RCT)

Here in addition to there being a controlled group, subjects are allocated to either group in a random fashion. This prevents selection bias and equally distributes possible confounders. If the process is blinded this also prevents observation bias. RCTs are considered the gold standard way to assess the efficacy of interventions.

Randomised Controlled Studies

An investigation without bias. Bias is the systematic deviation of the results of a study from the truth because of the way it has been conducted, analysed or reported

Randomised Controlled Trials

Controlled : Experimental & Control groups Randomised: Random selection of members of experimental and control groups. Randomisation Simple Randomisation Randomised blocks Blinding Single-Blinded Double-blinded

Randomised Controlled Trials

Time Direction of inquiry
Start
Non-participants Intervention

Finish
Measure outcome Lost to follow-up Control Measure outcome

Sample Population

Randomization to groups

Experimental/Interventional Studies Randomised Controlled trial (RCT)

Pros: Reliable measure of efficacy which can be used in meta-analyses. Cons: Expensive to run, ethical issues around not giving treatments to ill patients.

Experimental/Interventional Studies Crossover Trial

Here all subjects receive the same intervention. Then half through the study the subjects cross over to another treatment or placebo. This is useful when there are limited number of subjects such in rare diseases. Not for conditions which can be cured with the treatment or end up in death; only for chronic diseases (pain, asthma, arthritis, psychological problems).

Experimental/Interventional Studies Crossover Trial

Pros: Subjects behave as their own controls, useful in rare diseases. Cons: Risk of carry-on effects of treatments which require wash-out periods to eliminate the effects.

Crossover Trials
Treatment Eligible/Consent Randomize Control Treatment Control

Run-in period

Washout period

Systematic Reviews and Meta-analyses

Systematic reviews are studies that systematically search and analyse research articles based on a clinical question.

Systematic Reviews and Meta-analyses Reviews of particular importance are those dealing with RCTs of a specific intervention. Such reviews pool the results of several studies to produce results that are statistically more powerful, reliable and accurate than the individual studies.

Systematic Reviews and Meta-analyses Due to the value placed on systematic reviews they are considered as a goldstandard in evidence based medicine. The importance of systematic reviews has led researchers to try and standardise on how such studies should be reported in published literature.

Systematic Reviews and Meta-analyses The QUOROM statement (Quality of Reporting of Metanalysis) was produced during a conference of the same name in 1999. This has recently been superseded by the PRISMA statement (Transparent Reporting of Systematic Reviews and Meta-analyses). The statement sets out checklist of 27 items that authors should report in their review.

Systematic Reviews and Meta-analyses Pros: Statistically powerful measures of effect, Cons: Quality of meta-analysis limited by the quality of original studies included or available.

Study Designs : Summary