Appropriate Patient Census: Hospital Medicine's Holy Grail


Ellis Knight, MD, MBA, FHM, senior vice president for physician and clinical integration at Palmetto Health in Columbia, S.C., recalls conducting root cause analyses after every serious adverse event when he was vice president for medical affairs at a large teaching hospital. “For every one of them—it was just like a broken record—every one of them, the nursing staff or the physicians involved would start the recount by saying, ‘It was a very, very busy day; we had a very high census,’” Dr. Knight says. “When that happens, when you get those, what I call tsunami waves of patients coming into a unit or being admitted at one time, it can really wreak havoc and it can make even the best clinicians get rushed, take shortcuts, and make mistakes.”

Researchers have long studied the consequences of temporary and longer-term workload imbalances for other healthcare providers; a recent in-depth study of one hospital found that the risk of inpatient patient mortality increased during shifts with below-target nurse staffing or higher patient turnover.1

Few studies, however, have specifically examined the repercussions of a patient census that is either too high or too low for a hospitalist service. At many facilities, that census can be influenced by an increasing threshold for hospitalization, meaning that the average inpatient is becoming sicker and more complicated, requiring more time during a hospitalist’s daily rounds. HM providers might report having better or worse electronic health records, support staff, and other ancillary services; different schedules; and mixes of clinical, administrative, and teaching responsibilities.

Most hospitalists feel overburdened with patient census. One practical solution, according to experts, is to use NPs or PAs providers to do patient histories, physical exams, or discharges.

Even then, David M. Mitchell, MD, PhD, a hospitalist at Sibley Memorial Hospital in Washington, D.C., and a member of the SHM Performance Standards Committee, cautions that the ability of a doctor to churn through a higher patient count in no way ensures quality. “You don’t want to confuse efficiency with sloppiness,” he says.

In the absence of clear precedents and solid guidelines, hospitalist groups are struggling to come up with their own formulas for ensuring that workloads balance high productivity with sustainable quality—no easy feat. Nonetheless, first-hand accounts and survey data suggest that more providers are identifying common warning signs and devising tailored solutions to help the rapidly maturing field stay on track.

Numbers Game: Delineating a “Good” Average Census

15-20 The patient census “sweet spot”: Ellis Knight, MD, MBA, FHM

18 Optimal census for physicians at Inpatient Medical Services

15-18 Typical census for HM providers at IPC: The Hospitalist Company A “good” average census: David Mitchell, MD, PhD

15 Average self-reported census: online HM survey led by Henry Michtalik, MD, MPH

11-15 “Appropriate patient census” for full-time hospitalist: 51% of respondents to survey

Henry Michtalik, MD, MPH, assistant professor of medicine at Johns Hopkins University School of Medicine, led one of the only surveys that has directly asked hospitalists how they perceive their own workloads. The survey, conducted through an online community of hospitalists and first presented at HM11, revealed several intriguing findings.2

On average, hospitalists reported seeing about 15 patients per shift or day, not including nights, weekends, or holidays. Apart from a few outliers, the range extended from the low teens to the mid-20s, Dr. Michtalik says. According to the survey, 40% of physicians said that more than once a month, their typical inpatient census exceeded the level that they deemed safe and appropriate for specific work settings; 36.1% of physicians reported that was true more than once per week.

Providers often reported that their average workload contributed to incomplete discussions with patients and families, the ordering of unnecessary tests or procedures, a delay in admissions or discharges, worsened patient satisfaction, poorer handoffs, and other problems. “We might be in a situation where we’re focusing on increasing the number of patients being seen or having high census numbers, which could be, paradoxically, actually increasing the costs of healthcare,” Dr. Michtalik says.

For a recent survey posted on, 51% of respondents picked 11 to 15 as the most appropriate patient census for a full-time hospitalist, while another 35% selected 16 to 20. Far fewer deemed it appropriate to see either more than 20 patients a day or 10 or less, suggesting that hospitalists recognize the need for equilibrium.

A more evolved strategy is to develop hospitalist-only floors, which allow providers to see a higher volume of patients very effectively. Yet another technique is to assign a case manager to a specific provider instead of by disease or floor. That way, a hospitalist facing a high patient census can round with the same case manager and much more effectively direct management resources.

A “Resounding” Success Story

David Yu, MD, MBA, SFHM, FACP, medical director of the adult inpatient medicine service at Presbyterian Medical Group in Albuquerque, N.M., says there’s no “magic number” for an ideal daily patient census, and cautions against fixating on national averages and metrics.

“For example, seeing 15 patients in an inner-city hospital—like we are, where the patients are ill and they have really incredibly high levels of social and medical issues like placement—versus seeing 15 patients in an affluent suburban hospital, it’s comparing apples and oranges,” he says.

When Dr. Yu became medical director in January 2010, he says, “we were in crisis,” with the rounding team’s average patient census ranging from 18 to 20 per day. Some hospitalists weren’t seeing their last patients until 4 or 5 p.m., losing the opportunity for timely discussions with specialists to help reduce their patients’ length of stay. By neglecting to send patients home when appropriate, Dr. Yu says, the hospital was losing thousands of dollars in revenue through the failure to open up beds for new admissions. “That’s the classic example of dropping a dollar to pick up a quarter,” he says.

Dr. Yu and his team launched a comprehensive quality-improvement (QI) project that incorporated unit-based rounding centered on the hospital’s geography, and hired more full-time equivalents. As a result, the service now employs 46 FTEs, making it one of the largest nonacademic HM programs in the country. Meanwhile, the average daily census has dropped to a more manageable 11 to 13 patients, plus a few admissions.

We have clinicians who report that they don’t even get a lunch break. That’s not safe, and that’s not lending itself to a work environment that’s satisfying for the practitioners.

—Ruth M. Kleinpell, PhD, RN, FAAN, FCCM, professor of nursing, Rush University Medical Center, nurse practitioner, Mercy Hospital and Medical Center, Chicago

Most significantly, average length of stay has decreased from 4.9 to 4.6 days with increased patient satisfaction and no significant change in the readmission rate, even as the hospital has added $2.5 million to the contribution margin (the revenue minus the variable costs). “So we took the focus on productivity and just elevated it higher to overall organizational finance,” Dr. Yu says. “We answered the age-old question: Is it better and financially more productive for the organization to lower the average starting census and to pay for the extra physician? And the answer is a resounding yes for us.”

The Flip Side

Adam Singer, MD, CEO of North Hollywood, Calif.-based IPC: The Hospitalist Company, points out that an overly low census can prove just as problematic, contributing to revenue and efficiency concerns. A hospitalist’s core ability to drive a delivery system, he says, requires sufficient exposure to a facility’s range of patients and contact with enough other staff members to propel a process of positive change.

“If you only have a few patients and your rounds are done in an hour, how engaged are you?” he asks.

Dr. Singer says his company’s more than 2,000 HM providers see roughly 15 to 18 patients on any given day. Even so, he says, the appropriate census for each practice can vary widely based on its structure, patient population, and the quality and experience of individual providers.

In a recent survey, 40% of physicians said that more than once a month their typical inpatient census exceeded the level they deemed safe and appropriate; 36.1% said that was true more than once per week.

To ensure the numbers remain in the right range, Dr. Singer says, the company provides “complete transparency across the medical group, so that every doctor in the group sees exactly how many people everybody else is seeing.” If one doctor is seeing six patients and another is seeing 20, the group can self-regulate its census.

IPC also closely monitors a core series of clinical measures to ensure quality, ranging from ACE inhibitor use to length of stay and readmission rates. If one of the clinical measures starts to degrade, Dr. Singer says, the company can spot the problem and provide counseling or staffing assistance to right the ship. Hiring more doctors might be the most effective solution, but if a facility cannot afford more FTEs and quality is diminishing, he suggests collaborating with local primary-care physicians or even a less-busy hospitalist group to help share the load.

Safe Patients, Satisfied Providers

Ruth M. Kleinpell, PhD, RN, FAAN, FCCM, professor of nursing at Rush University Medical Center in Chicago and a nurse practitioner at Mercy Hospital and Medical Center, says each institution needs to do a self-assessment based on clinician feedback. Is the workload manageable? Can the providers take breaks? What do their satisfaction surveys suggest? What are the turnover and burnout rates?

“We have clinicians who report that they don’t even get a lunch break,” Kleinpell says. “That’s not safe, and that’s not lending itself to a work environment that’s satisfying for the practitioners.”

We took the focus on productivity and just elevated it higher to overall organizational finance. We answered the age-old question: Is it better and financially more productive for the organization to lower he average starting census and to pay for the extra physician? And the answer is a resounding yes for us.

—David Yu, MD, MBA, SFHM, FACP, medical director, adult inpatient medicine service, Presbyterian Medical Group, Albuquerque, N.M.

Dr. Mitchell has seen overwhelmed hospitalists defer the care of patients they could normally handle to specialists, which leads to higher costs. Ultimately, Dr. Mitchell says, group leaders, administrators, and staff can all help set the right tone. “In the group I’m with now, there’s positive peer pressure to do the right thing, to be efficient, to communicate,” he says, “and if someone doesn’t do it, then it kind of stands out.”

Truly overwhelmed hospitalists can’t continue working well at an unsustainable pace. “It’s an extremely tricky situation, and I think for me it comes down to working with doctors that I trust and working with an administration that trusts us to say, ‘This is what’s best for patient care,’” Dr. Mitchell says. “And you need to prove that by getting the patient feedback and staff feedback that says, ‘Hey these guys are doing a good job.’”

Dr. Yu says many medical directors see the administration’s chief financial officer as an adversary when they should be working together. That kind of collaboration means coming up with strategies, metrics, and models that a financial department can relate to.

“You can’t just complain,” he says. “If your hospital is losing money, your program is going to shut down. But if you provide bad care, the hospital is going to do badly. Both sides have very legitimate points, and one of the jobs of a good medical director is to bridge those two worlds.”

Once the administration is on board, though, each facility must devise the right remedy for a chronically frenetic workload. John Nelson, MD, MHM, FACP, medical director of the hospitalist practice at Overlake Hospital Medical Center in Bellevue, Wash., says facilities can relieve overworked doctors by relieving them of tasks that other staff members could easily do.

“There are places I go where the hospitalists are doing things like arranging follow-up appointments themselves. That’s just nuts,” says Dr. Nelson, a co-founder and past president of SHM, practice management consultant, and columnist for The Hospitalist. “Or the hospitalists themselves are tasked with printing out a copy of their discharge summary and faxing it themselves.”

Other solutions depend on the makeup of clinical teams. “Do you have the ability to integrate nurse practitioners or physician assistants into the team?” Kleinpell asks. “Because certainly they can help maximize the hospitalist’s efficiency by seeing patients who maybe are less severely ill, or new admissions.”

Calling upon other providers to do patient histories, physical exams, or discharges, she says, also removes some of the burden.

Geographical rounding at one facility where he still occasionally practices, Dr. Knight says, “has made all the difference in the world” in improved efficiency. Responsibilities can be subdivided based on more than geography, too. At Palmetto, a team of nurse practitioners does all of the day-to-day management of stroke patients, helping to provide more standardized, reliable care.

A more evolved strategy, Dr. Singer says, is to develop hospitalist-only floors, which allow providers to see a higher volume of patients very effectively. Yet another technique is to assign a case manager to a specific provider instead of by disease or floor. That way, Dr. Singer says, a hospitalist facing a high patient census can round with the same case manager and much more effectively direct management resources.

Like other hospitalists, Dr. Nelson says hard caps should be considered “only in the most dire circumstances or only when all other options have been exhausted.” Sending patients away during peak times, he says, does nothing to address unusually slow days. Apart from the economic consequences, instituting a cap also can fuel the perception that an HM group isn’t pulling its own weight and raises questions about who else will have to take the group’s patients.

There may not be any one-size-fits-all solution, but observers say they are seeing a growing maturity and sophistication in how hospitals are dealing with patient censuses. At first, facilities may view volume and production as the most important considerations.

“Over time, they realize that’s a self-defeating way to operate because it does lead to more errors, it leads to more complications, it leads to longer length of stay,” says Dr. Knight. Eventually, he adds, most organizations come around to the realization that a more modest number of patients, perhaps 15 to 20 per day, may be more realistic for achieving quality and efficiency.

“Common sense tells you that if you’re running around trying to see 40 patients a day, you can’t just pay attention to the things you need to provide high-quality and efficient care,” Dr. Knight says. “You’re just running around and putting out fires.”

Bryn Nelson is a freelance medical writer in Seattle.

for additional resources visit the free SHM Practice Management Online Resource at

An additive effect

Michael N. Huber, MD, president and CEO of Inpatient Medical Services Inc. in Fairlawn, Ohio, says the day-to-day unpredictability of HM can thwart even the best-laid plans. “You just don’t know day to day when you’re going to have a day where the average is exceeded or it’s far under the average,” he says.

Dr. Huber says the company tries to keep its physicians’ average census at 18, where metrics suggest they’re at their most efficient and productive. Experience consistently shows that once the daily census exceeds 20, it will likely keep climbing. “It’s because the discharge process gets delayed, the throughput gets delayed,” he says. “And then if we add a physician to that mix, either temporarily or permanently, then we see the number drastically go down.”

At one of the company’s 13 practices, patient volume increased so much that the daily staff grew from three to four hospitalists last March to at least five to six by December, requiring both flexible, short-term solutions and long-term planning. One Monday in December, for example, the five hospitalists on duty called for help after unexpectedly dealing with 117 patient encounters. The company intervened by bringing in a regional administrator the next morning, part of the “fireman team” that can step in temporarily put out fires. By that Thursday, total patient encounters had fallen to less than 100, dropping the daily census from more than 23 to less than 17.

“I can be reasonably certain that if we were to leave this practice for a length of time at five physicians, we would be at 120 encounters a day, and our report cards would start to suffer,” Dr. Huber says. Conversely, the added help has had a cumulative effect, making the whole team more efficient.


  1. Needleman J, Buerhaus P, Pankratz S, Leibson CL, Stevens SR, Harris M. Nurse staffing and inpatient hospital mortality. N Engl J Med. 2011;364(11): 037-1045.
  2. Michtalik H, Pronovost P, Driscoll B, Paskavitz M, Brotman D. Impact of workload on patient safety and quality of care: a survey of an online community of hospitalists. J Hosp Med. 2011;6(4):S50.

Next Article:

   Comments ()