Is the Hospital Bad for Your Health?

Decreasing the risk of readmission to a hospital is a hot area of research. It has been a target for quality improvement ever since a study in the New England Journal of Medicine found that just under 20% of all fee-for-service Medicare patients who had been hospitalized during 2003-2004 were readmitted within 30 days of discharge (. The estimated cost of all the unplanned readmissions (90% were assumed to be unanticipated) was a cool $17.4 billion.

The NEJM study was widely interpreted to imply that these patients received sub-optimal care, quite possibly downright poor care. Either they were discharged before their condition was adequately diagnosed or treated, or the plan for follow-up care was inadequate. Sometimes patients were admitted for one problem, say heart failure, but were found to have a second problem, say anemia, and the primary problem was addressed but the second one was not. Sometimes patients were sent home on the wrong medicines—an important medication had been accidentally omitted during the hospital stay and wasn’t restarted when the patient returned home, or a medication was “held” during the hospitalization for good reasons but should have been resumed on discharge at a lower dose. Or patients were simply overwhelmed when they went home and confused about exactly what pills they were supposed to take and what appointments they were expected to make. In response to the high readmission rates, countless programs and projects have been developed to decrease “avoidable” readmissions, which are both burdensome to patients and costly to society. Some of these programs have reported dramatic successes; most have moved the dial only a little bit.

Now two new reports in the medical literature shift the focus from “readmissions” to all hospitalizations near the end of life. For many patients, hospitalization entails risk (for instance, hospital acquired infection, acute confusion associated with a changed environment). As people approach the end of life, these risks may outweigh any benefit. The benefit/burden balance shifts even more decisively for nursing home residents with advanced dementia, to whom invasive medical interventions are incomprehensible, frightening and, given that they are close to death, at best minimally useful. One of the new studies looks at exactly this group of patients, individuals they for some mysterious reason refer to as having “cognitive issues” when in fact they have severe, endstage dementia. What the investigators found was that 19% of such individuals experienced at least one “burdensome transition” in the last 90 days of life.

The definition of “burdensome transition” is a bit complicated: it includes any kind of transition to a new form of care in the last 3 days of life, or multiple hospitalizations in the last 90 days of life, or lack of continuity in nursing home care in the final 90 days. I could quibble with some of these measures. For example, hospice enrollment in the last 3 days of life (occurring in 4.4% of the nearly 475,000 nursing home residents who died between 2000 and 2007) is less than ideal, but for dying patients surely it’s better than not enrolling in hospice at all. The degree of burden is very different for these patients, say, than for the 4.2% of patients who were hospitalized 3 or more times in the last 90 days of life. But the main point of the article remains valid: some nursing home residents who are close to the end of life are subjected to interventions that are painful (to them) and costly (to society) for little if any benefit to anyone.

The second study asks how many Medicare fee-for-service patients had an inpatient surgical procedure in the year before their death. What they found is that among the 1.8 million Medicare beneficiaries who died in 2008, 32% had an operation in the hospital during their last year of life. In fact, 18% had a procedure in the last month and 8% in the last week of life. Now, it’s important to remember that, as with the previous study, neither the physicians nor the patients could have known with certainty that death would come very soon. Most likely, surgeons offered the procedure in the hope, however slim, that it would improve quality of life (for example by repairing a fractured hip) or prolong life (by opening a narrowed coronary artery). And in fact, surgery may well have been life-prolonging in all those patients who underwent an operation but by definition didn’t make it into the study because they didn’t die. Nonetheless, the study points to the indisputable reality that a large number of older people had surgery, especially individuals in their seventies, only to die within a year.

The message most readers will take away from these studies is that American medicine continues to over-treat patients, fueled by the technological imperative and by our refusal to come to terms with our own mortality, and paid for by Medicare’s generosity. That’s an important message, though one we’ve heard before. But I think there’s a second message as well. If we want patients or their surrogates to consider forgoing hospitalization, patients who are very old and ill and in some cases severely demented, we need to offer them a reasonable alternative. Nursing home patients ought to already have such an alternative available to them: most were hospitalized for infections or dehydration, both of which should in principle be treatable in the nursing home. But evidently 19% of the time (and considerably more often, in some states), we don’t provide patients, even in the nursing home, timely assessment of an acute medical condition. We don’t send a physician or nurse practitioner to evaluate their fever or their lethargy (possible signs of infection and dehydration, respectively). We don’t offer simple blood tests and basic X-rays in the facility. We don’t routinely administer intravenous antibiotics or fluids. We don’t provide vigorous symptomatic treatment of shortness of breath and agitation, which commonly result from infection. Instead, we give patients the alternatives of hospital treatment or no treatment. “No treatment” is not an acceptable option.

Likewise for the case of surgery. If we want patients to contemplate forgoing surgery, we need to offer a viable alternative. We need to explain that medical therapy for a gallbladder infection has a good chance of treating the problem, that surgery to prevent recurrence makes sense for someone with a 10- or 20-year life expectancy, but not for someone with a 1- or 2-year life expectancy. We need to explain that medical treatment for coronary artery disease may well be successful and that placement of a stent or, even more invasive, coronary artery bypass surgery, may be excessively burdensome.

As physicians struggle to develop a better way to care for their elderly patients, they will need to develop creative strategies to keep patients out of the hospital and to prevent those who are admitted from recidivism. To achieve this end, physicians will need to design models of care that offer a robust alternative to the hospital. The troubling truth behind the reports in the NEJM and the Lancet is that much of American primary care offers no acceptable treatment for acute medical problems other than hospitalization. Only hospice provides care that is in time, comprehensive, and compassionate. For those patients who meet the criteria for hospice enrollment and who accept an approach focused exclusively on comfort, hospice is a wonderful program. For the many frail, sick, older individuals who are in the final phase of life but who are either not hospice eligible or who are not ready to forgo all potentially life-prolonging treatment, we need another approach. We have to provide better care in the home and the nursing home—with both qualified clinicians and the many low tech diagnostic and treatment interventions that can make a difference to patients—if we want to give high quality care outside the hospital.