Types of Study Designs

The study design is a framework of methods and procedures used to collect and analyze data in a particular research problem. Research study designs are of many types, each with its advantages and limitations. The type of study design used to answer a particular research question is determined by the nature of the question, the goal of the research, and the availability of resources. The research question should include 1) the defined population from which the study group is taken 2) the outcome that is measured 3) the time frame 4) Comparison group (in an analytical study) and 5) Intervention group (in an interventional study). Since the design of a study can affect the validity of its results, it is important to understand the different types of study designs and their strengths and limitations.

Study designs are broadly classified as observational and interventional studies.

• Observational studies.

Observational studies are those where the researcher is documenting a naturally occurring relationship between the exposure and the outcome. The researcher does not do any active intervention in any individual, and the exposure has already been decided naturally. For example, looking at the incidence of central retinal vein occlusion in those with and without hypertension.

In this study design, the researcher is just observing and analyzing the characteristics of particular groups of individuals. There is no active intervention in the form of either medicines or any surgical intervention.

This itself may be classified as a descriptive or analytical study.

DESCRIPTIVE STUDY

Descriptive (or non-analytical) studies merely describe the data on one or more characteristics of a group of individuals.[¹] These do not try to answer questions or establish relationships between variables. These types of studies can generate hypotheses. Examples of descriptive studies include case reports, case series, and cross-sectional surveys (please note that cross-sectional surveys may be analytical studies as well).

For example, we are observing a group of individuals with a corneal ulcer in our institution. We observe the characteristics of this group of patients. We can find out the number of patients with a fungal, bacterial, or viral corneal ulcer. The baseline characteristics of these individuals can be studied. There is no control group in this type of study.

Case report

Osler one of the fathers of modern medicine stated that “Always note and record the unusual…. Publish it”

A case report is a “detailed description of a few patients or clinical cases with an unusual disease or complication, uncommon combinations of diseases and unusual or misleading semiology, cause or outcome.”[²]

The value of case reports is usually underestimated when compared to other studies which are supported by evidence. Case reports have indeed led to discoveries and the introduction of treatments for novel diseases.

Case series

According to the latest version of the Dictionary of Epidemiology a case series is defined as a “ collection of patients with common characteristics used to describe some clinical, pathophysiological or operational aspects of a disease, treatment or diagnostic procedures”[²]

In the review article “Clinical case series: a conceptual analysis” published in the Journal of African health sciences, FM Abu-Zidan et al.[³] reviewed 586 papers and suggested that a case series should have more than 4 patients while 4 or fewer patients should be reported as case reports.

One of the reasons behind this is that, if data of a group of subjects are to be summarized statistically, a minimum number of subjects is needed to be valid. Five is the minimum number needed because the standard error of the mean, which is used for comparison will be much larger if the number is less than five.

Cross-sectional surveys

This type of study is a transverse type of study design which includes a larger number of patients who are evaluated at a single time without any follow-up. There is no comparison group. Both the cause and effect have already occurred. It can measure the prevalence of a disease or risk factor. It cannot find out the cause-and-effect relation. It is easy to conduct.

For example, a cross-sectional survey may be conducted to find out the prevalence of presumed ocular tuberculosis by studying a group of patients with the disease.

ANALYTICAL STUDY

This type of study design attempts to test a hypothesis and aims to find out a cause-and-effect relationship.[⁴] The researcher tries to find out the effect of exposure on the outcome or risk factor of a particular disease. They can be of different types.

Cross-sectional analytical study

As described earlier this is a transverse type of study design and the difference with the descriptive cross-sectional survey is that there will be a comparison group in an analytical cross-sectional study.

Cohort study

“Cohort” term refers to a group of individuals with a shared characteristic 9. In a cohort study, individuals with or without a risk factor are followed up over time to know about the occurrence of a disease or an outcome. So, it is a forward direction study (moving from exposure to the outcome) or “prospective study.” It helps to find out the risk of disease or outcome among those with the risk factor compared to those without the risk factor. (Relative risk-RR)

Example: Individuals with and without thyroid disorders are followed up for the occurrence of primary open-angle glaucoma.

Advantages:

1. Risk of outcome among exposed and unexposed can be studied.
2. It can be ensured that exposure occurred before the outcome.
3. Multiple outcomes to exposure can be analyzed.
4. It closely resembles interventional study except for the lack of random assignment of exposure and hence it has the strongest validity among observational studies.

Limitations

1. Long follow-up is needed in most cases.
2. Confounding factors may affect the outcome. In the above example where those with and without thyroid disorders were followed up for the development of glaucoma, the confounding factors may be diabetes, coronary artery disease, etc., which may also predispose to the development of glaucoma.

Variation of a cohort study:

Retrospective cohort study: Although cohort studies are prospective, in some situations retrospective cohort study can also be done. For example, in the above-mentioned study if the researcher includes patients with and without thyroid disorder from the hospital records 5 years back and then calls up these patients to see whether they have developed glaucoma, then this becomes a retrospective cohort. Here the direction is always forward i.e., from exposure to the outcome, but since the outcome has already occurred this becomes a retrospective study.

Case control study

The researcher selects cases (with disease or outcome) and controls (without disease or outcome) and then tries to find out the history of exposure among these two groups. These are backward-directed studies and are always retrospective as the outcome has already occurred.

Identification of controls is of key importance in the case-control study as this may influence the estimation of the association between exposure and outcome. The control should be similar to cases in all aspects except for the absence of disease.

Example: All preterm babies with (cases) and without ROP (controls) were included and the duration of exposure to oxygen during the neonatal period was analyzed in the two groups.

Advantages

1. Easy to perform
2. Less time consuming
3. Multiple risk factors can be studied.
4. It can be used for studying rare diseases.
5. Since it is easy to perform these are usually done as an initial study over which further complex studies like cohort or interventional studies can be done.

Limitations

1. Improper selection of controls can lead to selection bias.
2. Since it is a retrospective study the data collection depends on the credibility of the records.
3. Confounding factors may play a role here also.
4. It cannot be ensured that exposure occurred before the outcome of interest.
5. We cannot estimate the risk ratio, can measure only the odds ratio.

Variation of case-control study:

NESTED CASE-CONTROL STUDY:

It is a part of the cohort study. Two groups of individuals with and without risk factors are included and followed up for the development of an outcome or a disease. During the due course, some may develop the outcome of interest and they will be selected as cases and the other group who have not developed the disease will be selected as controls. Here the advantage is that irrespective of being a retrospective study it can be ensured that exposure occurred before the disease or outcome of interest.

Interventional study

In the interventional study design, the researcher is actively involved in the process by performing an intervention in some or all participants.[⁵] This design is prospective by its nature. Sometimes confusion may arise between a prospective cohort study and an interventional study. The only difference is that in interventional studies the researcher assigns each person for the intervention whereas in a cohort study the intervention or exposure is already decided.[¹]

Broadly speaking interventional studies can be divided into two:

1. Clinical trials
2. Community trials.

It can also be divided into preventive trials or therapeutic trials depending upon whether the intervention includes preventive measurements or therapeutic measures.

Several variations of interventional study designs are there which include:

1. Randomized controlled trials
2. Non-Randomized controlled trials
3. Interventional studies without concurrent controls
4. Before-After or Pre-Post studies
5. Factorial study design
6. Crossover study design
7. Cluster Randomized trials.

Randomized controlled trials

Participants fulfilling the inclusion and exclusion criteria are randomly assigned into two groups with different interventions or one group with any other group without intervention. Randomization means that each participant has an equal chance of being allocated into each group. The two groups are similar in all aspects except for the intervention so the two groups are comparable. Some additional features are usually added to this type of study design like allocation concealment, blinding, intend-to-treat analysis, measurement of compliance, minimizing dropouts, and ensuring appropriate sample size.

Nonrandomized controlled trials

The allocation to two groups is not Randomized. For example, the investigator may decide which participant is to be allotted to one group based on availability or affordability. So, the two groups may not be comparable and so the validity of this study is also low.

Interventional studies without concurrent controls

When a new intervention or a drug is invented the researcher can compare those who have received the drug or intervention to a group of individuals who were followed up in the past and have not received the treatment. This includes a lot of bias.

Before-after or pre-post studies

A variable is measured in the same group of participants before and after an intervention. There is no comparable group. The particular outcome may be related to other changes that might have occurred at the same time as the intervention. Thus, the outcome in such studies may not be attributed to the intervention making it a weaker study design than RCT. Some include this study design as an observational study design rather than an interventional study design.

Factorial study design

When there are two or more interventions for a particular disease, we can find out whether each intervention is efficacious or whether a combination of these two interventions is efficacious or not.

For example, if there are two interventions A and B, we can allot participants to any of the four combinations of interventions.

1. A alone
2. B alone
3. A and B
4. None of A or B.

So, we can find out the effect of interventions A and B, the combination of A and B, and compare with the controls (not receiving A or B). So, in this single study, we can compare the effect of two interventions and the effect of their combination too.

Cross-over study design

This is a type of study design in which the study participants intentionally switch over to the other arm of intervention. At the start of the study, participants are allotted to a particular intervention group based on random allocation and after some period they will be switched over to the other arm. The other type of study design where the participants don't have a crossover is called parallel-arm RCT.

The advantages of this type of study design

1. Each person serves as a control for himself.
2. Requires a smaller number of participants when compared to parallel arm study.
3. It is possible only in some types of diseases and interventions.

Cluster randomized trials

In some situations, a particular intervention cannot be applied to individuals but can be administered to groups. A cluster of individuals can be allotted to an intervention group and other clusters to the other arm.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.