Discharge Instructions

It’s 1AM, you’re a third of the way through your night shift. You finish at 7:30 but that seems a long way away. You’ve just seen a 25 year old male patient with some vague abdominal pain, nothing specific, and a normal examination. You don’t think it’s an appendicitis and you’re getting ready to discharge him.

At most hospitals you now have the option of creating a discharge information package for this fellow. It’s a document written by the hospitals medical staff and lawyers, usually many pages long, detailing every imaginable thing that could happen or go wrong. Usually it prints from the computer system and you fill in the blanks. It’s written in plain language, easy to understand as long as the patient has at least a pre-med degree.

Pre-printed discharge instructions are mostly useless. Most emergency departments place a waste container near the exit to handle these instructions as patients leave the building. Patients with wood-stoves or fireplaces may keep them to use as kindling, but they serve little other purpose. They are too long, too complicated, and too impersonal.

When you are discharging a patient, especially a patient with an uncertain diagnosis, do it yourself. Take a moment and sit at the bedside. Tell the patient that it is still uncertain what exactly is causing their problem. Tell them that you don’t think it is anything serious or life threatening or you would be keeping them in the hospital.

When you explanation is complete tell the patient; ‘There are a few things I want you to do.’ Then take out your prescription pad and write down a maximum of three discharge instructions. Three, that’s the number that most patients are going to be able to handle. If you need 15 instructions to get the patient home, the patient should probably not be going home, so you only get three.

Instruction number 1 is about time and place specific follow-up. ‘See your GP tomorrow morning.’ ‘See the orthopaedic surgeon next Tuesday.’ ‘See me, in the ER in 6 hours for a recheck.’ The person is specified, and the time is specified with no room for ambiguity.

Instruction number 2 gives situation specific cues for the patient to return to the emergency room immediately. Usually these are simple, if your pain is worse, if you have any weakness, if you are vomiting. This instruction always lists the place of follow-up as the emergency department.

Instruction number 3 always gives the same simple message; If you are worried, or if you think something is wrong, come back to the emergency department right away. We would be happy to see you again, anytime.

These instructions are simple and understandable. They minimize your medico-legal risk by minimizing the patients risk. They are personal, the patient knows that you created the instructions just for them, and just for this situation. Finally they are written on a prescription pad, a quarter sized piece of paper that only important things are written on.

Our patient has abdominal pain, maybe he has gas, maybe he has early appendicitis, the reality is that we just don’t know. So that’s what we tell him, and then we write a set of discharge instructions that list:

  1. Time and person specific follow-up
  2. Cues to return to the emergency department immediately
  3. An open invitation to return at any time if the patient thinks something is wrong or has any concerns

Sample discharge instructions

The chances are that our patient will be all better in the morning and go to work, instead of coming back for his re-check. But, for the occasional time that this patient will come back with appendicitis a good set of discharge instructions changes everything; instead of being the doctor who missed appendicitis, you will be the doctor who knew it might be appendicitis and arranged follow-up.

A scan to be sure…

Both doctors and patients like the idea of certainty. We want to be sure that a patient with chest pain isn’t having a heart attack, and that the patient who seems to have a kidney stone doesn’t actually have an aortic aneurysm. We are trained to think ‘worst first’ and to consider uncommon presentations of common diseases. Most of time this serves us well, but sometimes it can lead us down the rabbit’s hole into a world of testing as a surrogate for certainty. We will go to great lengths to prove something is or isn’t happening, but there’s a dirty little secret in medicine; we’re never 100% sure.

Even in our most certain moments, the chest pain patient with a ECG showing STEMI, the abdominal pain patient with CT proven appendicitis there lives uncertainty. Every test and examination we do in medicine has a sensitivity and specificity (though sometimes the s/s is un-studied and unknown). In emergency medicine we live in a world of uncertainty, and it can be difficult to communicate this with our patients.

“Doctor, how do you know it’s not my appendix?” asks the patient who vomited and has abdominal pain, but a completely reassuring examination “Shouldn’t I have a scan to be sure?”.

In some medical systems, driven by fear of litigation, the answer seems to be yes, the patient should be scanned in order to be sure. But, we should be clear, many of these definitive tests offer only the illusion of certainty, along with the reality of risk. For many of the tests we think of as offering certainty (imaging, angiography, etc) the risks of complications and of false positives of any sort are amplified when the test is applied to patients at very low (but not zero) risk of the disease. Some resource limited systems err in the oposite direction, encouraging clinicians to be overly certain of the lack of pathology in low risk patients.

Do we need to go down the rabbit’s hole of testing every-time? Should we simply tell the patient they don’t have the illness, when there is still a tiny bit of uncertainty? Is there a better way to communicate risk to the patient we think that the risk of disease is low? I think there is, and the path to get there is through better communication. When we see patients at low (but not zero) risk for a condition what we need is usually not more, or better tests, rather better communication. In any doctor-patient interaction there are 2 crucial questions that must be answered, the question of the disease entity the doctor is worried about, and the question of the disease entity the patient is worried about. The first is almost always answered, the second is surprisingly often both unknown and unanswered.

I find it useful to ask patients ‘What’s the thing you were most worried about today?” Sometimes it lines up with what I’m worried about and sometimes not, regardless that question is the reason the patient came to the emergency and the patient deserves to have that question answered and explained. Then I say to the patient; ‘When I first hear a story like yours I am worried about X. Today, when we really flesh the story out, and did a good exam I can tell you that I think you are at low risk for X and here’s why…”  Then I disclose to the patient medicine’s dirty little secret, we are never 100% sure, and I try to give a risk estimation to the patient. Then I try to gauge the patients personal risk tolerance and reconcile with the actual risk level, sometimes there are tests we should consider in particularly risk averse patients. Finally I give the patient a brief summary  of the cues that will bring them back to my care.

In summary:

  • Answer the concern that prompted the patient to come to the emergency department
  • Answer the concerns you had when you heard the patients story
  • Disclose the risk level as accurately as possible to the patient
  • Try to reconcile the risk level and the patients risk tolerance
  • Give the patient a brief summary of cues to return to the emergency department

There is always resistance to this approach, doctors believe it will take too much time or that patients won’t like this approach. In fact it probably saves time rather than placating patients with potentially un-needed tests, and it is definitely an approach most patients appreciate.

So the next time a patient with abdominal pain and a normal exam asks you if they shouldn’t have a scan, just to be sure, try taking the couple of minutes to have a discussion about shared risk. I think you will be surprised how effective this strategy actually is.

Cross Cultural Communication Lecture; Feedback Needed!

This is a video of a lecture I am giving next week on cross-cultural communication. It’s part of an interdisciplinary panel, and will be given to a room full of lawyers and mediators who are interested to hear about the physician approach.

A room full of lawyers…a lot different than the standard medical talk…

I thought I’d better crowd source some feedback on the lecture before I actually give it. Watch it, and let me know what you think, and how it can be tweaked.

The time and action specific consultation

As emergency physicians we spend a great deal of time interacting with consultants, but we don’t think about this as a skill. In fact we probably interact with fellow physicians more than any other specialty, but rarely is the interaction itself included in any type of formal curriculum. The well being of our patients often depends on our ability to get a particular consultant involved in a case.

Each emergency doctor has their own approach to consultation, from the buddy-buddy type of interaction to the frankly hostile. When junior physicians call consultants they often simply state what sort of case they have and see what the consultant has to say; ‘I have Mr. X here, a 65 year old man with chest pain….’ Unfortunately this leaves the consultant in the position of guessing what the emergency physician (and the patient) wants or needs. The consultant may respond with management advice when what the emerg doc really wanted was for them to come see the patient in consultation or accept the patient for admission.

A better system is the time and action specific consultation. This allows the emerg doc to clearly communicate their request to the consultant in a way that is non-confrontational, but also unambiguous. The phone call opens with an introduction; ‘Hi I’m Dr. J in the emerg, who am I speaking to?’ This allows both parties to know exactly who they are dealing with. Then (after the exchange of niceties that are an important part of collegial community practice) the emergency doc briefly summarizes the case and makes a specific request of the consultant and gives a specific timeframe.

An example might look like this:

“Hi it’s Dr. J in emerg, who am I speaking to?” “Oh hi Dr. Cardiology, happy new year! Listen I have a 65 year old man with chest pain, unresponsive to nitro and morphine. He has some ischemic ECG changes. I’d like you to come down and see him in consultation in the next 15 minutes.”

When you need advice rather than in person consultation it is equally important to say so:

“Hi it’s Dr. J in the emerg, who am I speaking to?” “Oh hi Dr. Opthalmology, thanks for calling me back. Listen, I have a contact lens wearer with a corneal abrasion. I’d like to put him on antibiotics, but he is allergic to quinalones. Can you recommend an appropriate non-quinalone antibiotic for corneal abrasion in a contact lens wearer and see him in follow-up tomorrow?”

The time and action specific consultation allows the emergency doctor, who has seen the patient, to set the pace of the consultation. This is appropriate since the emerg doc has the best idea of the actual acuity of the patient, but can be lost when the conversation lacks specificity. Most of the time your consultants will appreciate the brevity of your patient presentation, and directive approach to consultation.

Disagreements will occur and will mostly be about the need to see the patient or the timeframe within which they should be seen. When this occurs I suggest first identifying the conflict and then giving the consultant more clinical information to explain why you have made a specific request. Suppose the opthalmologist in the above example suggests a 1 week follow-up;

“Thanks for the advice Dr. Opthalmology, I will start the patient on gentamicin drops right away. I see we have a difference of opinion about when this patient should be followed up; this patient is a contact lens wearer and he has a large corneal abrasion right in the centre of his visual axis. It looks deep. I am worried about this patient so I think he should be seen tomorrow, rather than in a week.”

Most consultants will grant your request once more information is given. Occasionally the conflict will persist and negotiation will be required. When this happens remember to be polite, but also remember that your primary responsibility is to your patient.

Finally I suggest closing the loop, briefly repeat the plan back to the consultant and make sure you are both on the same page. Then document on the chart the consultants name, the action, and the timeframe.

“Thanks Dr. Opthalmology, I will put the patient on gentamicin drops and have him call your office in the morning for an appointment tomorrow.”

In summary a managed approach to conversations with consultants can benefit both parties, by expediting conversations, and making expectations clear from the outset. Managed conversations leave less room for ambiguity, assumption and error. The steps to the time and action specific consultation are :

  1. Identify yourself and the consultant
  2. Give a very brief summary of the case
  3. Ask the consultant to perform a specific action (advice, in person consultation, admission, etc.)
  4. Tell the consultant the time frame in which you need them to perform this action
  5. Identify and resolve any disagreement or conflict
  6. Close the loop by repeating the specified plan and time frame back to the consultant
  7. Document all of the above for the chart and the patient

Otitis Media: How to treat your patients right and do the right thing

In a busy emergency department nothing is as wonderful as a simple and straight forward case. Otitis media used to be that case; a child is crying, the ear is sore and red, hand over a prescription for penicillin and the job was done. Happy parents, happy doctor, happy child….oh wait, the child was still unhappy but 2 out of 3 isn’t bad.

Mounting evidence and experience in European countries has shown us that immediate treatment of otitis media with antibiotics is probably not warranted. In a modern immunized society the majority of cases of OM are essentially benign self limited illnesses that don’t require treatment. In many North American emergency departments physicians have slowly been backing off of antibiotic treatment for OM. This past January 2 separate studies were published in the NEJM that have again shifted us back towards antibiotic prescribing for OM. Separate studies by Tahtinen and Hobermen both seemed to show a benefit in treating children less than 2 with a 10 day course of amoxicillin-clavulanate. There was much analysis of these studies, perhaps the best is by Dr. David Newman, and in the end it seems that perhaps there is a small reduction in symptoms when children are treated with antibiotics at a cost of significant antibiotic associated side effects, most commonly diarrhea.

Physicians are confused, we don’t know what to do, we try to read the parent and sense what they want and do what we think will make them happy. Parents are confused, sometimes doctors tell them their kids need antibiotics, sometimes they withhold antibiotics, it doesn’t make sense to them. What was once a lovely and brief encounter is turned into an emotionally turbulent episode of hand wringing, half explanations and crying children; a doctor who is unsure what to do, and a parent who is aggravated by the seeming randomness of the situation and a conflict around antibiotics.

The problem with ear infections is that they hurt, they hurt in the same way it hurts when a sadistic dentist blows compressed air onto a freshly drilled tooth (why do they do that!?). Adults with an ear infection want a shot of morphine, kids get a pat on the head and an ‘it’ll be okay buddy’. The only reason that anyone takes a marginal reduction of symptoms in an OM study seriously is because the symptoms are terrible and because we don’t treat the symptoms. It seems ridiculous that we are willing to accept a 15% risk of diarrhea with amoxicillin-clavulanate for 12 hours less of symptoms, particularly when the symptoms can be treated (if only we decide to treat them).

Here’s what I’ve found helpful in turning otitis media back into a great emergency department encounter:

  1. When I walk in the room and see a crying child with a bright red and bulging TM I say Wow! That looks really sore, we need to do something to get you feeling better right away!
  2. I always carry a couple of sterile ampules of 2% lidocaine in my breast pocket, and as long as there is no allergy I put 2-3 drops in the ear (there are commercial products designed for this, but they are expensive and single use lido works just fine). Then I tell the parents, just to keep him still for a minute while I go get a sticker to distract the child.
  3. When I get back into the room no one wants to talk about antibiotics because the child is now happy and smiling. What they want to talk about is the magic pocket drops and where they can get some of their own.
  4. I then give the child 15mg/kg of po acetaminophen.
  5. Then I say to the parents; Now that he’s feeling better we can talk about what to do. These days in children who are immunized ear infections are not dangerous and get better with time, people sometimes worry that they might turn into meningitis but I can reassure you that that is not going to happen. The biggest problem is that ear infections are painful so the first thing we need to talk about is how to control the pain…I give them a talk on using proper weight based doses of acetaminophen or ibuprofen and a topical agent like the lidocaine drops or a commercial alternative and that typically symptoms last 5 to 7 days…When it comes to antibiotics the studies that are out there don’t show a lot of benefit and they do show a lot of children getting side effects like diarrhea, so I like to start with getting the child feeling better with pain control. I’m not dogmatic about antibiotics though, if you are worried he is not getting better we are always happy to see him again and reassess the situation or if you would like I will even write you a prescription for antibiotics you can fill in a couple of days if things are not getting better.
Almost no one takes the delayed antibiotics script, and often when they bring in another child for an ear infection the parents decline antibiotics and ask for pain control options instead. The whole encounter takes about 5 minutes. It is much faster and easier to control the pain and explain the options than to have a tense antibiotic associated stand-off to the tune of a screaming child.
And ahhhhhhhhh, that’s the sound I remember from the old otitis media visits, happy parents, happy doctor and happy child, oh wait this time we are 3 out of 3….