Average length of stay is calculated by dividing total inpatient days by total discharges over the same period. If a hospital recorded 3,000 inpatient days and 500 discharges in a given month, the average length of stay (ALOS) is six days. The formula itself is simple, but nearly every term in it hides decisions that change the final number, sometimes dramatically. How you count a “patient day,” which patients you include, and whether you adjust for how sick those patients were all shape whether the result is meaningful or misleading.
The Basic Formula and What Goes Into It
The standard calculation uses discharge data, not admission data, because a stay’s full duration is only known once the patient leaves. You take all the patients discharged during a defined period, add up every day each one spent as an inpatient, and divide by the number of discharges. Most hospitals compute this monthly or quarterly, and the result is typically reported in days.
The discharge-based approach means that a very long stay that started months ago still gets counted entirely in the month the patient is finally discharged. This can make a single month look unusually high if several long-stay patients happen to leave around the same time. Some analysts prefer an admission-based calculation for certain uses, but the discharge method remains the standard because it captures completed episodes of care.
How a “Patient Day” Actually Gets Counted
The most traditional method is the midnight census: if a patient occupies a bed at midnight, that counts as one patient day. It is straightforward but imperfect. A patient admitted at 11 p.m. and discharged at 6 a.m. the next morning would register as one patient day despite spending only seven hours in the hospital. A patient admitted at 1 a.m. and discharged at 11 p.m. the same day would register as zero patient days by midnight census, since they were never in a bed at midnight, even though they were there for nearly an entire day.
A study assessing the reliability of five different patient-day reporting methods found that overall agreement between routine hospital data and a carefully constructed benchmark was excellent, but two of the methods systematically underestimated patient days. In units with a high proportion of short-stay patients, one method’s bias reached roughly eight percent, enough to distort nursing workload indicators and other metrics derived from patient days.1PubMed. Midnight census revisited: Reliability of patient day measurements in US hospital units For most hospitals, the midnight census is “close enough,” but if you are trying to benchmark units where patients frequently come and go the same day, the counting method matters more than you might expect.
Observation Stays and the Classification Problem
One of the biggest distortions in ALOS comes from how hospitals classify patients. A patient placed under “observation status” is technically an outpatient, even though they may be lying in the same bed, on the same floor, receiving the same nursing care as an inpatient next door. Because observation patients are outpatients on paper, their hours do not enter the ALOS calculation, and their departure is not counted as a discharge.
This matters because observation stays are not brief. A study at an academic medical center found the average observation stay lasted about 33 hours, with roughly 17 percent of observation stays exceeding 48 hours.2PubMed Central. Hospitalized but not Admitted: Characteristics of Patients with “Observation Status” at an Academic Medical Center In children’s hospitals, the effect is even more pronounced. A large analysis of over 625,000 pediatric hospitalizations found that observation encounters made up about a third of all discharges. When observation stays were included alongside traditional inpatient stays, the risk-adjusted average length of stay dropped from roughly 75 hours to about 54 hours, and three-quarters of hospitals shifted by at least one ranking tier in length-of-stay benchmarks.3Pediatrics. Observation Encounters and Length of Stay Benchmarking in Children’s Hospitals
If you are comparing your hospital’s ALOS to a competitor’s, the first question should be how aggressively each facility uses observation status. A hospital that classifies borderline cases as observation will report a higher inpatient ALOS (because the shorter, easier cases are filtered out) but might actually be moving patients through faster overall. You can’t tell from the headline number alone.
Why the Mean Can Be Misleading
Length of stay does not follow a bell curve. Most patients leave within a few days, but a small number stay for weeks or months due to complications, lack of available post-acute care, or complex social situations. This produces a distribution that is heavily skewed to the right, and it means the arithmetic mean is pulled upward by a handful of extreme cases. A single patient with a 90-day stay can shift the monthly average for an entire unit.
Researchers have long recommended using the median rather than the mean for this reason, because the median is far less sensitive to those outliers.4PubMed Central. Hospital length of stay: A cross-specialty analysis and Beta-geometric model A study focused specifically on the statistical challenges of LOS data found that median regression, which identifies factors associated with the middle of the distribution rather than the average, was a suitable alternative to traditional approaches and avoided the need to trim outliers or transform the data.5Medical Care. Analyzing Hospital Length of Stay
In practice, many hospitals report both: the mean for official benchmarking and the median for internal operational discussions. If you are calculating ALOS for quality improvement purposes, looking at the median alongside the mean gives you a much more realistic picture of what a “typical” stay looks like. A large gap between the two is itself diagnostic, signaling that a small number of very long stays may warrant individual review.
Adjusting for Case Mix
A hospital that handles mostly routine outpatient-converted admissions will naturally have a shorter ALOS than a Level I trauma center receiving critically ill patients by helicopter. Comparing their raw ALOS numbers tells you almost nothing about efficiency. This is where case-mix adjustment comes in.
The idea is to weight each hospital’s patient population by the expected resource intensity of the conditions treated. In the United States, this is done primarily through diagnosis-related groups, which assign each admission to a category with an expected length of stay. If your hospital’s case-mix index is high, meaning you treat sicker or more complex patients on average, your expected ALOS will be higher. The interesting metric then becomes the observed-to-expected ratio: is your actual ALOS longer or shorter than what would be predicted given the types of patients you serve?
Constructing a credible case-mix index is not trivial. Early methods focused purely on diagnostic categories, but modern approaches also account for comorbidities, age, and procedure complexity.6PubMed Central. A method for constructing case-mix indexes, with application to hospital length of stay A systematic review of case-mix-based hospital performance measurement concluded that while the case-mix index is useful for risk-adjusted comparisons, it should not be treated as a standalone measure of efficiency.7Mongolian Journal of Economic Review. Diagnosis-Related Group – Based Case-Mix Index and Hospital Efficiency: A Systematic Review A hospital might have a favorable observed-to-expected ratio not because its clinical processes are excellent, but because it discharges patients to lower levels of care earlier, potentially at the cost of higher readmission rates.
The Readmission Trade-Off
Pushing ALOS down is not automatically a win. If patients leave before they are clinically stable, some will bounce back within days. A 14-year study across 129 Veterans Affairs hospitals found that hospitals with a mean risk-adjusted length of stay lower than expected had a higher readmission rate, with roughly a six percent increase in readmissions for each day shorter than expected.8PubMed. Associations between reduced hospital length of stay and 30-day readmission rate and mortality: 14-year experience in 129 Veterans Affairs hospitals Another analysis found that patients whose stays were at least one day shorter than the guideline length for their diagnosis had about a one percentage point higher risk of readmission within 30 days, and even shorter stays amplified the risk further.9Production and Operations Management. Sooner or Later? Health Information Technology, Length of Stay, and Readmission Risk
This creates a genuine tension, especially under payment systems that penalize both long stays and readmissions. Calculating ALOS in isolation, without also tracking 30-day readmission rates alongside it, gives an incomplete picture. Some hospitals now use bundled metrics that account for the total cost and outcome across the initial stay plus any return visits within a defined window.10PubMed. Measuring the hospital length of stay/readmission cost trade-off under a bundled payment mechanism
When Patients Die, the Numbers Get Strange
In-hospital mortality introduces a counterintuitive wrinkle into ALOS calculations. A patient who dies early in a hospitalization has a very short length of stay. If a hospital has high mortality among a particular patient group, those early deaths pull the ALOS down. A hospital with lower mortality keeps those same patients alive and in beds longer, which pushes the ALOS up. Without adjustment, a hospital could appear more “efficient” partly because its patients are dying sooner.
A large international study found that patients in the upper quartile of length of stay had about 45 percent higher odds of dying than those in the lowest quartile, and hospitals with high standardized mortality also had substantially higher proportions of prolonged stays.11PubMed Central. Evaluation of hospital outcomes: the relation between length-of-stay, readmission, and mortality in a large international administrative database An older but influential comparison between New York and California hospitals illustrated the problem sharply: average hospital stays for certain conditions were about twice as long in New York, and inpatient mortality was 25 percent higher. But when researchers looked at 30-day mortality instead, California actually had slightly higher death rates. The longer stays in New York were catching deaths that in California happened just after discharge.12JAMA. Assessing Hospital-Associated Deaths From Discharge Data: The Role of Length of Stay and Comorbidities The lesson is that ALOS and mortality should always be interpreted together, and ideally using post-discharge mortality windows rather than just in-hospital deaths.
Discharge Bottlenecks and Weekend Effects
Not all extra hospital days reflect clinical need. A patient who is medically ready for discharge on a Friday afternoon but cannot get a bed at a skilled nursing facility until Monday adds two days of non-clinical length of stay. Weekend discharge delays are a well-documented contributor to prolonged ALOS, driven by reduced availability of specialists, imaging, case managers, and therapy staff on Saturdays and Sundays.13PubMed Central. Reducing Weekend Hospital Discharge Delays Without Seven-Day Coverage by Leveraging Thursday and Friday Planning Surgical patients in particular are discharged far less frequently on weekends, and this pattern is associated with excess length of stay for those who ultimately leave on weekdays.14PubMed. Challenging weekend discharges associated with excess length of stay in surgical patients at Veterans Affairs hospitals
The availability of post-acute care in the community also shapes ALOS. A recent study found that within markets, each additional nurse staffing hour per patient-day at nearby skilled nursing facilities was associated with a roughly 3.5 percent shorter hospital length of stay.15JAMA Network Open. Skilled Nursing Facility Network Capacity and Hospital Length of Stay When waiver programs eliminated the traditional three-day hospital stay requirement before skilled nursing facility admission, the mean hospital length of stay for affected patients actually declined by about 0.7 days, suggesting that the rule itself had been keeping people in hospital beds longer than clinically necessary.16PubMed Central. Waiving the Three-Day Rule: Admissions and Length-of-Stay at Hospitals and Skilled Nursing Facilities did not Increase When you spot an unexpectedly high ALOS on a unit, discharge bottlenecks are often a more productive place to investigate than clinical decision-making.
How Hospitals Compare Internationally
If you want context for whether a particular ALOS figure is high or low, international benchmarks help set the range. Across 36 OECD countries with comparable data, the average length of stay for acute care in 2023 was 6.5 days. The shortest stays were in Turkey and Mexico, at 4.7 days; the longest were in Japan, at 15.7 days. Most countries have seen their ALOS decline since 2019, though the United States and the United Kingdom bucked the trend with increases of over half a day on average.17OECD Publishing. Health at a Glance 2025: OECD Indicators
The downward trajectory has been playing out for decades. In the United States, ALOS fell steadily through the 1970s and then accelerated sharply in the mid-1980s after Medicare adopted prospective payment, which replaced per-day reimbursement with a fixed payment per diagnosis, creating a direct financial incentive to shorten stays.18PubMed Central. Trends in length of stay for Medicare patients: 1979-87 Across 25 European countries over two decades, average LOS dropped from about 9.2 days in 2000 to about 7.2 days in 2019. The study also found that how a country’s health system is organized and how hospitals are reimbursed both independently affect ALOS, with social health insurance systems and certain payment models associated with shorter stays.19PubMed Central. Analyzing the 20-year declining trend of hospital length-of-stay in European countries with different healthcare systems and reimbursement models
These numbers are useful as benchmarks but are not directly comparable without adjustment. Japan’s much longer stays partly reflect a different model of care that includes more rehabilitation and convalescence within the acute hospital rather than transferring patients to separate facilities. Comparing raw ALOS between countries without understanding these structural differences is a common mistake.
Special Populations Need Different Benchmarks
Calculating ALOS for neonatal or pediatric populations requires different reference points entirely. A study of very low-birth-weight preterm neonates who survived to discharge found a median hospital stay of 24 days, with an interquartile range stretching from about 14 to 40 days. Gestational age, the type of initial treatment, and the presence of complications all significantly affected how long those infants stayed.20PubMed Central. Length of hospital stay and factors associated with very-low-birth-weight preterm neonates surviving to discharge a cross-sectional study, 2022 Comparing a neonatal intensive care unit’s ALOS to a hospital-wide average would be meaningless. For specialty units like NICUs, burn centers, and long-term acute care hospitals, you need population-specific and severity-specific benchmarks, and the same case-mix adjustment logic described earlier applies with even more force.
Predictive Models and Hospital-at-Home
Hospitals increasingly use machine learning models to predict length of stay at the point of admission, rather than only calculating it after discharge. These models draw on diagnosis codes, age, comorbidities, lab values, and other data available in electronic health records to estimate how long a patient is likely to stay. A study exploring this approach on a large open health dataset concluded that such models show practical utility for predicting LOS at the time of admission.21PubMed Central. Predicting hospital length of stay using machine learning on a large open health dataset The practical value is in flagging patients at risk for prolonged stays early enough for discharge planning to begin on day one rather than day five.
Hospital-at-home programs add another layer of complexity. Under Medicare’s Acute Hospital Care at Home initiative, patients receive hospital-level care in their own homes. An analysis by CMS found that these at-home episodes actually had a slightly longer length of stay on average than comparable in-hospital episodes for the same diagnoses. However, Medicare spending in the 30 days after discharge was lower for at-home patients across the majority of the most common diagnosis groups.22Centers for Medicare & Medicaid Services (CMS). Fact Sheet: Report on the Study of the Acute Hospital Care at Home Initiative As these programs grow, calculating ALOS for a hospital will increasingly require deciding whether to include at-home episodes in the denominator, and different choices will yield different numbers.
Practical Steps for Getting It Right
If you are responsible for calculating and reporting ALOS at your facility, a few decisions will determine whether the number you produce is actually useful or just technically correct:
- Define your denominator: Decide whether you are including only traditional inpatient discharges, or also observation stays. Document the choice and be consistent over time, because switching methods will create an artificial trend break.
- Report median alongside mean: The mean satisfies most reporting requirements, but the median tells you what is happening for the typical patient. If the two numbers diverge by more than a day or two, investigate the long-stay outliers individually.
- Segment by service line: A hospital-wide ALOS blends surgical, medical, obstetric, psychiatric, and neonatal stays into a single number that describes none of them well. Calculate ALOS by service line or diagnosis group for operational use.
- Use observed-to-expected ratios for comparison: Raw ALOS comparisons between hospitals are nearly useless without case-mix adjustment. If your hospital reports to a benchmarking database, use the O/E ratio rather than the raw number when comparing performance.
- Track readmissions in parallel: An ALOS reduction that comes with a readmission spike is not a real improvement. Report both together.
- Watch for discharge day-of-week patterns: If your ALOS is higher than expected, check whether weekend discharge volumes are significantly lower than weekday volumes. That pattern points to operational fixes rather than clinical ones.
Getting the formula right takes 30 seconds. Getting the context right around it is where the real work lives, and where the number becomes something you can actually act on.