| Home / Blog
Trends in critical care beds and use among population groups

Trends in critical care beds and use among population groups

Posted by-Lawerslog
Member Since-29 Dec 2015


All data from U.S. hospitals was obtained from the Healthcare Cost Report Information System Master Files. The Centers for Medicare & Medicaid Services (CMS) maintains HCRIS, which is a collection of annual submitted federally mandated hospital cost reports. We focused on pediatric and non-federal hospitals with CCM beds. Federal hospitals were not included (i.e. We excluded federal hospitals (i.e. Department of Veterans Affairs and Department of Defense) because they don't have to submit cost reports.

Only the fiscal year-end final cost report for each hospital was used. Annualization was done for cost reports with a duration of 12 or more months. CCM beds were not included in acute care hospitals. Inpatient days of less than 10 days or inpatient beds for a fiscal year were also excluded. To ensure that all cost reports were included, that reports were not duplicated and that the cost report data was accurate and complete, quality controls were conducted.

CCM and Hospital Beds

CCM and hospital bed data were extracted from CMS 255296 Worksheet S-3 Part I . The beds are the available beds. HCRIS follows the federal definitions of hospital and CCM beds. The following data are available for CCM and hospital beds. "Hospital" refers to all inpatient beds (adults, children, babies, and CCM). "CCM" contains aggregate data from five summary CCM groups: 1) total, 2) adult (intensive, coronary, and burn care), 3) child, (pediatric, and premature/neonatal), and 4) pediatric (pediatric), and 5) neonatal/premature/neonatal. In HCRIS, there are no intermediate, progressive or step down bed data. We used data from the US Census Bureau to determine the populations for five CCM beds: 1) US total (all ages), 2) US adult (>=18years), 3) US child (18 years), 4 US pediatric (1-6 years), and 5 US premature/neonatals (1 years).

Hospital and CCM Day Occupancy Rates

Similar data was taken from CMS 255296 Worksheet S-3 Part I for hospital and CCM inpatients and beds available. Inpatient days are the actual number of days used. "Bed days availability" refers to a hospital-based calculation of all possible bed days, based on the number of operational beds. By dividing inpatient days by available bed days, the occupancy rates were calculated.

Both traditional fee-for service and managed care (third-party) Medicare and Medicaid days were available at the hospital level. At the CCM level, however, only traditional fee-for service Medicare and Medicaid days were included. HCRIS only included fee-for services days. CMS also provided us with annual Medicare and Medicaid beneficiary enrollment totals.

CCM costs

We used the modified Russell equation to estimate CCM costs per-day. This is a top-down approach that does not consider patient-level details and allows for broad costing. The "Adjusted expenses for inpatient days" was used as the cost basis. This was calculated annually by AHA for its "nonfederal special hospital category". To represent the CCM-to-non-CCM cost ratio in all study years ( 1), and allow our CCM costs data to be longitudinally tracked over 25 years, we used the 3:3 value. CCM day costs were multiplied with CCM days per annum (HCRIS), to calculate annual CCM cost.

The CCM cost per year was compared to three major US financial indicators: 1) Hospital Care (HC), which is the cost of all inpatient and outpatient hospital care; 2) National Health Expenditures(NHE) – all health care spending; 3) Gross Domestic Product, (GDP), which is the primary indicator of a country’s economic health. We chose not to index the annual cost since the cost per annum was only calculated for 11 years.

Statistic Analyses

Analyses carried out

Health Data Insights (Las Vegas NV) created the HCRIS database using Microsoft SQL Server 2012 & Excel 2010. The variables listed above were summarized as follows: Annual, 5-year (2000-2005, 2005-2010), 11-year (2000-2010) After an initial graphic analysis of the data revealed a non-linear pattern that could lead to multiple outcomes, the 5 year cutoff was selected. The main metric for assessing change over time was the derived percent change in annual and 5-year differences. The average and standard deviation percent change over 11 years were also calculated.

Two ways were used to examine the relationship between CCM beds, US population and CCM beds. We first looked at the slopes that describe the change of beds relative to the change in US population over the five or eleven years. (slope = (beds2010-beds2000) / population2010 - population2000). We also determined the number per 100,000 inhabitants (beds/100K) for each year. The standard reporting capacity of beds/100K is beds/100K. However, the slopes give more detailed information about changes in population and beds over time. We also calculated the percentage of total CCM beds used by child, adult, and premature/neonatal beds.

CCM, Medicare, Medicaid, and total inpatient days were calculated using the slope to compare population/numbers of Medicare and Medicaid beneficiaries. This is a measure of population change and days. The total number of CCM and CCM days was also used to calculate the percentage of CCM, Medicare and Medicaid days.

We had a total of 11 years of data. The sample unit of measurement was the year. Our results are observational at the macro level and focus on potential patterns and trends on the national level. To make the study more clear, all computations were made using exact data from the sources. However, rounded values will be reported for clarity. SAS 9.4 (The SAS Institute Cary, NC) was used for all analyses.

Population changes and changes in CCM beds.

The slope of the CCM bed numbers to population changes is positive. This indicates that CCM beds have increased with increasing population (+576 beds for every million people). Adults had a similar slope (+451 beds for every million adults). The slope for children was higher than that of adults (+451 beds per million adults increase). This is due to slower growth in child population and a decrease in premature/neonatal populations growth. The premature/neonatal slope is positive in the period 2000-2005 because both population and beds grew. However, the slope between 2005-10 is negative due to the decline in premature/neonatal population growth (4.00 to 3.95M) and continued increase of premature/neonatal bed numbers (15,490-18,567).

Changes in Days by Payer Mix in Relation to Population Changes (Slope)

Medicare and Medicaid showed different relationships between changes in days and changes in population over the time. From 2000 to 2010, the slope of hospital Medicare days was negative (800K per million beneficiaries decrease). While Medicare beneficiaries increased between 2005 and 2010, Medicare days fell (62.8M to 53.4M), which resulted in a negative slope of -1.8M per million beneficiary increases. For Medicaid beneficiaries, the slopes were less dramatic. The overall slope for Medicaid days was positive from 2000 to 2010. It averaged 100K per million beneficiaries.


Our study shows that CCM costs, utilization and beds in the US increased from 2000 to 2010. This is the first national study to examine the allocation of CCM resources for age-specific populations. CCM beds increased in both the adult and preterm/neonatal groups between 2005 and 2010. While the increase of CCM beds in adult categories corresponded to an increase in adult population, there was a greater percentage increase in CCM beds in premature/neonatals despite a decrease in adult population between 2005-2010. These results suggest that CCM beds do not increase to accommodate an aging population. Nuanced analyses can be helpful in understanding aggregate data.

CCM beds for premature infants or neonates have seen a paradoxical increase in their numbers. This may be due to the survival rate of premature infants born with congenital anomalies or very low birthweights over the past 20 years. These changes have been made in the obstetric and infant-care practices to help premature infants survive, but they also require more intensive and premature/neonatal ICU beds. The increases in neonatal beds that may exceed their needs could also be due to non-clinical reasons. These include deregionalization and variability in state-based neonatal unit certification, possible overuse neonatal bedding based on special allowances for neonatal care styles or costs, and the inclusion multiple levels of neonatal mattresses (Levels I-IV), within the non-nuanced sheet neonatal bed reporting structure.


Between 2000 and 2010, the US saw an increase in critical care medicine beds and their use. CCM beds saw the greatest growth, despite the decline in this population. In addition, Medicaid beneficiaries are using more critical care services per capita than Medicare beneficiaries. This may explain why CCM beds have been increasing in number. As they deal with the growing number of critical care beds and associated costs, legislators, state-based health departments, and hospitals must account for the higher use of CCM among the Medicaid and premature/neonatal patient populations.