Do Services and Staffing in Residential Care Facilities Vary With Residential Needs?. 5.3.3. Staffing Levels With the Facility as the Unit of Analysis


Multivariate analysis was conducted to disentangle the determinants of staffing at the facility level. Only facility-level regression was estimated, as individual characteristics of RCF residents cannot be used to predict overall facility staffing rates. The dependent variable is total direct care staffing HPRD.

To predict the facility-level direct care staffing ratio, we estimated an OLS regression using the following model:

Direct care RCF staffing ratio = f (RCF size, RCF chain status, RCF profit status, RCF Medicaid participation, RCF urban status, % residents with short-term memory problems, high percentage of residents needing help ADLs, presence of a ADRD Unit/RCF serving only adults with ADRD) + error term,

where resident care mix includes: (1) the percentage of a facility's residents who have short-term memory impairments; (2) an indicator of whether more than half of all residents in a facility require assistance with bathing, eating, or transferring; and (3) an indicator whether a facility (a) has a distinct unit, wing, or floor designated as a dementia or Alzheimer's special care unit or (b) serves only adults with dementia or Alzheimer's disease. RCF characteristics include facility size (number of beds), ownership and chain status, rural or urban location, and the provision (or not) of long-term services and supports to Medicaid residents.

Table 6 provides descriptive data about facility characteristics for the variables used in the multivariate analysis.

About half of facilities nationwide are small (4-10 beds); extra-large facilities (100+ beds) represent only 7 percent of all facilities. Medium and large facilities together make up 44 percent of all facilities. Most facilities are for-profit (82 percent), and 38 percent are part of a chain. The NSRCF defines a "chain" as two or more facilities under common ownership or management. Most facilities (81 percent) are located in urban areas. Forty-three percent of all facilities serve at least one resident on Medicaid; half of all facilities serve an impaired population, in which half or more of all residents require help with bathing, eating, or transferring. Eleven percent of all facilities have a distinct unit, wing, or floor designated as a dementia or Alzheimer's special care unit or serve only adults with dementia or Alzheimer's disease.

Table 7 presents the results of the multivariate analysis to predict direct care staffing ratios in RCFs. Overall, the equation explains 13 percent of the variance.

TABLE 7. Multivariate Analysis: Predictors of Facility Direct Care Staffing Ratio

Variables   Beta Coefficient  
(total staff HPRD)
Intercept 4.08 ***
Facility Characteristics --- ---
Number of beds -0.05 ***
Facility is owned by a chain, group, or multifacility system -0.44 **
Facility has private, for-profit ownership 0.42 **
Facility serves Medicaid residents 0.20 ---
Facility is located in a MSA 0.39 ---
Case Mix Characteristics --- *
Percentage of residents with short-term memory problems <0.005 ---
Facilities with more than 50% of residents needing help with bathing, eating, or transferring 1.18 ***
Facility with an Alzheimer's disease and related dementia unit or that serves only adults with Alzheimer's disease and related dementias   0.26 *
Model fit R2 = 0.130 ---
SOURCE: RTI International analysis of the NSRCF.
NOTES: Stat. Sign.: statistical significance. Direct care hours include nursing, personal care aide, and direct administrator HPRD.
*p<0.1, **p<0.05, ***p<0.0001.

This multivariate analysis found that two characteristics are statistically significantly associated with a higher direct care staffing ratio at the 0.05 level or better. Facilities with more than 50 percent of residents needing help with bathing, eating, or transferring tend to have more than an hour average higher care staffing than facilities where such residents compose less than 50 percent of the total census. In addition, for-profit facilities are significantly more likely to have a higher direct care staffing ratio than facilities that are non-profit or government owned.

Two factors were found to be statistically significantly associated with lower staffing ratios. Larger facilities have a significantly lower direct care staffing ratio than smaller facilities. In particular, each additional bed is associated with a 0.05-hour decrease in the direct care staffing ratio, which suggests that a 20-bed increase in bed size is associated with a 1 hour per resident per day decrease in staffing. RCFs that are owned by chains have significantly lower direct care staffing ratios than do individually owned facilities.

Controlling for all other factors, there is no statistically significant difference in direct care staffing ratios between facilities that do and do not serve Medicaid residents. In other words, this analysis found no evidence that facilities serving Medicaid residents have lower staffing levels than facilities not serving Medicaid residents. Nor were there differences by rural or urban location, the proportion of residents with short-term memory problems, the presence or absence of a special dementia unit, or between facilities that do or do not exclusively serve persons with dementia.

View full report


"ResNeed.pdf" (pdf, 875.55Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®