Metric myopia: does reducing exploitation of medical trainees have 'little effect' on patients?

In "On Sleep Deprivation and Medical Education" Roy Poses wrote that:

It's a sad testimony to our training system that the experience described [overwork, sleep deprivation, psychological abuse, etc.] is little different from the one I went through just about 30 years ago.

Now here come studies that may be used to justify a continuation of such abuses:

Limited effect in reduced hours for MDs in training (link to article)
By Kathy Matheson
Associated Press

Cutting the grueling work hours of doctors-in-training had little effect on reducing patient deaths, according to two large studies.

Death rates dropped in one group of patients in veterans' hospitals but not in three others, researchers reported.

The results come from the most comprehensive national look at work-hour restrictions, which were implemented four years ago to reduce medical errors by tired physicians.

"We were a little surprised," said Kevin Volpp, the studies' lead author and a physician at the Philadelphia Veterans Affairs Medical Center. "We thought that mortality outcomes would improve more consistently."

The studies are in today's Journal of the American Medical Association.

The new work-hour rules limit doctors-in-training to 80-hour weeks. Critics feared they would hurt continuity of care for patients and create a shift-work mentality among doctors. Others supported the limits, saying they might lead to fewer deaths from mistakes.

Before the rules, medical residents often worked 100-hour weeks, with some shifts lasting 36 hours straight. The new limits are still roughly double what other full-time jobs require, but advocates say long hours for doctors-in-training are seen as a traditional trial-by-fire approach that give them needed, intensive experience.

The two studies included 318,000 VA patients and more than 8.5 million Medicare patients at hospitals nationwide. The researchers looked at deaths that occurred within 30 days of hospital admission in the years before and after the rules went into effect in 2003. They compared death rates among hospitals with a large number of residents and hospitals with few residents.

Here is my response, in the form of an Op Ed that is unlikely to be published:

To: inquirer.letters@phillynews.com
Date: 09/05/2007 10:55AM
cc: volpp70@mail.med.upenn.edu
Subject: Re: Limited effect in reduced hours for MDs in training

Sept. 5, 2007

Re: Limited effect in reduced hours for MDs in training

Dear Inquirer,

While I am relieved to see studies showing that “cutting the grueling work hours of doctors-in-training” did not increase patient deaths, I suggest that these studies reflect the old adage about “lies, damn lies, and statistics.”

An endpoint of “patient mortality” seems disingenuous. These studies will undoubtedly be used by reactive forces who wish to profit from the cheap labor provided by medical trainees. The medical “scut work” performed by these trainees, and the cost of labor required to replace it, would be considerable to any hospital.

In fact, non-fatal errors, not mortality rates alone, need a much more careful examination. Patient morbidity and related factors such as severity of illness, length of stay, and number of procedures and treatments required (especially in helping a patient recover from a medical error) need evaluation. Unpleasant but non-fatal adverse events are likely far more common than premature patient death due to errors of omission and commission by exhausted medical trainees on sleepless multi-day shifts.

Let me provide some personal examples.

I have personally seen overtired surgical residents make such errors as falling asleep at the operating table due to exhaustion while holding a large clamp retractor on the aorta, during a procedure to bypass blockages in the major arteries of the leg. This resulted in an aortic tear, massive bleeding, blood transfusions, and an extra-long hospital stay for monitoring. The patient survived.

In another case, I witnessed an exhausted surgical resident insert a long hollow spear too deep. The spear was used to make a canal under the skin for an artificial blood vessel, but was inserted under an anatomic landmark instead of above it, thereby piercing the patient’s abdomen and damaging their intestines. This required major exploratory abdominal surgery and prolonged the patient’s stay a month. The patient also survived.

In medicine, I observed overtired residents making subtle errors such as ordering an intravenous bag of saline mixed with insulin for an out-of-control diabetic patient. With no glucose (sugar) in the infusion, the patient was found comatose several hours later with a blood sugar of about 8 (yes, eight). A rapid infusion of a large amount of high-sugar liquid brought the patient back to consciousness. There were no apparent consequences, but this patient could have suffered major injury.

I’ve observed exhausted interns administer too much narcotic to a jaundiced elderly patient with liver disease, causing them to become unconscious and their blood pressure to fall precipitously, requiring rapid infusion of a narcotic blocker to avoid death. Again, the patient survived.

Need I go on? I am only one observer to the mayhem caused by exhaustion and exploitation of medical trainees. Many others have similar stories they have been disincented to share via the culture of medicine.

There is also the question of physician morale, which affects trainee interaction with patients – probably for the remainder of their career. Being broke, exhausted and abused by comfortable attending physicians does not make for the best long-term empathy and bedside manner. In my own medical internship at MCP [The Medical College of Pennsylvania in Philadelphia, now closed -- ed.], I was often times so tired that I could barely function. As an honors student, this certainly had nothing to do with ability.

Further, I experienced a level of abuse not uncommon in medical training programs. As an example, I was required by an attending to perform a pelvic exam and pap smear on any woman admitted for any reason at night to the medical service who had not had one in the prior year, “or else.” Performing pelvic exams at 3 AM on unfortunate patients with normal minds but severe physical deformities, admitted from Inglis House [a Philadelphia-area facility for people of normal mentation but with severe physical disability- ed.] for urinary tract infection or other non-gynecologic problems, was not my idea of humanistic medicine.

In another splendid example of the non-fatal but very real detrimental effects of cost-cutting, during my internship at MCP I rotated through the Philadelphia VA Hospital, the home of the studies’ lead author, for several months. An intern (one year out of medical school), a medical student, and two or three nurses covered perhaps fifty sick veterans. Many times in the middle of the night, veterans needed an IV replaced.

This task was the job of the intern. As it happened, the Philadelphia VA had only one length of intravenous catheter and it was quite rigid. Interns would often have to stick patients multiple times and waste many of these IV starters before a reliable IV could be placed in a fragile vein of an elderly vet. Being exhausted did not help the odds at all of this often tricky process. (I was so infuriated I put in an employee suggestion for change. Many months after I left the organization I was actually awarded a prize of $200 from the VA for my suggestion to get a larger variety of IV starters in different lengths.) [doing so saved them money -- ed.]

In conclusion, medical errors and the non-fatal “salvage” efforts they necessitate are a somewhat “taboo” subject. Reducing trainee abuses, especially exploitative work hours, are one effort at improving a patient’s chances of the shortest possible hospital stay and minimal discomfort. However, there are major financial incentives to continue the abuse and severe lack of a service mentality. Perhaps recent announcements by payers that “salvage care” due to medical errors will not be reimbursed will help even the financial balance sheet.

Significant work needs to be done to measure nonfatal morbidity-related issues resulting from cavalier practices in medical training. Mortality alone is a vastly inadequate statistic that can lead to further justification for abuses by reactionary forces in healthcare and those with financial conflict of interest.

One obstacle to performing such studies in any detail is the backwardness of Clinical Information Technology in healthcare organizations. We can judge the quality of our high tech manufacturing by internal QA records, but often cannot easily judge the quality of care delivered in acute settings at a fine level of “granularity.” Many of the causes for this Information Technology problem, another “taboo” subject related to inappropriate leadership structures for electronic medical records projects, are outlined in my web site “Sociotechnologic issues in clinical computing: Common examples of healthcare IT failure” at this link.

If my op ed is published, even in abbreviated form, it will be a miracle.

I note that one study did show this:

One study examined VA patients treated for a heart attack, stroke, gastrointestinal bleeding, or congestive heart failure. It found that two years after the [trainee work hour restriction] rules were implemented, mortality improved by 11 to 14 percent in major teaching hospitals, compared with hospitals with few residents.

It seems that less-tired residents may actually save lives, especially when compared to non-teaching hospitals where physicians can be scarce as hen's teeth, especially at night. One can only wonder what changes there were in morbidity-related issues.

I will probably be challenged for providing anecdotal evidence in my Op Ed. However, there is a principle from an unlikely source that should be considered when ignoring anecdotal evidence that many people can identify with:

Ignoring all Anecdotal Evidence
Example: I always get hives immediately after eating strawberries. But without a scientifically controlled experiment, it's not reliable data. So I continue to eat strawberries every day, since I can't tell if they cause hives.

I note that the studies' author is an Assistant Professor of Medicine and an Assistant Professor of Health Care Systems at the Wharton School. Perhaps he can undertake a study of the non-fatal consequences of the work hours of medical trainees that in other cognition-critical professions would probably be considered criminal.

Finally, the blog by "Panda Bear MD" is an excellent site for reading about the challenges of medical training in more depth.

-- SS