The Rivermead Mobility Index (RMI) is a standardized assessment of mobility originally designed for persons with either head injury or stroke but has subsequently been validated in other populations as well. Currently there are 2 versions available. The first is the original 15-item RMI, derived from the Rivermead Motor Assessment, which comprises 14 self-reported items and 1 direct observation item. The assessment includes such items as observing and grading the patient’s ability to stand for 10 seconds or their self-reported ability of picking up an article off the floor. Each item is scored as either 0, unable to perform or 1, able to perform. A maximum score of 15 points is possible with higher scores indicative of better mobility. The modified version, which includes an expanded scoring system, was developed in response to the original’s inability to detect small changes because of its two item scoring system (Lennon & Johnson, 2000). The m-RMI is now only a direct observational tool as well and is more stroke detailed with increased sensitivity to the effects of therapy (Lennon & Johnson, 2000). This is done by using an extended 6-point scoring system as well as a reduction on the number of items from 15 to only 8 that are stroke specific and include (1) turning over, (2) lying to sitting, (3) sitting balance, (4) sitting to standing, (5) standing, (6) transfers, (7) walking indoors, and (8) stairs. The m-RMI is scored as follows: (0) unable to perform, (1) assistance of 2 people, (2) assistance of 1 person, (3) requires supervision or verbal instruction, (4) requires an aid or an appliance, and (5) independent. A maximum score of 40 is possible with higher scores indicative of increased mobility. Both assessments take less than 15 minutes to complete.
Early studies of the original RMI by Collen, Wade, Robb, and Bradshaw (1991) concluded that the RMI correlated significantly with the Barthel Index (0.91) and aspects of the Functional Ambulation Category Scale at 0.89 and with its mobility category at 0.85, gait speed (0.82), balance (0.67), and 6-minute distance (0.63), but not with the actual number of falls (0.30). A study by Hsieh, Hsueh, and Mao (2000) of stroke patients (n=38) showed that the RMI scores were moderately to significantly correlated with the Barthel Index and Berg Balance Scale scores at > 0.60 and > 0.80, and that subjects improved by more than 3 points during their stay indicating that the RMI is sensitive to change over time. Lennon and Johnson (2000) established content validity and inter-rater and test-retest reliability by assessing subjects (n = 30) with two independent raters in two different settings, an elderly care unit and a stroke rehabilitation unit using the modified version of the RMI. Results showed that it was responsive to change as well as highly reliable between raters with intra-class correlation coefficients of r = 0.98 and high internal consistency with a Cronbach’s alpha of α = 0.93. The highest level of agreement was for standing and the lowest level was for walking indoors.
The two versions of the RMI are simple and quick assessments that target mobility items that are relevant to the clinician. The noted floor/ceiling effect in the original RMI have arguably been reduced in the modified version as it is considered more sensitive to change. Tentative research results support the use of both versions in practice as reliable and valid assessment tools. The Rivermead batteries only require a minimum of training to administer.
There is a limited amount of peer reviewed evidence pertaining to the m-RMI. Also, the suppression of 7 items in comparison to the original 15-item tool could potentially have an impact on its content validity (i.e., it now may be assessing a totally different phenomenon) suggesting that further studies are needed and among different client populations as well (Hsieh et al., 2000). A study by Hsueh et al. (2003) found the RMI to have a limited score range and a notable floor effect at ≤ 14 days post-stroke. Indicating that it may not be able to adequately characterize patient mobility in the early stages following a CVA, a problem not found in the m-RMI during the same study.
The m-RMI is an 8-item observational tool of functional performance in which the therapist grades the client’s ability to perform certain movements relative to balance that uses a 6-point Likert scale as follows: (1) unable to perform, (2) assistance of 2 people, (3) assistance of 1 person, (4) requires supervision or verbal instruction, (5) requires an aid or an appliance, and (6) independent. The maximum score for the m-RMI is 40. The original 15-item scale is composed of both observation and activities, employs only a 2-response scoring system (0 or 1), and has a maximum score of 15. Higher scores are indicative of better mobility for both scales.
1. Turning over
2. Lying to sitting
3. Sitting balance
4. Sitting to standing
7. Walking indoors
9. Walking outside (even ground)
10. Walking inside without an aid
11. Picking up an article from the floor
12. Walking outside (uneven ground)
Adapted from Franchignoni, F., Tesio, L., Benevolo, E., & Ottonello, M. (2003). Psychometric properties of the Rivermead Mobility Index in Italian stroke rehabilitation inpatients. Clinical Rehabilitation, 17(3), 277.