Skip to main content

Table 3 Cox multivariable regression

From: A novel integrated dressing to secure peripheral intravenous catheters in an adult acute hospital: a pilot randomised controlled trial

 

Hazard ratio (95% CI)

Univariable

Multivariable

Intervention group (referent control)

0.72 (0.46–1.11)*

0.51 (0.29–0.89)**

Age (1-year increments)

1.01 (1.00–1.02)*

^

Female (referent male)

1.66 (1.08–2.57)**

1.83 (1.09–3.09)**

Weight category (referent underweight):

Healthy

0.64 (0.34–1.21)*

§

Overweight

0.75 (0.39–1.44)

§

Obese

0.43 (0.19–0.99)**

§

Inserted on dominant side (referent no)

1.44 (0.92–2.25)*

§

Skin integrity: fair/poor (referent good)

1.64 (1.02–2.63)**

§

Surgical emergency admission (referent elective/medical)

0.56 (0.30–1.05)*

0.38 (0.17–0.86)**

Wound at baseline (referent no)

1.54 (0.98–2.42)*

2.17 (1.22–3.86)**

Device size 22/24 gauge (referent 18/20 gauge)

2.12 (1.37–3.29)**

§

Placement in accessory cephalic or medial antebrachial vein (referent other)

1.64 (1.04–2.59)**

§

Location: hand (referent forearm)

2.11 (0.96–4.62)*

2.79 (1.05–7.41)**

IV medication received:a

Fluids

1.73 (0.92–3.26)*

§

Ceftriaxone

0.34 (0.06–1.81)*

§

Piperacillin/tazobactam

1.52 (0.86–2.70)*

§

Nothing

0.31 (0.08–1.16)*

^

Device in usea

3.55 (1.09–11.6)**

3.75 (1.09–12.8)**

Additional securementsa:

Elasticised tubular bandage

0.22 (0.08–0.64)**

§

Nil

2.15 (0.84–5.48)*

^

  1. Covariates at univariate p ≥ 0.20 excluded; covariates breaching the proportional-hazards assumption at univariable analysis excluded
  2. CI confidence interval, IV intravenous
  3. *p < 0.20; **p < 0.05
  4. aVariable set up as continuous variable between 0 (never observed) and 1 (observed at every check)
  5. §Variable dropped at p ≥ 0.05 during multivariable model building; multivariable model: n = 227
  6. ^Excluded due to correlation with another covariate