By BOB WACHTER, MD

“Don’t get sick in July!” We’ve all heard patients and family members say this – part declaration, part wishful thinking – in reference to the perceived summertime risks of teaching hospitals. When I hear it, I usually respond with comforting bromides like “robust supervision” and “cream of the crop.” But deep down, if I had the choice of entering a teaching hospital in July and April, I’d choose the latter.

This preference comes partly from my recollections of my own training experience. The day before I began my residency at UCSF, my entire intern class gathered to meet our new bosses. We were on pins and needles – laughing at jokes that weren’t really funny, suspiciously eyeing our colleagues, whose admission to the program (unlike our own) could not be explained by clerical error. The chief of service, Greg Fitz, a brilliant gastroenterologist with a disarming “aw shucks” manner and a Southern drawl (he’s now Dean at UT Southwestern), stood to address us.

“I know you’re all nervous,” he said, catching our collective mental drift. “But don’t worry. If we can turn bread mold into penicillin, we can turn you guys into doctors.”

I was only partly reassured.

The next day, I began my internship on the wards at San Francisco General Hospital. I picked up the service from a graduating intern. His sign out to me was pithy. “Sucker!” he shouted gleefully, as he jammed his beeper into my abdomen like a football. Panicked, I managed to survive my first few weeks on the wards without killing any patients.

About two weeks in, I had my first outpatient clinic. I received a list of my new patients – people who were told that I was their doctor, many no doubt foolishly trusting that this would work out okay. When I endeavored to learn something about these new patients, I found that, while my predecessor had patched together a brief sign-out for a few, in many cases there was nothing: no case summary, no up-to-date problem list, no listing of lab tests or x-rays to follow up, no clues about which patients represented diagnostic dilemmas or were in abusive family relationships or never took their meds.…

That was 1983, and – if anyone had been measuring such things back then – it would not have surprised me to learn that my patients’ clinical outcomes weren’t very good.

We do better today. The internship orientation has lengthened (though it’s in danger of becoming an overstuffed suitcase of an exercise designed to ensure that we’ve met boatloads of regulatory requirements). We try to put our best attendings on the wards in July, and all the supervising faculty and residents are told to Be Afraid, Very Afraid. Outpatient schedules are eased to give new docs more time for chart reviews and first visits. Our interns receive a core “Doctor, Your Patient Has…_______” [shortness of breath, a fever, hypotension, died] curriculum. The attendings are far more available and, on the whole, more helpful to our frail young trainees.

Despite these improvements, I don’t think we’ve fully come to terms with the hazards of July. As my colleague and mentee John Young, a UCSF psychiatrist who is emerging as the nation’s top researcher in this area, has observed,

At year-end, teaching hospitals experience a massive exodus of highly experienced physician trainees who are also familiar with the working environment of the hospital. The ‘July Effect’ occurs when these experienced physicians are replaced by new trainees who have little clinical experience, may be inadequately supervised in their new roles, and do not yet have a working knowledge of the hospital system. It’s a perfect storm.

(Note that in the U.S., our moniker for the risks associated with new residents is “The July Effect,” since that’s when our trainees begin. In the UK, where I’m now on sabbatical, the transition occurs a month later and the phenomenon goes by a more macabre name: the “August Killing Season.”)

Over the past few years, we’ve started paying attention to the safety hazards of all kinds of handoffs. That’s a very good thing. But the ones we focus on – signing out a clinical service after a week on duty or discharging a patient to home from the hospital – are probably less risky than the Mega-Handoff that occurs on or around July 1st each year.

No other industry would tolerate having such crucial responsibilities shifted to an enormous cohort of rookies on the same day. Why do we? Because our teaching hospitals depend on trainees, who all make the transition from student to front-line caregiver (often of the sickest patients, 24-7) simultaneously. In the inpatient world, this means that the patient who was being cared for by a day-365 intern (supervised by a day-365 resident) on June 30th is suddenly in the hands of an intern who, just yesterday, was a medical student. Not only do these newbies know less about medicine (“dyspnea and clear lungs, what could that be?”), many – particularly the ones from outside med schools – are still figuring out where the bathrooms are and how to log into the computers. It’s a dangerous brew.

In the outpatient world, the hazards may be even greater. Here, the transition might not involve an end-of-year R1 handing her patients off to a brand-new intern, but rather a graduating R3 or R4 doing the same. For you chess fans, this isn’t a knight-for-pawn trade, it’s a rook-for-pawn trade, one that puts the whole board at risk.

This is on my mind not only because it’s July (and nearly August, lest I become ill myself while in London) but because Young has produced several important articles that have helped shape our understanding of these critical issues. Last year, he and I published a piece in JAMA describing the year-end transfer in the resident outpatient clinic. We estimated that nearly 13,000 trainees transfer their clinic populations to rookies each July, a handoff that affects as many as two million patients yearly. We laid out a series of recommendations for mitigating this harm, including providing better faculty supervision, promoting the use of web-based sign-out systems, and mandating discussions between outgoing residents and incoming interns about the most challenging patients and complex follow-ups. Gratifyingly, a number of training programs have taken these recommendations to heart, and the ACGME is considering integrating some of them into its accreditation standards.

In this month’s issue of the Annals of Internal Medicine, we published the most complete systematic review to-date of the July effect. In the study, which received wide press coverage (such as here and here), Young and several of our colleagues analyzed 39 English-language papers published between 1989 and 2010. While not every study found a July effect, the better-quality ones (ones with concurrent controls and appropriate risk adjustment) did, with an average 8 percent increase in mortality, along with some evidence of worsening efficiency (longer lengths of stay, higher charges).

In an editorial accompanying the Annals paper, Paul Barach of the University of Ultrech and Ingrid Philipson of the ACGME argued that the time has come to forcefully add this issue to the patient safety agenda:

Patients trust that the teaching faculty members are taking care of them, watching out for their welfare, and coordinating the clinical care delivered by all members of the team. Patients would not consent to care if they knew that the opportunities for residents to learn come at the expense of their welfare… The solutions – such as enhancing supervision, reducing the tempo of the uptake of clinical responsibilities in the first weeks of service, avoiding overnight responsibilities during that period, coupling experienced providers with inexperienced ones in a buddy system, and implementing even more involved interventions (for example, staggering the start dates of trainees over the year) – will greatly enhance individual learning, performance, and patient safety. Simulation, team training, and better ‘on-boarding’ of new trainees centered on the clinical microsystem will help transfer knowledge from departing trainees.

I think that this is a pretty good list of the things we need to start doing – either on our own (which seems unlikely) or prompted by phased-in requirements from either hospital (ie, Joint Commission) or training (ie, ACGME) accreditors. In addition to the procedural and educational changes, as ever much of the battle involves the thorny issues of culture. Since we don’t want and can’t afford a system in which the supervising resident or attending is peering over the intern’s shoulder every minute for a full month, we need to couple having supervisors available by phone/beeper with a culture in which the trainee doesn’t think twice about checking in when he or she is uncomfortable, for fear of “looking weak.” This is Hidden Curriculum 101, and changing this culture will take more than a fiat.

John Young’s work reminds us that there is a learning curve and that – absent system and culture changes, focus, and investment – it probably is more dangerous to be a patient in July than in April. While these are scary studies, they also represent the first step toward making next July a little bit safer, for both patients and their newly minted doctors.

Robert Wachter, MD, is widely regarded as a leading figure in the modern patient safety movement. Together with Dr. Lee Goldman, he coined the term “hospitalist” in an influential 1996 essay in The New England Journal of Medicine. His posts appear semi-regularly on THCB and on his own blog, Wachter’s World.