Nursing Home Compare Star Rating and Daily Direct-Care Nurse Staffing During Hurricane Irma

Abstract

Objectives:

Nursing homes (NHs) are affected by major hurricanes and other natural disasters. To mitigate adverse effects of a major hurricane, NHs often increase their direct-care nurse staffing levels to meet the needs of their residents. However, the quality rating of the NH may affect the resources available to obtain and retain staff. This data brief provides estimates of direct-care nurse staffing levels by quality star rating during Hurricane Irma.

Design:

Retrospective cohort study from September 3, 2017, to September 10, 2017.

Setting and Participants:

570 Florida NHs that sheltered in place during Hurricane Irma.

Methods:

We stratified NHs by their NH Compare overall quality star rating and then measured change in direct-care nurse staffing levels for registered nurses, licensed practical nurses, and certified nursing assistants.

Results:

We found that the NH Compare overall star rating was positively associated with a greater staffing level response during Hurricane Irma among registered nurses, licensed practical nurses, and certified nursing assistants. This change was largest for 5-star facilities and smallest for 1-star facilities.

Conclusions and Implications:

Higher-quality NHs may be more responsive and have the resources to be more responsive, to increased needs during a natural disaster. Our findings may serve as a platform for ongoing discussion on the role of the federal, state, and local governments in ensuring minimum staffing standards during natural disasters.

Nursing homes (NHs) are uniquely affected by hurricanes and other natural or human-made disasters. Several studies have reported an increased rate of hospitalization and mortality for NH residents following a major hurricane event.^1–6 NHs often increase staffing levels to adequately meet the needs of their residents during a natural disaster.⁷ However, less is known about the relationship between facility quality and staffing-related response during a major hurricane event.

NHs are rated by the Centers for Medicare & Medicaid Services (CMS) through its Five-Star Quality Rating System, results of which are made available on CMS’s NH Compare website. Higher star values suggest better quality on a set of objective components that include an overall quality rating from 1 to 5 stars. The overall rating represents facility performance in 3 categories that receive their own star ratings, health inspections, resident quality measures, and overall staffing comprising registered nurse staffing and total direct-care staffing.⁸ Research suggests that 1-star rated facilities have heightened risks of resident mortality⁹ and preventable hospitalizations.¹⁰ Although the 5-star system was developed to improve quality by increasing the public’s ability to choose a higher-quality provider, some have suggested it has widened disparities in resident health as consumers with greater resources differentially select higher-quality facilities.^11,12 This differential selection has led to larger financial resources flowing to high-quality NHs (eg, reimbursed through Medicare or privately paid), whereas low-quality NHs are reliant on the smaller margins of Medicaid reimbursement.^13,14 It is possible that having greater resources could play a role in an NH’s response to a disaster, as staffing-related expenses for natural disasters are not routinely covered by states.¹⁵

Therefore, we examined whether the NH Compare overall star rating was associated with direct-care nurse staffing levels in Florida NHs during Hurricane Irma. Hurricane Irma made landfall on the west coast of Florida on September 10, 2017. Given its track and size, the category 4 storm affected nearly the entire state. We hypothesized that higher-quality NHs exhibited a more robust response to the increased direct-care staffing needs during Hurricane Irma than lower-quality NHs.

Methods

We used data from Payroll-Based Journaling (PBJ), Certification and Survey Provider Enhanced Reports (CASPER), and the FLHealthStat reporting system. CMS provides daily staffing data through the PBJ database, which uses the payroll to validate staffing levels. CASPER provides facility-level characteristics such as profit status, chain status, whether a NH is part of a continuing care retirement community, availability of special care units, and bed size. FLHealthStat (now the Florida Emergency Status System) was an emergency reporting system used by the State of Florida for Hurricane Irma to track healthcare facilities’ evacuation status and is further described elsewhere.¹⁶ NH Compare star ratings were obtained from CMS on the NH Compare website from the December 2018 assessment, which would primarily use historical data from 2017.

Our sample included all Florida NHs reporting staffing information in PBJ during Hurricane Irma (n = 661). Excluded were 10 NHs without matching CASPER facility-level data or NH Compare data. Of the 651 facilities, NHs that evacuated were excluded from the main analysis and reported in supplemental analyses (n = 81). Evacuated NHs were excluded because they increased their staffing levels more dramatically than facilities that sheltered in place,⁷ and very few were rated 1-star (n = 2). This left 570 unique NHs in the final analytic sample. Nonadministrative direct-care nurse staffing [ie, registered nurses (RNs), licensed practical nurses (LPNs), and certified nursing assistants (CNAs), not including administrative RN, LPN, or director of nursing] was originally reported in unadjusted hours in PBJ. These hours included both employed and contract staff. To compare across facilities of different sizes, hours were converted to hours per resident-day by dividing by the resident census on the given day provided by PBJ. For example, 0.50 hours per resident-day is equivalent to 30 minutes of direct-care services per resident on average. To examine the effect of Hurricane Irma on NH staffing levels, change was estimated by extracting staffing levels on landfall (September 10, 2017) and subtracting staffing levels from 1 week prior (September 3, 2017) by the NH Compare overall star rating. Analysis of variance and χ² tests were used to compare means and categorical data for Tables 1 and 2.

Table 1.

Nursing Home Characteristics by Nursing Home Compare Overall Star Rating (N = 570)

Facility Characteristics	1-Star (n = 18)	2-Star (n = 120)	3-Star (n = 66)	4-Star (n = 143)	5-Star (n = 223)
	Mean (SD) or %	Mean (SD) or %	Mean (SD) or %	Mean (SD) or %	Mean (SD) or %
Bed size	123.00 (44.95)	131.95 (50.70)	132.20 (50.42)	124.95 (42.09)	118.26 (48.61)
CCRC	6	3	12	10	9
Chain	72	71	62	66	56
For-Profit	72	80	73	78	72
Any SCU	6	14	24	20	13

CCRC, continuing care retirement community; SCU, special care unit.

All 570 facilities sheltered in place. Omnibus P values from analysis of variance or χ² tests were either trending (chain: P = .045) or not statistically significant (bed size: P = .076, CCRC: P = .12, for-profit: P = .41, SCU: P = .08).

Table 2.

Direct-Care Nurse Staffing Levels 1 Week Prior to Hurricane Irma by Nursing Home Compare Overall Star Rating (N = 570)

Overall Rating	RN HPRD	LPN HPRD	CNA HPRD
	Mean (SD)	Mean (SD)	Mean (SD)
1-star (n = 18)	0.248 (0.306)	0.997 (0.265)	2.581 (0.237)
2-star (n = 120)	0.344 (0.226)	0.797 (0.224)	2.524 (0.223)
3-star (n = 66)	0.378 (0.254)	0.821 (0.251)	2.537 (0.254)
4-star (n = 143)	0.374 (0.228)	0.827 (0.226)	2.563 (0.222)
5-star (n = 223)	0.397 (0.241)	0.834 (0.279)	2.619 (0.301)

HPRD, hours per resident-day.

All 570 facilities sheltered in place. Staffing levels are reported in HPRD. Interpretation: 0.50 HPRD is equivalent to 30 minutes of direct-care services per resident on average. Data came from September 3, 2017, in the Payroll-Based Journaling data set.

Results

Of the 570 Florida NHs that sheltered in place during Hurricane Irma, 39% (n = 223) received an overall rating of 5 stars, 25% (n = 143) received 4 stars, 12% (n = 66) received 3 stars, 21% (n = 120) received 2 stars, and 3% (n = 18) received 1 star. Most facility characteristics did not differ appreciably by quality rating (Table 1). However, lower-quality facilities relied largely on Medicaid reimbursement and cared for larger proportions of young, racially and ethnically diverse, residents. Although resident acuity did not meaningfully differ by star rating, higher-quality facilities cared for a greater number of post-acute residents as measured by the number of admissions per bed (Supplementary Table 1). One week before landfall, RN staffing did not differ across quality ratings (P = .06), but lower-quality facilities provided higher LPN (P = .04) and lower CNA (P = .01) staffing (Table 2).

Change in Staffing Levels at Landfall

There was a positive relationship between the overall star rating and change in RN, LPN, and CNA staffing at landfall, with higher-quality NHs showing greater increases in their direct-care staffing levels (Figure 1). This change in staffing was greatest among facilities rated 5 stars and least among facilities rated 1 star.

Fig. 1.

Average change in direct-care nurse staffing levels during Hurricane Irma among facilities that sheltered in place, by Nursing Home Compare overall star rating (n = 570). Staffing levels are reported in hours per resident-day (HPRD). Interpretation: 0.50 HPRD is equivalent to 30 minutes of direct-care services per resident on average. Change was estimated by extracting staffing levels on landfall (September 10, 2017) and subtracting staffing levels from 1 week prior (September 3, 2017) by the Nursing Home Compare overall star rating. M_difference for RN from 5-star to 1-star: 0.11, 0.07, 0.06, 0.04, −0.004; M_difference for LPN from 5-star to 1-star: 0.24, 0.11, 0.23, 0.14, 0.07; M_difference for CNA from 5-starto 1-star: 0.67, 0.48, 0.57, 0.39, 0.23.

Supplemental Analyses

Supplemental analyses examining 81 evacuated facilities showed similar trends as the main analysis for RN staffing; higher overall star ratings associated with greater increases in RN staffing when Hurricane Irma made landfall (Supplementary Figure 1). No clear trends were found for LPN and CNA staffing.

Discussion

NHs increase their direct-care nurse staffing levels in response to major hurricane events to meet residents’ clinical and psychosocial needs. We found a positive relationship between the increase in direct-care nurse staffing and the overall star rating, suggesting that higher-quality NHs increased direct-care staffing more than lower-quality NHs in response to Hurricane Irma. This change, greatest among 5-star NHs and lowest among 1-star NHs, suggests that 5-star NHs may be more responsive, and have the resources to be more responsive, to increased needs during a natural disaster.

Facilities rated as 1-star provide poor quality of care on average, as exemplified by heightened risks of resident mortality⁹ and preventable hospitalizations.¹⁰ This research suggests that quality concerns in 1-star facilities extend to disaster preparedness and response in terms of availability of direct-care staff during a major hurricane. Prior research has found that NH residents are at heightened risk of morbidity and mortality during hurricane events.^1–6 Our findings suggest that residents in 1-star facilities may be among those at highest risk. One explanation for the greater change in staffing levels in higher-quality NHs could be the mass discharge of post-acute residents, thus inflating the staff-to-resident ratio. However, in sensitivity analyses, such a mass discharge was not evident (ie, the resident census stayed consistent). Therefore, it could be suggested that post-acute residents affected change in staffing levels predominantly through the more generous reimbursements of Medicare or private pay and not through occupancy.

Disparities exist by ethnicity and race among NH residents on a variety of outcomes such as quality of life,¹⁷ quality of care,¹⁸ and readmission rates.¹⁹ Long-term care residents who identify as black or Hispanic/Latinx may also be disproportionately affected by natural disasters because they reside in lower-quality facilities on average.⁹ In our study, 2-star facilities cared for the largest proportion of black and Hispanic/Latinx residents, and 1-star facilities cared for the youngest population. In addition to caring for larger proportions of black and Hispanic/Latinx residents, lower-quality NHs also serve a larger number of socioeconomically disadvantaged residents reimbursed predominantly through Medicaid and are often found in socioeconomically disadvantaged geographic locations.^13,20,21 Therefore, addressing low staffing levels during natural disasters may be especially impactful to black and Hispanic/Latinx residents and those of low socioeconomic status, likely through improved continuity of care (eg, attending to heat-related illness, managing medications and insulin, providing psychosocial support) that would ultimately affect the risk of resident hospitalization and mortality.

Prior work associates greater RN staffing levels (ie, a higher skill mix) with lower avoidable hospitalization rates,²² emergency department visits,²³ and health deficiencies in NHs.²⁴ Whether RN staffing affected hospitalization and mortality rates in NHs exposed to Hurricane Irma is not yet known and should be investigated further.⁴ In a 2001 report for the Congress of the United States, CMS argued, “Depending on the nursing home population, [nurse staffing] thresholds range between 2.4 – 2.8, 1.15 – 1.30, and 0.55 – 0.75 hours/resident day for [CNAs], licensed staff (RNs and LPNs combined), and [RNs], respectively. Although no quality improvements are observed for staffing levels above these thresholds, quality is improved with incremental increases in staffing up to these thresholds.“²⁵ Two decades later, direct-care nurse staffing levels continue to be lower than the proposed thresholds on average, and many NHs contend with fluctuating staffing levels each day and during natural disasters.^26,27 Though direct-care nurse staffing levels were lower than what was originally proposed by CMS as the gold standard in our study, these lower levels may be partially explained by the exclusion of administrative staff who may have engaged in direct-care services in addition to their administrative duties and by inaccurate timekeeping (ie, salaried direct-care staff may be erroneously recorded as 40 hours/wk in PBJ, despite working more hours).

Lower-quality NHs may need resources to access and recruit additional staff during a natural disaster. Work is needed to better understand why higher-quality NHs are able to recruit and retain staff during natural disasters, to what extent this is related to fiscal issues, and how workplace environment and culture may influence decisions to provide care during a natural disaster (eg, offering shelter to family and pets of staff, a culture that encourages workplace loyalty).^28–30 One factor that may influence retention of staff during natural disasters is the full-time or part-time status of the staff. Full-time RNs, LPNs, and CNAs may be more willing to work extra hours during a natural disaster and weather the storm with their main employer. Conversely, part-time staff may be less likely to remain with the facility, or may be asked by several employers concurrently. That is, a CNA working at several NHs may have to decide which facility to commit to, which could be partly influenced by hazard or overtime pay. This nuance of how part-time vs full-time status influences response during a disaster should be further investigated.

Effect of Evacuation

Evacuated facilities increased their staffing levels more dramatically during Hurricane Irma than those that sheltered in place,⁷ likely reflecting the workforce needed to physically transport medically complex residents. In our supplemental analyses, a substitution effect was seen among 2-, 3-, and 4-star evacuated facilities on change in LPN and RN staffing (Supplementary Figure 1). That is, higher-quality facilities (eg, 4-star) were more likely to increase RN staffing whereas lower-quality facilities (eg, 2-star) were more likely to increase LPN staffing. The overreliance on LPNs rather than RNs has previously been associated with a heightened rate of unplanned rehospitalization and emergency department visits.²³ Whether RNs mitigate adverse effects of transfer trauma during natural disasters deserves greater attention.

Limitations

This study was not without limitations. Only two 1-star NHs evacuated during Hurricane Irma, and few Florida NHs received 1 star (n = 20). This small sample size limits generalizability to other states or response to other natural disaster events, and largely prevents formal statistical analysis. In supplemental analyses elsewhere,⁷ the effects of weekday and weekend variability in staffing did not explain the increase in staffing during Hurricane Irma. Although we did not examine director of nursing hours or other administrative RN and LPN hours in PBJ, it is conceivable that some of the direct-care staff provided administrative duties (eg, compiling the Minimum Data Set). In addition to excluding administrative nurse hours in this study, PBJ may not accurately reflect salaried employees and therefore the RN staffing levels may be underestimated.

Conclusions and Implications

We found that the NH Compare overall star rating was associated with a greater staffing level response during Hurricane Irma. This change was especially pronounced for 5-star NHs and least pronounced for 1-star NHs. Our findings may serve to advance discussions on federal, state, and local government roles in ensuring adequate direct-care nurse staffing during natural disasters. Given that a disaster may further increase the challenges of keeping residents safe and meeting medication and other care schedules, regulators may need to consider increasing staffing requirements and expanding the availability of resources during a disaster. NHs with lower staffing levels during “normal” times may be especially underprepared during emergencies and disasters.