Different research designs and their characteristics in intensive care

Nedel, Wagner Luis; Silveira, Fernando da

doi:10.5935/0103-507X.20160050

Acessibilidade / Reportar erro

Brasil

Revista Brasileira de Terapia Intensiva

Español English

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

SPECIAL ARTICLES • Rev. bras. ter. intensiva 28 (3) • Jul-Sep 2016 • https://doi.org/10.5935/0103-507X.20160050 copy

Different research designs and their characteristics in intensive care

Authorship SCIMAGO INSTITUTIONS RANKINGS

ABSTRACT

Different research designs have various advantages and limitations inherent to their main characteristics. Knowledge of the proper use of each design is of great importance to understanding the applicability of research findings to clinical epidemiology.

In intensive care, a hierarchical classification of designs can often be misleading if the characteristics of the design in this context are not understood. One must therefore be alert to common problems in randomized clinical trials and systematic reviews/meta-analyses that address clinical issues related to the care of the critically ill patient.

Keywords:
Research designs; Randomized clinical trial; Systematic review; Meta-analysis; Cohort studies

Study design	Characteristics/applicability	Disadvantages
Randomized clinical trial	- Considered the gold standard for analyzing therapeutic interventions - The study groups differ only in terms of the intervention factor: there is direct inference of causality - Ideally, it is able to control selection bias and confounders that can affect the study outcome	- Expensive and time consuming - Not always feasible for ethical reasons - The absence of blinding, especially when it cannot be applied, may directly affect the study result - Subject to loss of patients during follow-up - Generally evaluates specific disease scenarios - Frequently conducted in academic scenarios, limiting the generalizability of the data
Nonrandomized clinical trial	- Performed when an RCT would be ideal but is not possible (due to cost or unacceptability to patients or managers) - Generally more practical than RCT - Used when attempting to demonstrate the efficacy of a new treatment (phase IV studies)	- Sample group is determined by convenience - Causality is not directly inferred - The effect of confounders may not be obvious
Before-and-after study	- Quick and easy to conduct - May be used to compare the ICU data or variables of a specific patient before and after an intervention is instituted - Useful in interventions that are difficult to perform using a clinical trial (e.g., to evaluate the efficacy of isolation measures in the transmission of nosocomial infections by multiresistant germs)	- Subject to the Hawthorne effect - Simultaneous interventions may alter the studied outcome and the interval between the measurement of outcomes - Subject to the “regression toward the mean” phenomenon
Cross-sectional study	- Measures the prevalence of a particular outcome - One of the first stages in investigating the cause of disease outbreaks - May be used to analyze frequencies of risk factors and study outcomes - Low cost and easy to conduct	- Measures of exposure and disease are performed at the same time, which reduces the study’s ability to establish a causal association (association does not necessarily imply causation) - Difficult to use for investigating low-prevalence conditions - Appropriate when evaluating diagnostic tests or disease outbreaks
Case series	- The characteristics of the study population are related - Useful when characterizing rare disease scenarios with a diagnosis or treatment that is not clearly established in the literature (for example: the characterization of outbreaks of rare micro-organisms)	- There is no comparison group (reference) - Sampling is not representative of the studied population - Causal associations cannot be evaluated
Case-control	- Cheaper and faster (than cohort studies) - Useful when investigating disease causes, especially when the incidence of the outcome (disease) is rare or has a long latency period, as the starting point is the outcome and a retrograde analysis of risk factors is performed	- There is no way of knowing the incidence of the disease - Prone to sampling bias among both cases and controls and to observation and recording bias
Cohort	- Provides the best information about the etiology, incidence and natural history of a disease by starting with risk factors and analyzing subsequent outcomes - Useful when risk factors have a low prevalence - A cohort study should be used when it is not possible to conduct an RCT for ethical reasons	- Expensive and requires long periods of observation - Subject to loss to follow-up - Unable to control all confounders - Subject to discontinuity of patient follow-up - Frequently presents results that are discordant with those of RCT that analyze the same research question
Systematic review with meta-analysis	- Clinical research with the lowest evidence level - Summarizes the results of primary studies using strategies that decrease the occurrence of random and systematic errors	- The quality of a meta-analysis depends on the quality of the studies it includes - The greater the heterogeneity of the included studies, the lower the reliability of the result obtained

Associação de Medicina Intensiva Brasileira - AMIB Rua Arminda, 93 - Vila Olímpia, CEP 04545-100 - São Paulo - SP - Brasil, Tel.: (11) 5089-2642 - São Paulo - SP - Brazil
E-mail: rbti.artigos@amib.com.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] Corresponding author: Wagner Luis Nedel, Unidade de Terapia Intensiva do Hospital Nossa Senhora da Conceição, Avenida Francisco Trein, 596, 1º andar, Zip code: 91350-200 - Porto Alegre (RS), Brazil, E-mail: wagnernedel@uol.com.br