Home About us Current issue Ahead of Print Back issues Submission Instructions Advertise Contact Login   

Search Article 
Advanced search 
Saudi Journal of Kidney Diseases and Transplantation
Users online: 5286 Home Bookmark this page Print this page Email this page Small font sizeDefault font size Increase font size 

DOCTORS DIARY Table of Contents   
Year : 2007  |  Volume : 18  |  Issue : 3  |  Page : 448-451
Diagnostic Decision Making in Medicine

Division of Nephrology, Hypertension and Renal Transplantation, King Abdulaziz Medical City, Riyadh; Clinical Professor of Medicine, King Saud Bin Abdulaziz University for Health Sciences, Saudi Arabia

Click here for correspondence address and email


Medical diagnostic decision making, like all decision making instances, involves a cognitive process. It starts with internalizing data, structuring it and generating hypotheses. In simple cases very often the experienced physician utilizes pattern recognition, intuition and retrieval in reaching a diagnostic decision. In more complicated cases more complex cognitive process takes place including the balancing of probabilities and the weighing of evidence. Apart from knowledge and experience, personal traits and perceptions and prejudices play some part in some situations.

How to cite this article:
Al Sayyari AA. Diagnostic Decision Making in Medicine. Saudi J Kidney Dis Transpl 2007;18:448-51

How to cite this URL:
Al Sayyari AA. Diagnostic Decision Making in Medicine. Saudi J Kidney Dis Transpl [serial online] 2007 [cited 2022 Sep 25];18:448-51. Available from: https://www.sjkdt.org/text.asp?2007/18/3/448/33769
Medical diagnostic decision making, like all decision making instances, involves a cognitive process. It starts with internalization of a set of data or observations, structuring them and coming out with limited number of decision making options. However, with experienced physicians, specially when an urgent decision has to be made, the process can be intuitive involving, pattern recognition approach to fit a set of observations into a specific category (categorisation). Myers, a psychologist, basing her theory on earlier work by Jung, developed four opposing sets of decision making inclinations or mind sets:

  • Extroversion versus introversion,
  • Thinking versus feeling
  • Sensing versus intuition and
  • Judgment versus perception

Depending on where you fit into this matrix you may be labeled as logical, emotional, intuitive, etc decision maker.

In complex situations, decision making process is a continuous process of refinement based on experience, data and evidence. We have learnt at medical schools that we make a diagnosis based on history followed by physical examination. After that we resort to instigations. Lucky for us often the history and physical examination with few simple supportive investigations suffice. Examples of this would include diagnosing diseases such angina, asthma, pneumonia, gall bladder disease, etc.

However, I wish things could be that simple all the time. We may be faced with what we call difficult historian or simply too complex a case to make a clear decision at the outset difficult or even impossible. Here comes the place of differential diagnosis and need for tests. How do we then decide on the diagnostic possibilities and what tests take priority and when and what, if any, empirical treatment should be given.

I must, though, first mention here that some junior staff, when pressed on a patient's history by there seniors resort to the old face saving adage of "but the patient is a poor historian" when in fact he/she isn't. This adage is also used to a presumed advantage during medical qualifying examinations. Some of us who practised in the West at some time of our careers may perceive our Arab patients as being poor historians. We may even conclude that this may be related to low level of education or low degree of health awareness. I personally disagree that our patients are poor historians. I have done a survey-based study (not yet published) comparing Saudi and Austrian patients. The survey was of medico­anthropological nature aimed at finding out if there are significant differences in the patients' perception and prioritisation of symptoms and signs and how they perceive disease state as a whole. It was amazing how different our patients are from the European patients. For example dizziness had higher priority over chest pain in our patients and the opposite in the Austrian patients. Is it, then, surprising that some of our patients with cardiac ischemic pain, may indicate dizziness as their presenting symptom and add pain as an afterthought.

Let us now consider the psychological aspect of the cognitive diagnostic reasoning process. There are two major components to this process- problem solving and decision­-making. As doctors, we should be aware that both components have their limitations and when we know this and also when we are aware of our prejudices we may indeed go a long way in avoiding errors.

For example we are taught, "Common things occur commonly." (Bayes's theorem of subjective probability). This is a valid and extremely helpful axiom. However we should be well advised to keep an open mind especially when things don't fit quite completely. A little hint from the history or a clue from the physical exam may well indicate a variance from this axiom. Experienced doctors are more likely than novices to pick up such clues. Recently we had what would, in virtually all respects, point to run of the mill diagnosis of end stage diabetic nephropathy, an unusually persistent bone pain led eventually to the diagnosis of multiple myeloma with cast nephropathy.

Yet there are times of uncertainty regarding diagnosis. Much research of psychological nature has been done on decision making in situations of uncertainty.

Factors such as accumulated long term experience and knowledge, perceptions and prejudices, recent frequent exposures, rareness or commonness of the disease all play part in diagnostic decision making. Then comes co­existing factors such the state of mind of the doctor, time available and attention span at the time. Available evidence to the clinician and his/her intuition also come into play.

A case of cerebral malaria will be immediately diagnosed in Khartoum and is likely to be initially missed in London. The opposite will happen in case viral encephalitis. Of course the case of cerebral malaria in London will have a better chance of an early diagnosis if one of the treating physicians happened to have practiced in Africa recently.

An experienced physician, while listening to the history and, almost from the outset, formulates a list of hypotheses in his mind. This list quickly shrinks as the history unfolds. He does that almost imperceptibly. The list is reduced further while he listens to the history such that be the end he reaches one or two possibilities which he confirms or differentiates by relevant tests. He may decide on the treatment immediately even before receiving confirmation of the diagnosis. He usually does this when being fairly sure of the diagnosis and/or when the condition of patient is considered serious enough as to not delay empirical treatment specially if it is of non-invasive nature.

In this connection the term "Primum non nocere" (first do no harm) axiom is invoked. This is commonly thought of as part of the Hippocratic Oath but it has nothing to do with it. It first appeared in the 18 th century. What Hippocrates actually said was "to help, or at least to do no harm" which is completely different as the former implies that we do not do anything that will cause harm to the patient which is an impossible fete in this day and age when virtually all forms of therapy carry some degree of harm. Of course what is really meant here is that we should not knowingly or intentionally do harm. This is sometimes referred to as "double effect" phenomenon.

In deciding treatment or intervention we traditionally look at the balance between doing good (beneficence) and causing harm (nonmaleficence). The outcome of estimating these opposing effects is called "utility." We estimate utility almost intuitively but to do it properly we should use evidence-based approach. The English philosopher Jeremy Bentham devised a method of calculating utility by giving a number to each benefit and harm expected and then subtracting the total number of "harms" from the total number of benefits." Another method is to calculate the ratio between the number of "harms" to that of "benefits."

Let us take an example

Suppose that a medical intervention A has an estimated benefit of 20 and estimated harm of 10 whereas an alternative intervention B has an estimated benefit of 6 and an estimated harm of 2. The treatment option you will choose will depend on whether you are:

1. The pessimistic type. (or the "non-dealer" in stock market terms) In this case you will follow the "Primum non nocere" doctrine and you will not attempt either intervention as there is possible harm from both options

2. The optimistic type (or the "bull" in stock market terms) In this case you will follow the Benthamian view and go for treatment A which has more benefit than B (20-10 = 10) versus 6-2= 4 for treatment option B

3. The Conservative type (or the "bear" in stock market terms).In this case you will choose the ratio approach and recommend treatment B over A since the benefit/harm ratio is 6/2 =3. versus 20/10 = 2 for treatment option A

Now let us assume that the diagnosis is not straight forward and requires some more investigations before a decision can be made (say a case of pyrexia of unknown origin with weight loss, aches and pains and anaemia).

Here again physicians are broadly of 3 types

1. Throw the book at him type . This type does all possible investigations early and at the same time. They justify this by the hope that the diagnosis will be reached earlier and treatment can then by implemented earlier. This is an expensive method and, if the investigations chosen are invasive, a risk to the patients is involved. This approach is justified if the patient is really sick and the situation is complicated and the list of differential diagnosis is long

2. Step at a time type . This applies to most physicians. In the case mentioned above they will start with septic screen, serology for various microbial antibodies then move on to echocardiography, CT scans and on to bone marrow examination and other invasive procedures

3. Get everybody involved type . In this specialized and defensive medicine age some of us may resort to early involvement of many relevant (and irrelevant) specialists early in the process. In the case described, why not call the haematologist, the oncologist, the infectious diseases people and, for good measure the rheumatologist. Apart from being an expensive approach I think this approach takes away the fun out of medical practice which centres around solving the diagnostic puzzle

As the diagnostic puzzle takes longer to solve, the patient becomes more despondent and a "burnout situation" develops in the physician's mind as he/she runs out of ideas. Here two options may be useful. If the patient is not really sick it may not be a bad idea to let him go home and re-admit, say in one month for a fresh look and repeat of investigations. May be a new physical sign or a repeat investigation will turn out some­thing new. Alternatively you may ask a colleague of yours known for this good clinical acumen and diagnostic abilities to have a fresh look at the patient. Don't be surprised if he turns up with an excellent clue you missed or even discover a physical sign or even a positive investigation in the patient's file that you and your team overlooked.

Talking of discovering a missed investigations lurking somewhere in the file, this always bothered me because it does happen too often for comfort. We as consultants depend entirely on the data imparted to us by the junior staff, when making diagnostic or treatment decisions. The junior staff can be often busy, tired or simply lazy. In the process we miss important and occasionally fatal clues. My advice is that we should ourselves check on history physical signs investigations. If you are lucky, like me, you may have in your own computer in your office the facility to look up all the results of your patients.

Dr. Saadi Taher, the Executive Director of Medical Services at the King Abdulaziz Medical City, Riyadh and a nephrologist himself, always reminds us of the pitfalls of deficient communications, fragmentation and rather primitive methodology of sharing and reporting data on patients during ward rounds. He has been striving to develop methods to overcome this in our institution.

Of course even when we have all the evidence available to us, we may still fall into error in our diagnostic decision-making. This may be related to things like

1. Lack of developing enough and correct hypotheses specially in difficult and complex cases

2. Misinterpreting evidence. This is not an uncommon occurrence Often we forget the patient and treat the results, We often ignore to pay attention to trends in results and concentrate on absolute values

3. Lack of frequent review and regeneration of diagnostic decisions and hypotheses on the basis of newly generated data. This occurs especially if the physician decides from the outset on a specific diagnosis and is inflexible in reviewing his position. Under these circumstances he will take into account any result supporting his hypothesis and belittle or even completely ignore any result opposing his hypothesis.

In short diagnostic decisions making is a cognitive process that builds on intuition experience, evidence, balance of probabilities pattern recognition, recent exposure, prejudices, perception and personality trait. It is likely that experienced physicians use a hypothesis gene­ ration and strategy formulation only with complex cases. In more simple and straight­ forward cases they depend largely on pattern recognition retrieval and intuition.

Correspondence Address:
Abdullah A Al Sayyari
Saudi Center for Organ Transplantation, P.O. Box 27049, Riyadh 11417
Saudi Arabia
Login to access the Email id

Source of Support: None, Conflict of Interest: None

PMID: 17679763

Rights and PermissionsRights and Permissions


    Similar in PUBMED
    Search Pubmed for
    Search in Google Scholar for
    Email Alert *
    Add to My List *
* Registration required (free)  


 Article Access Statistics
    PDF Downloaded355    
    Comments [Add]    

Recommend this journal