Bed Turnover

Bed Turnover, Turnover Interval and Hospital Capacity

Bed turnover belongs to a set of three linked bed statistics, with occupancy and length of stay, and health economists meet it in productivity comparisons, capacity planning and the costing of bed-days. Two measures share the name, the turnover rate (patients per bed per period) and the turnover interval (empty-bed time between patients), and both follow from occupancy and average length of stay. OECD, WHO Europe and NHS definitions differ in their inputs, the Pabón Lasso and Barber-Johnson charts read the measures together, and in an illustrative hospital the same rise in turnover has a different economic meaning depending on what happens to the freed beds.

Turnover rate and turnover interval

Two quantities are reported under this name, and they answer different questions. The turnover rate counts how many patients each bed served over a period, a measure of throughput per bed and a partial productivity ratio of one output, discharges, to one input, beds. The turnover interval measures the average time a bed stands empty between one patient leaving and the next arriving.

$$T = \frac{D}{\bar{B}}, \qquad \bar{B} = \frac{1}{n} \sum_{d=1}^{n} A_d$$

where $T$ is the turnover rate for the period, $D$ is the number of discharges in the period, including deaths, $\bar{B}$ is the average daily number of available beds, $n$ is the number of days in the period and $A_d$ is the number of beds available on day $d$. Averaging daily counts matters when wards open or close during the year; a mean of the opening and closing bed counts can misstate capacity if beds were closed for part of the period.

$$TI = \frac{n\bar{B} - U}{D}$$

where $TI$ is the turnover interval in days, $n\bar{B}$ is the available bed-days in the period, $U$ is the occupied bed-days and $D$ is the number of discharges as above.

How turnover, occupancy and length of stay fit together

The three statistics are not independent. The average length of stay page gives the throughput identity linking annual discharges to beds, occupancy rate and average stay, and dividing both sides of that identity by the number of beds gives the turnover rate. With occupancy written as $o = U / (365\bar{B})$ and average length of stay as $\bar{L} = U / D$, the relations are as follows.

$$T = \frac{365\,o}{\bar{L}}, \qquad TI = \bar{L}\,\frac{1 - o}{o}$$

where $T$ is the annual turnover rate, $o$ is average bed occupancy as a proportion, $\bar{L}$ is the average length of stay in days and $TI$ is the turnover interval in days. Adding the two gives $\bar{L} + TI = 365 / T$, since each bed cycles through one stay and one empty interval per patient.

These relations hold exactly only when all three statistics come from the same discharge, bed-day and bed counts, as when the OECD definition of average length of stay, bed-days divided by discharges, is applied to the same data as occupancy. National practice departs from this. The Slovak Republic divides treatment days by the average of admitted and discharged patients and removes temporarily non-occupied beds from available bed-days, while Austria counts all bed-days of a stay against its discharge and subtracts nights spent away from hospital. In a steady state they also follow from Little's law in queuing model theory, which states that the mean number of units in a system equals the arrival rate times the mean time each unit spends there; for beds, average occupied beds equal admissions per day times mean stay. When the three come from the same counts, turnover carries no information beyond occupancy and length of stay. A high turnover rate can come from short stays, from full beds or from both, so it is ambiguous on its own.

OECD, WHO Europe and NHS definitions

Statistical bodies define the inputs to turnover more carefully than turnover itself. What counts as a bed or a discharge carries straight into the ratio, so comparisons need matching definitions.

OECD Health Statistics collects curative (acute) care discharges, bed-days, average length of stay and occupancy, with occupancy defined as total bed-days in the year divided by available beds times 365, multiplied by 100. Health at a Glance counts a discharge as the release of a patient who stayed at least one night, including deaths in hospital. Country notes show that bed denominators still vary. Canada, for 1995/96 to 2003/04, used the mean of rated bed capacity at the beginning and end of the year, then, outside Quebec, moved to beds staffed and in operation as a more reliable measure; the Slovak Republic uses a mean of daily bed counts and subtracts bed-days for temporarily non-occupied beds. Luxembourg, until 2022, excluded day cases from discharges but included them when calculating its turnover rate, to stay consistent with its bed count where day-case beds could not be identified.

The WHO Regional Office for Europe's European Health for All database reports acute care beds, discharges, average length of stay and occupancy. Its occupancy indicator is utilised bed-days times 100 divided by available bed-days in the calendar year, and for countries in the joint Eurostat, OECD and WHO Europe data collection it adopts the OECD definition, so turnover derived from these series inherits OECD conventions.

The NHS Wales Data Dictionary, defining its bed indicators by reference to the QueSt 1 (QS1) return, calls the turnover rate "bed use factor" (indicator IP03), defined as deaths and discharges divided by the average daily number of available beds. Its bed turnover interval (IP04) is average unoccupied beds times days in the period, divided by deaths and discharges. In England, the quarterly KH03 collection from NHS organisations that operate beds records available and occupied bed-days by consultant main specialty, so a turnover rate has to combine KH03 bed-days with discharge counts from another source covering the same beds.

Turnover in capacity planning

Planners use turnover to translate expected admissions into beds, and the turnover interval is where the trade-off between efficiency and access shows. Empty-bed time is a buffer against day-to-day swings in emergency demand, and squeezing it out raises turnover while leaving no room for peaks.

Bagust, Place and Posnett simulated emergency admissions to a hypothetical English acute hospital with a discrete-event stochastic model. Risks of having no bed for a patient needing immediate admission became discernible when average occupancy exceeded about 85%, and regular shortages and periodic bed crises could be expected at 90% or more; they concluded that purchasers should bear the cost of spare capacity. In their experiments risk rose as variability in demand rose, with seasonality the most influential factor. At an average stay of 5 days, 85% occupancy leaves a turnover interval of 5 × 0.15 / 0.85 = 0.88 days, while 95% cuts it to 5 × 0.05 / 0.95 = 0.26 days, about six hours. Capacity planning models, including discrete event simulation, represent this variability instead of relying on averages.

Turnover in costing and economic evaluation

Turnover sits behind two steps in hospital costing. In top-down costing, the annual cost of a staffed bed that does not vary with patient numbers is spread over the patients it served, so this cost per case equals annual cost per bed divided by the turnover rate. Higher turnover lowers the average only because the same cost is shared among more patients.

The second step values the bed-days an intervention releases, where Page, Barnett and Graves argued that the economic cost of a bed day, rather than its accounting cost, is the better value for resource decisions, as set out on the average length of stay page. The turnover-specific point is that released bed-days become cash only if variable costs fall or beds close; otherwise their value is the opportunity cost, the benefit to the extra patients that higher turnover lets the hospital admit.

Turnover also appears in efficiency studies. Aloh, Onwujekwe, Aloh and Nweke, in Cost Effectiveness and Resource Allocation, used turnover, occupancy, length of stay and turnover interval to assess teaching hospitals in Southeast Nigeria, reporting mean occupancy of 42.14% against a benchmark of 80 to 85% and mean turnover of 21.27 patients per bed per year. Such ratios set one output against one input, whereas data envelopment analysis estimates technical efficiency from several of each. In Mehrtak, Yusefzadeh and Jaafaripooyan's study of 18 Iranian general hospitals, data envelopment analysis classed 61% as efficient against 44.5% under the Pabón Lasso model.

Reading two indicators at once with the Pabón Lasso and Barber-Johnson charts

Because any two of occupancy, length of stay and turnover fix the third, two charts plot a pair and read the rest from the geometry. Each point stands for all three values.

The Pabón Lasso model, first published in 1986, puts occupancy on the horizontal axis and turnover on the vertical axis, with lines at the average of each dividing the chart into four zones. Since $T = 365\,o/\bar{L}$, each line from the origin joins points with the same length of stay. Low occupancy with low turnover suggests surplus beds; low occupancy with high turnover is typical of obstetric and short-stay units and may reflect unnecessary admissions; high values of both indicate efficient use with few vacant beds; high occupancy with low turnover marks long stays, as in psychiatric care. Mehrtak and colleagues note Pabón Lasso's caution that single-specialty hospitals with extreme values can distort zones drawn at the mean.

The Barber-Johnson diagram plots average length of stay vertically and turnover interval horizontally. Occupancy appears as lines from the origin, since $\bar{L} / (\bar{L} + TI) = o$ is constant along each, and turnover appears as diagonals, since $\bar{L} + TI = 365/T$ gives equal intercepts on both axes.

Worked example of two routes to a higher turnover rate

The figures below are illustrative. A hospital has 400 available beds throughout a 365-day year, 124,100 occupied bed-days and 24,820 discharges, and a staffed bed costs £120,000 a year in costs that do not vary with the number of patients treated.

1. Baseline indicators. Available bed-days are 400 × 365 = 146,000, so occupancy is 124,100 / 146,000 = 0.85. Average length of stay is 124,100 / 24,820 = 5.0 days, the turnover rate is 24,820 / 400 = 62.05 patients per bed, and the turnover interval is (146,000 minus 124,100) / 24,820 = 0.88 days. As a check, 5.0 + 0.88 = 5.88, which matches 365 / 62.05.

2. Route A, shorter stays with freed beds refilled. Average stay falls by 15% to 4.25 days and extra demand keeps occupancy at 85%.

$$T_A = \frac{365 \times 0.85}{4.25} = \frac{310.25}{4.25} = 73$$

where $T_A$ is the annual turnover rate in route A, 0.85 is occupancy and 4.25 is the new average stay in days. Discharges rise to 73 × 400 = 29,200, or 4,380 more patients, and the turnover interval falls to 4.25 × 0.15 / 0.85 = 0.75 days. This bed cost per discharge falls from £1,933.92 (£120,000 divided by 62.05) to £1,643.84 (£120,000 divided by 73), but no cash is saved, and the gain is the health benefit of the extra patients, net of the variable cost of treating them.

3. Route B, shorter stays with the same patients. Without extra demand, discharges stay at 24,820 and occupied bed-days fall to 24,820 × 4.25 = 105,485. Occupancy drops to 105,485 / 146,000 = 0.7225, or 72.25%, turnover stays at 62.05, and the turnover interval rises to (146,000 minus 105,485) / 24,820 = 1.63 days. The 18,615 released bed-days (124,100 minus 105,485) save cash only on variable costs. Closing 60 beds restores 85% occupancy, because 105,485 / (365 × 0.85) = 340 beds, and lifts turnover to 24,820 / 340 = 73. The example assumes the full £120,000 per staffed bed is avoidable when beds close, treating it as semi-fixed, and that beds close in ward-sized steps rather than one at a time; on those assumptions closure avoids £7.2 million a year (60 beds at £120,000 each).

4. Interpretation. Both routes end with a turnover rate of 73, yet route A turns the shorter stay into extra activity and route B into a smaller hospital with a cash saving. The turnover figure alone cannot tell an evaluator which saving to count; the costing has to follow what was done with the capacity.

Where turnover misleads

Several mechanisms raise turnover without any gain. Each readmission is a new discharge, so premature discharge followed by return lifts turnover while adding cost, which is why turnover is best read with the readmission rate. Counting ward transfers or consultant episodes instead of whole stays inflates discharges, and counting day cases in the numerator while excluding day beds from the denominator overstates bed use.

Turnover also falls mechanically as the share of long-stay patients rises, so comparisons across hospitals need case mix index adjustment or comparison within specialties. Averaging hospital-level ratios does not reproduce the ratio of group totals, so pooled figures should come from summed discharges, bed-days and beds. Aloh and colleagues' means illustrate this, since 365 × 0.4214 / 8.15 gives 18.9 patients per bed, against the reported mean turnover of 21.27, because each mean is an average across hospital-years. A negative turnover interval means occupied bed-days exceeded recorded available bed-days, for example when patients occupy escalation beds not counted as available. Fixed benchmarks need care too. Aloh and colleagues cite a suggested ideal turnover interval of 1 to 3 days, but the right buffer depends on how variable demand is, as the Bagust simulation shows.

Sources

• Aloh HE, Onwujekwe OE, Aloh OG, Nweke CJ. Is bed turnover rate a good metric for hospital scale efficiency? A measure of resource utilization rate for hospitals in Southeast Nigeria. Cost Effectiveness and Resource Allocation. 2020;18:21. https://doi.org/10.1186/s12962-020-00216-w
• Bagust A, Place M, Posnett JW. Dynamics of bed use in accommodating emergency admissions: stochastic simulation model. BMJ. 1999;319(7203):155-158. https://doi.org/10.1136/bmj.319.7203.155
• Little JDC. A proof for the queuing formula: L = λW. Operations Research. 1961;9(3):383-387. https://doi.org/10.1287/opre.9.3.383
• Mehrtak M, Yusefzadeh H, Jaafaripooyan E. Pabon Lasso and Data Envelopment Analysis: a complementary approach to hospital performance measurement. Global Journal of Health Science. 2014;6(4):107-116. https://doi.org/10.5539/gjhs.v6n4p107
• Morera Salas M. Diagrama de Barber y Johnson para el análisis de la gestión de la cama hospitalaria en Costa Rica. Revista Costarricense de Salud Pública. 2013;22(1). https://www.scielo.sa.cr/scielo.php?pid=S1409-14292013000100008&script=sci_arttext
• NHS England. Bed Availability and Occupancy (KH03). Statistical work areas. Accessed 1 October 2026. https://www.england.nhs.uk/statistics/statistical-work-areas/bed-availability-and-occupancy/
• NHS Wales Data Dictionary. Indicator IP03 Bed use factor. Accessed 1 October 2026. https://www.datadictionary.wales.nhs.uk/WordDocuments/bedusefactor.htm
• NHS Wales Data Dictionary. Indicator IP04 Bed turnover interval. Accessed 1 October 2026. https://www.datadictionary.wales.nhs.uk/WordDocuments/bedturnoverinterval.htm
• OECD. OECD Health Statistics 2025: Definitions, Sources and Methods. Hospital aggregates: curative (acute) care. Paris: OECD; 2025. https://stats.oecd.org/wbos/fileview2.aspx?IDFile=efc0d5e1-1350-4d26-a1ff-7d17533dede5
• OECD. Health at a Glance 2025: OECD Indicators. Paris: OECD Publishing; 2025. Chapter 5, Hospital activity. https://www.oecd.org/en/publications/2025/11/health-at-a-glance-2025_a894f72e/full-report/hospital-activity_8d89d2a3.html
• Page K, Barnett AG, Graves N. What is a hospital bed day worth? A contingent valuation study of hospital Chief Executive Officers. BMC Health Services Research. 2017;17:137. https://doi.org/10.1186/s12913-017-2079-5
• WHO Regional Office for Europe. European Health for All database: Bed occupancy rate (%), acute care hospitals only (indicator E992913.T). Accessed 1 October 2026. https://gateway.euro.who.int/en/indicators/hfa_542-6210-bed-occupancy-rate-acute-care-hospitals-only/