Tuesday, 30 June 2015

Diagnosis, uncertainty and the hub of clinical decision making

Is diagnosis an authoritative statement about a patient’s condition? Does it involve 100% certainty that this is the correct label for the type of the patient’s problems, what the underlying causes are for the problems, that this is the sole problem, what the correct treatment should be, what the outcome is going to be with and without treatment?


In fact a doctor’s diagnosis is a summary of a series of possibilities that trigger further enquiries, not least is the diagnosis correct and should it be changed. These possibilities inform the doctor’s management plan such as what types of further tests and examinations the healthcare team should do and when.  What are these possibilities?

  • What kind of features does this diagnosis present with? The range of probabilities of how the diagnosis will present in terms of symptoms, signs, lab tests or other important clinical features. This will tell the doctor how to recognise the condition, what information to seek and how. It will guide the doctor in what questions to ask, how to interpret the answers, what tests to run and what pattern of results to look for. As noted elsewhere this is a pattern recognition exercise.
  • What if the diagnosis is wrong? The possibility that the diagnosis is correct (or its’ reliability) according to some external criteria – usually for most medical specialties, a test (such as biochemical test, pathology of diseased organ, imaging test) but could also be another expert doctor’s opinion. This possibility of the diagnosis made being incorrect means the doctor should continue to be vigilant for disconfirmatory information i.e. information that suggests the diagnosis is incorrect. This may be symptoms or signs or laboratory tests or not responding to a treatment that usually works or other differences in the clinical picture that are usually incompatible with the diagnosis or better explained by another diagnosis. If these arise then the doctor should revise (change) the diagnosis to one that fits the information better. The doctor keeps a differential diagnosis (or many differential diagnoses, or even the “null option” of no diagnosis/ no illness) in mind. For example, a patient with initial diagnosis of alcohol intoxication but then the doctor notices they have signs of a subarachnoid haemorrhage and changes the diagnosis to that instead.
  • What other illnesses could be present? The possibility that other healthcare problems are also present, known as co morbidity. Like birds of a feather, some illnesses are associated with other illnesses being present. An illness may increase the chances of developing another illness . Illnesses have similar underlying causes thus increasingly the probability of more than one illness. This alerts the doctor to look for signs or symptoms or run tests or investigations for these other potential diagnoses. For example if a patient has an unusual infection diagnosed such as pneumocystis carinii pneumonia this makes the doctor think that the patient may have an immunocompromised state (immune system that is very poor at fighting infections) so will run tests looking for causes of this such as HIV infection as well as taking precautions to reduce the chances of other infections.
  • What complications should I look out for? Apart from what the doctor has recognized and identified there can be other problems or features known to be associated with the diagnosis. This triggers the doctor to look for these other potential problems through asking the patient for symptoms indicating the problem is present, examining the patient to look for these problems and running further tests to identify these problems. Even if the problems are not present at the initial assessment, they can develop later so armed with this knowledge the doctor can be alert for them developing through history taking, clinical examination and running tests in the future. For example if a patient is seen with a heart attack (myocardial infarction) there are a variety of complications the doctor will be aware of and monitor for such as heart failure or cardiac arrest.
  • What treatment should I offer? The range of possibilities of treatment and the range of responses to treatment – what are the different alternatives of treatment (not just medication)/ help available, how likely the diagnosed condition will respond to treatment, how long this improvement will take, what are the signs of this improvement taking place and what side effects may occur with treatment and how to identify these. If the treatment fails and the diagnosis is correct what are the likely causes for treatment failure (e.g. maybe not taking the medication or not changing lifestyle factors) and how to identify them. Further treatment option knowledge is also triggered at this point. Notice how, as in the above examples, diagnosis links into other types of knowledge, in this case pharmacology and treatment guidelines. The doctor can discuss with the patient these important details and ask them what treatment choices they want to make. The doctor is reducing the information asymmetry deficit. For example if a patient presents with high blood pressure the doctor can discuss what the different options are e.g. lifestyle changes, medication and when to review them to see if succeeding or what changes in treatment are needed.
  • What do I expect to happen in people with this diagnosis? The range of possibilities of outcomes seen with this diagnosis. Most illnesses don’t have a fixed outcome but have a range of probable outcomes. This allows the doctor to discuss these details with the patient. The range of factors that are associated with better and worse outcomes are also linked with the doctor’s knowledge about the diagnosis and allow the doctor to look for the presence of them and modify them if possible. For example, if a case of cancer is seen the doctor will try and identify the size of the tumour, whether it has spread (metastasised) including to local tissues or to lymph nodes or to other areas of the body. He will also look for other factors that may affect outcome such as socioeconomic class, other factors that can be modified e.g. if smoking worsens the outcome of the cancer that can be targeted with smoking cessation. For some tumours genetic testing of the tumour can take place to identify targeted treatments.

It can be seen that diagnosis is provisional and is always open to review as more information comes to light. It is the hub of linked information that the doctor has learned about the illness that guides the doctors management plan: what tests are run, what signs or symptoms are observed for, how to recognize if the diagnosis is wrong, what other illnesses may also be present, what treatment options are discussed and chosen, what to expect in terms of outcomes and what further information is sought to refine the outcome and improve it. Doctors learn about medicine by learning about a diagnosis and the interlinked information as described above. 

Diagnosis can be thought of as a quantum reflexive hyperlinked seed. Quantum reflects the uncertainties as to whether the diagnosis is right, outcome, treatment and so on. Reflexive in that the information associated with the diagnosis includes procedures to monitor whether it is accurate or, like the phoenix, be immolated to be replaced by a new diagnosis. Hyperlinked in that the diagnosis is a hub that links to multiple domains of information as described above. The seed describes the compact nature of a simple term like the diagnosis yet jam-packed with the information and procedures needed to grow a management plan to help the patient with the right environment of a doctor with the ability to respond and act on the information from the patient.