Accuracy and precision.
Biometrics is the measurement of physiological parameters and is important in clinical practice as well as research (Strickland, 2004). Physiologic variables often are measured indirectly rather than directly because indirect measures are less invasive. For example, the most accurate measurement of blood pressure is directly within an artery using a catheter. However, this is invasive, and neither practical nor appropriate for repeated measurements in patients who are not seriously ill (Lewis, 2002). Direct and indirect measurement can lead to error variability as no measurement is error-free. How much error exists in physiologic measurement depends on the accuracy and precision of the instruments used to measure the variable (sphygmomanometers, blood glucose meters, etc.), as well as the variability of the people doing the measurement.
Accuracy is similar to validity and refers to how close the measured (or observed) value comes to the true or actual value (Stone & Fraser, 2005). Generally, accuracy of equipment is supported by ensuring the appropriate calibration is conducted at least as often as recommended by the manufacturer. Calibration is the comparing of a measurement device against a standard or known reference, and is handled by expertly trained personnel on a schedule outlined in the research plan. For sphygmomanometers, part of the calibration process ensures that the measurements begin from zero. If the starting mark is above and below zero, the measurement will be inaccurate (Lewis, 2002).
Accuracy of a measurement also is supported by using the best available equipment. As in the study, authors should indicate the manufacturer of the instrument and, if there are various types of devices, the specific type should be described. Any research about different devices in normal use and their comparable accuracy can be described briefly in the article with the rationale for the choice of device used in the study.
Precision is similar to reliability and refers to how close several measurements are to each other. The closer the data, the more repeatable the measure is and the more likely the results will be similar in the future. Good precision has predictive value; there is confidence in future results with the same methods (Rodrigues, 2007). Precision of a measurement is supported by careful, consistent measurement procedures and proper use of the physiologic measure or equipment. In addition, using the same instrument made by the same manufacturer rather than switching to different brands of instrument will help ensure the precision of the measurements.
Because the precision of a measurement is not related to its accuracy, it is possible to have a very precise group of measurements that are all inaccurate. The analogy of a dart board often is used to explain the difference between accuracy and precision. The closer you are to the bulls-eye, the more accurate you are. On the other hand, you can miss the bulls-eye completely but have all your darts land close together, which is precision. As a clinical example, a nurse obtains the same precise blood pressure measurement on a patient each time when she takes the measurements 15 minutes apart; however, if the nurse is using the wrong size cuff, the blood pressure still will be inaccurate.
In physiologic measurement, there is low accuracy, low precision; low accuracy, high precision; high accuracy, low precision; or high accuracy, high precision. Of course, we desire the highest accuracy and precision that reasonably can be obtained. Rodrigues (2007) outlined three types of error in physiologic measurements: personal or operator error (not using consistent procedures), method error (using the wrong size cuff), and instrumental error (using a sphygmomanometer that has not been calibrated recently).
Turner and colleagues (2008) had the clinical engineering department at their hospital calibrate the sphygmomanometers used in the study to ensure their accuracy. They also described the type and manufacturer of the instruments. All data collectors completed a retraining on blood pressure measurement and competency assessment using American Heart Association guidelines. The authors also described as part of their data collection the procedures that were used in this study, which included arm circumference measurement and selection of the appropriate cuff (using the equipment properly). All these activities address the accuracy and precision of their measurements, allowing the reader to have relative confidence in the study findings.
In some studies, investigators conduct inter-rater reliability assessments to examine how close all data collectors were in their measurements. This is not as practical as it may seem in a study on blood pressure measurements because blood pressure can change quickly and, as most clinicians know, taking a blood pressure often influences the next blood pressure measurement. Another way to increase the consistency in measurement is to have the same person take the measurements. However, as the authors noted in the discussion section, it is normal in clinical practice to have multiple data collectors. Even the same person can measure blood pressures differently at different times. The training and competency assessment is probably the best choice in this study. These procedures should be described in enough detail that others could replicate the training and the measurements.
Turner and co-authors (2008) do a commendable job of outlining their study limitations and what they did to minimize these limitations. For example, they listed the physiologic variations (stress, eating, smoking) that can affect blood pressure and described their use of the patients as their own controls, as well as randomly assigning subjects to the position variable to control the effect of these variations. In any study involving physiologic variables, researchers should consider the conditions that affect the measurement of the variables and develop plans for minimizing these effects.
Accuracy and precision of a measurement are not the only considerations for the selection of physiologic tests or instruments. Practicality as well as cost issues have to be part of the decision making. In addition, applied research requires consideration of what is clinically feasible for the study results to be translated to practice (Strickland, 2004).
Lewis, C. (2002). Checking up on blood pressure monitors. Retrieved on January 10, 2008, from http://www.fda.gov/FDAC/features/2002/502_hbp.html
Rodrigues, G. (2007). Defining accuracy and precision. Retrieved on January 10, 2008, from http://www.encyclopedia. com/doc/1G1-168286146.html
Stone, K.S., & Fraser, S.K. (2005). Measurement of physiological variables using biomedical instrumentation. In C.F. Waltz, O.L., Strickland, & E.R. Lenz (Eds.), Measurement in nursing and health research (3rd ed., pp. 295-325). New York: Springer.
Strickland, O.L. (2004). Ensuring the credibility of physiological measurements: Assessing error variability. Journal of Nursing Measurement, 12(2), 91-93.
Turner, M., Burns, S.M., Chaney, C., Conaway, M., Dame, M., Parks, C., et al. (2008). Measuring blood pressure accurately in an ambulatory cardiology clinic setting: Does patient position and timing really matter? MEDSURG Nursing, 17(1), 93-98.
Lynne M. Connelly, PhD, RN, is an Assistant Professor, University of Kansas, School of Nursing, and Clinical Nurse Researcher, University of Kansas Hospital, Kansas City, KS. She is Research Editor for MEDSURG Nursing.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Research Roundtable|
|Author:||Connelly, Lynne M.|
|Date:||Apr 1, 2008|
|Previous Article:||Measuring blood pressure accurately in an ambulatory cardiology clinic setting: do patient position and timing really matter?|
|Next Article:||Patient safety challenges in treprostinil therapy.|