People Don’t Heal Overnight, So Stop The Needless Next-Day Referrals


Stop making needless, expensive next-day referrals

Since graduating from my EM residency in the Paleolithic period of our specialty I have heard on multiple occasions the advice from “experts” to send patients discharged from the ED to their physician “tomorrow” or “in a day or two” for follow-up care. I’ve heard this from malpractice attorneys, ED group leaders and others who should know better.

As a general rule, this advice has always struck me as bad advice. It is primarily proffered to cover the butt of the emergency clinician, not to benefit the patient. Sure, there are occasional diagnoses where a follow-up in a day or two would be appropriate – a discharged febrile pyelonephritis patient treated in the ED with IV fluids and parenteral antibiotics, a low-risk chest pain patient with negative markers and ECGs but without a functional or anatomic study, a TIA patient who is, for some reason, unable to have a minor part of their work-up completed in the ED, a tight asthmatic who largely clears in the ED


But remember, 80% of most ED patients go home – and most have minor problems. Many of these patients will have a very predictable disease or injury trajectory. The patient with the sprained ankle can be expected to be hobbling for at least two weeks – what possible benefit can be achieved by having this person take off a day from work and see their primary care physician in a day or two? None. Yet emergency clinicians don’t seem to care about the inconvenience or expense they are generating – it is worth it in order to cover their butts (even though they bear none of the inconvenience or expense). And musculoskeletal back pain – what possible value would there be to see the PCP the day after an effective ED visit?

And what about the patient with a hacking cough – no fever, minimum productivity – clearly a viral illness for which nothing effective can be done (other than recommending some honey or Jack Daniels). Why in the world would we advise this patient to seek follow-up care in a day – just to have the patient bother their PCP? We know full well that this illness will last weeks. But instead of setting the appropriate expectations, we cover our butts acting on the delusion that unindicated and premature referrals will reduce our chance of being sued.

The list of diagnoses that are inappropriately referred to physicians for immediate follow-up is nothing short of incredible. One wonders what percentage of the 130 million ED patients gets this bad … and I would go so far as to say unethical, advice. Why is it unethical? Because we know with virtual certainty that the patient will get no benefit from the follow-up visit – yet there will be expense and time wasted – and we have taken the trust that patients have in our advice and used it to our own necessities.


Here are two papers that put into perspective how long it takes to get over a cough. These are not obscure studies from journals you’ve never heard of. The first is by a highly regarded researcher, well known to me, who has recently been added to the US Preventative Services Task Force. His group looked at the duration of cough illness in 19 studies of untreated controls (18 days) vs. how long patients felt the duration of a cough would typically last (median was 5-7 days). So there is a substantial disconnect between how long coughs actually last vs. how long patients think they will last. The inevitable consequence of this mismatch – patients return to their doctor because of perceived persistence of their cough. And what is the poor doctor expected to do? More than what was done on the first visit, of course. So out comes the prescription pad for the ubiquitous Z-pak.

Ebell, M.H., et al, Ann Fam Med 11(1):5, January-February 2013
BACKGROUND: Although acute cough illness (ACI) is generally viral in nature and self-limiting, many of these patients request antibiotics and rates of antibiotic prescribing in this situation remain unacceptably high (about 50% in 2006).
METHODS: The authors, from the University of Georgia, examined the relationship between the natural history of ACI in otherwise healthy adults as reported in the medical literature and patients’ expectations regarding the duration of ACI. They evaluated the duration of ACI in untreated controls (community-dwelling adults with undifferentiated ACI) participating in 19 clinical trials, and assessed responses to a survey of a random sample of adults residing in Georgia regarding their expectations about the duration of an ACI (493 respondents among 1,131 persons contacted, 43.6%).
RESULTS: The weighted mean duration of cough in controls participating in the 19 studies was about 18 days, while the predicted median duration of ACI among respondents to the population survey was about 5-7 days (mean, 6.5-9.2 days). Predictors of a longer anticipated duration of cough included self-reported previous use of antibiotics for ACI, a history of asthma or chronic lung disease, female gender and Caucasian race. Predictors of a belief that antibiotics are always or usually helpful for the treatment of ACI included nonwhite race, less education and previous use of antibiotics for respiratory tract infection.
CONCLUSIONS: These results are consistent with a substantial mismatch between the actual duration of ACI and patients’ expectations. The authors cite the importance of education of the public, the media and physicians about the natural history of ACI to reduce inappropriate use of antibiotics. 29 references ([email protected] – no reprints)
Copyright 2013 by Emergency Medical Abstracts – All Rights Reserved 6/13 – #28

The following study looked at 48 trials in children regarding how long it takes various infections to resolve. Symptom resolution occurred in 90% of children by:

  • Seven to eight days for those with acute otitis media
  • Between two and seven days for children with sore throat
  • Two days in children with croup,
  • 21 days in those with bronchiolitis
  • 25 days in the setting of acute cough
  • 15 days in children with a common cold

Surprisingly, it took 21 days for bronchiolitis to resolve in up to 90% of the children, but this long duration was also noted in the last study in our series.


Thompson, M., et al, Br Med J 347:f7027, December 11, 2013
BACKGROUND: Respiratory tract infections are generally self-limiting but account for more than one-third of pediatric primary care visits and often prompt unnecessary prescription of antibiotics.
METHODS: These multinational authors, coordinated at the University of Washington, reviewed the findings of 23 randomized controlled trials (RCTs) and 25 observational studies that provided information on the duration of symptoms of common respiratory infections in otherwise healthy children. The study included children residing in high-income countries and receiving placebo or control treatment.
RESULTS: The authors assessed the results of seven RCTs and three observational studies of acute otitis media (1,409 children), six RCTs and one observational study (344 patients) of sore throat, six RCTs and six observational studies (1,763 children) addressing cough, croup and/or bronchiolitis, and four RCTs and 15 observational studies (5,327 children) of the common cold. It was determined that symptom resolution occurred in 90% of affected children by seven to eight days for those with acute otitis media, between two and seven days for children with sore throat, two days in children with croup, 21 days in those with bronchiolitis, 25 days in the setting of acute cough and 15 days in children with a common cold.
CONCLUSIONS: This study provides information about the likely duration of symptoms of common respiratory infections in children and has the potential to influence parental decisions about whether to seek medical care, and clinician decisions regarding management. 55 references ([email protected] – no reprints)
Copyright 2014 by Emergency Medical Abstracts – All Rights Reserved 5/14 – #28

This last paper affirms that bronchiolitis can have a prolonged duration. The median interval to resolution of symptoms was 15 days, although symptoms persisted at 20 days in 25% of the infants and 11% were symptomatic at the end of the four-week follow-up.

Petruzella, F.D., et al, Pediatrics 126(2):e285, August 2010
BACKGROUND: It would be helpful if providers had accurate information about the natural history of bronchiolitis to give to parents or other caretakers, in order to avoid undue concern and unnecessary repeat visits for medical care.
METHODS: This prospective study examined the duration of cough in 95 infants below one year of age seen in the ED of Children’s Hospital of Wisconsin for a first episode of acute bronchiolitis. Caretakers were contacted weekly for up to four weeks in order to determine the interval from symptom onset to resolution of the illness.
RESULTS: Bronchiolitis was judged to be mild to moderate in these infants, and two-thirds were discharged from the ED after evaluation and treatment (with infants below the age of three months being much more likely to be admitted than older infants). Of the subset of infants who were tested, 74% were positive for RSV infection. The median interval to resolution of symptoms was 15 days, although symptoms persisted at 20 days in 25% of the infants and 11% were symptomatic at the end of the four-week follow-up. A history of eczema in the infant and a family history of atopy tended to be associated with prolonged illness duration, but documented RSV infection was not. Among the subgroup of infants attending daycare, a median of 2.5 days was missed as a result of the illness, while caretakers missed a median of 2 days of work or school. An unscheduled medical visit (typically to a primary care provider) related to bronchiolitis was documented for 37% of the infants.
CONCLUSIONS: The caretakers of infants seen in the ED for bronchiolitis should be advised of the likelihood of prolonged symptoms. 23 references ([email protected] – no reprints)
Copyright 2010 by Emergency Medical Abstracts – All Rights Reserved 12/10 – #28

Seems to me, the advice that patients ought get when being discharged from the ED, is to follow-up in some medically appropriate time as determined by the emergency clinician (as ascertained by the diagnosis and severity) and to return to the ED sooner than the specified time (like immediately) if symptoms worsen or if there are new symptoms. Some patients don’t need any follow-up — the patient with an occasional episode of cystitis, the pharyngitis patient (strep or not), the majority of low-grade extremity injuries, etc, etc.

I’m not a big believer that laymen of varying intelligence and educational backgrounds can be converted into reliable mini-doctors as the result of being handed a sheaf of aftercare instructions. That’s why I feel better having them come back immediately if there are any new or worsening symptoms (most people will understand “new” or “worse”). New or worsening symptoms bother me. I’m willing to acknowledge there may be some wasted time and money using this broad advice, but it will be a lot less than the ubiquitous advice to “see your doctor tomorrow” when the vast, vast majority of patients discharged from the ED will receive no benefit from following this advice and incur only expense and wasted time.


Dr. Bukata is the Editor of Emergency Medical Abstracts.


  1. For better or worse, physicians have been successfully sued for failure to refer. It happened to a medical school classmate of mine.

    Until the current malpractice/tort system is changed to resemble something more like what Canada or New Zealand has, verbal and printed instructions to follow up with your primary care physician or specialist within 24 hours will continue to be the standard.

    Always remember that to protect your patients you must first and foremost protect yourself and your license. Protecting the system is not your job.

    • Agree, change the legal standard and we can change practice. Ignore the legal standards at your peril. I try to do the right thing for every patient, every time. Courts feel that we have responsibility until the torch is passed, regardless of what patients do (or don’t do) between the ED visit and the follow up visit.

      Follow up ASAP is my advice.

  2. I see patients with cough/URI symptoms who come to me after three trips to urgent care in a short time frame. Urgent care changes their needless antibiotic three times for their bronchitis, then sends them to a “real” ED. Perhaps the UC docs can read this article and tell their patients up front that there is no quick cure for a cough.
    Meanwhile, when I discharge a patient, I receommend follow up to the primary care doc, not to be seen necessarily, but to at least call and let the office know that they were seen in the ED and ask when a prudent time for an office visit would be.

  3. I have to disagree, Dr. Cynic. While it’s not our job to “protect the system”, it is our job to try and make it better. What you just said was tantamount to “As long as you’re practicing bad law, I’m going to practice bad (or at least not beneficial) medicine”. We have to remember that most patients aren’t litigants – to punish the majority for the sins of a few isn’t right in my opinion. I agree with Dr. Bukata.

  4. Myra Buttreeks on

    While you’re at it – Only perform only necessary documentation: CC and diagnosis. Your risk management dept will love it!

  5. It doesn’t matter if you personally make the appointment for the patient to be seen on Monday morning. If you’re going to get sued, you will get sued with or without the Monday morning appointment. The jury knows, and the lawyers will mock you for asking the patient to follow up in 1-2days. Who can get an appointment in 1-2 days? How many times have you heard your patient saying that they called the office but can’t be seen for another 3 weeks. If you think it’s that important for them to see their doctor in 1-2 days, then you better make sure they’re seen by their doctor or by your co-worker in 1-2 days.

  6. Great Pollyanna position. Fix the legal system and I’ll stop the 1-2 day follow ups. Until then, business as usual. I don’t intend to forfeiture my license to some litigation happy patient.

  7. Well stated Dr. Bukata. This is indeed a big problem. I worry how the retail clinic / acute care telemedicine business may fuel this cognitive dissonance. Nobody should be deprived the opportunity to obtain a Z-pak for a viral illness! Perhaps those follow-up appointments will be busy managing iatrogenia?

  8. Wouldn’t it be nice if patients actually listened. You can tell them what you believe is wrong but in todays society everyone wants everything now, that includes to feel better immediately. People come in to an urgent care and have to pay for a visit and want instant gratification no matter how long or how much you try and explain things to them. Yes, you can ridicule the urgent care centers but at least we try to educate the patient before they leave, what are you doing in the E.R’s. This is a society of “me” and “now”, don’t know how much any of us can change things.

  9. It’s really the same as the offices we refer to, their phone machine tells everyone to call 911 if they think it’s an emergency. Don’t blame us, blame the lawyers. Common sense has taken a backseat to rare but costly lawsuits.

  10. Come on MJS, telling the patient to call their primary care physician and ” at least call and let the office know that they were seen in the ED and ask when a prudent time for an office visit would be” is even more of a cop out. How could the PC doc know when the prudent time would be, much less the receptionist the patient is talking to, if you don’t know? They would have to defensively reply, “ASAP”. Your suggestion is just aimed at spreading the liability risk to colleagues who haven’t even seen the patient for their ailment.
    I agree with JACK and PHXDOC (and Dr. Bukata) to be responsible in your recommendations. Inform the patient of reasonable expectations for recovery, tell him/her to return to you if worse (or certain specific symptoms occur), and suggest a visit to the PC at a realistic future decision-point time, if they are not better.

  11. By the same token, satisfaction-related patient call-backs, i.e., : “how are you doing?”, “are you feeling any better?”, “were you satisfied with your care at our facility?” would be a more meaningful quality marker if postponed for 48-72 hours after a visit. NOBODY is “better” the next day (excluding those who wanted something other than health care in the first place, such as medications or a work note). Lingering malaise undoubtedly influences patient satisfaction scores.

  12. I think Dr Bukata has a great point but it is hard to fight the clipboard nurse policy that says every pt MUST have a recommended follow up with a specific physician. Never mind many of our patients do not need follow up at all unless they have “new or worsening” symptoms, the nurses are told by the hospital that they can not discharge anyone without that. And now our physician group has a policy that EVERY patient with a BP over 140 systolic or 80 diastolic MUST be referred to a PCP within 2 weeks for pre-hypertension, if over 150 syst or 90 diast, within one week!

Leave A Reply