Sunday, January 13, 2013

Death by 1000 Needlesticks: The Nocebo effects of Hospitalization

When I read an excellent article by Krumholz in this week's NEJM, a paradigm that has been evolving in my mind and my practice patterns for several years congealed:  hospitals and hospitalizations are, to some extent, bad for you.  In this post, I will extend Krumholz's ideas to conjecture about several modifiable aspects of hospitalization that I think do more harm than good.  Many interventions employed by physicians are thought to be benign or beneficial, but there are at least three ways that they could be subversively harmful:  1.) because they prolong hospitalization which is harmful in the way described by Krumholz as the "post-hospital syndrome"; 2.) because there are unrecognized direct untoward effects of the interventions and the environment in which they are administered; and 3.) because patients have adverse psychological reactions to otherwise benign interventions - that is, nocebo effects of hospitalization.  (See also this recent NYT article on nocebos.)

One of the most notorious bugbears of hospitalized patients is the blood draw.  As my colleagues and I mused in this review of laboratory testing in the ICU, much laboratory testing is unnecessary or wasteful, and may even be harmful. Nonetheless, patients admitted for a whole spectrum of maladies have laboratory studies pre-ordered on admission at distinct intervals:  daily labs in the wee hours of the morning (don't get me started on that one - why on Earth we awaken patients at 4AM to draw blood, and how do we justify it?), interval labs throughout the day to "monitor" things (e.g., hemoglobin in gastrointestinal hemorrhage), labs ordered for that Sisyphean task of electrolyte replacement, and so on.  Beyond the oft discussed topics of wasted resources and the anemia-inducing effects of phlebotomy, what if the act of drawing blood has a nocebo effect on patients?  That is, what if they perceive psychologically or unconsciously that many blood draws implies a greater degree of sickness or a stalled, stuttering, or laggard recovery?  What if the pain of phlebotomy induces neurohumoral responses that impair recovery?  What if being awakened at 4AM every day for a painful nuisance not only disturbs sleep, but impedes response to specific therapy through a nocebo effect?

In this post on MIVF and NPO orders, I ranted about the silliness of routinely keeping patients NPO, but many other routine admission orders may be either directly harmful to patients or may have a nocebo effect.  As an example, consider "Bedrest", one of the most common admission activity orders.  Surely, most patients are harmed more by inactivity in the hospital than they are by getting out of bed, exceptions notwithstanding.  But bedrest needs to be reconsidered not only because of its promotion of deconditioning and catabolism (we have a saying in my ICU:  "Moving is for the Improving - Lying is for the Dying"), but also because it may have a nocebo effect - patients confined to bed in the "sick role" may become victims of a self-fulfilling prophesy - I'm acting sick and bedridden, so I am sick and bedridden.  In any case, it is my firm opinion that bedrest is an activity order that should be reserved for very specific purposes rather than as default.  I opine that "Out of bed TID and/or for all meals" should be the default, and indeed we often incentivise patients to get out of bed by not providing meals to capable patients unless they are sitting in a bedside chair.

Then there is the issue of sleep in the hospital (NYT article here) which has received vastly more commentary and lipservice than it has logical thought and activism.  In the case of sleep in the hospital, physicians have it totally wrong, and this is ironic because one of the most sleep-depriving environments in the hospital, the ICU, is often staffed by sleep specialists familiar with the concept of sleep hygeine, and because much of what is causing the sleep disturbances is happening on the orders of physicians.  Take, for example, the phlebotomy mentioned above that happens every morning at 4AM, or the Q4 hour vital signs, the Q2 hour blood sugar checks, the follow-up CT of the head with a low expected diagnostic yield that is ordered at 7PM and happens at 1AM, complicated polypharmacy medication administration schedules, etc, etc, etc.  All of these things disturb sleep in a way that is difficult or impossible to counteract - a patient simply cannot sleep if s/he is being awakened repeatedly throughout the night.  And our solution?  More medications such as Restoril (Arrest-n-Kill) and Ambien which disturb sleep architecture even in healthy patients, and which are the arch-nemesis of the elderly, the demented, the delirious, and the acutely ill.  For every patient that achieves restful sleep with one of these agents, there is another who enters a state of agitated delirium and runs the risk of a fall or other injury as well as many other untoward effects.  (Incidentally, the FDA just recommended a dosing change of some of these medications for women.)  My solution?  Patients should be given earplugs and a sleep blinder on admission, vital signs should be adjusted for changing acuity, unnecessary labs and tests scaled back, medication regimens simplified, doors closed, noise reduced. Each intervention ordered for hospitalized patients has a neglected but additive contribution to sleep disturbance.  MIVF means a pump is constantly running with inevitable occlusions and beeping;  telemetry, continuous pulse-oximetry and unrealistic ventilator and NIPPV alarms all promote a noisy environment; wires, tubes, and lines interfere with comfortable positioning during sleep; hunger from NPO status disrupts sleep. All interventions in the hospital should ideally undergo interval revision and scaling back to the minimum necessary to minimize sleep disruption and other nocebo effects.  Like Burt Munro's neighbors in the [excellent] movie The World's Fastest Indian, the answer to sleep deprivation from Burt's revving his motorcycle at the crack of dawn is not 30mg of Arrest-n-Kill, it is getting Burt to stop revving his motorcycle at the crack of dawn - or to put in some ear plugs.

So much of what we do in medicine has assumed value, but actual value is increasingly recognized as dubious.  The last decade of evidence based medicine has taught us as much about how what we're doing is harmful or useless as it has taught us about novel therapies.  The message from studies such as low tidal volume, sedation vacations, transfusion, hematopoesis stimulating agents, colloid and insulin administration seems to be that "less is more."  I would argue that this paradigm could and perhaps should be extended to even the most basic aspects of care such as activity orders, diet orders, the intensity of testing, monitoring, and vital signs, and the complexity of medication regimens.  If I'm right, the post-hospital syndrome might simply be an extention of the "in-hospital syndrome," a culmination of many small nocebo effects - that is, death by 1000 needlesticks.

4 comments:

  1. Very interesting article! How much of testing do you think is because hospitals fear missing something and being sued? How much does that play into copious testing?

    ReplyDelete
  2. There is a lot of fear especially among green or insecure physicians that they'll miss something and this does indeed lead to overtesting

    ReplyDelete
  3. My husband has spent many, many nights in hospitals in the last four years, even spent three nights in your ICU last month, and everything you've said about sleep deprivation in the ICU/hospital is right on the money. He has complained for years about not being able to sleep because of all the interruptions. Since we've been home, he is on Ambien, and I don't believe that helps much, and he says it only works for about four hours. Too bad we live in Washington and not Utah, we would very much prefer to keep you as his doctor.

    ReplyDelete
  4. Jackie, thanks for your comments. Feel free to have his doctor call me to discuss (politely.)

    ReplyDelete