For the prospective reader who is interested in potentially picking up this book

Symptom to Diagnosis sets out to do the impossible: summarize all of medical diagnosis in 600 pages.  Even in its ambition, it is remarkably successful at this. If one has at the onset an attitude that all the percentages are to give a general impression of test usefulness rather than need to be memorized, the book imbues a constant sense of statistical thinking alongside the most up to date decision trees that will serve any medical student well. It covers many high-yield clinical concepts alongside vignettes which grew to be more and more memorable as the discussion followed twists and turns you would likely see in the clinic. While following a more realistic process of generation, filtration and then ordering when presenting differentials would have helped the reader practice real world scenarios, I realize that would have extended the length of the book. An excellent read for someone wishing to consolidate their clinical acumen, I give Symptom to Diagnosis a 9.5/10 for being something I wish I had access to earlier in my career (so I bought myself my own copy!).

The below is a stream of consciousness series of commentaries on aspects of the book which may or may not be helpful but which was helpful for my own processing as I went through

My favorite section by far was the chapter on fatigue because so many things cause it and a large proportion, up to 75% of patients per the book, have psychiatric symptoms with fatigue. I would love to specialize in figuring out why people are fatigued one day as a psychiatrist. I think that would be a great way to combine my general medical knowledge and psychiatric knowledge with my interest in lifestyle medicine interventions. The differential is just challenging and takes a lot of wisdom to know what to do next.

I generally really enjoyed the manner in which they used percentages to help process what could be causing particularly challenging clinical scenarios. For example, a patient in a scenario had back pain after moving, and the first sentence in the discussion is that 97% of back pain in a primary care practice is nonspecific mechanical pain, which has a particular testing regimen you want to follow: 4-6 weeks to see if resolution occurs if there are no other concerning symptoms. Another example is how the book had a patient with both neurogenic and vascular claudication, which provided excellent discussion. 

I also appreciated how the book provided case examples where there weren't clear answers and the initial most likely scenario wasn't the correct final diagnosis. For example, a patient with likely psychogenic origin of ill-defined lightheadedness later returns in the case to being unsteady on her feet in more of a disequilibrium diagnosis, the discussion of which was enlightening. The book did fall into the trap, which is somewhat different than real life, where there is always a final diagnosis. In my discussion with Dr. Leeds, someone who has many years of experience in family medicine, diagnosing a large variety of illnesses, sometimes in the clinic no definitive answer can be had. I don't think they modeled that so much, which is only a minor critique. When the book asked to come up with a differential diagnosis, I would try and make one in my mind using bundling. This felt more instinctual, like what would I in my type 1 cognitive process think of as associated with this. I then switched to type 2 thinking moreso by stacking methods on top of each other. After using bundling which tended to be more system 1 and subconscious, I would transition to the VINDICATES acronym and a subsequent mental CT scan technique. Throughout the many vignettes in the book, I was able to come up with more robust differential diagnoses in this manner. It was fun seeing my skills progress as I went through the book, and by the end I was coming up with a larger percentage of the different diagnostic hypotheses they would provide for each case. 

The book continually had helpful tables, like Table 10-8, which compared community-acquired pneumonia, influenza, and acute bronchitis among the symptoms high fever, localized lung findings, shortness of breath, and seizure. This was greatly helpful for my learning. I remember a class in undergrad that I took about Apocalypse literature and one takeaway I had was that it is often the most similar groups that have the most vehement disagreements. The class specifically was referring to the Qumran dead sea community and "mainland" Judaism, but I also think of modern-day Republicans and Democrats, who in the grand scheme of things have a great deal in common with each other ideologically but tend to only focus on their differences in a sort of familial dispute to differentiate themselves from their siblings. I think taking this mindset as a physician when considering similar clusters of diseases can be very beneficial to exaggerate the differences in one's mind and create a better ability to distinguish them. Comparative tables like the Symptom to Diagnosis book provides are helpful in this. When they did have decision trees, they were hit or miss in regard to their helpfulness, like the vertigo section which I thought it was easier to just read descriptions of to get the same info. 

The section on dyspnea was a journey down memory lane for me in the Springfield Emergency Room, where we were evaluating someone with chest pain and difficulty breathing. I did not know what the cause of the breathing issues may be, but the attending I was working with deftly suggested we order a CT angiogram. The difficulty breathing was less pronounced, but the patient ended up having a large pulmonary embolism, without a history of recent immobilization. This story reinforced how being able to recognize clinical patterns, which in hindsight I think were clearer in this patient than they were initially presenting to me in the moment, is important. 

While reading I frankly most benefited when there was a paragraph discussion of the differential process in sentence after sentence. This is, I think, similar to how our minds function, where we consider one thought at a time with particular reasons for comparing and ruling out certain diagnoses.

I like how the book had some flexibility in its decision-making explanations. For example, a patient who had a Wells score of 3 was chosen in the clinical vignette to get a CTA instead of a D-dimer because pulmonary embolism is the most likely diagnosis. This is not necessarily how they presented the optimal decision tree, but I enjoy that they did this because it emphasizes how there is flexibility and clinical judgment is needed (the way the vignette was written there were very few potential alternatives to PE). To what degree is such judgment appropriate? What if the vignette was written the other way and there was an unnecessary test that was ordered? Dr. Leeds and I pondered this thought and I hope as I learn more about differential diagnosis in my career, I can have a better understanding of this gray area. 

One issue I had was with the statement, “Total cholesterol and HDL are minimally affected by eating and can be measured in fasting or nonfasting individuals.” While of course a review book needs to make general statements for brevity and clarity, I am not sure this reflects the most accurate understanding. For example, the American College of Lifestyle Medicine promotes primarily plant-based diets with a high proportion of whole grains, fruits, and vegetables to help reduce heart disease risk. I also somewhat disagreed with the diagnostic algorithm on page 93 due to a contradiction. The book paragraphs state that the acid-fast bacilli smear is not very sensitive for diagnosing HIV-associated TB pneumonia. However, in the algorithm, the smear test is placed before the culture test. This layout suggests that if the smear test is negative, the more sensitive culture test would not be performed, which is inconsistent with the information provided. Few errors like this were seen throughout, however, and otherwise it was a very tight text.

Interesting learning points that I will apply in my practice (more helpful for me personally to reinforce my own learning)

• There's a 25% prevalence of osteoporosis in women aged 65 or older (compared to 5% in men).

• One of the greatest differentiators of abdominal pain is the time history whether it has occurred in the past or is occurring for the first time.

• Diabetic ketoacidosis is an important differential to consider in abdominal pain.

• Orthostatic vitals are important to take in a patient with abdominal pain. 

• It is not the norm for fever to present alongside perforated appendicitis.

• Dark urine in a patient with symptoms suggestive of cholecystitis is very likely call choledocholithiasis since the conjugated bilirubin is coming out through the urine.

• In autoimmune thrombocytopenia, it is not necessary to order a bone marrow examination unless there is atypical features like splenomegaly or lymphadenopathy.

• Ultrasound is not helpful for diagnosing cholelithiasis as it is insensitive.

• For high-risk patients with suspected choledocholithiasis, you can go ahead and do an ERCP which is diagnostic and therapeutic whereas if there is moderate risk from 10 to 50 percent try and do an MRCP or endoscopic ultrasound which can avoid unnecessary ERCPs. On my surgery rotation, this was always a challenging area for me to understand and reading this book helped greatly.

• Hemoconcentration with a hematocrit above 50% can predict severe pancreatitis because of the inflammatory release of water from the vessel lumen.

• It is helpful to consider cardiac history and atrial fibrillation in someone who presents with umbilical-oriented abdominal pain even if you suspect it's a bowel obstruction.

• Someone with bowel obstruction may have orthostatic hypotension that is significant because of dehydration. This wouldn't necessarily be caused by an abdominal aortic aneurysm which is also on your differential of midumbilical abdominal pain.

• Visible peristalsis is a more specific finding in large bowel obstruction than the presence of constipation because there can be significant stool emptying distal to the obstruction during the time course of the abdominal pain.

• A patient with significant abdominal pain and a history of atherosclerosis may be experiencing acute mesenteric non-obstructive ischemia because of a recent medication change, like a beta blocker that is causing hypotension.

• The most common misdiagnosis of someone who has a ruptured abdominal aortic aneurysm is renal colic. This is important to diagnose abdominal aortic aneurysm because around 30% of people die before getting to the hospital and overall mortality for a ruptured AAA is 80%.

• Type 2 diabetes accounts for 12 to 47 percent of diabetic ketoacidosis presentations, which is higher than I thought it was.

• When deciding between Irritable Bowel Syndrome and Inflammatory Bowel Disease, consider that IBS more commonly has irregularly irregular diarrhea, whereas IBD tends to have consistent diarrhea.

• Pregnant patients and those on SGLT2 inhibitors can oftentimes have normal blood glucose in an episode of diabetic ketoacidosis.

• Supplemental potassium is indicated in diabetic ketoacidosis treatment when the potassium falls below 5.3. This is normally a high potassium level!

• Cardiogenic shock is an important differential in the presentation of lactic acidosis.

• Distinguishing acute respiratory acidosis from chronic respiratory acidosis is important when someone is presenting with respiratory distress and a pH that is low. Elevated bicarbonate levels indicate chronicity.

• It is important to measure the anion gap in a patient who has obvious respiratory acidosis from something like COPD in order to make sure they are not septic as well.

• Some patients with short bowel syndrome can have colonic bacteria metabolize carbohydrates into D-lactic acid which is not the physiological L-lactic acid. You also need special testing in order to evaluate for this because normal lactate labs do not give the dextro-rotatory configuration.

• I thought the figures regarding contraction alkalosis were exceptionally helpful to my understanding of the processes involved.

• Hypothyroidism can cause a non-megaloblastic macrocytic anemia.

• Monitor viral load and CD4 counts in those with HIV every three months.

• Give two ART medications for prophylaxis and three simultaneously for treatment of HIV.

• In a patient with AIDS, you want to give Prevnar 13 and then Pneumovax 23 two months afterwards, in that order 👆.

• The differential around the abnormalities in lumbosacral radiculopathies was incredibly helpful to me as this differential diagnosis was one of the most challenging throughout my medical school journey.

• A previous history of breast cancer has a very high pos likelihood ratio of 14.7 for vertebral metastasis.

• With vascular claudication, there's a fixed walking distance before the onset of symptoms, whereas in neurogenic claudication, there's a variable walking distance before symptoms start to show.

• Wide-based gait has a high positive likelihood ratio for spinal stenosis.

• Romiplostim and eltrombopag are treatments for refractory immune thrombocytopenia.

• Known cirrhosis can cause hypersplenism, which can be a cause of thrombocytopenia.

• A clinical score is indicated for suspected heparin-induced thrombocytopenia because the specificity of antiplatelet factor 4 antibodies is low and otherwise you may over-diagnose it.

• Response to nitroglycerin should not be used as a diagnostic test to evaluate esophageal vs. cardiac origin of chest pain.

• Radiation to both arms has a very high likelihood ratio for myocardial infarction.

• You can induce a vasospastic angina with IV administration of Ergonavine.

• You can pretty much rule out influenza in a patient presenting with fever and cough if they are not presenting in December through May.

• Routine testing with culture is optional in community-acquired pneumonia.

• Amantadine can increase dopamine in cough reflexes which can prevent aspiration pneumonia with a low number needed to treat.

• Fever that increases over several days is not indicative of influenza and more indicative of pneumonia.

• The sensitivity of influenza testing is somewhat low at 60%, but more specific.

• When someone has a productive cough it is more likely other causes than pertussis.

• It was helpful to go over the different causes of delirium, but I must say I continually came back to the acronym I WATCH DEATH which sticks with me for some reason 💀.

• Only a small percentage of patients with delirium end up fully recovering which is shocking and something I want to take with me when I'm in the hospital.

• The Confusion Assessment Method is one of the best ways to diagnose delirium.

• Those with alcoholic hallucinosis tend to have strong performances on Confusion Assessment Method tests in comparison to those with true delirium.

• Memory loss reported by a family member is more likely Alzheimer's whereas reported by the patient is more likely depression.

• Depression is prevalent in around 50% of patients with Alzheimer's disease.

• Hemoglobin A1c can be seen in iron deficiency.

• 50% of hemoglobin A1c is determined in the past 30 days, even though the test is for the previous 90 days.

• Patients with HHS have about a 4 to 5 times increased risk of death compared to patients with DKA.

• I thought it was helpful to break down diarrhea into infectious gastroenteritis, infectious inflammatory, and non-infectious.

• Antibiotic-associated diarrhea is only caused by C. diff 15-20% of the time.

• Tests should be done to rule out syphilis when considering a diagnosis of Meniere disease.

• 30-50% of patients with multiple sclerosis have episodes of vertigo.

• Lhermitte sign is an electric-like sensation when the neck is flexed, while Uthoff phenomenon is when in warm environments there is a worsening of MS symptoms likely due to decreased nerve conduction in heat.

• Visual changes like diplopia or field defects are present in 69% of cases and is the most common symptom of vertebro-basilar insufficiency, even more than vertigo 🥴!

• Multiple sensory deficits was a unique differential I had not considered for dizziness, and it's something that is easily correctable by getting the correct pair of prescription glasses, improving lighting, etc.

• I feel as though on my internal medicine rotation we put a greater emphasis on jugular venous distention in the ascertainment of heart failure, but in reality an S3 gallop has a higher specificity and greater positive likelihood ratio.

• Those with moderate to severe heart failure have a potential 50% risk of death per year.

• Cirrhosis and thyrotoxicosis can cause wide pulse pressures, so such a finding on physical exam is not specific for aortic regurgitation per se.

• The most common thrombophilia is Factor V Leiden, which is caused by a mutation in Factor V that prevents cleavage by activated protein C. I have a hard time remembering this mechanism for some reason.

• Patients with COPD should be screened for heart failure and pulmonary embolism because both of those diagnoses are often missed on acute presentations of COPD exacerbation.

• A better electrocardiogram finding of pulmonary embolism than S1Q3T3 is actually T-wave inversions in V1 through V4 with an associated T-wave inversion in lead III, which has 99% specificity and 88% sensitivity. 🤯 I guess this is a case where a classic boards factoid is not necessarily the most helpful in the real world.

• Reviewing the complex sensitivity and specificity workflows around the diagnosis of pulmonary embolism and DVT and their associated treatments was extremely helpful, I felt like, in Chapter 15.

• It was surprising to me that different types of lung cancer like small cell and squamous cell carcinoma are relatively equal in prevalence on biopsy and that, when thinking of a differential, you should keep this in mind.

• A loud S2 heart sound does not have a very high positive likelihood ratio for pulmonary hypertension.

• NSAIDs can cause drowsiness. 

• Edema, when present in a patient with fatigue, signifies potential significant kidney or liver disease.