Measures
What is a measure?
A measure is an explicit, specific, clear, observable, and realistic piece of evidence.
What makes a good measure?
Measures can sometimes be difficult to develop, because we are often going from an intangible concept (such as the outcome: healthy newborns) to something that is specific and measurable (the number and percentage of newborns born weighing at least 5.5 lbs). Sometimes, a concept (or outcome) will require multiple measures, as is the case with healthy newborns. In this case, possible measures could be newborn weight, responsiveness, or survival.
Most importantly, measures need to be measurable and clear. In the example above, newborn responsiveness at birth would be a vague measure that would be interpreted differently by different people. A better measure would be the number and percentage of newborns with an Apgar score of 7 or higher, which is unambiguous and well defined.
Additionally, measures need to be realistic and reasonable within the time span of an evaluation. It is important not to hold a program accountable for measures beyond its control. For example, if a program has a two-year grant to improve diabetes care in a clinic, using the number of diabetes deaths in that city as a measure of program success is probably not a good idea because 1) many of the diabetes deaths will be for people who have never had any contact with the program’s clinic, and 2) it might not be reasonable to expect the program to have a large impact on diabetes deaths during the short time frame of the grant. In this case, it is probably more appropriate to focus on shorter-term outcomes for those patients served by the clinic. Measures might include: the number and percentage of patients with diabetes who received a nutrition referral, the number and percentage of patients with HbA1c levels at or below 7, and the change in patient eating habits as measured by a standardized instrument.