Fong 2013.

Study characteristics
Methods	Multi‐centre, randomised, sham‐controlled RCT Setting: rehabilitation hospitals, Hong Kong
Participants	Stroke patients: (1) cerebral vascular disease diagnosed by computed tomography scan or magnetic resonance imaging in a medical report and compatible with unilateral hemispherical involvement; (2) evidence of unilateral neglect from neurological testing or Behavioural Inattention Test (BIT) conventional subtest score < 129; (3) right‐handedness; (4) within 8 weeks after stroke; (5) ability to understand verbal instructions and to follow 1‐step commands; (6) severe to moderate unilateral upper limb paresis, defined as levels 1 to 5 in the Functional Test for the Hemiplegic Upper Extremity (FTHUE) with the range of some beginning voluntary motion of the hemiplegic shoulder and elbow to beginning ability of the hand to combine components of strong mass flexion and strong mass extension patterns Age, years: intervention = 66.2 ± 14.8, control = 68.6 ± 10.6 Sex: men/women: intervention = 16 (84.2)/3 (15.8), control = 9 (56.2)/7 (43.8) Hemisphere damaged: right, basal ganglia, internal capsule, corona radiate, or other Time since stroke, mean and SD, days: intervention = 24.3 ± 18.5, control = 22.3 ± 12.0
Interventions	Experimental group wore a wristwatch device emitting vibration cue with actometer for 3 consecutive waking hours on weekdays for 3 weeks. In the experimental group, patients had to press the acknowledgement button on the device with their right hand as soon as possible after each cue. Patients were told to follow every sensory cue with customary consecutive movements of the hemiplegic arm. They were instructed to perform the movement consecutively 5 times after each sensory cue. Cues came at intervals of 5 minutes for 3 hours. There were 2 kinds of customary movements tailored for participants according to the severity of their arm impairment: patients in levels 3 to 5 of the FTHUE who had partial use of their shoulders or arms were instructed to flex or extend their elbows; those in levels 2 to 3 of the FTHUE who had some voluntary motion of the shoulder were told to flex or abduct their shoulders Control group wore a sham wristwatch device with actometer
Outcomes	BIT‐C Fugl‐Meyer FTHUE FIM Outcomes were reported as follow‐up but only 3 weeks following end of intervention, and therefore not meeting our definition of ‘persisting’ for this review
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Computerised random number generators”
Allocation concealment (selection bias)	Low risk	“Allocation to treatment sequences [was] concealed from all investigators”
Blinding of participants	High risk	Sham clearly ineffective: lost 20% of controls and none from active group for “lost interest”
Blinding of personnel	High risk	Ineffective sham
Blinding of outcome assessment (detection bias) All outcomes	High risk	Ineffective sham
Incomplete outcome data (attrition bias) All outcomes	High risk	Lost 45% of control group before follow‐up assessment
Selective reporting (reporting bias)	Low risk	Consistent reporting of key outcomes
Other bias	Low risk	Nothing obvious