You are about to have the honor and great pleasure of working with a group of patients I have come to know and respect over the years. While I cannot tell you how to practice medicine, I feel no reluctance to tell you what made it so worthwhile for me.  

First, my style is my own. While I have had teachers and mentors (and patients) from whom I learned more than I can describe, I discovered early that one cannot simply copy another clinician; one has to observe, study, test, adapt, internalize, in order to ultimately have a style of one’s own. This has been a four decade process for me, and the evolution shows no sign of abating. We can always learn, adapt, and improve. 


My True North is the patient. Do not let the simplicity of this deceive you; simple is not the same as obvious, comfortable or easy. My commitment to work for the patient is constantly being challenged by a host of outside pressures: time, revenue imperatives, algorithms, Meaningful Use (Meaningless Abuse) metrics, guidelines, insurance limitations, the insanity we call the electronic health record, and employers who will try to convince you you work for them, not for the patient. 

(I work for the patient and have a business arrangement with my employer. There’s a difference. My patients get it when I talk to them.  My employer did not.) 

I don’t have a recipe or a list of rules for how I do what I do. I have a vast armamentarium of tips and tricks, but sharing them would be of little value unless they are woven together with the threads of the principles from which they spring. So, I will talk about the six basic principles. 


Know the patient in the patient’s context.

Health care does not happen in the office or hospital.  It happens wherever the patient is. A little back of the envelope math shows that even a chronically ill patient is unlikely to spend more that 0.01% of their time in our medical settings. They live in their universe, not in ours. If we want to help them, we need to know them in their universe, learn to think and speak in their language, and help them design an approach to health that meets their needs, fits with their values, and works with their resources in their universe. 

This goes far deeper than simply not identifying patients as their disease: the gall bladder in bed 2, or the woman with mitochondrial myopathy. We need to know who they are, what they do, what the feel, what they fear. We need to understand their strengths and weaknesses. This cannot be done in one visit. Depending on the checkboxes and dropdowns provided by the EHR for a social history is as delusional as spitting in the ocean to make the tide come in. Ask lots of questions. Simple questions. Open-ended questions. Allow plenty of time for them to answer. Then ask for more. After they tell you where they work, ask them what their job is.  Then ask them to describe it. Then ask them how they feel about it. Develop a collection of questions that encourage them to describe things from their perspective: How is that for you? How does that make you feel? What would you like to change? Learn to be comfortable waiting while they answer. AFTER they answer, continue to wait. (Almost) never interrupt an answer - it sends the message you didn’t really want them to tell you. 


Collaborate as equals. (Not really!)

Two people simply cannot collaborate as true equals when only one is at risk for the benefits or harms from their actions. Start by understanding that clinicians do not give patients the right to collaborate or elevate their status to that of an equal. It’s the other way around. Patients give us the opportunity to collaborate and elevate our status from spectators to participants in their universe. 

You will know more than many of your patients about many common. self-limited, and episodic problems you see: otitis media, ingrown toenail, overuse tendinitis, sinusitis. Don’t let this go to your head. The rarer the condition and the longer (or more often) the patient has had it, the greater the chance that the patient knows more than you do.  Sometimes MUCH more. This is most easily seen in the rare condition (Klinefelter’s, empty sella syndrome, Budd-Chiari syndrome) where the patient has only one subject to research, joins an internet support group, and keeps up with cutting-edge research.  Patient-based medical expertise is also common in diseases like migraine, diabetes or restless leg syndrome: patients may be very strongly motivated to understand and manage their condition and become quite expert.  This, in fact, is what you want. There is data to show that many diabetics know more about managing sugars and the wide range of medications than their primary care clinicians. 

The patient almost always knows more about the patient, and the patient’s context, than you possibly can. Listening to the responses to open ended questions can, over years, give you considerable insight into the intra- and interpersonal life of a patient - but you will never see and experience the world as they do. We can’t do that with our spouse or children, so it is silly to think we can ‘know' the patient.  

You can provide context, experience, guidance, and lots and lots of medical information (but don’t discount theirs). 

Collaboration, then, means that you and the patient bring two different but overlapping piles of information to the discussion, you work together to understand both the medical and patient-centric nature of the problem, you establish a goal or goals, and a plan to achieve the goal or goals, after which you facilitate and provide coaching and feedback.  

Don’t pay obeisance to the algorithmic overlords. Remember that our medical science is incomplete, provisional, and our evidence at best applies to large populations rather than individuals. Help the patient understand and use the evidence to inform decisions, but do not commit the ethical breach of sacrificing patient autonomy by letting the evidence make the decision for the patient.  

Pretty much everything in medicine is preference sensitive. Our job is not to tell patient what to do, but to tell them both what the evidence suggests, and how good the evidence is. 

Some examples of things to discuss with patients, instead of simply telling them what to do, include the following: 

(Note: due to the constant growth in information, these numbers and approaches make good sense to me the week that I am writing them.  I would not be surprised if things have changed in a month, and I would be surprised if things had not changed a lot in 2 years. Lesson: whatever you learned the last time you reviewed a subject may not be valid anymore. Don’t rest on your laurels - they are not comfortable to sit on.) 

  • Strep throat. Strep resolves on its own in 4-6 days. At least half of what we diagnose as strep throat is actually viral, with the strep present but not causing disease. We treat strep to shorten what is usually a 4-5 day illness by less than one day, to reduce contagion, and to prevent truly rare complications. There is no evidence that treatment prevents local complications like abscess, and the evidence says we have to treat millions of children to prevent one case of rheumatic fever. Complications from antibiotics are MUCH more common than benefits. We do not need to treat strep throat with antibiotics.  and
  • Otitis media. The diagnosis of acute bacterial otitis media based on a red ear is very flimsy; a solid diagnosis requires an appropriate history supported by a red, opaque drum that does not move with pneumatic otoscopy. Pneumatic otoscopy is easy, uses inexpensive equipment, and almost no PCPs do it.  At least 3 of 4 healthy children with AOM will resolve without antibiotics. It is reasonable to give a written prescription and wait before filling.
  • Hypertension. The guidelines for treatment are based on evidence from studies that often exclude patients like the one in the room with you. The goal is not a prettier number. The goal is to prevent complications like heart attack, stroke, or kidney failure. The decision whether or not to treat should be collaborative and based to the extent possible on the evidence as pertains to a patient of that age, race and co-morbidities. Never forget the law of diminishing returns. Remember that the numeric targets in guidelines are arbitrary lines in the sand: going from 142 to 138 in Mr. Jones helps meet guidelines and protects your bonus, but provides no meaningful benefit, while going from 168 to 142 in Mr. Smith fails to meet guidelines or protect your bonus, but may provide significant benefit to the patient.
  • Hyperlipidemia. Ignore numeric targets. Use global risk instead. Remember that the harms of statins occur in ~ 10% of all patients, regardless of baseline risk, but that the potential benefit in a given patient is determined by their baseline risk. Remember that there is NO evidence that statins as primary prevention save lives - though they do prevent cardiac events in a very small number of patients. Satins increase the risk of diabetes. (Don’t be seduced by relative risk.) Remember that the Mediterranean diet coupled with regular exercise provides at least as much protection as statins for the great majority of patients. (Link)
  • Diabetes. The guidelines were developed including data that is old enough to predate much of our current treatment, and is population based rather than individual patient based. There is little data that most subsets of patients benefit from pushing the A1c below the 8.5 - 9.0 range. There is no evidence that a normotensive diabetic with well controlled diabetes benefits (in terms of preventing real disease) from an ACE. Similarly, the evidence that a normotensive, non-smoking well-controlled diabetic with normal or mildly abnormal lipids benefits from a statin is, at best, shaky. Most of our current diabetic armamentarium improves proxy measure like A1c but has not been shown to improve outcomes. Exercise, metformin and insulin clearly improve outcomes. 
  • Pneumovax. PPSV23 is directed at strains that have decreased dramatically in prevalence. When the prevalence was still 80% of pneumococcal disease, the NNT to prevent invasive pneumococcal disease was several thousand.  The current prevalence is ~ 60% and then NNT has not been restudied but is clearly much higher. The PCV13 has not been around long enough to have actual outcome data, but appears to be directed at a more pertinent group of strains. As best we can estimate, it has an NNT of 500-1500.
  • Influenza vaccine. I’m a fan of influenza vaccine. The fewer people in a community who are susceptible, the fewer cases will occur, and the fewer truly at risk patients (immune deficient, cystic fibrosis, chemotherapy) will be exposed. That said, it is a mistake to tell a patient that they will probably benefit. They may benefit, but probably not. What do I mean. The NNT for flu vaccine in a good year is 37: vaccinate 37 people only one of whom benefits. Her chance of benefitting is 1/37 or 3%.  In a bad year, the NNT is 77, so if you vaccinate 77 people, 1 will benefit, or 1.2%.  Not what most of us would call a probable benefit.

Go look at a table of NNT values, perhaps here:  Humility should set in. Most of our ‘no-brainer’ treatments benefit only a small fraction of recipients or only to a small degree.  

There are immense gaps and discrepancies between what our guidelines and metrics incentivize and what the evidence actually says throughout medicine. We need to be self-aware and recognize how marginal many of our treatments are, how great the risk of harm may be, and not pretend with patients that we have powerful, reliable, safe tests and treatments that do not require discussion or thought. 

Do not tell patients they should do something. (A patient of mine calls this ‘shoulding on patients’.) Instead, say things like:  "I think the benefits outweigh the risks” or “I think this is worth doing." 


Use the chart as a tool.

The chart is not your personal diary, a journal of your your perceptions and recollections of the visit. Despite what your employer, the billing office, or third party auditors may say, the chart is a clinical tool, not primarily a billing or regulatory tool. Use it as such. Properly used, it is a collaborative tool that you and the patient use to collect, assess, and use information to make collaborative medical decisions. This means several things. 

  • Nothing should go in the chart that the patient does not know about.  How can the the patient engage in self care, make decisions, or collaborate if they don’t see what you see?
  • As much as possible, do your documentation out loud and in real time, in the presence of the patient. This helps ensure that you and the patient are quite literally on the same page. It gives the patient an opportunity to modify, correct or expand on the information when they hear how it is being recorded. It demonstrates that you are listening and paying attention and accurately recording the information, and attests to the fact that you consider it valuable. And it keeps you honest: you cannot shade what they say to fit better with your hidden perspective. It increases the likelihood that you have those difficult but essential conversations with your patients about differences in perspective; this is especially important in settings like chronic pain, behavioral or mental health issues, self-destructive (from the clinician’s perspective) behaviors, somatic focus, problems of living. 
  • You, your patient, your colleagues and your staff can all use the chart note to know what happened, the next day or the next year.
  • The patient can take the note home and use it to remember, review, and stay on task with the collaboratively created plan.


Be yourself and don’t be afraid to share.

When I started in medicine, I struggled to project the image of a competent physician. It took a long time to understand and resolve my imposter syndrome. ( Over time, I have become more at ease with myself and more comfortable being real during patient encounters. It turns out, patients want a clinician who is human, who feels and cares. They do not want a machine or an internet interface. There is nothing wrong with showing your joy or sadness, your uncertainty, your concern - AS LONG AS IT IS ABOUT THE PATIENT. It is not a competition, where you need to match or surpassed your patient’s grief or stress.  

I have found that sharing things from my life, past and present, helps the patient know that I am human and unlikely to think less of them for their struggles. But this kind of sharing requires careful self-awareness. When I tell of events or feelings from my life, it is not about how I succeeded in overcoming a problem like theirs, it is about why I understand what they are experiencing.  To the extent that I share my solution to a problem (which I only rarely do), it is framed as ‘I found a solution that was unique for me and my setting. We need to talk about what will work for you.' 

I don’t mind that they know I am married, have children and grandchildren and ask about my family. I don’t mind that they know I like to hike and play the guitar, and ask about it or share their experiences. This is a way for us to increase commonality and trust, and is valuable - as long as it doesn’t take up so much time that other issues are shortchanged.  

I am comfortable with touch and accept hugs. 

The bottom line: the most commonly prescribed drug in your armamentarium is YOU. As Balint ( “The doctor is a drug." 


Get comfortable with limits.

Know your personal  limits and the limits of medical science. Be comfortable admitting them and know how to use them. It is much easier to be certain than to be right. And being wrong feels exactly the same as being right. (insert link  Kathryn Schulz on being wrong 

Three of the most powerful words in medicine are: "I don’t know." Especially if they are followed by: “So we need to find out." 

Don’t overestimate the accuracy of our diagnostic tools or the power of our treatments. Most of our best tests are wrong (false positives plus false negatives) a quarter of the time.  

Don’t be seduced by the false security of medical science.  All knowledge is provisional, or as one of my favorite patients says: The Truth keeps changing.  


Love your job. All of it.

If you don’t consider what happens in the exam room magic, if you don’t thrill to a complex problem you have never seen before, if you cannot find satisfaction in traveling a tough road with a patient and his incurable progressive disease, if you do not love the fact that you have no idea what will present itself in the next room, if you do not have the curiosity and courage to look under rocks and behind closed doors - then primary care is probably not where you belong. 

However, if you love this, if you get up every morning thankful for the chance to be invited into their lives by your patients, you are blessed. Your patients will thrive. As will you. 


Links to more on this topic: