Description

Samsa et al modified the scoring of the Medication Appropriateness Index of Hanlon et al (second author in current paper). This allows for a single, summated score for each medication being reviewed. The authors are from the VA Medical Center (Durham), University of North Carolina and Duke University in North Carolina.


 

Questions of the Medication Appropriateness Index: 10 questions

 

The weighted scores were determined based on responses from a survey asking the relative importance of each measure.

Parameter

Response

Weighed Score

indication (1)

inappropriate

3

 

appropriate

0

effectiveness (2)

inappropriate

3

 

appropriate

0

dosage (3)

inappropriate

2

 

appropriate

0

correct directions (4)

inappropriate

2

 

appropriate

0

drug-drug interaction (6)

inappropriate

2

 

appropriate

0

drug-disease interaction (7)

inappropriate

2

 

appropriate

0

practical directions (5)

inappropriate

1

 

appropriate

0

expense (10)

inappropriate

1

 

appropriate

0

duplication (8)

inappropriate

1

 

appropriate

0

duration (9)

inappropriate

1

 

appropriate

0

 

where:

• It would be interesting to see if expense would have a higher score now, considering the greater emphasis on cost containment.

 

summated score =

= SUM(points for all 10 parameters)

 

Interpretation:

• minimum score: 0

• maximum score: 18

• The higher the score the more inappropriate the medication use.

• A score of 18 indicates completely inappropriate.

 

Performance:

• The score was found to be relatively reliable and valid.

• The estimated intraclass correlation coefficient was 0.74.

 

Modifications:

• On page 894 the authors describe item weighting and scoring.

• Scores are given only for inappropriate reviews.

• A modified score might be to allow for a 4 or 5 point Likert scale with the ends scored according to the weights above and the interval states are given points based upon the percentage across the interval.

 


To read more or access our algorithms and calculators, please log in or register.