Ïã¸ÛÁùºÏ²Ê¼´Ê±¿ª½±

Traumatic Brain Injury: (Whyte et al, 2003; n=223 patients with TBI in model systems rehabilitation centers; mean age 37 (17.9) years; mean GCS 8.4 (4.2); Rancho Level of Cognitive Functioning >III; tested 3-17 days after admission to acute rehabilitation; mean time post injury tested= 11.4 days)

• Excellent correlation of ratings by OT and PT (r=.64); OT with tendency to rate patient attention lower than PT on average

Traumatic Brain Injury: (Whyte et al, 2008; n=149 patients with TBI at Rancho Level IV or higher; Mean age 43.3(197) years; mean ED GCS score 9.8(4.5))

• Ratings by 52 staff for 924 ratings demonstrated adequate to excellent interrater reliability for rehabilitation disciplines of OT, PT, SLP (ICC=.69-.78 at initial rating and 0.67-0.72 at discharge)
• Nursing ratings were in the adequate range from 0.59-0.68 at initial rating and 0.48-0.59 at discharge from rehab

Traumatic Brain Injury: (Whyte et al, 2003)

• Excellent internal consistency (a=0.95)

Predictive validity: 

Traumatic Brain Injury:  (Hart et al, 2009; n=107; mean age 41.5(19.4) years; GCS at ED admission mean 8.6(5.1)) 

• MARS score was a significant predictor of outcome (Disability Rating Score) by follow-up at one year, better than severity of injury measures or a battery of attention subtests.

Traumatic Brain Injury: (Hart et al, 2006; n=372 patients with TBI at model systems centers; mean age 37(17.9) years; mean GCS 8.4 (4.2); Rancho Level of Cognitive Functioning >III)

• Factor analysis resulted in factors that have excellent to adequate correlations as follows: sustained attention/restlessness r=.75; sustained attention/initiation r=.66; initiation/restlessness r=.46) 
• Changes in MARS scores (transformed to logits) were significantly associated with a composite measure of attention and cognitive FIM scores (Hart et al, 2009)

• A pilot study was conducted with an initial 53 item research version of the test, which was reduced to a 45 item version of the MARS (Whyte et al, 2003).
• Following Rasch analysis with a sample of 223 patients with moderate to severe TBI, person separation and separation reliability values met the criterion levels. Misfit items were analyzed to determine whether they should be retained for further testing (Whyte et al, 2003). 
• Following item analysis and exploratory and confirmatory factor analysis, a final version of 22 items was retained. Half of the items on the MARS can be used to generate factor scores in the following areas (Hart et al, 2006): 
  • 1- restlessness/distractibility (5 items) 
  • 2- initiation (3 items) 
  • 3- sustained/consistent attention (3 items)

Initial items were chosen based on literature review, expert consultation and use of focus groups with expert clinicians. Initial items were pilot tested on 10 patients by multiple rehab disciplines to create a draft version of the test for research. This was tested further with another sample of 20 patients with TBI prior to conducting a larger study using Rasch analysis. (Whyte et al, 2003).

A subsample of studygroup (n=104) demonstrated MARS score increases of 9.9 on average during rehabilitation, with score changes from 7.8 (OT) to 13.1(nursing) (Whyte et al 2008).