How Many Patients Can a Hospitalist Safely Manage? Evidence on Census, Workload & Patient Outcomes

1. Elliott et al.: probably the most important direct census study
The study I would cite first is “Effect of Hospitalist Workload on the Quality and Efficiency of Care,” published in JAMA Internal Medicine.
Researchers analyzed 20,241 admissions managed by a hospitalist group in a large U.S. academic community hospital system. They measured workload using both daily patient encounters and RVUs.
The most interesting finding involved length of stay.
When hospital occupancy was 75%–85%, LOS increased more sharply once hospitalist census reached approximately 15 patients. At occupancy below 75%, LOS increased from about 5.5 to 7.5 days as workload increased. Costs also rose with increasing workload.
However, the study did not find workload-associated changes in mortality, 30-day readmissions, rapid-response activation, or the other measured outcomes.
What this study actually tells us
It provides evidence that pushing hospitalist workload upward can eventually reduce efficiency.
But it does not prove:
“15 patients is the maximum safe hospitalist census.”
The ~15 figure was a point where LOS became particularly sensitive to workload under certain hospital-occupancy conditions—not a universal patient-safety threshold.
2. Michtalik et al.: hospitalists frequently perceive excessive census as unsafe
Another major U.S. study surveyed 506 hospitalists from academic and community settings.
A striking 40% reported that their typical inpatient census exceeded safe levels at least monthly, and 36% of that group said this happened more than once per week.
The survey also found that 71% worked in programs with some mechanism for high patient volume. The most common approaches included staffing augmentation and fixed census caps.
Hospitalists associated excessive workload with problems including incomplete discussions with patients and families, delayed admissions/discharges, poorer handoffs, delayed follow-up of important test results, and treatment or medication errors.
This is important evidence, although it is survey evidence based on physician reports, not proof that a specific census directly caused those outcomes.
3. What determines a “safe” hospitalist census?
A related study examined potential predictors of safe attending workload. The research reinforces why a single nationwide census number is problematic: hospitalist workload is influenced by characteristics of the patients, physician, hospital, and care team—not simply the number of names on the rounding list.
A meaningful workload assessment therefore needs to consider factors such as:
• starting census;
• number of admissions and discharges;
• patient acuity and complexity;
• ICU responsibility;
• geographic dispersion;
• cross-cover;
• procedures and rapid responses;
• teaching responsibilities;
• APP, pharmacy, case-management, and other support.
Two hospitalists with a census of 16 can consequently have dramatically different workloads.
4. Hospitalist continuity also affects outcomes
Census isn’t the only staffing variable that matters.
Goodwin and colleagues studied 114,777 medical admissions at 229 Texas hospitals using Medicare data. Patients cared for by hospitalists whose schedules allowed greater continuity had better outcomes, including lower 30-day post-discharge mortality, lower readmissions, a greater likelihood of discharge home, and $223 lower 30-day post-discharge costs.
This wasn’t a census study, but it makes an important point for hospitalist program design:
How we staff hospital medicine services can influence outcomes—not just how many patients each physician carries.
5. Census and hospitalist burnout
Workforce sustainability matters as well.
In one academic hospitalist survey, 62% reported burnout. Among respondents, high patient census was cited by 94% and unrealistic workload by 83% as contributors to burnout.
This doesn’t establish a safe patient threshold, but it supports considering workforce effects when hospitals design census expectations.
So what census does the evidence support?
A reasonable interpretation of the U.S. literature is:
Around 12–15: A defensible planning range for many traditional daytime rounding services, assuming ordinary acuity, limited additional admission burden, and adequate support. This is an operational interpretation—not an official national standard.
Around 16–18: Not automatically unsafe. But acuity, admissions, discharges, geography, ICU responsibility, cross-cover, and support become increasingly important.
Persistent 18–20+: There is not enough evidence to call this universally unsafe, but I would consider it a strong reason for a hospital medicine program to monitor workload-related outcomes closely—especially if physicians are simultaneously admitting, managing high-acuity patients, covering ICU patients, or handling complicated discharges.
20+ plus significant admissions/high acuity: Headcount alone understates the workload. This is where an acuity- and task-adjusted workload measure becomes far more meaningful than census.
The 2014 JAMA study is particularly useful because it provides objective outcome data showing that workload can affect LOS and cost, but even its accompanying commentary emphasized the difficulty of finding a single “magic number” for hospitalist workload.
Bottom line
The strongest evidence does not support saying:
“The safe national hospitalist census is 15.”
A more scientifically accurate statement for CME Getaways would be:
U.S. research suggests that increasing hospitalist workload can adversely affect healthcare efficiency and physicians’ ability to deliver optimal care. A census around the mid-teens is an important evidence-informed reference point, but no universal safe census has been established. Patient acuity, admissions, discharges, cross-cover, ICU responsibility, geography, teaching, and clinical support must all be considered when determining an appropriate hospitalist workload.
That is a much stronger position medically—and it would make an excellent evidence-based CME Getaways article because you can target searches such as “safe hospitalist census,” “hospitalist census studies,” “hospitalist workload patient outcomes,” “how many patients should a hospitalist see,” and “hospitalist census patient safety” without overstating the evidence.