In this month's PSNet perspective, I interview Dr. David Meltzer, Chief of the Section of Hospital Medicine and Director of the Center for Health and Social Sciences at the University of Chicago. The focus of our conversation is a pilot program that David launched several years ago. In it, providers referred to as Comprehensive Care Physicians (CCPs) treat patients in both inpatient and outpatient settings.(1)
As one of the founders of the hospitalist movement in the mid-1990s, this seems like a good opportunity to describe the origins of hospital medicine, place the Comprehensive Care Model in context, and reflect on some of the key advantages and challenges associated with this new organizational model. In 1996, I coined the term "hospitalist" in a New England Journal of Medicine article written with Dr. Lee Goldman.(2) At that time, the dominant model for providing care to hospitalized patients primarily involved primary care physicians (PCPs). In other words, a patient's PCP remained the physician-of-record when a patient needed to be hospitalized.
We made the point that while this traditional structure had the advantage of providing inpatient–outpatient continuity (the same doctor cared for the patient in both settings), it also had profound disadvantages.(3,4) First of all, with average panel sizes of 1800–2000 patients, PCPs were consumed by their office practices, with little time to spend with their hospitalized patients except for quick visits very early in the morning or late in the day. This challenge continued to grow as the threshold for hospitalization increased; PCPs began to have fewer hospitalized patients, making it even harder to carve out the time to see them in the hospital. In the 1980s, the average PCP had about 10 hospitalized patients at any given time, allowing him or her to block out much of the morning for hospital rounds. By the mid-1990s, the number of hospitalized patients per PCP had fallen to about 2–3 patients on average, meaning that PCPs had to schedule a full day of outpatients without any block of time specifically reserved for hospital rounds.(4)
Second, the mid-1990s witnessed the rise of managed care. Hospitals and physician groups faced new pressures to manage inpatients efficiently. "Bed-days-per-thousand" became the mantra, and the question of whether patients were lingering in the hospital because the PCP was unable to drive inpatient care forward became increasingly germane. Early studies of the hospitalist model clearly demonstrated a significant association with decreased length of stay and lower hospital costs, supporting a strong economic rationale for the new model.(5,6) In fact, a portion of these savings became the source of funds used by hospitals and medical groups to invest in their new hospitalist programs.
Finally, growing pressures to improve quality and safety created a new paradigm which emphasized systems thinking. It seemed logical to believe that improving hospital care might require physician engagement and leadership, which in turn might benefit from a cadre of physicians whose primary clinical focus was hospital-based care and who spent the entire day (and sometimes the night) there.
Implementing the hospitalist model was not without challenges. The most common critique concerned the discontinuity it created between ambulatory and inpatient settings. While hospitalist program leaders and researchers emphasized the importance of improving handoffs, the fact remained that the hospitalist model's benefits (on-site hospital presence, professional focus on hospital care, and improved efficiency) literally required the creation of the inpatient–outpatient discontinuity. It was a feature, not a bug.(4)
One more aspect of the hospitalist model was potentially problematic: the fact that the inpatient physician didn't know the patient. This, of course, was another price of the inpatient–outpatient divide and was not completely new. In modern medical care, we have come to expect that a stranger will often care for us when we require specialized care in the hospital—patients rarely know their emergency medicine physician, their anesthesiologist, their surgeon, or their intensivist before receiving the services of these physicians. While we worried about the tradeoff from the familiar PCP to the hospitalist, we were reassured by early studies that showed that overall levels of patient satisfaction were as high under hospitalist models as they were under the traditional PCP-based model of inpatient care.(7) Although some patients missed their PCP, many others found that the availability of the hospitalist on-site more than compensated for the PCP's absence.
Hospitals, health plans, and most physicians embraced the hospitalist model—making it the fastest growing specialty in modern medical history. Today there are more than 50,000 hospitalists in the United States.(8) Hospitalists are now well-respected members of most medical staffs, and many have taken on local and national leadership roles. Under the Obama Administration, Medicare's top physician leader and the Surgeon General were both hospitalists. At the University of California, San Francisco, the Chief Medical Officer, Chief Quality Officer, Chief Health Information Officer, and Chair of the Department of Medicine are all hospitalists.
With all the successes of the hospitalist model, why did Meltzer even consider creating a new model that reinstated inpatient–outpatient continuity? And is this, in fact, a nostalgic recapitulation of the past, or is this new wine in an old bottle? The answer, as described by Meltzer in the interview, is that much here is new. Meltzer recognized that the inpatient–outpatient discontinuity—which was tolerable for the vast majority of inpatients—was problematic for a small cohort: patients who are hospitalized frequently, often resulting from a combination of chronic medical problems and social comorbidities. Such patients, sometimes referred to colloquially as frequent flyers, may not be well served by the hospitalist model, since any given hospitalist doesn't know them well, and the "voltage drop" between inpatient and outpatient settings may be especially dangerous for these patients. This recognition led him to develop and test the Comprehensive Care Physician Model.
The key to the model is that it is not a wholesale rejection of the hospitalist model, but rather involves a holistic approach to the management of a small number of the most frequently admitted patients. In fact, Meltzer does not challenge the value of the hospitalist model for the vast majority of inpatients. Instead, he has created a system for the sickest and most often hospitalized patients—ones for whom having care provided by the same physician in both the inpatient and outpatient setting might be particularly beneficial. While we await publication of the results from his model in a peer-reviewed journal, early data, including those shared in a recent New York Times Magazine article, indicate that the model is achieving its goals of fewer hospitalizations, lower costs, and high patient satisfaction.(9)
I am not particularly surprised by these findings. Like Meltzer, I believe that frequently admitted patients may not be well served by the hospitalist model. Not only do some of their core social needs go unaddressed but the lack of any one physician seeing the big picture for these patients creates special difficulties, perhaps even a bit of a revolving door. With only about five physicians serving as CCPs in Chicago's program, one wonders how much the success of the model actually depends on the personal characteristics of the doctors themselves.
As I think about how this model can be scaled and disseminated (without the extramural grant funding that helped support the pilot program at Chicago ), I am most concerned about the economics. Part of what motivated the advent of hospital medicine was the need for PCPs to carry full panels of patients, seeing them in 15–20 minute intervals throughout the day in order to survive economically. This workload allows neither the time necessary to provide chronically ill and socially complex patients the care they need in the outpatient setting, nor the time to spend several hours a day in the hospital serving as the physician-of-record for inpatients. In order to make the Comprehensive Care Program work, office visits need to be far more generous in length, outpatient staffing (social worker, advance practice nurse, and the like) needs to be robust, and the physician needs to be able to block out several hours—perhaps an entire afternoon—to round on his or her inpatients.(1) One can do the math: if the average PCP has a panel of 1800 patients to create an income of $250,000 per year, a panel of 200 patients (the size of the CCP panels) would generate approximately $28,000. Meltzer makes the point that this figure is an underestimate, since these particular patients require frequent visits of high complexity, which generates more RVUs [relative value units] than their raw numbers might predict. But regardless of which way you cut it, time costs money, and the money must come from a health care system or health plan that sees a sufficient economic benefit from keeping patients out of the hospital—either because they are poorly insured (and thus cost the hospital money; this is particularly germane if the hospital is full and patients with better insurance are being blocked out) or because the system bears risk for the costs of hospitalization.
The bottom line is, sadly, the bottom line. While I await the publication of the full results of the experiment, to me the CCP model seems like a good idea for a small, very select group of patients at the highest risk for hospitalization. Far from being a threat to the hospitalist model, I can easily see health systems embracing a hybrid model, in which the vast majority of the inpatients are cared for by hospitalists, while a few patients—those at the greatest risk for inpatient admission—are cared for by CCPs. Creating and sustaining such a program will require hiring and retaining a truly special group of physicians, and—probably most importantly—making the economics work.
Robert M. Wachter, MD Professor and Chair, Department of Medicine Holly Smith Distinguished Professor in Science and Medicine Marc and Lynne Benioff Endowed Chair University of California, San Francisco
1. Meltzer DO, Ruhnke GW. Redesigning care for patients at increased hospitalization risk: the Comprehensive Care Physician model. Health Aff (Millwood). 2014;33:770-777. [go to PubMed]
2. Wachter RM, Goldman L. The emerging role of "hospitalists" in the American health care system. N Engl J Med. 1996;335:514-517. [go to PubMed]
3. Wachter RM. An introduction to the hospitalist model. Ann Intern Med. 1999;130(4 pt 2):338-342. [go to PubMed]
4. Wachter RM. Does continuity of care matter? No: discontinuity can improve patient care. West J Med. 2001;175:5. [go to PubMed]
5. Wachter RM, Katz P, Showstack J, Bindman AB, Goldman L. Reorganizing an academic medical service: impact on cost, quality, patient satisfaction, and education. JAMA. 1998;279:1560-1565. [go to PubMed]
6. Auerbach AD, Wachter RM, Katz P, Showstack J, Baron RB, Goldman L. Implementation of a voluntary hospitalist service at a community teaching hospital: improved clinical efficiency and patient outcomes. Ann Intern Med. 2002;137:859-865. [go to Pubmed]
7. Wachter RM. Response to David Meltzer's paper "Hospitalists and the doctor-patient relationship." J Legal Stud. 2001;30:615-623. [go to Pubmed]
8. Wachter RM, Goldman L. Zero to 50,000—the 20th anniversary of the hospitalist. N Engl J Med. 2016;375:1009-1011. [go to PubMed]
9. Tingley K. Trying to put a value on the doctor-patient relationship. New York Times Magazine. May 16, 2018. [Available at]