ÄÛBÑо¿Ôº

Schizophrenia: (Haslam et al., 2010; n = 20; Mean Age = 44.3 (8.49) years)

• Acceptable test-retest reliability (ICC = .60-.80)

Psychiatric Disorders: (Pan and Fisher, 1994; n = 60; Mean Age = 37.9 (14.9); Sample included diagnosis of affective disorders, delusional disorders, schizophrenia, or alcohol hallucinosis)

• Excellent test-retest reliability (ICC = .95)

Schizophrenia: (Haslam et al., 2010)

• Adequate internal consistency (Cronbach's alpha = .79)

Predictive validity

Schizophrenia (Haslam et al., 2010)

• Significant ability of AMPS Global Process Skills to predict employment levels for people living with schizophrenia (p = .03)

• Significant ability of AMPS Sustaining Performance Process Skill to predict employment levels for people living with schizophrenia (p = .05)

Psychiatric Disorders: (Merritt, 2011; n = 8556; Mean Age = 55.1(17.9) years; Subset of data from AMPS Project International database)

• Poor ability of AMPS Motor score to predict ADL motor performance (AUC = .68)

• Adequate ability of AMPS Process score to predict ADL process performance (AUC = .77)

Psychiatric Conditions associated with cognitive impairments: (McNulty & Fisher, 2001; n = 20; Mean Age = 58 (16.05) years)

• Excellent ability to predict motor and ADL process ability measures to participants¡¯ safety at home (r = .75)

Discriminant Validity

Psychiatric Disorders: (Pan & Fisher, 1994; n = 60; Sample includes diagnosis of affective disorders, delusional disorders, schizophrenia, or alcohol hallucinosis)

•  Significant ability to differentiate between persons with psychiatric disorders and persons without (p ¡Ü 0.001)

Stroke: (Bernspang & Fisher, 1995; n =230; Individuals with history of RCVA (n = 71), history of LCVA (n = 76), and nondisabled (n = 83))

• Process scores below the 1.0 log-odd probability units (logit) scale indicate poorer process functioning

• Motor scores below the 2.0 log-odd probability units (logit) scale indicate poorer motor functioning

Stroke: (Fisher & Bray Jones, 2010 as cited in Poulin et al., 2013; n = 8801; subset of AMPS Project International database; adults with hemispheric stroke)

• Excellent test-retest reliability: (Motor: r = .88 - .91; Process: r = .86- .90)

Stroke: (Fisher & Bray Jones, 2010 as cited in Poulin et al., 2013)

• Excellent: Cronbach's alpha for motor score = 0.92
• Excellent: Cronbach's alpha for process score = 0.91
• *Scores higher than .9 may indicate redundancy in the scale questions.

Concurrent validity

Stroke (Marom, Jarus & Josman, 2006; n= 30; Individuals in their first week home during stroke recovery)

• Moderate associations with the Large Allen Cognitive Levels (r = 0.57 for motor; r = .66 for process)

Predictive validity:

Hemispheric Stroke: (Merritt, 2011; n = 17568; Mean Age = 61.7 (20.6); Subset of AMPS Project International database: Individuals with hemispheric stroke ( n = 8801) and individuals with other neurological conditions ( n = 8767) 

• Adequate ability of AMPS Motor score to predict ADL motor performance (AUC = .82)
• Adequate ability of AMPS Process score to predict ADL process performance (AUC= .82)

Stroke: (Dickerson, Reistetter & Trujullo, 2010; n = 46; Mean Age = 71.67 (10.76); Community sample referred for driving assessment)

• Excellent: Scores on AMPS correctly predict those who fail on-road evaluation or needed additional evaluation 87% of the time (p < 0.001)

Convergent Validity

Stroke (Kizony & Katz, 2002; n = 30; Mean Age = 71.3 years; Inpatient acute care, 4-5 weeks Post-Stroke)

• Moderate relationship between various cognitive assessments and AMPS process scores

Stroke: (Bjorkdahl et al., 2006; n = 58; Assessed at discharge, three weeks, three months, and one year after discharge; Swedish sample)

• Adequate, positive changes were found between discharge and one year after discharge for both motor scores (x? = 1.45, 2.18) and process scores (x? = 1.00, 1.55)

Stroke: (Bernspang & Fisher, 1995)

• Significant differences between individuals who had a previous stroke (RCVA and LCVA) when compared to non-disabled individuals for IADL performance (p ? .05)

Geriatric: (Doble, Fisk, Lewis & Rockwood, 1999; n = 55; Mean Age = 77.9 (7.0) years; Community-dwelling elderly adults)

• Measurement error accounted for 22% of the differences in subjects¡¯ ADL ability measures

Geriatric: (Fioravanti et al., 2012; n = 54; Mean Age = 80 (8.6) years; Mean Length of Stay = 24 (12) days; Canadian sample in a geriatric and neuro-oncology inpatient rehabilitation unit

Geriatric: (Doble, Fisk, Lewis & Rockwood, 1999)

Geriatric with cognitive impairments: (Doble, Fisk, Lewis & Rockwood, 1999; Rockwood, Doble, Fisk, MacPherson, & Lewis as cited in Fisher, 2003)

• Excellent test-retest reliability: (Motor Scale r = 0.88 - 0.9; Process Scale r = 0.86 - 0.87)

Geriatric: (Doble, Fisk, Lewis & Rockwood, 1999)

• Excellent test-retest reliability: (Motor: r = 0.88; Process: r = 0.86)

Concurrent validity:

Older adults: (Wales, Clemson, Lannin & Cameron, 2016; Mean Age > 70 years; Analysis of 56 papers with RCT design detailing functional assessments for older adults)

• Moderate concurrent validity with AMPS motor and FIM motor scores (r = 0.54, r = 0.62 p <.001).

• Moderate concurrent validity with AMPS process and FIM cognitive ratings (r = 0.56, r = 0.62, p<0.001).

Geriatrics: (Fioravanti et al., 2012)

• Moderate correlation between the FIM motor scores and AMPS motor measures (r = 0.54, p<0.001) and between the FIM cognitive scores and the AMPS process measures (r = 0.56, p<0.001). 

• Weak correlation between the FIM motor and AMPS motor measures at time of discharge (r = 0.29, p = 0.035).

• Moderate correlation between FIM cognitive scores and the AMPS process measures (r = 0.48, p < 0.001).

• No significant correlation in comparison of the amount of change detected by each instrument.

• No significant correlation for changes in FIM motor scores changes detected using the AMPS motor scale (r = 0.01, p = 0.951).

• No significant correlation for changes detected by the FIM cognitive scale and those detected by the AMPS process scale (r = 0.19, p = 0.180).

Geriatric with Memory Impairments: (Robinson & Fisher, 1996)

• Low significant correlation between CAMCOG and AMPS motor ability (r = -0.04, p = .38)

• Low significant correlation between MMSE and AMPS motor ability (r = -0/01, p = 0.47)

• Moderate correlation between FIM Motor and AMPS motor ability (r = 0.62, p<0.001)

• Low significant correlation between FIM Cognition and AMPS motor ability (r = -0.07, p=0.32)

• Moderate correlation between CAMCOG and AMPS process ability (r = 0.65, p<0.001)

• Moderate correlation between MMSE and AMPS process ability (r = 0.67, p<0.001)

• Low significant correlation between FIM Motor and AMPS process ability (r = 0.13, p=0.19)

• Moderate correlation between FIM Cognition and AMPS process ability (r = 0.62, p<0.001)

Older Adults: (Wales, Clemson, Lannin & Cameron, 2016)

• Moderate support for construct validity in older adult populations

Geriatric with Memory Impairments: (Robinson & Fisher, 1996; n = 51; Mean Age = 75.4 (9.56) years)

• No ceiling effects found for the AMPS motor and process scales

Geriatric: (Fioravanti et al., 2012)

• Comparison of sensitivity to change from admission to discharge at inpatient rehabilitation:
  • AMPS motor: Effect size (d = 2.14), Large positive effect
  • AMPS process: Effect size (d = 1.59), Large positive effect

Older Adults with Dementia of the Alzheimer¡¯s Type (DAT) (Hartman, Fisher & Duran, 1999; n = 788; Independent Older Adults ( n = 329, Mean Age = 70.5 (5.9)), Older Adults with minimal DAT ( n = 167, Mean Age = 71.2 (9.7)), Older Adults with moderate DAT ( n = 292, Mean Age = 74.5 (8.4)); Sample selected from AMPS database)

• Scores below 2.0 logit for motor scale indicate increased need for assistance to live in the community (correct classification rate = 64%; n= 788)
• Scores below 1.0 logit for motor scale indicate increased need for assistance to live in the community (correct classification rate = 94%; n= 789)

Older Adults with Dementia of the Alzheimer¡¯s Type (DAT) (Hartman, Fisher & Duran, 1999)

Concurrent validity:

Geriatric with Alzheimer¡¯s disease: (Doble, Fisk & Rockwood, 1999; n = 26; Mean Age = 76.8 (6.6) years; Canadian sample)

• Moderate agreement between AMPS ADL process scale and Older Americans Resources and Services (OARS) (k = 0.36).

Older Adults with Dementia (Fisher & Jones, 2012; n = 5417)

• Moderate correlation between AMPS and FIM (r = -0.62)

Predictive

Dementia (Merritt, 2011; n = 2488; subset of AMPS Project International database)

• Adequate ability of AMPS Motor score to predict ADL motor performance (AUC = .78)
• Excellent ability of AMPS Process score to predict ADL process performance (AUC = .92)

Older Adults with Dementia of the Alzheimer¡¯s Type (DAT) (Hartman, Fisher & Duran, 1999)

• Significant main effect between AMPS motor score and functional levels of individuals (p < 0.0001)

• Significant main effect between AMPS process score and functional levels of individual (p < 0.0001)

School-Aged Children with Identified Disability: (Atchinson, Fisher & Bryze, 1998; n = 54; Mean Age = 4.0 (0.7) years; Students receiving occupational therapy for an identified disability (n = 32) and typically developing students as comparison group (n = 22))

• Although ideal values are MnSq = 1.0 and z = 0, MnsQ < 1.4 and z < 2 are used because the values are based on criteria to develop the AMPS

School-Aged Children with Identified Disability or At-Risk: (Munkholm, Berg, Lofgren & Fisher, 2010; n = 984; Age Range 3-13; Students from North America, Australia, New Zealand, United Kingdom and Nordic countries)

• DIF (Differential Item Functionality) has no difference between regions if (-0.55) < logit < (0.55)

School-Aged Children with Identified Disability: (Atchinson, Fisher & Bryze, 1998)

• Strong intrarater reliability. (MnSq = 1.0, z = 0)

School-Aged Children: (Fingerhut et. al, 2002; n = 42; Age Range 5 ¨C 7; Kindergarten students from five public schools)

• Adequate predictive ability of Peabody Developmental Motor Scale -- Fine Motor (PDMS-FM) when using school AMPS motor and process scale (r = .45, r = .35)

Children with No Known Disabilities: (Peny-Dahlstrand, Gosman-Hedstrom & Krumlinde-Sundholm, 2010)

• Low significant age by region interaction effect; F = 1.455 (df 12), p = 0.133 but there was a significant difference between the regions F = 30.80 (df 1), p = <0.001 with Nordic children having the higher measures.

• Low significance age by region interaction effect, ANOVA comparison of variance for ADL process ability F = 1.086 (df 12) p = 0.367 and

• Low significant differences between the regions, F = 1.88 (df 1), p = 0.170.

Developmental Delays (Kang et al., 2008; n = 33; Mean Age = 6.1 (1.9) years; Korean sample)

• Moderate correlations between AMPS-M and COPM-P (r = 0.67)

• Moderate correlations between AMPS-P and COPM-S (r = 0.64)

Children With or Without Mild Disabilities: (Gantschnig, Page, Nilsson & Fisher, 2013; n = 10,998; Mean Age = 8.7 (3.2) years; Sample selected from the international AMPS database)

• Poor significance between groups in mean ADL motor ability (p = .308, t = 0.503)

• Excellent significance main effect for age (p < .001, t = 26.187)

• Excellent significance for Age x Group interaction effect (p <.001, t = -6.612)

• Excellent significance between two groups in mean ADL process ability measures (p < .001, t = -4.296)

Children with No Known Disabilities (Poulson, 1996; n = 162)

• Goodness of fit; 90% to AMPS-M, 95% to AMPS-P

Children With or Without Mild Disabilities: (Gantschnig, Page, Nilsson & Fisher, 2013)

• Small Change (Cohen¡¯s d = 0.17) for 4 year olds in ADL motor ability

• Moderate Change (Cohen¡¯s d = 0.32 to 0.50) for 5-8 year olds in ADL motor ability

• Big Change (Cohen¡¯s d = 0.81 to 0.98) for 12-15 year olds in ADL motor ability

• Big Change (Cohen¡¯s d = 0.83 to 1.26) for 6-15 year olds in ADL process ability

Children with No Known Disabilities: (Peny-Dahlstrand, Gosman-Hedstrom & Krumlinde-Sundholm, 2012; n = 4613; Age Range 3-15 years; Subset of the AMPS Project International Database, North American (n = 2239) and Nordic (n = 2374) children )

• Confidence interval of 95% = (+ 0.49) and (+ 0.39) logits respectively. To be considered relevant, difference should exceed + 1.96 SEM.

• No age group had differences in logits for ADL motor ability larger than 1.96 SEM (+ 0.49)
• No significant difference between regions

 Community Dwelling Adults (Merritt, 2011; n = 38,540, Randomly-selected subset of AMPS Project International database)

• Process scores below the 1.0 logit scale indicate higher need for assistance (sensitivity = .81, specificity = .7)
• Motor scores below the 1.5 logit scale indicate higher need for assistance (sensitivity = - .67, specificity = .72)

Community Dwelling Adults (Merrit, 2011)

Mixed Population: (Gantschnig, Page & Fisher, 2012; n = 145489; Mean Age = 54.06 (24.43) years; Sample from the international AMPS database)

Mixed Population: (Fisher & Jones, 2012; n = 148158; Age Range = 3 - 103; Sample from international AMPS database)

• Excellent test-retest reliability of the AMPS at predicting AMPS-motor (r = 0.9-0.91)
• Excellent test-retest reliability of the AMPS-process scores (r = 0.87-0.90)

Community-Dwelling: (Goto, Fisher & Mayberry, 1996; n = 10; Mean Age = 28.9 (3.98) years; Mean time living in United States = 12.4 (8.8) months; Japanese sample living in the United States for less than 3 years)

• Excellent interrater reliability (proportion of individual misfitting ratings was less than 2.5% (t = 3) )

Community-Dwelling: (Fisher, Liu, Velozo & Pan, 1992; n = 20; Mean Age = 28.5 (3.32) years; Non-disabled Taiwanese sample living in United States for less than 3 years)

• Excellent intrarater reliability (r = 0.93)

Mixed Population:  (Fisher & Jones, 2012)

• Excellent internal consistency. ADL motor (r = 0.92)
• Excellent internal consistency. ADL process (r = 0.91)

Mixed Population: (Fisher & Jones, 2012)

• Moderate-Excellent correlations between AMPS and SIB (r = 0.62 - 0.85)

 Community Dwelling Adults (Merritt, 2011)

• Fair discriminating value for determining community independence for AMPS motor (AUC = .78)
• Good discriminating value for determining community independence for AMPS process (AUC = .84)

Mixed Population: (Fisher & Jones, 2012)

• Extensive literature review; filming and observation of wide range of ADL tasks; Rasch model
• Acceptable goodness-of-fit of tasks, skill items and participants

Mixed Population: (Gantschnig, Page & Fisher, 2012)

• Only one ADL item, Aligns, demonstrated differential item functioning (DIF), but did not result in differential test functioning (DTF)
• AMPS is free of cross-regional bias when used in middle Europe

 Community Dwelling Adults (Merritt, 2011)

• Clinically significant associations, large effect size with AMPS motor scores and global functional level (p < 0.01; Cohen¡¯s d= .213)

• Clinically significant associations, large effect size with AMPS process scores and global functional level (p< 0.01; Cohen¡¯s d= .331)

Multiple Sclerosis: (Doble et al., 1994; n = 44Mean Age = 44.5 (12.3) years, Mean Duration of Self-Reported MS = 19.9 (12.4) years)