Feedback for learning clinical motor skills: Good or bad?

Heather Carnahan, PhD
Professor, Dept. of Surgery, University of Toronto
The Wilson Centre
Toronto, ON, Canada

Deliberate practice is critical for motor learning

Learning a technical clinical skill is a complex process that not only involves learning to coordinate the muscles, but also acquiring higher level processes that deal with more cognitive issues involved in performing psychomotor tasks.  Thus, the process of motor learning involves both facilitating physical repetition of a skill, but also helping the learner to acquire the appropriate self evaluation skills that are necessary for successful independent performance after the practice session.

The most important variable for improving motor skills is deliberate practice.  However, this practice must be purposeful and guided by the use of feedback to be optimally effective.¹ While it is clear that feedback is important for learning, feedback needs to be appropriate for the learning scenario (ie, what type of feedback, how often should it be delivered, when should it be delivered?)

The interesting issue with feedback is that more is not always better.  In fact, there are many issues related to motor learning that are counterintuitive, thus it is important to not give feed-back based only on intuition, or what has been done in the past, but also consult with the research literature for confirmation of your educational approach.

Too much feedback can impair learning

It has been demonstrated that there are situations in which feedback can actually impair motor learning.  To make sense of this statement, a distinction has to be made between the practice phase and the retention phase of skill acquisition.  During practice, when instructors are typically interacting with and instructing trainees, there are many temporary effects of feedback that will influence performance.  One of these temporary effects is the guidance effect in which large amounts of feedback act like a crutch.  When feedback is removed, as happens in a retention phase where the trainee is performing independently outside of the training environment, performance deteriorates.  It is hypothesized that the ability to self-evaluate one’s own performance never develops because the excessive feedback during practice provided too much informational support.²  However, many instructors have not observed this effect, because they only interact with the students during the training session, and don’t formally evaluate technical performance outside of the practice.  However, this observation is actually good news, since in reality--with limits to faculty resources--formal training sessions are often interspersed with practice trials, some with feedback and some without.

So what is the optimal prescription for giving feedback?  The question of the optimal frequency for the delivery of feedback during practice is still unresolved.  However, a recent study showed that when learning a simulated wound closure skill involving suturing and knot tying on a skin pad, practice with an instructor/trainee ratio of 1 to 2 or 1 to 4 was better than a ratio of 1 to 12.³ The conclusion drawn from this study was that a faculty to student ratio of 1 to 4 was optimal (more frequent feedback did not improve learning, and less frequent feedback resulted in poorer performance).  However, there are other ratios that deserve investigation and that could further optimize resources and maximize learning (e.g., 1 to 6).

Provide both terminal and concurrent feedback

When trainees are learning how to suture, they may not be experienced enough to realize they are not rotating their hand correctly when suturing or that they do not have the correct spacing between the sutures.  In this case, external feedback from an instructor is required.  This feed-back can come in many forms, the most commonly studied form of which is referred to as “knowledge of results.”  This type of feedback refers to verbalized terminal or post-movement feedback that describes the outcome of the movement in reference to a predetermined goal. However, instructing the trainee to rotate the wrist as they make the suture is an example of “knowledge of performance” or concurrent feedback.  This type of feedback is more commonly used in clinical instruction, but is less studied in the scientific literature.

In a recent study, medical students were learning to perform a wound closure skill using an interrupted suture on an artificial bench top model.⁴  Trainees in this study either received feedback about the process of their movement (akin to knowledge of performance), or received feedback after they completed their movements (knowledge of results).  The study results showed that the “knowledge of results” group actually outperformed the “knowledge of performance” group when evaluated on a retention test one month later.  The researchers hypothesized that the knowledge of results allowed trainees to generate some error during their performance, and thus they developed the self evaluation skills necessary for later independent performance.  This study is another example in which less is more in terms of feedback.  Knowledge of performance that guides how a motion is carried out is still very effective, particularly for difficult motor tasks; however, it is important to not provide such feedback during every practice trial.  Also, verbal cues should be short and concise phrases that direct the trainee to an environmental situation, or to some aspect for their movement. For example, the cue “watch where to take the bite with the suture” is an environmental cue and “rotate your wrist” is a movement cue.  Verbal cues can be used in conjunction with demonstration.

Observational learning: Let trainees watch each other and experts demonstrate prior to practice

A common strategy when practicing skills in groups is for the instructor to demonstrate the skill--either live or with video--and then allow the trainee to take a practice turn.  However, the effectiveness of this approach depends on the characteristics of the skill being learned.  If the task to be learned involves a new pattern of coordination, then observational learning is quite effective.  In fact, having trainees watch each other practice is an effective way to facilitate learning, particularly if feedback is given to the trainee performing.  The observer then effectively learns from the performer’s mistakes.  This process can be enhanced even further by providing the observer with a checklist and then using this checklist to provide feedback to the performer.  In terms of the timing of demonstrations by an expert, the basic motor learning literature shows that several demonstrations should precede practice, and there is no benefit to additional demonstrations interspersed throughout the practice session.⁵

Precision of feedback: Not too much, not too little, just the right amount of information

What kind of feedback do you provide about the end result of a skill?  One option is to simply tell the trainee that they got it right or wrong!  However, there is benefit to providing information about the direction of errors produced and information about the magnitude of error.  Keep in mind that it is possible to get carried away with too much precision.  There is no advantage to providing extremely precise feedback; individuals ultimately round off very precise feedback to a meaningful level of precision.

Another approach is to provide feedback only when performance is outside a specified band-width.  Thus, no feedback is effectively providing the message, “you are doing well.”  With this approach, there is a tendency for trainees to attempt to repeat what they just did rather than correct their performance.  One caution of providing too much feedback with too much precision is that trainees are always altering their performance in response to this feedback.  This alteration can lead to inconsistent and variable performance during practice and can impair learning.⁶  It is preferable to tell trainees they are doing well and then ask them to repeat what they just did.

When should feedback be delivered: Make the trainee wait

An instructor might be compelled to provide feedback immediately after the performance of a skill so that the trainee gets immediate information before there is any forgetting of what has taken place.  Yet, increasing the time period between performance of a skill and delivery of feedback has no major effect on motor learning.  In fact, if the feedback is delivered too quickly it can have a detrimental effect on learning.  Still, when learning surgical skills, this scenario would be most likely to occur only from computer based feedback; human instructors typically are not fast enough to provide instantaneous feedback.

Another way of delaying feedback delivery is by providing a summary of feedback after the performance of a number of trials.  Attempts have been made to determine the optimal summary length; collectively, results have shown that a summary length of five trials is better than both feedback after every trial and longer summary lengths of 15 trials.  The optimal summary length for providing feedback during the learning of surgical skills has not yet been addressed.

This summary supports the notion that less is more.  The delivery of feedback should be thoughtful and appropriate, not providing an excess of information, but serving to guide the learner to develop their own reflective self-evaluation skills.

References

¹ Ericsson KA; The influence of experience and deliberate practice on the development of superior expert performance.  Cambridge handbook of expertise and expert performance,  Cambridge, UK: Cambridge University Press.  2006: 685-706.

² Salmoni AW, Schmidt RA, Walter CB; Knowledge of results and motor learning: A review and critical reappraisal.  Psychological Bulletin, 95: 355-86, 1984.

³ Dubrowski A, MacRae H; Investigation of optimal instructor to student ratios for teaching fundamental clinical skills to medical students: A randomized, control study.  Medical Education, 40: 59-63, 2006.

⁴ Xeroulis GJ, Park J, Moulton C, Reznick RK, LeBlanc V, Dubrowski A; Teaching suturing and knot tying skills to medical students: a randomized controlled study comparing computer-based video instruction and (concurrent & summary) expert feedback.  Surgery, 141: 442-49, 2007.

⁵ Magill RA; Motor learning and control: Concepts and applications (8^th Ed).  McGraw Hill College.  2006.

⁶ Lee TD, Carnahan H; Bandwidth knowledge of results and motor learning: More than just a relative frequency effect.  Quarterly Journal of Experimental Psychology, 42A, 777-89, 1990.

 

Online December 5, 2007