Skip to main content
. 2020 Oct 20;68(12):2721–2726. doi: 10.1111/jgs.16883

Table 2.

Logistic Regression Models of Nursing Homes with PPE Shortage and Staffing Shortage, June 15 through July 19, 2020 (n = 13,445)

Variable Short PPE at Least 1 wk Short staff at least 1 wk
Odds Ratio P value Odds Ratio P value
State in which facility is located
In Northeast PPE Consortium
No [1.00]
Yes 0.82 .021
Has LTC teams with a focus beyond testing
No [1.00] [1.00]
Yes 0.82 .035 0.72 .170
Expanded Medicaid payments to LTC
No [1.00] [1.00]
Yes 1.20 .108 1.01 .939
Adjusted LTC licensing or training requirements
No [1.00]
Yes 0.89 .473
Increased payments or incentives for LTC staff
No [1.00]
Yes 0.89 .623
Offers LTC job match assistance
No [1.00]
Yes 0.70 .064
County in which facility is located
Change in COVID‐19 cases per 100,000 population between May 24, 2020, and June 14, 2020 1.00 .184 1.00 .435
Population >75% urban
No [1.00] [1.00]
Yes 0.95 .493 0.70 <.000
Nursing home characteristics
Nursing home size
Small (<50 beds) [1.00] [1.00]
Medium (≥50 and ≤150 beds) 0.94 .495 0.99 .960
Large (>150 beds) 0.97 .792 1.02 .834
Ownership
Nonprofit [1.00] [1.00]
Government 0.86 .306 1.22 .149
For profit 1.46 <.001 1.05 .619
Chain status
Not part of a chain [1.00] [1.00]
Part of a chain 1.03 .600 0.90 .058
In‐hospital status
Free standing [1.00] [1.00]
Part of a hospital 0.67 .015 0.80 .109
Share of revenue from Medicaid (quartiles)
1 (Lowest) [1.00] [1.00]
2 1.03 .673 1.08 .259
3 0.98 .853 1.09 .385
4 (Highest) 0.96 .653 1.28 .001
Black residents >25% of all residents
No [1.00] [1.00]
Yes 0.89 .068 1.36 .003
CMS overall five‐star rating
Missing rating 0.79 .239 1.16 .447
1 (Lowest) 0.95 .676 1.57 <.001
2 1.02 .837 1.32 <.001
3 0.99 .900 1.20 <.001
4 1.06 .394 1.10 .104
5 (Highest) [1.00] [1.00]
Change in COVID‐19 cases per bed between May 24, 2020, and June 14, 2020 0.68 .318 4.45 <.001
Information from LTCFocus.org missing
No [1.00] [1.00]
Yes 1.12 .493 0.99 .978

Note: The logistic regression models were estimated using robust standard errors clustered at the state level. Information from LTCfocus.org (chain status, in‐hospital status, Medicaid revenue quartile, and Black resident share) was missing for 754 facilities. Facilities missing LTCfocus information were assigned a value of zero for these variables, and the models included an indicator variable for missing LTCfocus information.

Abbreviations: CMS, Centers for Medicare and Medicaid Services; LTC, long‐term care; PPE, personal protective equipment.