Reduce data errors using feedback systems

“Feedback is a central feature of life.” – M. B. Hoagland and B. Dodson, The Way Life Works, 1995. There are multiple examples of feedback systems in biology such as homeostasis which is the tendency of an organism or cell to regulate its internal environment and maintain equilibrium, usually by a system of feedback controls, so as to stabilize health and functioning (Lutz, 2013).

In engineering, feedback system is one in which the output signal is sampled and then fed back to the input to form an error signal that drives the system. Control systems maintain the environment, lighting and power in our buildings and factories; they regulate the operation of our cars, consumer electronics and manufacturing processes; they enable our transportation and communications systems; and they are critical elements in our military and space systems (Astrom and Murray, 2016).

Feedback is one of the typical aspects of Agile approach to software development. The teams of various Agile methodologies use feedback to improve the quality of their products and to create new versions of them by conducting user acceptance tests. During these tests the final users detect all bugs in the software products and estimate them according to their own acceptance criteria (Pierce, 2016).

External feedback provides essential information for successful learning. Multiple studies show importance of feedback in early childhood, as toddlers strongly rely on external signals to determine the consequences of their actions. In adults, many electrophysiological studies have elucidated feedback processes using a neural marker called the feedback-related negativity (Meyer et al, 2014).

Feedback is a powerful tool that can reduce errors in clinical research. Current research practice introduces feedback at the time of submission to journals through the peer-review system. Clinician scientists and researchers could benefit from receiving feedback earlier in the clinical research process. In this post, I suggest two areas in the research process where feedback systems can be put in place to reduce data errors:

Study Design Phase:

• Create your chart review form (CRF) iteratively and systematically receive feedback from collaborators on which data points would produce high yield. This process will not only help create a well-designed CRF but it also builds relations between collaborators. Ensure that you have a systematic way of keeping track of collaborators’ comments and input.

Data Collection Phase:

• Ask feedback from researchers (residents, medical students or coordinator) who are collecting the data. It would be too late if you find out that there were data errors at the end of a 5-year study due to data abstractors’ misconceptions or wrong assumptions on data definitions.
• Count and quality matter. It is useful to have real-time information on data points’ frequencies and percentages. This gives the investigator a sense of how much data is missing during the study duration. It is also important to include measures of data quality in the feedback system to ensure that all data can be used for analysis. For example in imaging studies, it is possible that although 100% of subjects have scans only 60-80% can be used for analysis.
• Data integrity queries embedded in the design of data capture system. This can be achieved in different ways such as including data format constraints or defining validation rules for the data values.

It is crucial to get data feedback prior to analyses and throughout the duration of the study to reduce data errors. 

“Errors using inadequate data are much less than those using no data at all” – Charles Babbage