We conducted a literature search to identify all relevant publications on the psychometric properties of the Rivermead Mobility Index (RMI) in individuals with stroke. We identified twelve studies. The RMI appears to be responsive in clients with stroke.
Franchignoni, Tesio, Benevolo, and Ottonello (2003) verified the floor effects for the RMI in 73 individual with sub-acute stroke. Participants were assessed at admission to a rehabilitation program and then again after 5 weeks. A poor floor effect was found at admission with 22% of patients scoring 0. When the re-assessment was performed the RMI showed an adequate floor effect with 9% of patients scoring the minimum score.
Hsueh, Wang, Sheu, and Hsieh (2003) examined floor and ceiling effects for the RMI, the Modified Rivermead Mobility Index (Lennon & Hastings, 1996) and the Stroke Rehabilitation Assessment of Movement (STREAM – Daley, Mayo, Wood-Dauphinee, Danys, & Cabot, 1997) in 57 clients with stroke. Participants were assessed at 4 time points: 14, 30, 90 and 180 days after stroke. Within 14 days after stroke, the RMI demonstrated a poor floor effect, with 23% of participants scoring 0 and an excellent ceiling effect, with no participants reaching the maximum score. At the thirtieth and ninetieth day after stroke, the RMI showed an adequate floor effect of 6% and 1%, respectively as well as an adequate ceiling effect of 2% and 3%, respectively. The RMI, when measured 180 days after stroke, demonstrated an excellent floor effect and an adequate ceiling effect of 2%. The MRMI and the STREAM showed an excellent floor effect at all points in time and the ceiling effect ranged from excellent at day 14 to adequate at day 30, day 60 and day 180, with 3%, 6%, and 7% of patients scoring the highest score, respectively.
Franchignoni et al. (2003) administered the RMI to 73 patients two months following a first ever stroke and found the internal consistency of the RMI to be excellent, with a Chronbach’s alpha = 0.92.
Roorda, Green, De Kluis, Molenaar, Bagley, Smith et al. (2008) administered the English and Dutch version of the RMI to 420 and 200 clients with stroke, respectively. The internal consistency of both measures was found to be excellent with a reliability coefficient of 0.96 for the English version and of 0.97 for the Dutch version.
Green, Forster, and Young (2001) evaluated the test-retest reliability of the RMI in twenty-two clients with chronic stroke. Participants were re-assessed with a 1-week interval by the same rater and under the same conditions. Agreement for total scores was investigated using Bland and Altman technique and agreement between items were assessed with kappa statistics. For the RMI total score the agreement was excellent (mean difference = 0.3). Kappa statistics were excellent for turning in bed (kappa = 1.00), walking inside with no aid (kappa = 0.89), walking outside on uneven ground (kappa = 0.83), bathing (kappa = 0.81), and picking objects off the floor (kappa = 0.79), and adequate for stairs (kappa = 0.68), lying to sitting (kappa = 0.64), sitting balance (kappa = 0.64), transfers (kappa = 0.64), walking up and down 4 steps (kappa = 0.67) and walking outside on uneven ground (kappa = 0.49). Kappa values were not provided for the remaining items (sitting to standing, standing unsupported, walking inside with aid, running).
Note: When performing a Bland and Altman analysis, a mean difference close to zero indicates higher agreement between measurements.
Antonucci, Aprile and Paolucci (2002) verified the test-retest reliability of the RMI in 308 clients with subacute stroke. Participants were assessed at admission and discharge from a stroke program of a rehabilitation hospital (the specific time-frame between the two evaluations was not specified). Test-retest reliability was calculated using Rasch Analysis, a type of item-response theory. Rasch Analysis allows verifying whether the item difficulty is the same across repeated measures. The RMI demonstrated item stability when performed at admission and discharge in that the most difficult and the easiest items remained the same. These findings suggest the RMI scores across testing occasions can be compared.
Chen, Hsieh, Lo, Liaw, Chen, and Lin (2007) examined the test-retest reliability of the RMI in 50 clients with chronic stroke. Participants were assessed twice by the same rater with a 7-day interval. The test-retest reliability of the RMI, assessed with the Intraclass Correlation Coefficient (ICC), was found to be excellent (ICC = 0.96).
No studies have examined the intra-rater reliability of the RMI.
Collen, Wade, Robb, and Bradshaw (1991) estimated the inter-rater reliability of the RMI in 43 patients either with stroke (n = 9), head injury (n = 13) or neurosurgery (n =1). Agreement as calculated using the Bland and Altman Technique was excellent (Coefficient of reliability = 2.0/15).
Note: When using the Bland and Altman analysis, the coefficient of reliability is double the standard deviation and indicates, in this study, that between raters, total scores in the RMI can range a maximum of 2 points out of 15.
Hsueh et al. (2003) assessed the inter-rater reliability of the RMI in 40 patients with stroke at a rehabilitation unit. The RMI was administered by 2 examiners within 24 hours of each other. Examiners were blinded to each other’s scores. Inter-rater reliability on individual items was calculated using weighted kappa and the inter-rater agreement of the total score was analyzed with ICC. Inter-rater reliability on individual items ranged from poor to excellent (weighted kappa = 0.37 to 0.94) and inter-rater agreement on the total score was excellent (ICC = 0.92).
Content validity with Guttman scaling is evaluated on the extent to which total scores predict the number of consecutive items passed. In a study of 38 patients with subacute stroke, critical values for two indices, coefficient of reproducibility (> 0.9) and coefficient of scalability (> 0.7), were all exceeded. The results of this study confirm the existence of a valid, cumulative, and unidimensional Guttman scale (Hsieh, Hsueh & Mao, 2000).
In a study by Hsueh et al. (2003), the concurrent validity of the RMI was examined against the Modified Rivermead Mobility Index (MRMI – Lennon & Hastings, 1996) and the Stroke Rehabilitation Assessment of Movement (STREAM – Daley et al., 1997) in 57 individuals with stroke. Correlations were calculated at 4-points in time (14, 30, 90 and 180 days after stroke) using Spearman’s rho and Intraclass Correlation Coefficient (ICC). Correlations between the RMI and the MRMI were excellent for all time points (rho = 0.78; rho = 0.90; rho = 0.90; rho = 0.93), as well as between the RMI and the STREAM (rho = 0.69; rho = 0.87; rho = 0.82; rho = 0.85). When the ICC was used, adequate correlations between the RMI and MRMI (ICC = 0.50; ICC = 0.59; ICC = 0.53; ICC = 0.55) and between the RMI and STREAM were found (ICC = 0.59; ICC = 0.71; ICC = 0.68; ICC = 0.68) at all times.
Hsieh et al. (2000) estimated the ability of the RMI measured at admission to a rehabilitation program to predict Barthel Index (Mahoney & Barthel, 1965) scores at discharge. Predictive validity of the RMI measured in 38 patients with acute stroke using Spearman’s rho was excellent (rho=0.77).
Note: In this study, admission scores were obtained on average 24 days after stroke. Discharge scores were collected on average 60 days after stroke.
Sommerfeld & von Arbin (2001) examined whether the RMI, Barthel Index (Mahoney & Barthel, 1965), sensory ability, aphasia, type and side of brain lesion, previous stroke, social status, living with another person, gender and age measured 10 days after stroke were able to predict an early discharge home, within three months after stroke onset. Length of stay in hospital was recorded from medical charts. Predictive validity of the RMI was assessed in 115 patients with acute stroke, 65 years and older. Compared to the other variables, a RMI score >4 was the best predictor of an early discharge home, followed by a Barthel Index score >35 and living with another person.
Hsueh et al. (2003) analyzed if the RMI, the MRMI (Lennon & Hastings, 1996), and the STREAM (Daley et al., 1997) measured at 14, 30 and 90 days after a stroke were able to predict the Barthel Index (Mahoney & Barthel, 1965) scores measured at 180 days after stroke in 57 individuals using Spearman’s rho. Within 14 days after stroke, adequate predictions regarding the Barthel Index scores were estimated from the 3 mobility measures. At day 30, the RMI was an adequate predictor of the Barthel Index scores while the MRMI and the STREAM were excellent predictors. At day 90, all three measures were excellent in predicting the Barthel Index scores measured 180 days after stroke.
Collen et al. (1991) estimated the convergent validity of the RMI with the Barthel Index (Mahoney & Barthel, 1965), the Berg Balance Scale (Berg, Wood-Dauphinee, Williams & Maki, 1989), the 6-Minute Walk Test (Butland, Pang, Gross, Woodcock, & Geddes, 1982), gait speed and number of falls in 43 patients either with stroke (n = 9), head injury (n = 13) or neurosurgery (n =1). Excellent correlations were found between the RMI and the Barthel Index (r = 0.91), gait speed (r = 0.82), the Berg Balance Scale (r = 0.67) and the 6-Minute Walk Test (r = 0.63). The RMI and number of falls had a poor correlation (r = 0.30).
Hsieh et al. (2000) assessed the convergent validity of the RMI by comparing it to the Barthel Index (Mahoney & Barthel, 1965) and the Berg Balance Scale (Berg et al., 1989) in 38 inpatients with subacute stroke. Correlations as calculated using Spearman’s rho were excellent between the RMI and the Barthel Index (rho = 0.70) and between the RMI and the Berg Balance Scale (rho = 0.85).
Franchignoni et al. (2003) evaluated the convergent validity of the RMI with the motor and cognitive scales of the FIM (Keith, Granger, Hamilton, & Sherwin, 1987), the leg section of the Motricity Index (Demeurisse, Demol, & Robaye, 1980) and with the Trunk Control Test (Collin & Wade, 1990) in 73 patients with subacute stroke. In this study, the correlation using Spearman’s rho was excellent between the RMI and the Trunk Control Test (rho = 0.89) and the motor scales of the FIM (rho = 0.73), adequate between the RMI and the leg section of the Motricity Index (rho = 0.49), and poor between the RMI and the cognitive scales of the FIM (rho = 0.10).
Hsueh et al. (2003) analyzed the convergent validity of the RMI by comparing it to the Barthel Index (Mahoney & Barthel, 1965) in 57 participants with stroke. Correlations were calculated using Spearman’s rho at 4-points in time: 14, 30, 90 and 180 days after stroke. Excellent correlations between the RMI and the Barthel Index were found at all times (rho =0.72, rho = 0.88, rho = 0.86, rho = 0.88), respectively.
Roorda et al. (2008) examined the convergent validity of the Dutch version of the RMI by comparing it to the Dutch version of the Barthel Index in 91 clients. Correlations as calculated using Spearman’s rho was excellent (rho = 0.84).
No studies have examined the known groups validity of the RMI.
Hsieh et al. (2000) assessed the ability of the RMI to detect minimal clinically important differences in 38 individuals with acute stroke. In this study, a clinically important difference was defined as an improvement of 3 or more points on the RMI. From admission to discharge, 76% of participants improved by more than 3 RMI points, suggesting the RMI was able to detect a minimal clinically important difference.
Franchignoni et al. (2003) estimated the responsiveness of the RMI. Seventy-three clients with subacute stroke were assessed at admission to a rehabilitation centre and then again five weeks later. The RMI demonstrated large responsiveness with an effect size of 0.89.
Hsueh et al. (2003) verified the responsiveness on the RMI, the MRMI (Lennon & Hastings, 1996) and the Stroke Rehabilitation Assessment of Movement (STREAM – Daley et al., 1997) in 57 participants with stroke. Responsiveness as calculated using Standardized Response Mean (SRM) was assessed between day 14 and 30, day 30 and 90, day 90 and 180, and finally between day 14 and 90. Except for the time-frame between day 90-180, where a small responsiveness was found (SRM < 0.5), all the 3 mobility measures showed a large responsiveness (SRM > 0.8), suggesting that the RMI, the MRMI, and the STREAM were able to detect change.
Antonucci, G., Aprile, T., & Paolucci, S. (2002). Rasch Analysis of the Rivermead Mobility Index: A study using mobility measures of first-stroke inpatients.
Arch Phys Med Rehabil, 83, 1442-1449.
Berg, K.O., Wood-Dauphinee, S., Williams, J. L., Maki, B. (1989). Measuring balance in the elderly: Validation of an instrument. Physiotherapy Canada, 41(6), 304-311.
Butland, R. J., Pang, J., Gross, E. R., Woodcock, A. A., & Geddes, D. M. (1982). Two-, six-, and 12-minute walking tests in respiratory disease. Br Med J (Clin Res Ed), 284(6329), 1607-1608.
Chen, H.M, Hsieh, C.L., Lo, S.K., Liaw, L.J, Chen, S.M, Lin, J.H. (2007). The test-retest reliability of 2 mobility performance tests in patients with chronic stroke. Neurorehabil Neural Repair, 21, 347-352.
Collen, F.M., Wade, D.T., Robb, G.F., Bradshaw, C.M. (1991). The Rivermead Mobility Index: a further development of the Rivermead Motor Assessment. Int Disabil Stud, 13(2), 50-54.
Collin, C., Wade, D. (1990). Assessing motor impairment after stroke: A pilot reliability study. Journal of Neurology, Neurosurgery, and Psychiatry, 53, 576-579.
Daley, K., Mayo, N.E., Wood-Dauphinee, S., Danys, I., & Cabot, R. (1997). Verification of the Stroke Rehabilitation Assessment of Movement (STREAM). Physiotherapy Canada, 49, 269-278.
Demeurisse, G., Demol, O., & Robaye, E. (1980). Motor evaluation in vascular hemiplegia. European Neurology, 19(6), 382-389.
Forlander, D.A. & Bohannon, R.W. (1999). Rivermead Mobility Index: a brief
review of research to date. Clinical Rehabilitation, 13, 97-100.
Franchignoni, F. Tesio, L., Benevolo, E., Ottonelo, M. (2003). Psychometric properties of the Rivermead Mobility Index in Italian stroke rehabilitation inpatients. Clinical Rehabilitation, 17, 273-282.
Green, J., Forster, A., & Young, J. (2001). A test-retest reliability study of the Barthel Index, the Rivermead Mobility Index, the Nottingham Extend Activities of
Daily Living Scale and the Frenchay Activities Index in stroke patients. Disability and Rehabilitation, 23(15), 670-676.
Hébert, R., Carrier, R., & Bilodeau, A. (1988). The functional autonomy measurement system (SMAF): description and validation of an instrument for the measurement of handicaps. Age Ageing, 17, 293-302.
Hsieh, C.L., Hsueh, I.P., Mao, H.F. (2000). Validity and responsiveness of the
Rivermead Mobility Index in stroke patients. Scand J Rehab Med, 32, 140-142.
Hsueh, I.P, Wang, C.H., Sheu, C.F., Hsieh, C.L. (2003) Comparison of Psychometric
properties of three mobility measures for patients with stroke. Stroke, 34, 1741-1745.
Keith, R. A., Granger, C. V., Hamilton, B. B., Sherwin, F. S. (1987). The Functional Independence Measure: A new tool for rehabilitation. Adv Clin Rehabil, 1, 6-18.
Lennon, S. & Johnson, L. (2000). The modified Rivermead Mobility Index: validity and reliability. Disability and Rehabilitation, 22(18), 833-839
Lennon, S.M.& Hastings, M. (1996). Key physiotherapy indicators for quality
of stroke care. Physiotherapy, 82, 655-664.
Mahoney, F. I., Barthel, D. W. (1965). Functional evaluation: The Barthel Index. Md State Med J, 14, 61-5.
Roorda, L.D., Green, J., De Kluis, K.R.A., Molenaar, I.W., Bagley, P., Smith, J., Geurfs, A.C.H. (2008). Excellent cross-cultural validity, intra-test reliability, and construct validity of the Dutch Rivermead Mobility Index in patients after stroke undergoing rehabilitation. J Rehabil Med, 40, 727-732.
Sommerfeld, D.K. & von Arbin, M.H. (2001). Disability test 10 days after acute stroke to predict early discharge home in patients 65 years and older. Clinical
Rehabilitation, 15, 528-534.