Imagine taking a measurement on a patient and not having an idea about the extent of measurement error associated with the measurement. Would you be confident that the measurement was really meaningful and would you use it for making important clinical decisions? For example, let's say that your patient was being treated for a head injury and you were interested in quantifying the extent of spasticity in the patient's upper extremities. Would you be comfortable using a spasticity measure that your colleague recommended but that you had never used or read about? Without having a good idea of how much error is associated with a measurement, we do not know how confident we can be in a measured value or the clinical decisions based on the measurement.
In this chapter, we provide a foundation for answering the question “How confident can I be about the score obtained for the outcome measure on my patient?” We show how the results from reliability studies can be used to provide estimates of measurement error. We cannot be confident that a measured value conveys useful information if we don't know the amount of error associated with the measurement. We begin our discussion by commenting on sources and types of measurement error. We then identify two frequently reported reliability coefficients and demonstrate how they are used to estimate measurement error. We conclude this chapter with an illustration of how the results from a reliability study can be applied to answer the question “How confident can I be about the outcome measurement on my patient?”
The Concept Behind This Question
You'll notice that the title of this chapter (and subsequent chapters) is written in the form of a question. This chapter's question is based on the reality that error is unavoidable and present to some extent in all measurements no matter how much research and development has gone into constructing instruments. Sources of measurement error include the patient, examiner/rater, environment, and measuring instrument. Table 4-1 provides examples of potential sources of measurement error applied to the measurement of knee flexion range of motion.
Table 4-1Potential Sources of Measurement Error Applied to Knee Flexion Range Measurements |Favorite Table|Download (.pdf) Table 4-1 Potential Sources of Measurement Error Applied to Knee Flexion Range Measurements
|Source ||Example |
|Patient || |
• On repeated knee flexion efforts the patient may provide different joint angles.
• The patient may try harder on some occasions compared to others.
|Examiner/Rater || |
• Inherent variation in the examiner's senses.
• The instructions or amount of verbal encouragement may vary from one measurement session to the next.
End digit preference
• The examiner may have a tendency to report specific numbers rather than the actual measured values (e.g., 0s and 5s).
• Knowing the previous ...
Pop-up div Successfully Displayed
This div only appears when the trigger link is hovered over.
Otherwise it is hidden from view.