Table 4 General characteristics of identified models and methods used to determine required numbers of hospital beds
Model/ method (Reference) | Country | Description |
|---|---|---|
Michigan’s Bed Need model [31] | United States | • Adopted in 1997 by the State-Wide Health Planning Commission • Based on the examination of demographic changes by age group and age-specific rates of hospital care use • Use of the ratio-based method • Suitable for areas and regions (sub-areas and sub-regions) |
Canada | • Presented in a study by The Manitoba Centre for Health Policy (MCHP) • Based on changes in population size • Considers that per capita utilization of hospital services is constant • Considers that changes in hospital bed utilization rates are equal to changes in population size (e.g. a 4% increase in population size should increase bed numbers by 4%) | |
Current Use Projection Model | Canada | • Presented in a study by MCHP • Based on demographic changes (population size, age and sex composition, and region of residence), and on current hospital bed utilization rates (based on three years of data) • Use of the ratio-based method |
Canada | • Presented in a study by MCHP • For the next 10 years, based on demographic changes (population size, age and sex composition, and region of residence) and trends in utilization of hospital services • The revised version of this model cannot project beyond 3 years • Considers that average length of stay and inpatient admission rates are decreasing • Use of Poisson regression | |
Israeli model [34] | Israel | • Similar to the Trends in Acute Care Bed Use model • Based on demographic changes (population size and growth, age and sex composition, and region of residence) and current patterns of hospital service utilization |
Scotland | • Combines top-down and bottom-up approaches • Bottom-up approach: identification of 14 clinical groups by examining care pathways and models of care • Top-down approach: Study of the following eight criteria: performance improvement (hospital goal to become a “top” hospital), bed occupancy rates by specialty, demographic changes (particularly age distribution), shift to community facilities (e.g. for patients with long lengths of stay), waiting times, emergency care (and new methods for emergency patients), increase in number of emergency patients, and geographic flows (patterns of patient flow between hospitals in different regions) | |
Switzerland | • Presented in 2000 and revised in 2009 • Consists of two stages: development of scenarios by area (canton), and estimation of future needs of hospital care based on Diagnosis-Related Groups (DRG) • Development of different scenarios based on key uncertainties (admission rates, average length of stay, demographic changes) • Considers that average lengths of stay will decrease in the next 10 years • Use of the ratio-based method • Suitable for determining bed requirements at the regional level | |
Lausanne University Hospital (CHUV) model [32] | Switzerland | • Modeled after the Swiss Health Observatory (SHO) model • Based on scenarios and key uncertainties (admission rates, average length of stay, demographic changes) • Use of the ratio-based method • Suitable for determining bed requirements at the hospital level |
Basic scenario model [32] | Switzerland | • Uses scenarios based on demographic changes • Use of the ratio-based method • Suitable for determining bed requirements at the regional level |
Capacity model [22] | New Zealand | • Based on a mathematical iterative linear equation, the examination of current hospital bed utilization rates, and factors affecting future rates • Considers trends of demand for services, factors related to demand (population growth, disease prevalence, transfers to or from the private sector), supply-side factors (technological advances, changes in funding, length of stay and patients’ transfers), external factors like inter-regional flows and sub-regional equitable access (SREA) • Prediction of bed requirements based on the cumulative impact of the above factors on baseline bed use for each service • Use of Monte Carlo analysis |
France | • Based on a score constructed with three parameters: bed occupancy rate (measure of efficiency), number of transfers due to lack of beds (measure of clinical effectiveness), and number of days without the possibility for unscheduled admissions (measure of accessibility) • The number of beds is optimal when the mean and standard deviation of this score is the lowest • The number of beds is optimal if the following parameters have a low value: the number of days for which the number of unoccupied beds exceeds a given threshold (efficiency), the number of patients transferred due to the lack of bed availability (clinical effectiveness), and the number of days without the possibility for unscheduled admissions (availability) • Using this model increases availability and clinical effectiveness, but reduces efficiency • Application of a simulation method using software | |
France / UK / Iran / OECD countries | • Introduced by Jung and Streeter in 1977 • Based on the ratio of the total length of stay (average length of stay × number of patients) to period duration | |
United States / United Kingdom / France / Switzerland / Iran / Greece / Brazil / Canada | • Introduced in 1984 • Based on target occupancy rates (80–85% for large hospitals and 45% for small hospitals) • Calculated by dividing the total length of stay (average length of stay × admission rate × projected population size) by (period duration × target bed occupancy rate) | |
Method using the distribution of present patients [21] | France | • Based on the distribution of occupied beds. For example, the proportion of days in which 0–5 beds, 6–10 beds, 11–15 beds, etc. are occupied, and the number of beds occupied on most days, indicates the number of beds needed |
Iran / Canada | • Based on admission rates, discharge rates, average length of stay, and distribution of occupied beds for each day • Used alone or in combination with other methods | |
Canada /Singapore | • Based on the number of occupied beds (dependent variable) as a function of independent variables such as occupied beds in past weeks, admission rates, length of stay, and emergency admissions |