Making clinical decisions

 

The emergency physician is faced with the challenging task of making many clinical decisions very rapidly during a shift.  In order to do this properly and without endangering patients, the emergency physician (EP) must not only have a solid base of medical knowledge but also a repertoire of decision-making skills in order to make good decisions while preventing mental fatigue.  These decisions can include aspects of patient care pertaining to triage, diagnostic testing, disposition, cost effectiveness, and patient satisfaction.  By better understanding how decisions are made, the EP can become more efficient and more effectively teach others how to make better decisions.

 

Review by Evan Schwarz
Column organized by Evan Schwarz

Chapman DM, Char DM, Aubin CD.  Clinical decision making.  Rosen’s Emergency Medicine:  Concepts and Clinical Practice, ed 6.  2006, pg. 125-133.

 

One of the most important decisions facing the emergency medicine physician is what is the diagnosis, or at least, what life-threatening event do I need to rule out.  In order to do this properly, the EP must use clinical reasoning which involves both medical inquiry and clinical decision making.  Medical inquiry refers to the skills or techniques used to gather data such as the history and physical and labs.  Clinical decision making refers to using that information to make a decision concerning the diagnosis or treatment of the patient.  The EP has developed four particularly useful strategies to help them make many rapid decisions concerning diagnosis and management:  pattern recognition, using an algorithm, hypothetico-deductive, and a naturalistic, event driven process.

When using a pattern recognition process, the EP is allowed to make decisions without using much conscious effort.  This allows for greater efficiency and proficiency while allowing the EP to build a mental reserve for more difficult situations.  This strategy is based on the memorization of many facts.  Over time the EP learns to look for these facts, such as specific vital signs or key aspects of the patient’s history and physical and group them into a specific pattern in order to make a “doorway diagnosis.”  To properly use this strategy, not much needs to be known about the pathology or physiology of the disease process.  However the practitioner must have had enough clinical encounters to identify the clinically relevant signs and symptoms and know the important consequences of the pathology.  Pattern recognition corresponds to the lowest level of the clinical decision making hierarchy. 

The EP can also make decisions using rules or algorithms.  To do this involves a greater understanding of pathologic and physiologic processes and so is higher on the hierarchy than simple pattern recognition.  The EP can use this strategy when faced with an atypical patient encounter or unusual chief complaint.  Algorithms and heuristics (rules of thumb) are used to classify the unusual signs and symptoms into previously defined diagnostic or therapeutic groups.  This is the case, for example, when using ACLS.  The EP may not initially know the underlying disease process but understands enough to use one of the pre-determined algorithms to act in these high stress and critical situations where higher levels of decision making may not be possible or practical.

 

The hypothetico-deductive process is at the top of the decision-making hierarchy.  In this strategy, the EP must create new solutions by using previously gained knowledge to find an answer to a difficult situation.  This requires a conscious, analytic process of stored knowledge.  This type of strategy involves hypothesis generation, hypothesis evaluation (data that is collected from the history, physical, and labs/radiology is used to confirm or exclude a previously generated hypothesis), hypothesis refinement (as one hypothesis is excluded, more may be added), and hypothesis verification.  To do this properly, the EP must be open to generating new ideas and expanding their differential while being aware of premature judgment or closure (not labeling a patient with a certain diagnosis before you have all the information to make the proper diagnosis).

 

In the naturalistic or event-driven process, the EP acts before a definitive diagnosis has been made.  This is most common in an unstable patient where the EP does not recognize an existing pattern or algorithm that can be used but must take some course of action.  To simplify the EP switches from an evaluation of possible diagnoses to an evaluation of possible actions or therapeutic options.  Here the EP may focus on ruling out the worst case scenario and must at times be willing to accept a good or likely diagnosis rather than a definitive one.

There are some strategies that can be used to make the EP more efficient while at the same time providing maximum patient satisfaction and hopefully decreasing length of stay.  Optimally we should see patients on arrival because time between arrival and seeing a physician is a major determinant of patient satisfaction.  Using complaint based templates with specific prompts, bed side charting, and on-line entry allow for multitasking and increased throughput times.  Trying to understand why the patient is coming to the ED and what they really want is imperative to patient satisfaction.  Other rules of thumb for the EP are to sit at a patient’s bedside to collect a thorough and uninterrupted history and physical.  Only order those tests that will affect diagnosis or disposition and avoid tests when possible by using available algorithms.  These will aid throughput time.  To help keep the EP mentally fresh, and therefore able to make good decisions, remember to only carry a maximum of 4-5 “undecided” patients before making dispositions.  After evaluating a patient allow a few minutes to process the encounter and work the patient through to disposition if possible.  Try to generate life threatening and most likely diagnoses first and then gather data to confirm or exclude them. 

 

It is vital that we understand how our thought process works in order to prevent mistakes and adverse consequences to our patients.  There are three broad categories of adverse events.  Affective errors occur because of physician attitudes or perceptions and not because of deficient knowledge.  Here errors can be made due to attribution error (we blame the patient for their illness) or attitudinal bias where our own stereotypes and perceptions negatively interfere with our management.  In this bias or own pre-conceived ideas about the patient could affect our perceptions and might cause the ED physician to under treat someone’s pain because we think they are dramatic or not consider a more serious diagnosis because they come to the department every week.  Psychomotor errors occur due to an incorrectly performed procedure or due to a procedure being withheld due to physician lack of skill.  Preparation and practice are key in avoiding this type of error.

 

Other errors occur in the cognitive process of the EP.  This can occur through the use of faulty data collection in the history and physical or improper ordering of diagnostic tests.  Confirmation bias occurs when we lock onto a diagnosis because of preconceived ideas and then misinterpret the diagnostic test.  In this category errors can also be made due to lack of understanding of how to apply an algorithm or clinical decision rule as well as because of lack of knowledge about certain possible diagnosis.  Without adequate knowledge of the signs and symptoms of a disease process, the EP can not perform adequate pattern recognition to arrive at a diagnosis.  Other errors in this category include a situational bias (they waited until after work to come to the ED so they cannot be sick), considering only the obvious diagnosis, not considering that there could be a second diagnosis, blaming the symptoms on an already diagnosed psychiatric condition, accepting previously diagnosed conditions even if it does not sound correct, and not doing a more complete work up because a prior work up is nearly complete.

Even in the hectic environment of most emergency departments there are strategies that can be used to help eliminate errors.  Keep in mind that particularly in the sick patients, making the correct management decision may be more important than making the correct diagnosis.  Be aware of tools such as online texts and PDAs to help when a disease process or diagnosis is unknown.  In decision analysis the benefits and risks of therapeutic options are weighed mathematically to help make difficult clinical decisions.  While not having much of an impact in the ED currently, it probably will have a large impact in the near future.  Most importantly physicians can improve by being aware of their own bias and errors in their own thought process.  We all have a subset of patients that are difficult because of our own personal reactions to them.  By being aware of this before the encounter, we can make sure to keep an open mind.  Other rules of thumb to prevent adverse events include paying attention to vital signs and the EMS notes, being aware of high risk times such as sign-out or when the department is exceptionally busy, being highly suspicious of return visits for the same complaint, and being aware when the presumptive diagnosis does not adequately explain all the signs and symptoms.