22713 PHYSIOLOGICAL AND PSYCHOLOGICAL DATA COLLECTION METHODS SUSAN K. FRAZIER AND CAROL GLOD ■ INTRODUCTION Evidence-based practice (EBP) is the clinical application of research findings that have been synthesized, replicated in appropriate populations of patients, evaluated for scientific rigor, and found to be the most effective approach to a clinical problem; clinical expertise and patient preferences also contribute to EBP. The intent of EBP is to optimize outcomes, whether individual clinical outcomes, or system-wide outcomes, or community outcomes. It is essential that clinicians are skillful in the interpretation, evaluation, and critique of quality research studies to determine their scientific soundness and appropriateness for application to clinical practice. This requires a rigorous evaluation of the physiological and psychological instruments used for data collection. ■ PHYSIOLOGICAL DATA COLLECTION METHODS Biomedical instruments collect data about individual physiological status, biological functions and processes, and the consequences of disorders, injuries, and malfunctions. Clinicians use biomedical instrumentation daily in practice to acquire data about patient condition, to monitor patient progress, and to evaluate the efficacy of management. However, biomedical instruments used in the clinical setting and in research studies must be an accurate reflection of reality. Research data and investigator interpretation may be doubted when instruments used to collect physiological data were inappropriate. Thus, published reports must include information about physiological instruments to demonstrate the appropriateness, accuracy, and reliability to ensure sufficient research rigor. Evaluation of these data requires a clear understanding of the principles of biomedical instrumentation, the physiological variables measured, the characteristics of biomedical instrumentation, the procedures for measurement of biological variables, and the criteria for evaluation of accuracy and precision of biomedical instruments. Biomedical instruments extend human senses by measurement of physiological variables, detection of changes over time in these variables, and amplification and display of these data so our senses can detect them. Many physiological changes are miniscule and not detectable with our senses. For example, the electrocardiogram (EKG) captures millivolt changes in electrical activity that occur in cardiac myocytes, amplifies the signal to volts, and displays the signal audibly as a sound, or visually as a waveform on a screen or a printout from a computer. 228■ CLASSIFICATION OF BIOMEDICAL INSTRUMENTS Biomedical instruments are broadly classified as in vivo and in vitro. In vivo (within the living) instruments are applied directly on or within a living organism. In vitro (in the glass) instruments study living cells or tissues after removal from the body, most often in a laboratory. A cardiac monitor is an in vivo instrument, as sensing electrodes are applied directly to the individual. In vitro instruments may measure intracellular activity in cells once they are removed from the body. For example, blood samples removed from an individual are analyzed for electrolyte concentrations. In vivo instruments are further categorized as invasive or noninvasive. Invasive instruments enter a body cavity or break the skin barrier for measurement. An arterial catheter connected to a pressure transducer for measurement of arterial blood pressure is an invasive measure, while a 12-lead EKG and a bedside cardiac monitor with electrodes applied to the skin surface are noninvasive biomedical instruments. The selection of an invasive or noninvasive instrument is a central consideration in clinical practice and research. The use of noninvasive instruments is often preferred because they are associated with fewer patient risks like tissue ischemia, hemorrhage, infection, and subsequent sepsis that may be associated with the intra-use of arterial or IV catheters. Another important consideration for instrument selection is the type and number of potential mechanical issues associated with the instrument and measurement technique, such as proper placement, potential infection, or possible obstruction. Prior to selection of a biomedical instrument, the investigator or clinician should consider the necessary frequency of data collection, either intermittent or continuous. Intermittent or cross-sectional measures provide data at only one point in time, and rapid and clinically important alterations in a variable may be missed. For example, indirect measures of arterial blood pressure using an automated oscillating blood pressure cuff provide intermittent data; the frequency can be altered by reprogramming the instrument. However, continuous measurement of arterial blood pressure with an intra-arterial catheter connected to a transducer provides a continuous direct measure of intra-arterial pressure with beat-to-beat responsiveness to rapid alterations in blood pressure that could be clinically important. Thus, the frequency of data capture is a major consideration in selection of an instrument. The choice of an indirect measurement technique requires the investigator or clinician to select the most accurate and reliable technique available for data collection. For example, indirect blood pressure measurement may be obtained by auscultation, oscillometric technique, finger probe technique using photoplethysmography to evaluate pressure fluctuations in the finger, an ultrasound technique, and tonometry (Muntner et al., 2019). The accuracy of indirect blood pressure measures is dependent on the technique used and the fidelity to recommended measurement procedure. Indirect blood pressure measurement by auscultation using a mercury sphygmomanometer is the gold standard indirect measurement for research and clinical care (Muntner et al., 2019); however, the use of mercury has decreased and in some countries has been banned because of its toxic nature and potential environmental contamination (O’Brien, 2000). Oscillometric measures of blood pressure with automated cuffs are common in clinical practice but often vary considerably from auscultory measures (Wohlfahrt et al., 2019). Investigators identified systolic pressure was 1 to 26 mmHg higher and diastolic was 1 to 25 mmHg higher when oscillometric technique was compared with auscultation; differences were greater in individuals older than 65 years. Because of questionable validity of oscillometric measures of blood pressure, an independent database (www.dableducational.org/sphygmomanometers) provides 229validation data for review and evaluation by consumers. Devices are evaluated using established standards and categorized as recommended, not recommended, or questionable. Unfortunately, this type of database is not available for other instruments. Clinicians and researchers must evaluate the manufacturer’s specifications and research findings focused on the validity, accuracy, precision of the biomedical measure needed. Errors in measurement may occur and invalidate physiological data. For example, the American Heart Association guidelines for measurement of auscultory blood pressure recommend measurements made with the cuff placed over the bare arm (Muntner et al., 2019). Prior investigators suggested that measures taken over clothing are also valid (Pinar et al., 2010), but Ozone et al. (2016) found that accurate measurements were made only over the bare arm. Recently investigators found that when compared to auscultory measures, automated systolic measures were on average 10 mmHg lower, and prevalence of hypertension varied from 31.5% with the automated technique versus 40.1% with the auscultory technique (Wohlfahrt et al., 2019). Thus, the investigator or clinician should select the most accurate technique and validated equipment and follow the established procedure for valid measurements. Although in vitro biomedical instrumentation does not pose direct risks to an individual in clinical practice or research, there must be consideration of the sample required for measurement. Sampling techniques may be associated with complications like bleeding, hematoma formation, pain, infection, and disability. The burden to the individual clearly will vary depending on the type of sample needed, and the instrument that provides the most accurate measurement with the least burden to the individual should be the first choice. In vitro measures are also be categorized as direct or indirect measures. For example, a direct measurement of catecholamine concentration uses blood or plasma, whereas an indirect measure evaluates the breakdown products of catecholamines in urine; and elevated carbon monoxide (CO) levels due to cigarette smoking are directly measured in blood with a CO-oximeter, and indirectly measured in exhaled air using an ecolyzer. ■ CATEGORIES OF PHYSIOLOGICAL VARIABLES DETECTED BY BIOMEDICAL INSTRUMENTS Physiological variables are commonly measured in hospitals, clinics, and community and home settings to evaluate health status, diagnose diseases, determine efficacy of therapeutic regimens, and provide metrics for goal setting. To ensure accuracy of physiological data, the components of the organism–instrument system, the subject, stimulus, and biomedical instrument, must be understood. A variety of physiological variables can be measured using biomedical instrumentation. ■ COMPONENTS OF THE ORGANISM–INSTRUMENT SYSTEM The components of the organism–instrument system include the participant, the stimulus, the sensing equipment, the signal conditioning equipment, the display equipment, and the recording and data-processing equipment (Figure 13.1). Each component requires evaluation to ensure that the data obtained are an authentic reflection of reality. However, these components must function together as a system to produce high-quality, accurate, and reliable data. 230FIGURE 13.1The organism–instrument system. Source: From Polit, D. F., & Hungler, B. P. (1987). Nursing research: Principles and methods (3rd ed.). Lippincott. Subject The subject is the individual, either healthy or ill, from whom data will be obtained. The demographic and clinical characteristics of the individuals included in a study are dependent on the research purpose and specific aims, and careful attention to the selection of demographic and clinical variables is essential in any research study. A number of nursing science studies also test interventions intended to improve health outcomes in a group of individuals with selected characteristics. However, there are research purposes and specific aims that cannot be studied in humans because of the need for invasive in vivo measures that would be inappropriate and/or unethical in humans, as well as associated with high potential for serious adverse reactions and risks. These studies may be performed in animal models, in vitro cell cultures, or with computer models. Stimulus The specific aims of a research study will determine the variables to measure. Once the variables to measure are identified, an experimental stimulus may be selected. This could be a nursing care procedure like endotracheal suctioning or position change, which alters the physiological variables to be measured. Other stimuli might include minute electric shocks to elicit an electromyogram (EMG), auditory stimuli like environmental noise in critical care, or tactile stimulation by touching the skin of a premature infant. In these examples, a stimulus elicits a response in selected physiological variables, which is then measured by biomedical instrumentation. In an experimental study, the stimulus can be altered by changing its duration, intensity, or frequency. For example, the effect of environmental noise in an intensive care unit lasting for 5, 10, or 15 minutes (duration) can be measured at 20 and 50 decibels (intensity) every 30 minutes (frequency) in a number of physiological variables that might include heart rate, blood pressure, intracranial pressure, respiratory rate, and catecholamine concentration in the blood. 231Sensing Equipment Sensing equipment is required to detect alterations in a physiological variable. Transducers and recording electrodes are types of sensing equipment commonly used in research and clinical practice for the measurement of physiological variables. Transducers sense a nonelectrical signal and convert it to an electrical signal; electrodes sense an electrical signal. Transducers A transducer is a device that converts one form of energy, like pressure, temperature, or partial pressure of gases, and simultaneously produces an electrical signal in volts proportional to the change in the variable. Conversion to volts is required because a biomedical instrument is an electronic device that will only respond to changes in electrical output. There are a number of different types of transducers. There are several considerations when using a transducer to collect physiological data to ensure accuracy and reliability of the measures. Pressures must be measured using specific reference planes. The right atrium, the reference plane for blood pressure measures, is located in the supine position at the fourth intercostal space in the midaxillary line (Figure 13.2). The pressure transducer balancing port is positioned so that it is perfectly horizontal to the right atrium of the individual using a level (Avellan et al., 2017; Jacq et al., 2015). Leveling a transducer using the appropriate reference point is vital; for each inch (2.5 cm) of difference between the balancing port and the right atrium, the blood pressure varies 2 mmHg (Jacq et al., 2015). If the position of the individual is changed, then the transducer must be releveled to the reference position to obtain valid measures. The right atrium is also the reference for indirect blood pressure measurement; the middle of the blood pressure cuff on the upper arm should be at the level of the right atrium (the midpoint of the sternum) for valid measures (Muntner et al., 2019). The reference point for intracranial pressure measurement is the level of the ventricles of the brain, which is in line with the foramen of Monro (Thompson, 2011). In the supine position, the foramen is level with the tragus of the ear or the outer canthus of the eye; in the lateral position, the foramen is level with the midsagittal line between the eyebrows. Position change again requires releveling for accurate measurement (Song et al., 2016). After the appropriate level of the pressure transducer is achieved, it must then be balanced and zeroed by opening the balancing port and exposing the sensing diaphragm to atmospheric pressure. This procedure establishes the strain gauge at zero voltage with respect to atmospheric pressure. In the transducer, four strain gauges, or resistances, are mounted to a sensing diaphragm; these resistances are connected to form a Wheatstone bridge circuit. In the strain gauges, as 232pressure increases, two stretch and two contract; the sensitivity of the transducer is then increased fourfold. When the balancing port is exposed to atmospheric pressure, the strain gauges are balanced or equal, and the voltage output is set at zero. When the balancing port is closed and the arterial catheter connected to the pressure transducer, the actual pressure changes occurring in the blood vessel cause the sensing diaphragm to move inward and outward, which changes the resistance in the wires and the voltage output. Transducers that are not correctly zeroed may systematically add or subtract from the actual values, and introduce an error into each measure made. FIGURE 13.2Right atrial reference plane for cardiac pressure measures. Pressure transducers must also be calibrated against a column of mercury (Hg) or water (H2O), depending on the range of pressures to be measured. Known values of pressure in increments of 50 to 250 mmHg for arterial pressure, 5 to 50 mmHg for pulmonary arterial pressure, or 5 to 25 cm H2O for central venous or intracranial pressure are applied to the transducer to determine whether the electrical output is linear. Linearity is the extent to which an input change is directly proportional to an output change. Thus, for every 1 mmHg change in pressure there is a 1 mmHg change in the measurement; this linear response must be consistent through the range of possible pressures. This procedure verifies that changes in blood pressure are proportional to the voltage output. To ensure the accuracy and reliability of research data, the transducer should be calibrated before, during, and after data collection. The same principles of balancing, zeroing, and calibrating apply to temperature and biochemical transducers. Recording Electrodes Recording electrodes sense naturally occurring electrical signals, most often from the heart (EKG), the brain (electroencephalogram or EEG), and muscle (electromyogram or EMG). Natural electrical signals occur because of ion currents produced when positive and negative ion cellular concentrations change, as during a cardiac action potential (Kennedy et al., 2016). There are three types of electrodes: surface electrodes, indwelling macroelectrodes, and microelectrodes. Surface electrodes are most commonly used in clinical practice and in research with humans. An example of indwelling macroelectrodes is a needle electrode, where a needle is placed in subdermal tissue to detect electrical signals. Microelectrodes are often placed in a single cell outside of the body to measure ionic currents directly at the cellular level. Surface electrodes placed on the skin surface record the sum of electrical potentials. Signal Conditioning Equipment Signals produced by a transducer or detected by electrodes are usually measured in millivolts, and must be amplified to volts to drive a display unit. Amplification of the signal is referred to as “increasing the gain.” Once a signal has been amplified, the frequency of the signal in cycles per second is modified to eliminate noise or artifact. An example of artifact is the muscle movement seen on an EKG or 60-cycle (Hz) noise from environmental electrical interference. Electronic filters are a component of the signal conditioning equipment that controls this noise or artifact by rejecting the unwanted signals. Artifact can also be separated, diluted, or omitted by adjusting the sensitivity control on a biomedical instrument. Display Equipment Display units may be an oscilloscope, a computer, or a graphic recorder. Most display units require an input voltage of 5 to 10 V (Enderle & Bronzino, 2012). Once a physiological signal has been 233modified and amplified by signal conditioning equipment, the display equipment converts the electrical signals into visual or auditory output that our senses can detect and evaluate. In the clinical setting, computers and computer screens display the voltage change or waveforms by time, and most often display multiple measured variables. Heart rate, cardiac rhythm, blood pressures, respiratory rate, and oxygen saturation are common physiological variables that may be displayed continuously and simultaneously on a computer screen. These data are automatically stored, and can be retrieved to evaluate trends over time, and to compare changes in physiological variables at different points in time. Computers with specialized software acquire, store, and analyze a wide variety of research and clinical data. However, the rate of data acquisition must be sufficient to ensure that sufficient values of physiological data are captured. Software systems also convert analog signals into digital values that can be transported into a data spreadsheet. ■ CHARACTERISTICS OF BIOMEDICAL INSTRUMENTS Measurement of physiological data using biomedical instruments is fundamental to many nursing science studies. However, not every instrument functions equally well for all circumstances, and inattention to any one component of the organism–equipment system may result in the collection of data that are not an accurate reflection of reality. Thus, there are basic characteristics of the data collected that must be considered for every physiological variable. These include the validity, accuracy, and precision or reliability of the measure. Validity is the extent to which the biomedical instrument measures the actual variable of interest, and accuracy is the degree to which the measured value reflects the actual value. For example, the validity of cardiac output measure taken with bioimpedance can be evaluated by inspection of the waveform obtained by the instrument, and the degree of accuracy by comparison with a “gold standard” measure like the Fick equation or a thermodilution measure. Reliability refers to the accuracy of the measure over time. When using new measures and new biomedical instruments, validity, accuracy, and reliability should be evaluated using a gold standard measure and reported. Biomedical instruments also have specifications that demonstrate their ability to measure the variable of interest. These specifications are important in the decision-making process for selection of the instrument and are typically reported in the instrument specifications provided by the manufacturer. Important characteristics of an instrument include the range, the frequency response, specificity, stability, linear response, and the signal-to-noise ratio (Table 13.1). ■ USING BIOMEDICAL INSTRUMENTATION FOR RESEARCH PURPOSES Any biomedical instrument used to obtain physiological data for research purposes must be thoroughly evaluated prior to selection and use for research purposes, as it directly influences the rigor of the study and the value of the findings. Bedside pulse oximetry is described as an illustrative example that considers the required characteristics of biomedical instrumentation. Oxygen Saturation The gold standard measure of oxygen saturation is obtained with an invasive arterial blood gas sample and determined in vitro using a CO-oximeter. An alternative noninvasive measure of 234oxygen saturation is obtained by pulse oximetry (SpO2). A clear understanding of the instrumentation and technique of these two measurement techniques is vital to ensure that the data obtained with a pulse oximeter are appropriate for a study and are an accurate reflection of reality. TABLE 13.1 Specifications to Determine the Accuracy and Validity of Biomedical Measures SPECIFICATION DEFINITION EXAMPLES Range Complete set of values that an instrument can measure Scale range 0 to 100 grams Cardiac monitor 1–250 beats per minute Thermometer 0 to 60o C Frequency response Capacity of the instrument to respond equally well to rapid and slow components of a signal EKG 0.5–100 Hz Pulse oximeter 0.66–15 Hz Sensitivity Degree of change in the physiological variable that the instrument can detect Instrument chosen for a study should have the degree of sensitivity that responds to the research purpose Scale may weigh within 1 gram or within 0.01 gram Stability Ability to maintain calibration over a given time interval Biomedical instruments develop gradual loss of calibration or calibration drift Equipment specifications describe stability over time, and indicate manufacturer’s recommendation for recalibration frequency Linearity Extent to which an input change is directly proportional to an output change Linearity should be evaluated for the entire range of the instrument For every 1 degree of actual change in temperature, there is a 1-degree change recorded by the thermometer Signal-to-noise ratio Relationship between the signal strength and the amount of noise or artifact detected The higher the signal-to-noise ratio, the fewer the artifacts and the clearer the signal obtained Measurement of Oxygen Saturation Measurement of hemoglobin saturation may be either a fractional or a functional measurement. A fractional measurement of hemoglobin saturation requires a blood sample and is performed with a CO-oximeter, typically associated with a blood gas analyzer. The fractional oxygen saturation is the ratio of oxygen-saturated hemoglobin or oxyhemoglobin (HbO2) to the total number of hemoglobin molecules. Thus, the status of total hemoglobin is evaluated with a fractional measurement. A pulse oximeter provides a functional measurement of hemoglobin saturation, which is the ratio of oxygen-saturated hemoglobin to the total amount of hemoglobin available for binding to oxygen. This type of measurement does not include evaluation of hemoglobin that is not available for binding with oxygen, either carboxyhemoglobin (HbCO) or methemoglobin (Hbmet). Newer pulse-oximetry technology uses eight wavelengths of light and is capable of making noninvasive fractional measures of HbCO and Hbmet. The primary use of these instruments is rapid detection of carbon monoxide poisoning. Compared with CO-oximetry values, this 235technology has a reported bias and precision for HbCO of 0.1% and 2.5%, respectively; thus, compared with gold standard measures, measures will be between –6% and 4% of actual value. Thus, an actual measure of HbCO of 10% could range between 4% and 14% using this noninvasive technique (Zaouter & Zavorsky, 2012). Validity of Oxygen Saturation Measurement by Pulse Oximetry The fractional measurement of oxygen saturation by a CO-oximeter using four or more wavelengths of light is the gold standard measure of oxygen saturation. Numerous studies have compared functional values obtained by pulse oximetry with simultaneous fractional measurement of oxygen saturation by CO-oximeter. In the range of 70% to 100% oxygen saturation, there is a very strong correlation between these values (range of correlation coefficients, r = .92 to .98). Within this range of values, pulse oximetry accurately reflects functional hemoglobin saturation. Most current manufacturers report a bias of 2%, which indicates that the actual value of saturation may be 2% higher or lower than the actual value. However, when oxygen saturation values are less than 70%, pulse oximetry may provide a falsely high value because of the calculation algorithm used (Nitzan et al., 2014). This is due to the development and testing of these instruments in healthy adults only; ethical considerations precluded developers from reducing oxygen saturation in healthy volunteers to levels developed by acutely and critically ill patients. SpO2 may be a clinically useful indirect indicator of oxygen transport; however, both the clinician and the researcher must remember that the use of a functional measurement of oxygen saturation does not reflect tissue oxygen delivery. Accuracy of Pulse Oximetry Certain clinical and technical phenomena may reduce the accuracy of saturation values obtained by pulse oximetry (Shamir et al., 2012; Table 13.2). Bias for pulse-oximetry values vary depending on the degree of hypoxemia; so, as oxygen saturation decreases, bias increases. Bias for pulse-oximetry measures ranges from less than 0.5% to as much as 10%. Thus, the measured values could be seriously inaccurate, particularly at critical values. Pulse-oximetry values in general are reported to have a margin of error or bias of ± 2% of the actual SaO2 value. This degree of error provides a wide range of potential values if SpO2 values are normally distributed. Clinicians and investigators must determine whether measures with this degree of potential error provide sufficiently accurate data to address the clinical and research objectives. Precision and Reliability of Oxygen Saturation by Pulse Oximetry Pulse oximetry can detect a 1% change in oxygen saturation. However, the speed of response by the pulse oximeter decreases as actual SaO2 values decreases. A statistical measure of the reproducibility of pulse-oximetry measures is precision. This value is obtained by calculating the standard deviation of the bias measurement. The precision measure is analogous to the scatter of data points in measures made over time. Precision measures for pulse oximetry are reported to range from 2% to 4% (Perkins et al., 2003; Wilson et al., 2010). Pulse-oximetry measures are generally consistent over time (Nitzan et al., 2014); a majority of reliability evaluation studies evaluated consistency of measurement over time using different probe types (reusable or disposable, finger, ear, and nose). The development of motion artifact appears to be the primary influence on the reliability of pulse-oximetry measures. However, other threats to accuracy also influence the reliability of this type of measurement. 236TABLE 13.2 Clinical and Technical Factors That Reduce the Accuracy of SpO2 Values CATEGORY FACTOR CLINICAL SITUATION Individual Weak arterial pulsation Shock states, hypothermia with shunting of blood flow from the periphery, or increased systemic vascular resistance may result in significantly reduced or absent light absorption detection by the sensor Venous pulsation Right ventricular failure or a partial obstruction to venous outflow. In the presence of both arterial and venous pulsatile flow, the SpO2 value is a composite value and will be lower than the actual arterial saturation Anemia May be secondary to the scattering of light in the plasma, which produces a shift in the degree of red light absorbed (Nitzan et al., 2014) Presence of significant portion of hemoglobin unavailable for oxygen binding (HbCO, Hbmet) At a CO partial pressure of only 0.1 mmHg, hemoglobin is 50% saturated with CO, but a functional measurement of saturation by the pulse oximeter may indicate very high oxygen saturation, as the remaining 50% of hemoglobin may be fully saturated with oxygen. High levels of HbCO increases the affinity of hemoglobin for oxygen and reduces oxygen unloading from hemoglobin at the tissues Hyperbilirubinemia and highly pigmented skin The majority of studies that compare SaO2 with SpO2 in the presence of hyperbilirubinemia (bilirubin up to 46.3 mg/dL) suggested that high bilirubin levels do not interfere with the accuracy of SpO2 when saturation is above 90% (Nitzan et al., 2014) Reduced at low saturations in individuals with highly pigmented skin, with errors up to 10% (Feiner et al., 2007) Use of systemic dyes Indigo carmine, indocyanine green, and methylene blue absorb light at wavelengths similar to those used by the pulse oximeter (660 nm) and alter the accuracy of SpO2 values Hyperlipidemia From endogenous lipids or administration of exogenous lipid solutions, in conjunction with total parenteral nutrition, may produce an artificially lower SpO2 value Artificial nails Fingernail polish does not influence the accuracy of SpO2. Use of artificial nails may reduce the transmission of light and influence the measures made (Rodden et al., 2007) Technical factors Motion artifact May be interpreted by the photodetector as arterial pulsation Ambient light High-intensity, high-quantity ambient light like heat lamps, surgical lights, and fluorescent lights may reduce the accuracy of SpO2 values, as ambient light may be detected by the pulse-oximetry photodetector. Photodetector receives information from both the LEDs and the ambient light source; SpO2 value is then a composite value and is likely inaccurate Optical shunt Some of the light from the LEDs is transmitted to the photodetector without passing through a pulsatile vascular bed; degree of red and infrared light received by the photodetector is again a composite of light exposed to hemoglobin and light not exposed to pulsatile blood 237Optical cross-talk Sensor placed in proximity to another instrument also using red and/or infrared light; light emitted by the secondary instrument received by the pulse-oximetry photodetector will result in a composite value for the SpO2 Excessive signal noise/electrical interference Impedes signal acquisition; signal processing may be disrupted with resultant delayed values that may be inaccurate Guidelines to Increase the Utility of Pulse Oximetry for Research Purposes If SpO2 values are used as research data, the investigator must ensure that these data are valid, precise, and reliable. Milner and Mathews (2012) evaluated 847 pulse-oximeter sensors used in 27 hospitals, and found that 11% of these contained electrical malfunctions that reduced accuracy, 23% of the oximeters emitted light spectra different from that reported by the manufacturer, and 31% of the oximeters were not functioning as expected. None of the inpatient facilities had a procedure or the equipment available for evaluation and calibration of these oximeters. Dugani et al. (2011) found that the use of an oximeter tester could identify faulty electronics and demonstrate the degree of error in SpO2 measurements, but only 65% of biomedical engineers responding to a survey self-reported accuracy testing of pulse oximeters in their facility. These data indicate that the accuracy of clinical values of SpO2 may be questionable in many facilities. The following guidelines improve the likelihood that the acquired SpO2 data will be useful. ■Select a pulse oximeter with indicators of pulsatile signal strength and ability to observe a pulse waveform to ensure that adequate, appropriate signal quality is available. ■Ensure that probe type, size, and position are optimal to detect arterial pulsation without technical interference from ambient light, optical shunt, or cross-talk. ■Assess the association between the apical heart rate and the heart rate detected by the pulse oximeter. These values should be the same. ■Evaluate the individual for the presence of dysfunctional hemoglobin, hyperbilirubinemia, hyperlipidemia, and anemia prior to data collection to ensure that these factors do not influence SpO2. ■Stabilize the probe so that motion artifact is not a significant confounding factor. ■Analyze the relationship between SpO2 and SaO2 obtained by CO-oximetry regularly. These values should be highly correlated with minimal bias and precision. Calculate the bias and precision to evaluate the accuracy and repeatability of the data. ■Perform instrument calibration and accuracy testing prior to each experimental use of the biomedical instrument, and evaluate the equipment using a known standard concentration. 238Physiologic Measurement in Community Settings Used for Research Until recently the use of biomedical instruments for clinical care and research were confined to clinical and laboratory settings. Currently, biomedical activity monitoring devices are widely marketed to individuals in the community as a strategy to improve overall health through regular measurement of daily activity and fitness. These devices are one component of self-care and a support for behavior change, as they provide rapid feedback to the user, and may be useful in evaluation of specific goals. These are easy to use, interface with smartphones or other electronic devices, and readily track activity and other measures like heart rate, oxygen saturation and sleep using technology. This technology may provide useful data for clinicians and researchers. However, as with all biomedical devices, it is necessary to evaluate the appropriateness, validity, accuracy, and reliability of these devices before collection of data. Activity monitoring devices may measure a number of variables that include heart rate, steps taken, distance traveled, minutes of activity, activity level (light, moderate, vigorous), expected calories expended during activity, oxygen saturation, and estimates of sleep time. More advanced activity models may be global positioning system (GPS)-enabled for precise tracking of location and distance, and may evaluate activity by the hour, calculation of sedentary time for a specified period, and provision of reminders to move if sedentary time reaches a set threshold. Many activity monitors contain a three-way accelerometer, an altimeter, and a gyroscope. An accelerometer measures the magnitude of individual acceleration in one to three planes of movement; the altimeter measures the altitude of the individual and changes in altitude over time; while the gyroscope measures orientation and rotation of the individual during activity. Algorithms that use data from each of these technologies calculate activity frequency, extent of time of activity, estimate of the intensity of the activity and subsequent patterns of activity over time (Broderick et al., 2014). For activity devices that provide sleep measures, an actigraph is one component of the device. An actigraph detects movement and may provide an estimate of sleep time in individuals with normal sleep patterns. However, these devices do not evaluate sleep stages or quality of sleep, and sleep time is based only on lack of movement during a designated sleep period. As with any biomedical device, demographic and clinical characteristics of the study participants will influence the choice of equipment. For example, a study of patients with heart failure may require cross-sectional measurement of heart rate, body weight, and activity level during the 24-hour period and could use a device that would evaluate activity frequency and degree of intensity, measure heart rate, and connect wirelessly to a smart bathroom scale to upload body weight data. There are a number of activity devices that use photoplethysmography to measure heart rate. This technique requires an optical emitter that projects a minimum of two wavelengths of light into the skin, a digital signal processor that captures refracted light, an accelerometer to detect motion, and specific algorithms to take the data from the processor and the accelerometer and calculate a motion-tolerant heart rate measurement (Carpenter & Frontera, 2016). The inclusion of activity detected by the accelerometer is vital to the accuracy of these algorithms, as activity increases motion artifact in the signal, which reduces the accuracy of the measure. There are a number of clinical and device characteristics that may reduce the accuracy of the heart rate data captured; these include skin pigmentation, presence of tattoos or the use of henna, skin temperature, severe peripheral vascular disease, hypotension, dysrhythmias, and variations in skin pressure where the sensor is placed. These characteristics may result in inaccurate heart rate calculations. Poor adherence to self-monitoring behaviors may also result in extended periods of missing data (Chung et al., 2017). 239Before the data collected by these devices is included in research studies or used to guide clinical care, it is vital to evaluate the validity, accuracy, and reliability of these devices. Heart rate measures with these devices have demonstrated fair to poor accuracy when compared to gold standard measures with mean absolute percentage of error of 9.17% ± 10.9% for an activity monitor compared with a gold standard measure (Lee et al., 2016). In older adults, the pace and speed of ambulation is often slower and may require the use of assistive devices. Unfortunately, most activity detection devices are not able to sense this, which results in increased error in calculations of movement and distance. Floegel et al. (2017) evaluated accuracy of four activity monitors in impaired and non-impaired older adults, and found that the number of directly observed steps taken were underestimated by 2.6% to 26.9%, and in impaired older adults, steps were underestimated by 1.7% to 16.3%. In those older adults using a cane or walker, directly observed steps were underestimated by 1.3% to more than 11.5%. Leininger et al. (2016) compared one commercial activity monitor to a research grade actigraph and found that the commercial device significantly overestimated the number of steps taken, and did not accurately discriminate between light and vigorous physical activity. Byun et al. (2016) identified a mean absolute percentage of error of 9.2% for sedentary time, 70.1% for moderate to vigorous activity, and 14.5% for total physical activity for an activity monitor compared with a research grade actigraph in healthy preschoolers. Thus, a high degree of error is common currently in many of these commercial devices. Although commercial activity monitors have become ubiquitous over the past decade, their accuracy, precision, and reliability have not been clearly demonstrated by rigorous study. These evaluations are required before the data from these devices can be useful in research and clinical care. However, early research findings supported the use of consumer-based activity monitors because of their positive effects on self-monitoring behaviors, and their utility to support lifestyle changes and improve health (Chung et al., 2017). With improved technology and rigorous evaluation of the accuracy, precision, and reliability, these devices may provide important data for investigators in the future, and they clearly may be useful adjuncts to interventions to support behavior changes. ■ SUMMARY When evaluating and critiquing the appropriateness of research studies to implement EBP in the clinical setting, knowledge about biomedical instrumentation is essential to determine the validity, accuracy, and reliability of the physiological data acquired for those studies and to determine the degree of certainty about the findings. Consideration must be given to the ability of any biomedical instrument to provide valid, precise, and reliable data to ensure that the conclusions derived from research studies are from rigorous measures, and are meaningful and useful. ■ PSYCHOSOCIAL DATA COLLECTION METHODS Psychosocial measures are instruments that researchers and advanced practice registered nurses (APRNs) use to measure variables related to psychological, emotional, behavioral, and related areas in a study. In general, instruments focus on certain topics or content domains such as depression or anxiety. Psychosocial data collection methods are important to guide the use of evidence in nursing practice. They are often used for approved research studies; however, increasingly, these tools are used to evaluate patients clinically or to assess treatment response. 240Where do the instruments exist, and how does a researcher or APRN locate them? One place to start is with other established researchers in the field. Another source is relevant articles located and reviewed in preparation for the study. An examination of the methods and instruments section of earlier studies often gives a detailed account of which measures were used and some of their key characteristics. Since most published manuscripts do not include the actual scales, the reference list should contain a citation for the original instrument. Permission to use a new or original scale may have to be requested from the author(s). Libraries and internet sources also contain compilations of standardized instruments that can be obtained by searching for key words that reflect the concept under study. Finally, databases and websites such as the Patient Reported Outcome Measurement Information System (PROMIS; www.nia.nih.gov/research/resource/patient-reported-outcomes-measurement-information-system-promis) can be accessed for potential instruments for clinical practice and research on patient reported outcomes (PROs). There are also books that include primarily tools for measuring concepts (Frank-Stromborg & Olsen, 2004). Some instruments are copyrighted and can be purchased for a fee. The fee may be a one-time purchase fee or a per-copy or per-use fee. To purchase some instruments, one may need to have certain credentials, such as a PhD. A variety of well-established scales of different types exist to measure concepts such as depression and anxiety and to answer different research questions. These concepts may be measured using interviews, whether structured or semi-structured, or questionnaires that are administered by the researcher and/or completed by the participant. Other methods, such as observation and checklists, may be used along with standardized tests to collect data on behaviors in order to validate observations. For example, a study that examines the sleep of hospitalized cancer patients may include nursing observations and an established patient self-report scale. ■ SELECTION OF INSTRUMENTS The selection of a measurement tool for psychosocial variables depends upon the research question, variables of interest, ages of participants, and other factors. A general rule is that the research question dictates the broader method to be used, whether qualitative, quantitative, or mixed method. Having a research question and key variables that are well defined helps to direct the selection of a method. Before considering various established instruments, the researcher should think about several important questions, including the study purpose, characteristics of the sample, the concept or content to be measured, and practical considerations. In order for the measure to be suitable, it should have established reliability and validity (Box 13.1). ■ INSTRUMENT DEVELOPMENT Occasionally, an investigator who is interested in a certain concept or area of study may find that there are no available scales or instruments that reflect that specific research problem. Creating new scales or questionnaires can be tempting; however, their development requires a deliberate and systematic approach. Beginning researchers may think it is a simple process to create a new scale for psychosocial variables; however, there are several steps involved. An initial step is often to bring together a group of experts in the field; these may be patients who experience a certain 241diagnosis or response to a problem, or experienced nurses who know the topic well and can serve as content experts. For example, a nurse researcher interested in immigrant mothers’ health practices with their children may bring together six or eight representative mothers to generate potential items for a questionnaire during a meeting that lasts from 1 to 2 hours. Next, content experts should review the draft questions, which should also be pretested, revised, and tested for validity and reliability. Overall, the creation of a new instrument requires extensive time and effort, but is appropriate when existing measures are unavailable or inadequate for the purpose. It is generally more feasible to use existing scales. Then results can be compared across samples. BOX 13.1 Factors to Consider When Selecting an Instrument Does the instrument measure the concept being examined? What are the psychometric properties of the instrument? Reliability Stability Equivalence Homogeneity Validity Content validity Criterion-related validity Construct validity Is the instrument feasible? Instrument availability Costs of data-collection tools Nature of the study sample Source: From Mateo, M. A., & Kirchhoff, K. T. (1999). Using and conducting nursing research in the clinical setting (p. 263). W. B. Saunders.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
