Rejected by JAMA in March 2013.

To the Editor:

JAMA’s laudable effort to upgrade medical abstracts [1] represents only a syntactic improvement in communicating quantitative results. It was proposed by editors faced with reviewing multitudes of abstracts submitted to research meetings.

Of greater use to JAMA’s general medical readership, and especially to the innumerable members of the public who read JAMA abstracts online via the Pubmed system, would be a graphical flow chart describing each study’s design. Instead of syntactic sugar, this would provide at-a-glance understanding of what is often the most innovative part of a study.

Such charts are now familiar to readers, having been part of JAMA’s instructions to authors since at least 1998 [2]. However, because they are often laden with details [3], they are themselves candidates for abstraction.

Although Pubmed already supports graphics in its abstract pages, authors would be better specifying these proposed abstract-flow-charts declaratively, e.g. with an XML data description language. Pubmed could then, someday, support searches based on details of study design, thereby fulfilling the hope expressed during JAMA’s introduction of structured abstracts in 1991: to “allow more precise computerized literature searches” [4].

[1] Bauchner H, Henry R, Golub RM. The restructuring of structured abstracts: adding a table in the results section. JAMA. 2013; 309: 491-492.   Pubmed 23385278

[2] Anonymous. JAMA instructions for authors. JAMA. 1998; 279: 69-72.

[3] Paradise JL, Bluestone CD, Colborn DK, Bernard BS, Smith CG, Rockette HE, Kurs-Lasky M. Adenoidectomy and adenotonsillectomy for recurrent acute otitis media: parallel randomized clinical trials in children not previously treated with tympanostomy tubes. JAMA. 1999; 282: 945-53.

[4] Anonymous. Structuring abstracts to make them more informative. JAMA. 1991; 266: 116-117.

First published on WSJ.com on Feb. 22, 2013

What we do right

Like other flight surgeons in the Air National Guard, a sizable chunk of my time has been devoted to occupational health practices. The best thing we do is ensure a safe workplace by instituting protections from noise, radiation, chemicals, particulates and other environmental hazards. Peoples’ jobs should not hurt them – even if they’re training for combat.**

The next best thing we do is give flu shots and other immunizations. After having been laid low in college with viral influenza and its muscle aches, headache, teeth-chattering chills and, of course, cough, it is deeply satisfying to give people armor against that experience and its potentially serious complications. Without doubt, immunizations are one of the greatest inventions in human history and their safety profile, while not perfect, is outstanding.**

What we could do better**

Unfortunately, we do not do everything in our power to frustrate and combat tobacco use. We should, and so should all employers. All employers should prohibit both smoking and chewing on their entire premises, at all times. All employers should provide strong smoking cessation programs and assistance.  If some employees are under age 26, strong programs should discourage the initiation of tobacco use.  All other general health concerns are secondary. Perks, incentives, and similar methods should be used to their utmost. Every employee who quits smoking, or who never starts, is doing the single best thing possible for his or her health. Employers win, too, because absenteeism decreases and health care expenditures decrease.**

Finally, employers should be mindful of their incredible influence on the mental hhealth of their employees (and, therefore, on their substance use, too). Being prescriptive on this topic is difficult, but the Golden Rule would seem to be excellent guidance.**

The opinions stated herein are the author’s, and do not necessarily represent those of the Department of Defense.

 

First published on WSJ.com on Feb. 21, 2013

Curiously, under the old unlimited-work-hour approach, no resident covering a medical ward ever spoke of working a “shift.” Each morning we would go to the hospital, and we would stay until our work was done, whenever that might be. Each day, one of the three interns (first-year residents) was “on call”—admitting new patients, covering the entire ward overnight and maybe, if efficient and lucky, sleeping. With responsibilities so clear and pointed, it was enormously maturing.  Physicians worked “shifts” only in the emergency room, where a different modus operandi—“meet ‘em and street ‘em”—prevailed.

Now that a shift-work mentality pervades teaching hospitals, I wonder what type of physician it produces. Certainly it produces a more self-centered physician, who must be given his or her rest time at the appointed hour, with the status of the patient being secondary. And certainly it provides an opportunity for laziness to bloom, knowing that the wall clock, and not the completion of work, governs one’s departure time from the hospital.

No one would ever say that a well-rested resident is a bad thing. But many people would say there are worse things than a tired resident. Foremost is a resident who does not know what is going on with his or her patients.  The more time a resident spends outside the hospital, the less the resident knows about the patient. This includes potentially pivotal information that rarely goes into the medical record or rises above the subconscious, like whether a patient is a complainer or a stoic, or how the patient’s facial expression changes when she is in distress.

It gets worse. Every time a resident goes home, he or she must hand off the care of his or her patients to another resident. Typically, the outgoing resident will speak to the incoming resident for 30 to 60 seconds about each patient. Inevitably, information is lost during these handoffs, just like making a Xerox copy of a Xerox copy of a Xerox copy. All physicians — and patients — know this problem of “handoff entropy.” Smart people have been trying hard to solve it, but they won’t: like physical entropy, handoff entropy is inherently unsolvable.

Let’s jump to the bottom line. What do studies show about the new schedules and their raison d’etre, patient safety? According to the 2011 report of the ACGME[1] (the people who promulgate resident work-hour requirements), the short, disappointing answer is: No detectable improvement in patient outcomes occurs under the new schedules. Thus, I doubt that even shorter work shifts will produce benefits exceeding the consequent increases in handoff entropy and patient unfamiliarity.

Most worrisome is that we have no way to measure whether the new residency schedules are turning out an inferior long-term product. In addition to increasingly evanescent book knowledge, residency teaches four essential eternities not taught in medical school:

  • Clinical reflexes–recognizing when a patient is seriously ill and needs urgent attention,

  • Humility in the face of nature’s incredible subtlety and complexity–hubris is the physician’s greatest sin,

  • Subservience to the patient’s needs—nothing better teaches that the patient comes first than being awakened from a sound sleep, and

  • Knowing when to seek help—the single most important attribute of a good resident, or any physician.

None of these essential lessons are measurable with standardized tests and none are strengthened by cutting hours.

Inexperienced doctors—indeed, all doctors—will make mistakes no matter how rested they are. Far better, therefore, for medicine to focus on designing work systems that intercept mistakes before harm comes to patients, rather than sacrificing essential elements of residency training to an apparently inscrutable and capricious god of short-term improvements, who has so far rejected the reformers’ offerings.

[1] Chapter 11 of the report is at:http://www.acgme.org/acgmeweb/Portals/0/PDFs/jgme-11-00-69-74[1].pdf

First published on WSJ.com on Feb. 19, 2013

Just to be sure we’re all on the same page, “screening” means we are looking for a disease that already exists in a person, but has not advanced to the stage where it is causing symptoms. Usually, but not always, this means we are using some kind of machine or instrument to look for the disease. Regardless, all screening tests cost money in some way and all have potentials for both medical benefit and medical harm.

Those three qualities – financial cost, medical benefit, and medical harm – are the pillars for any discussion about screening.

If we remove cost from the equation – let’s say our patient is wealthy and happy to pay for anything – questions of over-screening vs. under-screening simply boil down to the patient’s individual taste for balancing the benefit and harm potentials. Ideally, the physician communicates these potentials to the patient clearly and accurately, then lets the patient decide.  Some decisions are easy, some are hard, as shown in Figure 1.

In this idealized, cost-ignoring scenario, over-screening and under-screening really do not exist. Some of us might question a patient’s decision-making, but, ultimately, all of us walk in our own moccasins.  In the real world, however, communicating benefits and harms is not easy[1] and most patients do not want to make decisions completely on their own: They want guidance from a physician, who is accustomed to thinking about medical trade-offs and living with medical complications.

People who make policy about screening have a much more complicated chore, because they cannot ignore cost – it is added as a third dimension to Figure 1. And, just as different individual persons have different tastes for weighing benefit and harm, so, too, individual societies (or segments of societies) have different tastes for weighing reasonable costs vs. excessive costs.

So, here again, it’s really not possible, on an absolute scale, to say whether there is too much or too little screening.

Certainly, however, support for improving our preventive health capabilities is universal.  PSA screening leaves much to be desired because it cannot distinguish between life-threatening prostatic cancers, and those that will ultimately prove harmless. Mammography would be better if it could more reliably distinguish benign from malignant lesions. Although knowledge of the patient’s genome will someday improve the performance of these tests, better biochemical testing is more likely to move us closer to the Holy Grail.

Finally, it is also worth remembering how much screening has already brought us. Apart from the invention of antibiotics, the greatest public health triumph of the 20th century was the reduction in death and disability from heart disease, which had reached epidemic proportions by the 1950s – a rate three-times its level today[2].  No one innovation is responsible for this near-miracle, but the practice of taking blood pressure at every physician visit and the adoption of widespread cholesterol testing were big contributors.  In fact, these screening techniques have been so successful that they have expanded the concept of “disease” to encompass what were once called “risk factors.”  Screening is now, and forever will be, a major part of medicine.

[1] The Foundation for Informed Medical Decision-Making has pioneered better ways to communicate the potentials for risk and benefit to patients.

[2] See figure 1 in: Nemetz PN, et al. “Recent trends in the prevalence of coronary disease: a population-based autopsy study of nonnatural deaths.”  Archives of Internal Medicine.  2008; 168(3): 264-170. Available online at: http://archinte.jamanetwork.com/article.aspx?articleid=413914

Rejected by JAMA in October 2012

To the Editor:

Art succeeds only when different people see different things in the same artwork.

Thus, in Clarence Carter’s striking “Port Huron” painting Dr. Torpy sees “abstractly designed clouds” of ovoid shape and uncertain genesis that raise questions about death, infinity, and consciousness (1).

Meteorologists, however, will see clouds of the uncommon “mamma” type, as found on the underside of an unstable “anvil” cloud. The ovoid, down-hanging shape inherent in their name arises from powerful downdrafts that have reached the base of the anvil. Supercell storms frequently display such features (2), which can serve as a warning to aviators. The internet has many photographs of mamma clouds (3).

(1) Torpy JM. The cover. JAMA. 2012; 308: 744.

(2) Wilcox EM. Clouds. London: Duncan Baird Publishers, 2008. Pages 41-43, 160.

(3)
  http://www.crh.noaa.gov/pah/?n=may2410isolated
  http://dnr.ne.gov/floodplain/mitigation/FEMA1590.html
  http://www.srh.noaa.gov/jetstream/synoptic/mam.htm
  http://www.crh.noaa.gov/ind/other_links.php