Skip to main content

Table 1 Coefficients derived from MELR and ENET models for strains discriminated at the 80% similarity threshold

From: Acquisition and persistence of strain-specific methicillin-resistant Staphylococcus aureus and their determinants in community nursing homes

Covariates Elastic net (ENET) Mixed effects logistic regression (MELR)
Estimate Estimate (SE)a Bootstrapped 95% CIa OR (95% CI)a (odds for Intercept) p-value
T01 (outcome)
 (Intercept) −3.42 −3.45 (0.29) (−4.46, −2.89) 0.03 (0.01, 0.06) NA
C01 −5.83 NA NA NA NA
Facility 0.09 0.58 (0.00, 1.37) NA NA
C01 (outcome)
 (Intercept) −2.53 −2.10 (0.30) (−3.04, −1.41) 0.12 (0.05, 0.24) NA
T01 −5.04 −20.30 (28.36) (−82.16, −13.96) 0.00 (0.00, 0.00) <0.001
Cogn −0.88 −0.71 (0.30) (−2.95, 0.72) 0.49 (0.05, 2.05) 0.010
Comorb 0.37 0.31 (0.16) (−0.11, 1.04) 1.36 (0.89, 2.84) NAb
Dev 0.26 NA NA NA NA
Wnd 1.11 1.02 (0.21) (0.31, 1.77) 2.77 (1.37, 5.87) <0.001
Facility 0.15 0.66 (0.16, 0.92) NA NA
T03 (outcome)
 (Intercept) −4.83 −5.64 (0.76) (−98.09, −4.25) 0.00 (0.00, 0.01) NA
Comorb 0.46 1.11 (0.59) (−0.99, 91.84) 3.03 (0.37, >1000) 0.047
Facility NA 1.01 (0.00, 5.55) NA NA
C03 (outcome)
 (Intercept) −10.47 −5.14 (0.73) (−29.75, −4.01) 0.01 (0.00, 0.02) NA
AB 0.36 NA NA NA NA
Cogn 1.15 1.34 (0.35) (−24.83, 2.54) 3.81 (0.00, 12.70) <0.001
Comorb 0.63 0.80 (0.34) (−0.50, 23.66) 2.23 (0.61, >1000) 0.018
Dev 1.10 1.40 (0.36) (−20.04, 2.34) 4.06 (0.00, 10.40) <0.001
Func 5.92 NA NA NA NA
Hosp −0.43 NA NA NA NA
Facility 0.14 1.48 (0.32, 3.23) NA NA
C12 (outcome)c
 (Intercept)c −7.40 −7.15 (1.35)/−34.55 (95.6) (−783.01, −5.29)/(−715.76, −26.84) 0.00 (0.00, 0.01)/
0.00 (0.00, 0.00)
NA
Cogn 2.33 2.41 (0.88) (−447.79, 772.85) 11.16 (0.00, >1000) 0.005
Comorb 1.95 28.59 (95.60) (21.23, 704.66) >1000 (>1000, >1000) 0.003
Facility c −0.06 1.30/1.75 (0.00, 6.81) /
(0.00, 7.39)
NA NA
  1. Covariates of each outcome variable (strain-specific acquisition or carriage events) represent Markov blankets of these variables. Covariates that were included in the Markov blankets of the outcome but not selected by the ENET or not statistically significant at the 95% confidence level in the MLER models are denoted by NA. The values in the OR and p-value columns that correspond to random effects terms of MELR models and the p-values for intercept terms are also denoted by NA’s. The p-values were obtained from the likelihood ratio test comparing two nested models, with and without the respective term
  2. Logistic regression coefficients of the fixed effects terms (potential risk factor terms and acquisition and carriage of MRSA) are provided for level 1 (reference level is 0). Facility was used as a covariate in ENET models, and its coefficient is provided in the corresponding column. Standard deviations of random intercepts are provided in the columns that correspond to MELR
  3. aMELR estimates and standard errors were derived from fitting the models to the data. Bootstrapped 95% CI’s are percentile CI’s obtained from multilevel bootstrapping with 5000 replications. OR’s are exponentiated model estimates, and respective 95% CI’s are exponentiated bootstrapped 95% CI’s
  4. bUnable to estimate the p-value from the likelihood ratio test comparing two nested models because the model without the term does not converge
  5. cThe model with both Cogn and Comorb included did not converge. Two simpler models, each including one of the terms, were fitted. The first entry in the (Intercept) and Facility columns represent the values from the model with Cogn, and the second entry, separated by “/”, represent the values from the model with Comorb
  6. T01, acquisition of non-USA300; C01, carriage of non-USA300; T03, acquisition of USA300; C03, carriage of USA300; C12, carriage of USA1200;
  7. AB, antibiotic use in the past 3 months (0 = Non-exposed, 1 = Exposed); Hosp, hospitalizations in the past 3 months (0 = Non-exposed, 1 = Exposed); Dev, presence of invasive device (0 = Non-exposed, 1 = Exposed); Wnd, presence of wound (0 = Non-exposed, 1 = Exposed); Comorb, comorbidity (0 = Non-severe, 1 = Severe); Func, functional status (0 = Non-severe, 1 = Severe); Cogn, cognitive status (0 = Non-severe, 1 = Severe);
  8. OR Odds ratio, CI Confidence interval, SE Standard error