How Much of the "Differential Diagnosis" Should Practitioners Share with Patients?

Should doctors share the “differential diagnosis” (DDX) with patients?

The diagnosis that a doctor or practitioner thinks the patient has is called the “working diagnosis.” The list of other possibilities that might also be possible is called the “differential diagnosis” (DDx) list.

I’m often asked by trainees whether clinicians should share the list of differential diagnoses and how much information to share.

The answer is that it really depends on the patient and all of her or his past experiences in life and experiences with the hair loss and all the likely experiences that lie ahead for the patient. The decision to share depends on the seriousness of the issues on the list and likelihood of those issues actually being present.

A Broad DDx Is Important to at Least Consider at First

First, let me say that a broad thinking process is essential for clinicians to have. One needs to be open to consider an array of possibilities at any time and then narrow those possibilities down and cross off the list those issues that are not likely. Of course, the speed and confidence that one does this “narrowing down“ depends on the level of expertise of the clinician. The inability to quickly narrow down the differential diagnosis leads to time consuming patient visits. One must acquire skills to quickly narrow down the DDx. Nevertheless, a broad differential diagnosis is important for the clinician to consider and whether or not this is to be discussed with patients should not influence or hamper the clinician’s overall thinking process.

Ad we will see below, there are situations where a discussion  of the differential diagnosis is appropriate and some where such a discussion woulld not be considered appropriate.

1) PATIENT ACCEPTANCE OF FURTHER TESTING: Is the patient OK to proceed with other tests?

Sometimes the decision to go into details with the patient about the DDx depends on the patient’s acceptance to undergo additional testing with a more limited explanation of the differential diagnosis.

Example 1: Patient consents to further testing

Consider a 43 year old female with extreme health anxiety who is quite overwhelmed by today’s visit. We have discussed doing a biopsy and discussed getting some blood tests and chatting on the phone in 3 weeks when everything comes back. A full review of the differential diagnosis will NOT change our current plan in any way at all .. and may only serve add to her stress. We will speak broadly about possible diagnostic categories but not dwell on a given possibility today. The patient leaves the office without knowing my full differential diagnosis. The full differential is recorded in the patient’s notes - either in broad categories or specific categories 

Example 2: Patient does not consents to further testing without more detailed understanding

Consider the 56 year old with unusual patterns of hair shedding. Labs are abnormal. I fear there is some serious issue going on like an endocrine issue or possibly cancer. In a situation whereby the patient will not agree to more testing without a comprehensive understanding of the differential diagnosis or perhaps I want to start the ball rolling with the patient and his or her doctors thinking about some deeper investigations - I may share my thoughts on the broader differential diagnosis in the event we can get necessary dialogue going right away. In general, if sharing the differential diagnosis has the potential to significantly impact the patient’s ultimate outcome in some way I may be more inclined to do so. If it does not, I may have less  of a discussion.

Example 3: Patient consents to further testing

Consider a 57 year old patient with a prior diagnosis of breast cancer at age 42 who presents with a single slightly reddened alopecia areata-like patch of hair loss on the scalp. You wonder about breast cancer metastases as the cause as you see many dilated blood vessels in the patch and no clear features of classic alopecia areata. 

You advise the patient that this could be many things and a biopsy is needed. The patient agrees. At this point there is no need to necessarily discuss the differential diagnosis. You should always have a good history in the chart including all details about the patient’s health. Here, I would ask about other bumps on the scalp, other scalp issues, fatigue, weight loss, bone pain, abdominal pain, lab abnormalities in the last few years (especially liver enzymes and blood counts). I would ask about type of breast cancer, how it was treated and whether follow up with cancer doctors has taken place.

If the patient asks if I am worried about breast cancer, I would most certainly advise at this point that it could be but I would be sure to say there are other possibilities too. The discussion of the differential diagnosis is never about withholding information. It’s always about delivering information at the appropriate time.

Example 4: Patient does not consents to further testing

Now suppose we have a situation where the patient does not want a biopsy. She states that she has been through a lot in the past years and she does not want another procedure that reminds her of past trauma. You come to understand that this is a source of significant potential trauma. Now is a situation where it is important for the patient to understand that breast cancer metastases are on the list of the differential diagnosis. It’s always up to the patient to consent to having the biopsy procedure done, but this information must be conveyed. Should the patient not wish to procedure, I would ask permission to discuss the differential diagnosis with the primary care physician or oncologist. A full evaluation is needed.

2) How likely are the entities in the Differential Diagnosis List?

Sometimes the decision to expand on the list of possibilities in the differential diagnosis with the patient depends on the overall likelihood of a given diagnosis.

Example 5: Long List of Possibilities with None Very Likely

Consider as another example a 32 year old female patient with hair loss who has irregular periods.  Periods were regular in the past… so this is new. You know that women with irregular periods need a variety of tests to exclude several hormonal issues including Cushing syndrome, polycystic ovarian syndrome, prolactinomas and early transition to menopause and cancers.

Should you explain all the conditions listed in the “differential diagnosis list”? At this point, no. These are screening tests. It’s appropriate to let the patient know “there are a series of standard tests we order to better evaluate potential causes of irregular periods.” Of course, should the patient wish a detailed explanation of all causes that’s different. 

Example 6: Long List of Possibilities with One or More Being Likely

Now suppose the patient with hair loss tells you she has irregular periods, headaches and milky discharge from the nipples. In this case, I would convey that I am concerned about a prolactinoma (a benign brain tumor) and that a variety of tests are needed to evaluate this. It is imperative that the patient appreciate the potential likelihood and seriousness of these issues.

Example 7: Long List of Possibilities with None Very Likely

As yet one more example, consider a patient comes to clinic with a friend or partner or family member. The differential diagnosis includes a sexually transmitted disease such as syphilis or HIV but the likelihood is very low. I may order tests and simply let the patient know these are our standard tests we always order in this situation.

Example 8: Long List of Possibilities with One or More Being Very Likely

Now consider a patient comes to clinic with a friend or partner or family member. The differential diagnosis includes a sexually transmitted disease such as syphilis or HIV but the likelihood is now reasonably high. I may ask the accompanying member to step out of the room or take the patient to another room to examine or ask more sensitive questions. In this case, a more broad discussion of the differential diagnosis with the patient is important.

3) POTENTIAL SERIOUSNESS OF THE DIAGNOSIS: How serious and likely are the issues being considered?

As we could see in the example above, sometimes the decisions about potential likelihood are closely related to potential seriousness.

Example 8: Potentially Serious Diagnosis but Very Low Likelihood

Let’s consider as another example a 48 year old woman with new onset androgenetic alopecia. It’s been quite rapid in onset. She has trouble with oily skin and some hair on the face. Her testosterone comes back 1.5 times the upper limit of normal.

In this case, I might consider an androgen secreting ovarian tumor in my differential diagnosis. I’ll ask about weight loss, bloating, fullness after eating, abdominal pain, back pain, pelvic pain, family history of cancer but all this is part of general questioning. In this case, all this questioning came back negative.

Doctors learn in their training that androgen secreting ovarian tumors are more likely when testosterone levels are three times above normal. Here the testosterone levels are only 1.5 times above normal which makes such a diagnosis unlikely. Here, I might advise the patient to have these labs tested again before seeing me in 6-12 months and I might advise the patient to review these labs with the primary care physician as well so that a full physical examination can be done. From my perspective, there is nothing to really suggest an ovarian tumor. I may have it on my differential diagnosis and keep it on my radar but there is nothing to suggest this is the cause. If I had prior labs that showed this has always been the result, I would probably remove it from the list completely. Perhaps this patient has idiopathic hyperandrogenism. In short, we need to follow this lab and base our next investigations on the other features of the history or examination (which the primary care doctor will know best about his or her patient).

Example 9: Potentially Serious Diagnosis and Change of Diagnosis Becomes More and More Likely

Now suppose the patient with hair loss mentioned she has abdominal pain, bloating and back pain. Her mother died of ovarian cancer. Her testosterone level was actually noted to be low in the past ,,,,and now it’s 1.5 times above normal. It’s still below the “three fold” cut off that practitioners learn about in textbooks. However, in this case I might express concern about this as something that needs further evaluation. I am not going to present this information to the patient as a strong likelihood and I am going to advise the patient that I am not the best suited to further evaluate this. However, I am going to let the patient know that this is indeed something that is on the list of possibilities (the differential diagnosis) and this should be discussed further with another doctor. I will let the patient know we need to explore this possibility further with doctors best suited to explore this. Perhaps the primary care will order an ultrasound after a good history and examination is performed.

Thinking Out Loud is Rarely Appropriate

A broad differential diagnosis is important for the clinician to have and whether or not this is to be discussed with patients should not influence the clinician’s overall thinking process. It is always appropriate for a clinician to consider many many conditions when evaluating a patient. However, it is rarely appropriate to “think out loud.”

Not a day goes by that I don’t say to myself “could this be a manifestation of some sort of cancer?” could this patient have another serious autoimmune disease?” “could this patient have some type of brain tumor giving hormonal issues?”

Thinking it is necessary, yes, yes.

Talking out loud with a running narrative is not always appropriate, no, no.

For many patients, I may consider dozens of possible diagnoses. I may order blood tests, or other diagnostic tests to further explore the possibilities. For some patients, it is appropriate to state “our standard work up for patients with this condition would be this particular set of tests” without going into exhaustive details.

4) ENHANCING COMMUNICATION

Sometimes, the discussion of a more broad differential diagnosis is needed as a communication technique to enhance and strengthen the therapeutic alliance.

Example 10: Sharing the DDx to Open up Needed Lines of Communication

Suppose the patient is a 32 year old patient who only wants oral minoxidil for what they feel is androgentic alopecia. They have come in to the clinic today specifically to get this prescription. That is the tone of the patient from the start.  The hope you give 5 mg as their friend got 2.5 mg and it didn’t work well. The patient wants 5 mg!! You note patient has itching and burning and a few pustules on the scalp. They have tufting of a few hairs and some perifollicular scaling. 

I may say to the patient “this could be androgentic alopecia as you suspect but have you considered that this could be lichen planopilaris or folliculitis decalvans. Have you considered these possible diagnoses? I don’t think it’s discoid lupus or early dissecting cellulitis but of course a biopsy will go along way to sort things out. We need to consider a number of issues here and unfortunately oral minoxidil might not be your best bet if any of these were the diagnoses. I am worried about a few issues actually so let’s first turn our attention to getting the right diagnosis before we focus our attention on the right treatment.”

Of course, I fully expect the patient to look at me blankly when I use the terms “lichen planopilaris” or “folliculitis decalvans.” I do not use this technique often as it can come off as quite arrogant and condescending. That is not the intention. But this can be very effective to help center the patient. The hope is that the patient joins me and together we hit the “pause” button for a moment before considering oral minoxidil! I hope the patient realizes that there may be more diagnoses out there that he has never considered. Sharing the differential diagnosis helps to center the patient.

Conclusion

The art of medicine is knowing when and how much of the information in the “differential diagnosis” to share. No, it is not always appropriate for a clinician to share his or her entire thinking process leading to the differential diagnosis. Sometimes it would even be considered malpractice. Good clinical judgment is what is needed. These 10 examples illustrate scenarios where the DDx is shared and where it might now be shared.