L'Ecuyer 1996 mbc.

Methods	Study design: Cluster Randomised Controlled Trial. Object of randomisation: nursing divisions. Three‐armed trial with separate control groups
Participants	USA. Nursing personnel from general, medical, surgical and intensive‐care units performing intravenous therapy. Number studied: 73,454 patient‐days (980,392 productive hours worked). Intervention two n = 3840. Control n = 2487 patient‐days.
Interventions	Use of needleless intravenous device metal blunt cannula. Passive system no need for activation. Control groups used standard IV needle systems.
Outcomes	Reported needlestick injures per 1000 patient‐days and 1000 productive hours worked.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Four groups of nursing divisions were prospectively randomised to use one of the two safety devices"
Allocation concealment (selection bias)	Unclear risk	No information about allocation concealment is available.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No information available on blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Selected nursing division were assigned to either the intervention (MBC then replaced by PBC, and 2‐way). The MBC was replaced after four months due to staff dissatisfaction. Authors reported all outcomes data for the intervention and control group.
Selective reporting (reporting bias)	Unclear risk	"Intravenenous‐therapy related injuries were categorized further as follows: low‐risk injuries involved needles without direct blood contact; intermediate risk injuries involved needles likely to have occult blood present and high risk injuries involved needles in direct contact with blood." However, there is no information available based on this categorization stipulated in the method section.
Similar recruitment of groups	High risk	The nursing divisions selected to participate to the study were from the same hospital. The recruitment time period of 2‐way device differed from the PBC. The PBC was selected to replace the MBC (after four months) due to staff dissatisfaction.
Adjustment for baseline differences	Unclear risk	The demographics of the workers (age, sex, years of experience) are not reported. The adjustment for baseline differences is not reported in the analysis.
Other bias	High risk	"Study participants generally have ready access to the traditional devices, which may contaminate the evaluation, so much attention must be focused on appropriate experimental device distributions and traditional device removal prior to study initiation." NSI reported in the study group may have been caused by the use of the traditional device. Based on the information available, it is not possible to separate NSI caused by the new devices or traditional ones.