Decision analysis can be useful in helping to translate information gleaned from the literature, from your own research, or both into action.

There are many forms of decision analysis, but one of the most readily understandable is that of Multi-Attribute Utility Analysis (MAUT). The following example is adapted from Posavac & Carey, (1989).

There are ten steps in the MAUT method:

  1. Identify the decision-maker.

    This is important. The task is to cast the decision in terms of the perceptions of the person (or organization) actually making the decision.

  2. Identify the issue or issues to be addressed. For example, let's say you have a patient with high cholesterol, and he or she is interested in your help in bringing it down.
  3. Identify the choices or options to be evaluated. (Note: I am not medically-trained, so this example may be a bit off. If you wish to comment, you can contact me by e-mail.)

    Someone with high cholesterol has three choices (or combination thereof):

    • Diet
    • Exercise
    • Medication
  4. Identify the relevant dimensions of value.

    In this example, let's say that the relevant dimensions are minimizing financial cost, minimizing cost in terms of time consumed, and maximizing acceptability in terms of physical comfort.

  5. Rank the dimensions in order of importance.

    Let's say that our patient in the example ranks these three dimensions in the following order:

    1. Minimizing cost in terms of time consumed
    2. Maximizing acceptability in terms of physical comfort
    3. Minimizing financial cost
    4. Rate the dimensions in importance, preserving ratios.

      The least important dimension is assigned a value of 10. This is arbitrary but empirically a useful starting point for fine grading of subsequent dimensions while retaining integer numbers. Let's say our patient assigns the following ratings:

      • Minimizing financial cost = 10
      • Maximizing acceptability in terms of physical comfort = 30
      • Minimizing cost in terms of time consumed = 40

      In other words, physical comfort is three times as important to this patient as is minimizing cost, and minimizing time is four times as important as cost.

    5. Sum the importance weights, divide each by the sum, and multiply by 100.

      This is purely a computational step so that the importance wieghts will sum to 100. It is called "normalizing".

      The sum of these importance weights is 80. Then:

      • Minimizing financial cost = (10 / 80 ) * 100 = 12.5
      • Maximizing physical comfort = ( 30 / 80 ) * 100 = 37.5
      • Minimizing time consumed = ( 40 / 80 ) * 100 = 50
    6. Determine the probability for each option, for each dimension of value, that the option will maximize that dimension.
      • Diet

        Minimizing financial cost = .7
        Maximizing physical comfort = .3
        Minimizing time consumed = .9

        In other words, our patient thinks that going on a diet to control cholesterol is highly likely (subjective probability = .7) to result in minimizing the financial cost of getting his/her choleterol down, not very likely to result in maximum physical comfort (p = .3), but quite likely to minimize the amount of time she or he will have to spend on the project of lowering her/his cholesterol.

      • Exercise

        Minimizing financial cost = .5
        Maximizing physical comfort = .2
        Minimizing time consumed = ..2

        Our patient doesn't seem too sure about financial cost (p=.5, or fifty-fifty). Perhaps she or he can foresee, e.g., medical expenses for injuries or perhaps the cost of equipment. The expectation that physical comfort will be maximized is quite low (p = .2), and also for time consumed (p = .2). Our patient evidently expects a decrease in physical comfort and an increase in time consumed.

      • Medication

        Minimizing financial cost = .1
        Maximizing physical comfort = .9
        Minimizing time consumed = .9

        Our patient sees this option as quite likely to be costly (p = .1 for minimizing financial cost), but highly likely to result in maximizing physical comfort (p = .9) and minimizing time consumed (p = .9).

    7. Calculate utilities.

      Here is a chart to help you visualize MAUT:

      Looks like our patient likes medication, diet, and exercise in that order, does it not? (I couldn't agree more ! :-> )

      Return to research flowchart.