Table 3.
Cox proportional hazards regression (n = 160)
| Univariable | Multivariable | |
|---|---|---|
| HR (95 % CI) | HR (95 % CI) | |
| 10 mL volume (referent: 3 mL) | 1.56 (0.97–2.50)* | f |
| 6 hr frequency (referent = 24 h) | 1.58 (0.99–2.53)* | f |
| Interaction of volume and frequencya | 0.54 (0.21–1.42) | f |
| Age (one year increase) | 1.00 (0.99,1.02) | f |
| Gender (female) | 1.90 (1.18,3.05)*** | 2.17 (1.33,3.55)*** |
| Weight (over/obese) | 1.29 (0.81,2.05) | f |
| Comorbidities (one category higher)b | 1.03 (0.78,1.35) | f |
| Insertion on dominant side | 1.67 (1.01,2.76)** | f |
| Infection at baseline (any) | 1.56 (0.93,2.59)* | f |
| Antibiotics at baseline (yes) | 0.99 (0.60,1.64) | f |
| IV treatment at baseline (yes) | 0.95 (0.43,2.07) | f |
| Insertion at hand/posterior wristc | 1.67 (1.04,2.70)** | 1.66 (1.02,2.69)** |
| Insertion at forearma | 0.58 (0.35,0.97)** | e |
| Insertion at cub. fossa/ant. upper armc | 1.01 (0.53,1.93) | f |
| Inserted by IV servicec | 0.95 (0.59,1.52) | f |
| Inserted by nursec | 2.01 (1.11,3.63)** | f |
| Size (20 g versus 22 g) | 1.14 (0.70,1.87) | f |
| Vein quality (poor) | 1.26 (0.79,2.01) | f |
| Multiple insertion attempts (yes) | 1.32 (0.78,2.23) | f |
| Extension tubing (yes) | 1.25 (0.78,1.98) | f |
| 3-way tap (yes) | 1.33 (0.79,2.24) | f |
| IV treatment at device removal (yes) | 2.61 (1.28,5.29)*** | e |
| IV treatment rate at device removald | 1.40 (1.07,1.82)** | e |
| IV antibiotics at device removal (yes) | 2.07 (1.27,3.38)*** | e |
| IV antibiotics rate at device removald | 1.33 (1.09,1.62)*** | e |
| Medication types administered: | ||
| - antibioticsc | 1.12 (0.67,1.87) | f |
| - cephalosporinsc | 0.80 (0.49,1.32) | f |
| - penicillin combinationc | 0.75 (0.40,1.41) | f |
| Episode rate (medications and flushes) | 1.22 (1.08,1.38)*** | 1.25 (1.10,1.41)*** |
HR hazard ratio, CI confidence interval, IV intravenous
*p value <0.2; **p value <0.05; ***p value <0.01
aModel for the interaction term results included the main effects
bOrdinal: 0 = none/one, 1 = two/three, 2 = four or more
cVersus all others
dPer day, ordinal: 0 = none, 1 = one/two, 2 = three/four, 3 = five or more
eExcluded due to correlation with other covariate
fCovariate not entered into multivariate model (at p ≥ 0.2), or removed during model building at p > 0.05. Only variables p < 0.2 on univariable analysis entered into the multivariable model