Archive for category EBM
I saw a very disheartening quote in a patient chart recently:
“…consider curbsiding ID for antibiotic recommendations…” Followed a few pages later by “Follow ID recs…”
It was disheartening in several ways: firstly that although someone had obviously thought us worthy of asking for advice they hadn’t actually called me about the patient; and secondly that they thought this question was worthy only of a “curbside”.
Regular readers and followers will already know my thoughts on curbsides – but I didn’t really delineate what I actually do. I addition, what should I expect of a consult and what should a consulter expect of the consultant?
A consult is triggered, at least on the inpatient side, when a physician asks a question of a colleague in another speciality. Now this happens informally all the time – we live in a learning environment after all – but a question along the lines of “I have a patient…” generally means that there are healthcare decisions being made on a real patient, and these are the real issue.
Most of the time a consult requires a specific question – I have had colleagues in fact pause for thought and tell me “I’m not really sure what question I have for you…” and then not bother formally consulting. In my mind, it’s very simple – if you are concerned or uncertain enough to seek out a subspecialty colleague for advice on a specific patient, that by itself is valid grounds for a consult. I have done more than my share of consults for “please assist with antibiotic management or further workup”. It’s not failure or weakness, nor a waste of my time – it’s what I do and it’s in the best interests of the patient.
I do expect a somewhat valid reason – a specific question is ideal, but even if I’m asked about antibiotic treatment I will usually try to go beyond that and include alternative diagnoses, testing, followup recommendations etc. I also expect timeliness – a consult should be called early, both early in the disease course and early in the day! I prefer to avoid problems than have to dig people out of them. I also like time to go to the lab before seeing the patient…and recently I had to fish culture plates out of the trash in order to help a colleague with antibiotic recommendations. Had they waited another day we would have had to use unnecessary antibiotics to treat more broadly than we needed to. Calling a consult prior to knowing all the information is ideal – I can often get preliminary results faster than the lab will report them and add on additional workup days ahead of time.
When I get asked a “simple” question like what antibiotics are recommended I go through the following steps:
Initial shoot-from-the-hip thoughts – what info do we have so far? How sick is this kid, risk factors, early lab results and prelim microbiology workup. I will ask for CBC (with diff, always with diff), CRP and sometimes ESR, renal and hepatic function as appropriate. Urinalysis results, and CSF findings if an LP was done. Empiric bugs will depend on likely source of infection and likely site of infection – the two might be quite different! Then I ask what drugs the kid is on and line up the likely antimicrobial activity of those drugs against the likely bugs and see if there are holes in coverage. I do this last step mentally over the course of a few seconds (usually) but I do it every time. Typically when I do the same process with residents or students it’s a several-minute slog through the bug/drug matchup table. Just because it’s easy for me doesn’t make it worthless though, it just makes me efficient. If the kid is “safe” then I tell them to stay put, if not I suggest additional empiric coverage before I even see the patient.
Second stage – chart review. I read the chart. Yeah, I really try to read it. ER records, admission note, progress notes, operative notes and even resident signout notes… I will scroll through every lab in the computer even if I don’t transcribe every one into the chart. Every microbiology lab, positive, negative or pending is recorded. I personally looks at x rays, scans, even ultrasounds (although my ability to read those things is pretty near useless). I always review the scans myself and THEN read the radiology report. Sometimes I’ll go down to radiology and review it with the radiologist.
At this stage if there are questions or additional workup obviously needed I will call the lab or let the team know, and if I get the chance I will physically walk over to the lab to look at the cultures myself. What’s the value in that you might ask? Well more than once has an initial “gram positive” gram stain turned out to be a gram-negative bug, and in some cases a gram stain alone can, with the right eyes and expertise, result in a diagnosis all by itself. On the culture plates, bugs like proteus, klebsiella, strep viridans, listeria, E. coli and pseudomonas have a characteristic appearance (and smell!) that may jump start the management a day before the Microscan or Vitek machine gives you the formal identification. A visual peek at a urine plate reported out as “flora” might reveal a predominant organism that you can point to and ask to get worked up.
Lastly – the patient. Go to the bedside, lay on eyes and hands. Talk to the parents and patient and find out the little nuances in the story that others missed – the dog bite in a kid with fever, the recent dental visit in a kid with bacteremia, the rash in a mom that started the day after delivering her baby… Test hypotheses, confirm or refute suspicions.
Sometimes with all of this I rethink my initial plan – which only goes to show how unreliable the shoot-from-the-hip curbside is. I may back off from my broad empiric coverage, or I may rethink a diagnosis completely and expand both therapy and workup. It’s not unusual for me to be consulted on disease X and have to tell the team that it’s really disease Y all along. A curbside cannot possibly do that. Any result requires a conversation with residents, students, nurses and colleagues – all of it educational and a two-way process.
So if you add all this up, what does it mean? It usually means, at minimum, a level four inpatient consult. Consults come in five levels – short of a life-threatening condition this is about as high as you can realistically go.
Let that sink in for a bit. A simple question of “what antibiotic should I use?” is justifiably as difficult as a decision to do elective surgery. In fact, based purely on asking me to review the evidence running up to the decision, there’s enough work to bill at that level even if I say “you’re doing a fine job, I recommend no changes or further workup”.
Yes, Infectious Disease specialists do all sort of other cool stuff too – we diagnose rare diseases, can help with resistant organisms or diagnostic dilemmas – but fundamentally we’re trained in how to best manage all the routine stuff as well. That’s not to say we need to get called on every pneumonia, meningitis or urinary tract infection – but if you get stuck with a question or concern with any of these it’s ok to ask for help.
And if you’re going to ask for help, don’t, just don’t, assume it’s not worth anything. No-one would dare ask a surgeon to operate for free…why treat your ID colleagues any differently?
I joke that, as a Peds ID doc, it is my duty to say this at least once a day…
Ok, I may not literally be slapping people upside the head, but there are certainly times when I’m doing it in my mind. The situation is common enough – a patient, parent or doctor, faced with symptoms consistent with an infectious disease, considers using antibiotics to treat bacteria. After all, we know that bacteria kill people, right? But in many of these situations the patient really has a viral infection – and viruses aren’t affected by antibiotics. So at the very least we’re wasting money and drugs. Worst case scenario? We’re promoting drug-resistant bacteria, antibiotic allergies and side effects – that in some cases can be life-threatening.
But aren’t there clues to help us make the distinction? Real clinical signs and symptoms? Well, lets review a few.
White pus on the tonsils
Everyone is familiar with the feeling of an awful sore throat, and having a doctor peer down and having you say “Ahhhh…” What are they looking for. Probably something like this:
This is a classic appearance of “Strep Throat” – a bacterial infection that aside from being painful in its own right can go on to lead to serious complications, such as rheumatic heart disease, kidney disease, a form of arthritis and a weird neurologic disorder called “Sydenham’s Chorea”. Fortunately it has no drug resistance so simple penicillin/amoxicillin will kill it (so if your doc tries to give you “stronger” antibiotics please feel free to slap them).
The trouble is, this isn’t a picture of strep throat. I grabbed this from an article on “Mono”. Infectious Mononucleosis can be indistinguishable from strep throat, but antibiotics do nothing for it. The “pus” you see isn’t really pus, it’s just a nasty-looking white gunk your tonsils make. A bad sore throat can be caused by influenza, adenovirus, RSV, metapneumovirus, rhinovirus….you get the idea. It can be hard to tell strep throat from any of the other many possibilities, but in general if you DON’T have a runny nose or a cough, and the lymph nodes in your neck hurt then it’s PROBABLY strep. But it could be a virus. Strep tests and cultures help – and holding off on treatment until the test comes back is a sensible plan.
What about ear infections? Another common bane of pediatrics (almost every young child I see with a prolonged illness has at some point been diagnosed with an “ear infection” before arriving at the correct diagnosis – I once saw a kid with a brain tumor get that diagnosis…). The symptoms are notoriously non-specific (ear pulling, fussiness, fever) and a good ear exam in a small, squirming child can be difficult! A crying baby can turn their ear drums pink…and voila! An ear infection! But even assuming your exam is good and the ear drum really does look nasty, how do we know its a bacterial infection? Despite the appearance of a rip-roaring otitis media (bright red, bulging ear drum, fluid behind it) it can be a viral infection too. Most of what you see is the BODY’S response to the infection remember. Clinical trials of antibiotic use have shown with without antibiotics, ear infections tend to get better just as quickly as with them. Complications from untreated bacterial infections do exist, and can be quite serious, but are rare. It is prudent to consider a “wait and see” approach to ear infections to see if it gets better by itself. I don’t want your kid to get mastoiditis any more than you do, but if it does happen I want it to be treatable with the best antibiotics!
Most of the time when we’re treating ear infections we’re not even treating the child…we’re allowing the adults in the house to get a good nights sleep…;-)
Cough, fever, patches on chest x-ray
Pneumonia? Guess what. Usually a virus, at least in kids, before they become immune to everything. Without proper testing though this can be harder to tell apart, and we’re getting into the realm of “sick kid” here. Almost every doc will feel a little weird ignoring a possible bacterial pneumonia, even if they really do think its viral. But the high rate of viral infections, along with the risk of increasing drug resistance, is why the current recommendations for antibiotic treatment of pneumonia in children start with plain old amoxicillin. RSV, metapneumovirus, influenza, adenovirus – they can all cause pneumonia. In the Bad Old Days viruses like measles and varicella could also do it, and they were quite nasty! With symptoms like a runny nose, rash, lots of sick contacts, the chances of it being a viral infection are quite high. Sitting it out for a few days is again a reasonable option – because you know if you see a doc and get a chest x ray they’ll start you on antibiotics, and we don’t want that, right?
Very high fevers, difficulty breathing, chest pain with pneumonia, coughing up junk – always worth getting checked out.
All of us have at some point experienced symptoms of a sinus infection. Fever, pressure, tons of snot, headache. They are truly miserable things. I hear all the time how “we knew it was bacterial because he had green snot”. Sorry, but that’s not all that helpful. The greenness of snot comes from the cells your body is sending in to kill the infection, which will tend to be neutrophils whether it’s a virus or bacteria. (Neutrophils don’t really kill viruses, but they’re just reacting to the inflammation there). Neutrophils have the awesome ability to create highly-reactive chemicals, one of which is called “superoxide” which gets converted to hydrogen peroxide which then reacts with chloride ions in salt to produce….bleach. The green color you see is actually the neutrophils and the enzyme they are using to create the bleach (myeloperoxidase), not the infection itself. You’ll get green snot regardless of what’s causing the infection, and it’s a good sign – a sign that your immune system is in full swing.
Severe sinusitis will produce lots of snot, for sure, but lots of snot doesn’t necessarily mean its a severe sinusitis, and certainly doesn’t prove it’s bacterial. If symptoms have lasted for a couple of weeks with no improvement, that’s a red flag for something non-viral.
Fever is a normal immune response which effectively suppresses bacterial and viral infections. It hurts them far more than it hurts the patient. A fever by itself won’t necessarily cause any harm at all – and high fever may or may not indicate bacterial infection. A fever is just a clue – a reason to look and figure out what’s going on. One you’ve figure out it’s a virus based on symptoms (runny nose, viral rash etc) then you’re good. And don’t worry if fever keeps coming back, it will do that until the infection is gone, which may take a week or more.
The height of the fever is only slightly predictive of the risk of bacterial infection – but influenza, adenovirus, EBV can all cause pretty good-going fevers of 102F and up. I’m far more interested in what ELSE is going on in addition to the fever.
Febrile seizures, convulsions caused by fevers in young children, are more closely associated with a rapidly rising fever than a high fever itself. If your child has a fever of 104.5F and has sat there for an hour, chances are good they’re not going to seize from that.
Addendum – Mark Crislip recently posted on fevers over at Science Based Medicine!
So that’s a rough overview of the various common viral infections. It really is surprising how often we do get sick from something that will simply run its course. Our immune system is pretty robust. That’s not to say that in exceptional circumstances viruses can’t or shouldn’t be treated (herpes, influenza, chickenpox, measles, adenovirus, CMV and EBV all have some form of treatment to try even if the therapies are nowhere near as effective as antibiotics are on bacteria) but for respiratory infections in particular we would be far better served by reassurance that our symptoms are more consistent with a virus than a bacteria, and that most of the time it will sort itself out. A large chunk of the inappropriate usage of antibiotics stems from over-treatment of viral respiratory infections – so next time you see your doctor for something like this consider asking about tips for symptomatic relief rather than an antibiotic prescription.
A few other studies: prescribing antibiotics doesn’t necessarily save time.
Antibiotic overuse, even based on physician diagnosis, worse with criteria-based diagnosis.
Understanding why physicians overprescribe – many different reasons.
Good advice can be found on the CDC website.
I have been told that I must credit my wife for originally coming up with the idea for the “IT’S A VIRUS” slapping Batman meme, and Quickmeme helped me create it.
The simple reason why I practice patient-centered care is that it’s better for the patient. But before I go into the details of that, we need to step back a bit.
Firstly, I’ll clarify what patient-centered care isn’t. It isn’t pandering to a patient’s or parent’s wishes and doing whatever they want, as a mere provider of healthcare. That is patient-LED care. I don’t think that’s always a good idea – most people after all have NOT gone through medical school and several years of practical training (something like 20,000 hours of supervised patient-care in my case) in order to make informed decisions on their health. Even though the Internet has leveled the information playing field considerably, you still have to know how to interpret that information in the appropriate context and with the correct background knowledge. There are places where patient-led care does play a role, but it is quite distinct from patient-centered care.
I define patient-centered care as “practicing medicine taking into account the patient’s concerns, expectations and understanding.” You may not find that definition anywhere else put quite like that, but to me it makes sense. It also follows a three-step process of “discover, validate, address” that I iterate through an encounter so that by the end we’re all on the same page.
When I was in medical school I was lucky enough to be asked to pilot a new curriculum element called Preparing for Patients (my sole legacy to Cambridge University is that I was the first to coin the abbreviation PfP – which obviously was no great intellectual feat, but I think worthy of a footnote in the annals of history). I was not yet seeing patients on the wards, and felt quite unprepared having spent much of my work experience during high school in various labs – examining things like different plant species, fiber glass tensile strength and drug purity.
PfP was an intensive program back then, a couple of weeks of daily sessions where we explored our own fears and thoughts on medicine and patients, then got to experience and practice actually talking to patients about their illnesses. The real beauty was in the use of standardized patients: actors and actresses who could give a consistent experience to everyone and respond to questions, even off-line, in character. I got to tell someone they had cancer with a 50% mortality several times before I actually HAD to tell someone they had cancer (which as it happens was at 2am one morning on call as an intern, by myself – that’s deserving of a blog post all for itself…). The experience was invaluable, and provided a toolset of behaviors, questions and actions (71 skills all told) that I could bring into play when I needed them during an encounter. I got to try out this new-fangled curriculum and provide feedback to the course creators on the process and content.
The course itself still stands, albeit in a greatly modified form. It is now a fully integrated part of the Cambridge curriculum, from the first year of pre-clinical science, and those 71 skills are the benchmark by which all Family Medicine (aka General Practice) docs in the UK are assessed for their board exams.
What I learned from that was invaluable – it turns out that talking to patients is a lot more than simply asking questions about their symptoms. Patients are people – they have preconceived ideas about their illness, they have worries, they have ideas on what needs to be done. Sometimes they’re wrong, in which case our job is to educate and reassure (or sometimes not…), but often they’re right and our job is to help get things done. I learned that illness (what a patient experiences) is different from disease (what a doctor treats). A tension headache is an awful illness, but a minor disease that the doc can do little about beyond over the counter pain meds. High blood pressure on the other hand usually has no symptoms whatsoever but serious effects on the body so that we want to treat it. The question was posed – how do you convince someone to treat something that isn’t making them sick right now?
What I also learned was that there was actually research to back up this approach – patient complaints and concerns about medical care (including well over half of all malpractice lawsuits) usually stem from communication failures or unresolved issues. Issues often were unresolved because the doc either didn’t allow the patient to bring it up, or didn’t explain things fully. Patients do not tend to bring up what medics would consider the most important issues first – for all sorts of reasons – and yet they are often cut off early in the rush to get them out the door and see the next one. Something as simple as asking “What are you concerned about?” early on in the encounter can save a ton of time, as you can focus in on what they’re most worried about right away. (Of note, you’ll get different answers asking that than if you ask “what are you worried about?”, which I find fascinating…) Making someone feel at ease is one way to encourage them to talk about embarrassing symptoms or scary possibilities, and there’s an entire skillset devoted to building rapport and trust for precisely this reason.
My general approach is the “discover, validate, address” thing I mentioned earlier. Discovering concerns may be as simple as asking them what they are, but there may also be subtle hints – a family history of cancer, a perseverance on a particular topic or symptom, facial expressions and other body language. You may focus in on something you notice, or use open-ended questions to hear things in their own words. It may be an ongoing process through the encounter, but ideally you get most of it done early on to avoid the “by the way Doc…” question as you’re wrapping up.
Validation isn’t simply agreeing with them – after all, people often get misled or misunderstand things. Validation is acknowledging that from their perspective what they’re feeling about something is entirely appropriate. They may be angry that their prior Doc didn’t treat symptom XYZ, but if, medically, it didn’t need treatment, then their Doc did nothing wrong. But if I can commiserate with them, ask about how it’s affecting their daily life, explain that this kind of symptom isn’t one we can treat – this often goes a long way to fixing the issue.
Addressing a concern may be already covered by just acknowledging its existence, but may require an explanation of why treatment or testing isn’t necessary, or it may require convincing someone to undergo a treatment plan that they’re really not all that keen on! It’s important to offer options – there is always the *option of doing nothing*, even though that’s not necessarily the best option. It’s critical to explain YOUR thinking about something – admit your biases, your own concerns about the patient – they’re more likely to follow through on your recommendations if they know why you’re sending them for blood work, x rays or a cardiac stress test – or asking them to pop a pill every day for the rest of their lives!
What this approach does is help the patient have more control over their medical care than an old-school paternalistic approach, but with more education and understanding than a patient-led approach. If you train doctors to talk to patients this way an amazing thing happens – the patients do better. Improved communication can improve management of diabetes and blood pressure, but also reduce followup visits and tests, lower reported pain levels, and some surprising things like reduced costs in the ICU. Others have already listed the main references. To me this proves two things.
Firstly – there are clearly deficits with doctor-patient communication that need to be and CAN BE addressed.
Secondly – YOU CAN TEACH COMMUNICATION SKILLS. I cannot overemphasize this enough. One of the largest myths in medicine is that you either have a good ‘bedside manner’ or you do not, and if you don’t you’re stuck with it. That simply isn’t true. You CAN teach medics of all levels – from medical students to consultants – new skills and demonstrate changes not just in their behavior, but in their PATIENTS’ behaviors. This is an astonishing finding, and the skills can persist for years. The only thing more astonishing than this finding is that we’ve known about it for decades. Communication skills are being given greater emphasis in medical school these days, finally, but testing is haphazard and unhelpful a lot of the time (feedback 3 weeks after a standardized encounter is nowhere near as helpful as an immediate conversation and a chance to do-over the visit) and training is often limited to lectures rather than structured practice sessions. It is difficult to teach it properly, and it is certainly difficult in an area traditionally taught through lecture format, and which is increasingly moving towards online self-directed educational formats. Carving out a 1-2 hour block of time every week to sit down in small groups with a trained facilitator and one or two trained standardized patients is what’s probably necessary, but I doubt many course organizers think that they’re able to do that – my argument would be that we need to find a way to make it happen, not that we avoid trying because it’s difficult. I am living proof that you can take someone who honestly was pretty socially inept and turn them into someone who can not only practice patient-centered care, but teach it to others. Throughout my residency and fellowship I led a group of child-life specialists, Residents and Attendings in weekly sessions with the pediatric clerkship students teaching a modification of the Calgary Cambridge Guide.
One common criticism about teaching patient centered care or communication skills is that it somehow detracts from the teaching of “real” medicine – the mass of signs, symptoms, risk factors, tests and treatment options that we basically have to rote memorize, as well as the practical application of all that knowledge with real, sick patients. My counter to that is: who says the two are mutually exclusive? You can learn medical facts during the practice sessions, you don’t need to know them beforehand. You can integrate the two aspects of medicine – and in fact you probably need to integrate them or else risk maintaining the mental block between “real” medicine and “communication skills”. Real medicine relies on communication skills to elicit a history and convey a plan – how else do you think this can be achieved? Telepathy? Flash cards? Who says you can’t run a code in a simulation then afterwards have the “breaking bad news” simulation with the manikin’s “relatives”?
And finally, doctors that have a disease-focused approach are more likely to experience patients as “difficult”, and those patients are more likely to have additional (unnecessary?) visits, than if the doctor had a more patient-centered approach. Patient-centered docs are happier docs.
So, to me, effective communication skills are an absolutely integral aspect of patient-centered care, and patient-centered care is a way to dramatically improve patient outcomes. These skills can be taught, and I argue they should be taught if we truly want the best for our patients.
If any readers of this actually do rotate through ID with me, remind me to discuss the process of an encounter as much as the content…I tend to forget!
As a consultant my expertise is sought out in largely two ways – a formal consultation (a request to see a patient, obtain a history and perform a physical examination, review laboratory tests and recommend further evaluation or treatment), or a curbside question (a quick hypothetical or general question with the expectation of a simple answer).
An example of a curbside question might be “How many pneumococcal serotype responses would you expect to be normal in an immune evaluation..?”. The answer is 5-10 depending on the age and immunization status of the child, but in reality the correct response is “why the heck are you ordering an immune evaluation on a kid that I know nothing about…?”. The indications for performing an immune evaluation (frequent or unusual infections) are generally the sort of thing an Infectious Disease specialist should have been consulted on!
People often start a curbside question with “This isn’t a consult, but…” as if a consult is a bad thing. It isn’t. A consultation isn’t an inconvenience, it’s what I get paid to do (salaried or not, divisional revenues ARE based on the consults I get called to see). It’s what I ENJOY doing – if it wasn’t I wouldn’t be in the job in the first place. And even if I AM busy, tied up in clinic, or off-site taking call from home, it’s in the patient’s best interest.
No matter how well you quiz someone over the phone, there is no way they can adequately convey the entire medical history and physical exam, the concerns of the patient and family, trends in lab values, recent antibiotics and other meds, and the simple gut vibe of a case… A complete consult, done properly, can take up to an hour and may involve field trips to radiology and the micro lab to check things out for yourself. That is a considerable chunk of time (certainly more than a curbside question) but the value of having a subspecialist see the whole picture cannot be overstated.
The dangers of answering a curbside about a specific patient are legion – you may miss drug allergies or interactions, co-existing diseases or subtle clues in the history or exam that would point towards a specific diagnosis, you will tend to overtreat “just in case”, lacking the reassurance of seeing the patient for yourself, but may just as easily undertreat an infection that had been missed or misdiagnosed. Worse, for the consultant, chances are good that their name will end up in the chart “case discussed with ID”, which medico-legally puts us in a bit of a spot. Then the onus is on you to show that you had no medical obligation or responsibility to the patient should something bad happen…a hassle and horrific waste of time at best.
The other issue is “added value”. Even when I’m called to answer a specific question, I almost always end up offering something else. If I’m asked about best treatment options, I will offer alternative diagnoses. If the question is what this disease could be, I will recommend empiric therapy as well. Every consult is a teaching opportunity, whether about a specific disease or a general bit of advice on ID. For THAT patient I want the docs who consult me to know as much about the disease as I do.
That’s all in theory – what about the evidence? One study of mandatory ID consultation for outpatient IV antibiotic therapy found that 39 of 44 patients had a change of therapy (!), meaning that 88% of the time the current plan was not ideal. 39% of the patients were sent home on oral instead of IV antibiotics, 13 patients (30%) changed medications, 5 patients changed dose, 3 changed planned duration, and 1 patient was stopped entirely. Cost savings were $500 per patient EVEN TAKING INTO ACCOUNT THE CONSULT FEE. In Germany and the US, ID consults have been linked to significantly reduced mortality from staph infections. In Italy, formal ID consultation on ICU patients reduced cost, mortality, ICU stay, length of mechanical ventilation – all due to improvements in antibiotic usage. A financial analysis of curbside consultations suggested that close to $94,000 in revenues were lost in a year by giving advice over the phone without performing (and billing for) an appropriate level of consult. With antibiotic cost savings and increased revenues to the hospital, consults really are a win/win situation.
So what’s really happening when you say “This isn’t a consult, but…”? You’re putting your patient at risk of being treated for the wrong diagnosis, or being wrongly treated for the right diagnosis, you’re increasing hospital costs and increasing patient mortality, and you’re passing up the opportunity to learn something yourself. It’s not good medicine – it’s not good for anyone.
Say it after me: “I’ve got a consult for you…”
This post may or may not have been inspired by the fact that I have had an inordinate number of consults this week which started out as curbsides that would have led to inappropriate care….