Difference between revisions of "2.1.4 Purpose of research"

From EQIPD
Jump to: navigation, search
Line 73: Line 73:
 
   
 
   
 
'''​PLEASE DO NOT FORGET'''
 
'''​PLEASE DO NOT FORGET'''
 +
 
To consider adding this subject to a training program for new employees or refresher training (if appropriate)​
 
To consider adding this subject to a training program for new employees or refresher training (if appropriate)​
  

Revision as of 16:56, 9 February 2021

​​​​​​​​​​​​A. Background & Definitions

Several commonly acknowledged risks can bias the research results (Hooijmans et al. 2014)​.

There are modes of research that can tolerate a certain level of uncertainty, and while not leading to a formal knowledge claim, such work is an essential part of the research process. It may be used to generate hypotheses, to provide evidence to give the investigator greater confidence that an emerging hypothesis is valid, or to “screen” compounds for potential effects prior to more formal testing.

There are also modes of research where researchers want to minimize the risks of failing due to inadequate control of the risks of bias.

For every study, EQIPD recommends scientists to apply protection against risks of bias and to be transparent about the protective measures applied.

EQIPD requires that the maximal rigor possible is applied (and exceptions explained / documented in the study plan) to research that is conducted with the prior intention of informing a knowledge claim (Glossary​). This will usually (but not always) involve some form of null hypothesis statistical testing or formal Bayesian analysis. Hypotheses are articulated in advance of data collection, with pre-specified criteria defining the primary outcome measure and the statistical test to be used. Depending on the purpose for which the knowledge claim will be used, different research strategies are appropriate. A single well conducted preclinical study may be considered sufficient to convince others that the phenomena are real enough to justify their attention, but may not be sufficient to justify major research investment such as a clinical study.

Examples of research requiring the maximal rigor possible include:

  • Experimental studies to scrutinize preclinical findings through replication of results alongside investigations into boundary conditions and robustness through conduct of additional (control) conditions and multicenter studies​ (Kimmelman et al. 2014)
  • Research aimed to generate evidence that enables decisions such as critical studies that, dependent on the outcome, will trigger a chain of activities and events associated with significant resource and time costs (e.g. a decision to initiate a new drug development project or to initiate GLP safety assessment of a new drug candidate)
  • Studies for which any outcome would be considered diagnostic evidence about a claim from prior research​ (Nosek and Errington 2020)
  • Labor-, resource- and/or time-intensive studies that cannot be easily repeated


​B. Guidance & Expectations

One of the Core Requirements of the EQIPD Quality System is that investigators must assert in advance whether a study will be conducted to inform a formal knowledge claim. This statement should be recorded in the 2.1.1 Study (experimental) plan (see below for a template).

If uncertain about your study being knowledge-claiming or not, please check FAQ​ page or contact the EQIPD Expert Team for an advice.

Efforts to minimize the risks of bias (Hooijmans et al. 2014) should be applied to all studies that aim to inform a knowledge claim (table inspired by Dirnagl 2016)​:


​​ All research Research informing a formal knowledge claim (i.e. research requiring maximal rigor)​*
2.1.1 Study (experimental) plan should be defined and documented before starting the experiments must be defined and documented before starting the experiments
Study hypothesis advised to define must be pre-specified
2.1.7 Blinding advised to implement should be implemented, exceptions must be justified and documented
2.1.8 Randomisation advised to implement should be implemented, exceptions must be justified and documented
2.1.6 Sample size and power analysis advised to define and document before starting the experiments must be defined and documented before starting the experiments (e.g. included in the study plan)
Data analysis advised to define and document before starting the experiments must be defined and documented before starting the experiments (e.g. as a formal statistical analysis plan and/or included in the study plan)
2.1.9 Inclusion and exclusion criteria advised to define and document before starting the experiments must be defined and documented before starting the experiments (e.g. included in the study plan)
Deviations from study (experimental) plan advised to document must be documented
2.1.11 Preregistration no should be implemented​
  • Please refer to Glossary for explanation on the EQIPD use of the verbs "must" and "should"​


​PLEASE DO NOT FORGET

To consider adding this subject to a training program for new employees or refresher training (if appropriate)​


C. Resources

Template to create a study (experimental) plan based on the above guidance - 2.1.1 Study plan.docx, 2.1.1 Study plan with macro.docm​


Literature:

  • Hooijmans CR, Rovers MM, de Vries RBM, Leenaars M, Ritskes-Hoitinga M, Langendam MW (2014) SYRCLE’s risk of bias tool for animal studies. BMC Medical Research Methodology 14:43 [1]
  • Kimmelman J, Mogil JS, Dirnagl U (2014) Distinguishing between exploratory and confirmatory preclinical research will improve translation. PLoS Biol 12(5):e1001863 [2]
  • Nosek BA, Errington TM (2020) What is replication? PLoS Biol 18(3): e3000691 [3]
  • Dirnagl U (2016) Thomas Willis Lecture: Is Translational Stroke Research Broken, and if So, How Can We Fix It? Stroke 47(8):2148-53 [4]



back to Toolbox

Next item: 2.1.5 Pre-specification