Comparison of pulse oximetry measures in a healthy population.
Non-invasive measurement of oxygenation via pulse oximetry is an accepted practice in most non-critical care areas (DeMeulenaere, 2007). Such non-critical care areas may include acute, step-down, and intermediate care hospital units as well as provider offices and patient homes. Nursing and medical interventions are provided and subsequent care is planned based on the readings provided by a pulse oximeter. For example, the Centers for Medicare and Medicaid Services pay for home oxygen therapy if the patient qualifies with a pulse oximetry reading of 88% on room air (Noridian Administrative Services, 2011).
Literature was researched from the PubMed and AccessMedicine databases in 2008. Initial studies of pulse oximetry were published in the 1970s. These studies were reviewed but not considered for inclusion in this review. Because initial work in both pulse oximetry for general use and education for pulse oximetry was performed in the 1970s and 1980s, however, some later studies from this period were included for background information.
The inclusion criteria for pulse oximetry studies reviewed for this article required the study population to be adult and non-critical care. Abstracts and studies about pulse oximetry were from peer-reviewed clinical and nursing journals. Primary source studies of pulse oximetry in critical care, operative, anesthesia, and emergency patients were excluded for two reasons. First, pulse oximetry may have up to a 20-second lag time from reading to display. Due to signal averaging (obtaining and calculating an average of multiple measurements of oxygen saturation), the oxygen saturation currently displayed may not be the current saturation (DeMeulenaere, 2007). Medically stable patients are served better by pulse oximetry than are physically unstable, critically ill patients because of the known delay in measurement report. Cardiac arrhythmias also may prevent a true reading of the pulsatile signal (Hill & Stoneham, 2000), and accurate pulse oximeter readings may not be obtained during severe or rapid oxygen desaturation (DeMeulenaere, 2007; Jensen, Onyskiw, & Prasad, 1998). Finally, non-critically ill patients are under-studied in regard to pulse oximetry assessment sites. Studies on the use of pulse oximetry in pediatric populations were excluded due to physiological differences between children and adults.
Pulse oximetry assessment using a finger sensor is an accepted method for monitoring the oxygen saturation of a stable adult medical or surgical patient (DeMeulenaere, 2007). Only one study used a stable adult medical-surgical population to determine whether a reading from a pulse oximeter finger sensor placed on the ear is equivalent to a reading obtained from the finger (Haynes, 2007). The Haynes study suggested a pulse oximeter finger clip used on the ear of a patient undergoing a pulmonary function test does not provide clinically reliable readings of pulse oximetry. PubMed and the AccessMedicine database were searched with the broad terms oximetry and pulse oximetry. No studies have compared finger and ear pulse oximeter readings in a healthy population. Review of all items published in the previous reports revealed no studies comparing pulse oximetry readings from differing sites on the same healthy individual. More research is required to determine if using the finger sensor on an alternate site for pulse oximetry allows safe, effective patient care.
Scope and Consequences of Oximetry Assessment
No studies were found that specifically addressed the cost of being unable to assess oxygen saturation via pulse oximetry. However, an informational article by DeMeulenaere (2007) indicated pulse oximetry has become an important screening device in cases of possible oxygen desaturation. Studies conducted by Hill and Stoneham (2000) and Howell (2002) supported the use of a pulse oximeter as a screening tool. The use of a pulse oximeter may reduce the number of arterial blood gas tests as well as other invasive and costly interventions.
Knowledge Gaps and Analysis
Two major themes were identified in the pulse oximetry studies and articles that met the criteria for this review. When pulse oximeter technology was new in general clinical practice, journals initially published explanatory and educational articles about pulse oximetry (Attin et al., 2002; Hill & Stoneham, 2000). These articles were replaced quickly by those with the first identified theme: lack of clinician knowledge about the uses and limitations of pulse oximetry (Elliott, Tate, & Page, 2006; Howell, 2002). The second theme involved use of alternative pulse oximeter sensor sites (Clayton, Webb, Ralston, Duthie, & Runciman, 1991; Fernandez et al., 2007). Studies in the 1980s and 1990s such as those by Severinghaus, Naifeh, and Koh (1989) and Carter and colleagues (1998) dealt with the accuracy of pulse oximetry when used on critical care patients. These studies on critical care and pulse oximetry were excluded from this review for the reasons previously stated.
Understanding the principles of pulse oximetry is important to assure that the instrument is properly placed on the finger. Correct placement of the finger sensor is essential for accurate pulse oximetry readings in critical care patients (DeMeulenaere, 2007; Stoneham, Saville, & Wilson, 1994). A literature review by Elliott and co-authors (2006) supported the conclusion that insufficient clinician knowledge about pulse oximetry principles and applications may lead to increased risk for patients due to incorrect pulse oximetry readings.
Knowledge level of clinicians. In an early study by Stoneham and colleagues (1994), responses from a pulse oximetry questionnaire suggested medical and nursing staff had a low level of understanding about the uses and limitations of pulse oximetry. A later study found the average number of correct answers on the test of pulse oximetry knowledge increased significantly after educational intervention with nursing staff (66% to 82%, p<0.01) (Attin et al., 2002). An audit by Howell (2002) also found clinical staff generally has a poor understanding of the limitations and uses of pulse oximetry.
Alternate sites for pulse oximeter sensor placement. The finger, toe, pinna, and lobe of the ear are typical sites for pulse oximetry assessment (Haynes, 2007; Oximetry.org, 2002). In a meta-analysis of 74 studies conducted from 1976 to 1994, Jensen and co-authors (1998) concluded research supports pulse oximeter finger sensors as more accurate than ear sensors. However, the criteria for this meta-analysis included use of pulse oximeter studies in the critical care population. This meta-analysis substantiated the findings of Clayton and colleagues (1991), whose experimental study found pulse oximeter finger sensors were better than ear sensors in patients with poor peripheral perfusion. Another study concluded that a dedicated forehead oximeter was more accurate than a finger sensor in cases of peripheral vasoconstriction (Fernandez et al., 2007). Finally, Haynes (2007) reported a pulse oximeter finger sensor does not provide a reliable reading when placed on the ear of patients undergoing a pulmonary function test.
Only one study has addressed placement of a pulse oximeter finger sensor on the ear to obtain a reading (Haynes, 2007). However, the author did not study stable, healthy participants. Few studies on pulse oximetry meet the criterion of a non-critically ill adult population.
Design and Setting
A non-experimental, descriptive, correlational design was used for the study. The setting consisted of four nursing programs located in the Midwest. The investigational review board (IRB) of the University of Illinois at Chicago and of Black Hawk Community College (Moline, IL) provided approval for this study. The deans of Clinton and Scott Community Colleges in Iowa provided letters of support from their respective institutions. Informed consent was obtained by providing an explanatory pamphlet regarding the study. Neither IRB required signed releases due to the non-invasive nature of the study, and because the study design prevented linking of collected data and data source.
Data were collected in the months of November and December 2008, and January 2009. Data collected included date, time, location of data collection, ambient room temperature, and result of a coin toss performed by the participant. Participant-specific data included age; self-identified sex; dominant side; health status on a descriptive scale of excellent, good, fair, and poor; and pulse oximeter readings from the ear and finger of the dominant side of the participant.
Inclusion criteria included students, faculty, and staff members who agreed to participate and who had tympanic temperatures of 95.7[degrees]-100[degrees] F (35.4[degrees]-37.8[degrees] C). This is the reported range of normal body temperature when measured with a tympanic thermometer (Sund-Levander, Forsberg, & Wahren, 2002). Individuals were excluded if they had known peripheral vascular or arterial disease (DeMeulenaere, 2007); known cardiac arrhythmias (Hill & Stoneham, 2000); physiologic deformities, such as joint distortion from osteoarthritis or rheumatoid arthritis (Hill & Stoneham, 2000); tympanic temperature outside the range identified above; or wore dark nail polish, such as black, blue, or green (DeMeulenaere, 2007).
Sample Size and Power
The convenience sample was comprised of healthy undergraduate students, graduate students, and college faculty and staff. Age and enrollment type were not barriers to study inclusion. Non-traditional students were encouraged to participate. A sample size of 89 was needed for a moderate effect size, power of 0.80 with an alpha of 0.05 (Polit & Beck, 2008).
The tympanic thermometer used was a Genius 2[TM] (Covidien; Mansfield, MA). It is a medical-grade device approved by the U.S. Food and Drug Administration (FDA) and commonly available in retail markets. The pulse oximeter was an Ohmeda Tuffsat[TM] (General Electric; Fairfield, CT), also approved by the FDA. The environmental thermometer was a Precisetemp[TM] (Springfield; Las Cruces, NM).
Data Collection Procedures
The first author (researcher) used the same pulse oximeter and thermometer to obtain all readings. Ambient temperature readings were obtained from a portable environmental thermometer and used at all study sites. The same watch was used to note time at all study sites and the same coin was used for the coin toss in all instances.
Following informed consent of the participant, the researcher assessed and recorded the ambient temperature to assure it was in the range of 66[degrees] F to 76[degrees] F (inclusive) to minimize the risk of ambient temperature influence on body temperature. Also, the date, time, study site, dominant side, age, and sex of participants were recorded. It became apparent that ear jewelry did not need to be removed routinely for participant comfort during assessment of ear oximetry. If a reading could not be obtained due to the presence of earlobe jewelry, the participant was asked to remove the jewelry; alternately, the pinna of the ear was used.
Participants were requested to sit at rest with legs uncrossed and hands on the table. Legs were uncrossed to facilitate whole body blood flow. Hands were at table level to prevent possible variations in readings from different hand heights and to preserve participant comfort.
The researcher applied a new probe cover to the thermometer and then measured and recorded the tympanic temperature from the participant's dominant side. A coin toss was used to determine the site of initial pulse oximetry measurement. Heads meant the ear measurement was done first, while tails meant the finger measurement was first. The researcher measured and recorded the two pulse oximetry results in the order determined by the coin toss.
The earlobe reading was obtained by using the finger sensor in a dorsal-ventral placement with the open end of the sensor pointed cephalically. If the pinna was used, the sensor may have been pointed cephalically or caudally. The finger reading was obtained by placing the sensor in the dorsal-ventral position on the forefinger on the dominant hand. Alcohol swabs were used to clean the pulse oximeter sensor after obtaining measurements from each participant.
Data Analysis Plan and Rationale
Descriptive statistics (frequency, percentages, mean, standard deviation) were used to illustrate the sample characteristics (SPSS version 16.0 [Chicago, IL]). As the data were negatively skewed, Spearman's Rho test was selected to examine the correlation between the measurements. Level of significance was set at p[less than or equal to]0.05.
Description of Sample
A poster display was used to invite people to participate in the study. Data were collected for 112 individuals; there were no refusals once data collection began. Data for eight participants were not used because their tympanic temperatures were below the cut-off point of 95.7[degrees] F. Ear oximetry results were not measureable in 15 additional participants.
The remaining 104 participants were ages 18-67 (M=33.8; SD=13.9). Forty participants (38.5%) were male and 64 (61.5%) were female. The majority (n=96; 92.3%) were right-handed. Participants rated their overall health on a four-point scale. The majority believed their health to be good (n=64, 61.5%). Thus, the typical participant was a 33-year-old, right-handed female who perceived herself to be in good health. See Table 1.
Both ear and finger oximeter measurements were obtained from 89 healthy adult participants. Results suggested no statistically significant association between the pulse oximeter results obtained from the ear and the finger of the same person in the sample ([r.sub.s] = -0.03, p=0.81). In other words, the oximetry measurement from the pinna was not likely to be more accurate than the measurement from a finger.
Current observed clinical nursing practice is to use a finger sensor probe to obtain a pulse oximetry measurement on a patient's ear if the finger is not suitable. The purpose of this study was to determine if finger and ear oximetry measurements are correlated highly in healthy adults. If such a positive correlation existed, a common nursing practice would have a stronger clinical basis supported by research findings. Results indicated otherwise. No statistically significant positive correlation was found between finger and ear oximetry measurements in this sample of healthy adults using a finger sensor probe.
Great care was taken to ensure rigor was maintained in this study, and results thus may be considered trustworthy. Several precautions were taken to ensure precise measurements. Though a convenience sample was used, great care was taken to minimize confounding variables. The researcher was the only data recorder for this study, and the only person who performed data entry. The same portable environmental thermometer, tympanic thermometer, and pulse oximeter were used for all data collection. All ambient temperatures were in the selected range of 66[degrees]-76[degrees] F. (inclusive). Further, the ambient temperature was monitored closely to avoid any influence of ambient temperature on oximetry measurement.
Participants were selected from similar sites (institutions of higher education). The clear majority (90%; n=79) of participants reported themselves to be in good or excellent health. The order of oximetry measurements was assigned randomly and taken via coin toss performed by the participant.
Participants' tympanic temperature was obtained because a subnormal tympanic temperature can result in an artificially low oximetry result (DeMeulenaere, 2007). Ear jewelry was not removed by participants as no participants reported discomfort from having oximetry assessed while ear jewelry was in place. Tympanic and finger oximetry measurements were obtained from the participant's self-identified dominant side.
Ear oximetry measurements were obtained preferentially from the earlobe; if the earlobe was too small to provide a measurement, the oximeter sensor was placed on the pinna of the dominant ear. A new tympanic thermometer probe cover was used with each participant. The pulse oximeter sensor was decontaminated and allowed to air dry after assessment of finger and ear oximetry on each participant. Therefore, the lack of statistically significant correlation between finger and ear oximetry readings must be explained in other ways. These explanations may involve the local environment, procedures, equipment, and population.
Confounding environmental factors included participant positioning and cold weather data collection in the months of November-December 2008 and January 2009. Average daily temperatures for these months in this part of the Midwest range from 24[degrees]42[degrees] Fahrenheit (Accuweather, 2011). Although ambient and tympanic temperatures were addressed, the researcher did not control for skin temperature. Participants entering from the outdoors may have experienced peripheral vasoconstriction which may have affected the validity of the finger and ear oximetry measurements. The manufacturer's statement regarding reliability and validity of the device is silent on the matter of skin temperature (GE Healthcare, 2009). Additionally, multiple data collection sites were utilized. These sites varied relative to entrances and exits from the buildings. In one case, the data collection table was immediately adjacent to the building's main door, which may not have allowed sufficient time for participants' skin temperature to normalize with room temperature.
Perhaps the most important potential influence on results was the participant's position. All data were collected with participants in a seated position. However, in clinical situations, finger pulse oximetry is measured most often while patients are lying in bed. In the experience of the researcher, postoperative or other hospital inpatients are almost always supine or in another non-upright position during assessment of vital signs. During the planning of this investigation, the literature was not informative on the point of patient position during pulse oximetry measurement. As a result, no mechanism was made for any participant position other than seated in a chair with feet fiat.
The discrepancy between finger and ear pulse oximetry results was noted early in the data collection process, and the researcher questioned if the participant's position may have some influence on results. Some participants at one site commented on initial results and offered to be reassessed while supine on a nearby lounge sofa. Anecdotal experience demonstrated that wide disparity between finger and ear oximetry measurements disappeared when the participant assumed a horizontal position. These additional readings were neither recorded nor reported. Such observations would have to be validated by further investigations. However, the omission in the literature of a statement concerning patient position during oximetry measurement needs to be addressed.
Another explanation of findings relates to equipment. In several instances, the finger sensor of the pulse oximeter was too short to capture enough ear tissue to generate a signal.
Attempting to gather enough earlobe or pinna tissue became uncomfortable for some participants. The manufacturer of the pulse oximeter does offer other sizes of finger sensors to accommodate larger or smaller fingers. Using sensors of different sizes may have been more comfortable for some participants. However, using different sensors would have increased the potential for error in data collection by introducing another variable. The researcher chose at the outset of the study to minimize variables and thus used only one size sensor for data collection.
The visual display of the pulsatile signal on the pulse oximeter also may have been a factor. Accurate oximetry readings depend on adequate signal (Oximetry.org, 2002). The oximeter provided a visual readout of the strength of the pulsatile signal; during data collection, the researcher believed the pulsatile signal was strong enough to provide accurate oximetry measurements. The signal's actual strength was not recorded during data collection. In addition, an average minimum or maximum time was not established for the pulse oximeter to signal a reading. The manufacturer's instructions contained no guidance on this point. However, such issues related to the equipment used in this project may have contributed to findings.
In this study, finger oximeter readings ranged from 80% to 100% oxygen saturation of hemoglobin (M=97.8, SD=2.13). Normal pulse oximetry measurements range from 96% to 100% (DeMeulenaere, 2007). The ear oximetry was measureable in only 89 participants; measurements could not be obtained for the remaining 15 volunteers. Because the researcher had not anticipated inability to obtain ear oximetry readings, the lack of reading was not established as a criterion for exclusion from the study. Thus, the ear oximetry statistics reflect only 89 participants. The ear oximetry readings ranged from 46% to 100% (M=89.4, SD=10.9). Perhaps a better measure of central tendency for the ear oximetry reading in these 89 participants is the median, which was 93%.
The population was relatively homogenous in that all potential participants were associated with Midwestern colleges or universities. Furthermore, most participants indicated they were in excellent or good health. Lack of variation in the sample may have contributed to the lack of a statistically significant correlation between finger and ear pulse oximetry measurement. Moreover, such a correlation may have required a larger number of participants than the final sample size of 89.
Study Limitations and Recommendations for Future Research
This study was conducted in winter in a Midwestern locale and should be replicated in other areas in other seasons. Sample homogeneity could be lessened by seeking participants of diverse races and educational levels. Researchers may want to consider collecting data in one session at one location, and allowing time for the surface temperature of participants to normalize if they enter the building from outdoors. However, the location is not the primary weakness of the study.
Of greatest importance is having a participant horizontal instead of sitting upright. This will require increased measures to preserve participant privacy and dignity, including possible use of a portable privacy screen. The adjustment to participant positioning almost certainly will yield data of a higher quality.
Subsequent studies also should address the length of the finger sensor used to obtain pulse oximetry measurements. Pulse oximeter attachments are available with longer and wider finger sensors, which may facilitate comfortably gathering enough ear tissue to obtain a reading. Future researchers also must establish minimum acceptable pulsatile signal strength, and set a minimum and maximum time allowed to obtain a reading.
Clearly, results of this study do not support using the ear as an alternate site for pulse oximetry measurement when using a dedicated finger sensor. Nurses in the clinical setting should review procedures for obtaining pulse oximetry readings, especially if the patient's fingers are not suitable. Correct use of appropriate equipment is vital to obtaining clinically supportable readings. Findings from this project do not preclude the use of the appropriate ear oximetry sensor, which is the appropriate equipment for obtaining measurement from a patient's ear.
Pulse oximetry measurement is now a part of basic nursing assessment. The education of future nurses is grounded in evidence-based practice. Although further study is needed, nursing curricula should discourage the use of a finger sensor to assess pulse oximetry anywhere other than a finger. The oximetry measurement obtained during clinical assessment also must be recorded with note of possible variances (e.g., site, patient position, type of sensor).
Findings from this study also have relevance for health care administrators. Nurse administrators should evaluate current policy carefully to assure use of a finger sensor on a patient's ear is not supported. Staff education on correct practices for pulse oximetry measurement needs to be consistent across the facility. Additionally, quality improvement studies based on current practice in health care institutions need to be conducted.
Many health care institutions have implemented the use of an electronic medical record (EMR). The EMR may not provide for notation of the source of the oximetry reading, although a comment option may be available. The outcome of this study suggests it is important to record the site (finger or ear) and equipment used (finger or ear sensor) to obtain oximetry measurement. Anecdotal results also suggest the position of the patient is relevant. Providers who utilize EMR should insist that programming be modified to allow oximetry site, equipment, and patient position to be recorded easily and routinely.
The current clinical nursing practice of using a finger sensor to obtain pulse oximetry measurement on a patient's ear as opposed to the finger is not supported by study findings. There are clinical, educational, and administrative implications from this study. This study needs to be replicated with the participant in a recumbent position.
Accuweather.com. (2011). Climatology. Retrieved from http://www.accuweather.com/
Attin, M., Cardin, S., Dee, V., Doering, L, Dunn, D., Ellstrom, K., ... Olson, L. (2002). An educational project to improve knowledge related to pulse oximetry [Electronic version]. American Journal of Critical Care, 11(6), 529-534.
Bilan, N., Behbahan, A.G., Abdinia, B., & Mahallei, M. (2010). Validity of pulse oximetry in detection of hypoxaemia in children: Comparison of ear, thumb and toe probe placement. East Mediterranean Health Journal, 16(2), 218-222.
Carter, B.G., Carlin, J.B., Tibballs, J., Mead, H., Hochmann, M., & Osborne, A. (1998). Accuracy of two pulse oximeters at low arterial hemoglobin-oxygen saturation. Critical Care Medicine, 26(6), 1128-1133.
Clayton, D.G., Webb, R.K., Ralston, A. C., Duthie, D., & Runciman, W.B. (1991). Pulse oximeter probes. A comparison between finger, nose, ear and forehead probes under conditions of poor perfusion [Electronic version]. Anaesthesia, 46(4), 260-265.
DeMeulenaere, S. (2007). Pulse oximetry: Uses and limitations [Electronic version]. The Journal for Nurse Practitioners, 3(5), 312-317. doi:10.1016/j.nurpra.2007.02. 021
Elliott, M, Tate, R., & Page, K. (2006). Do clinicians know how to use pulse oximetry? A literature review and clinical implications [Electronic version]. Australian Critical Care, 19(4), 139-144.
Fernandez, M., Burns, K., Calhoun, B., George, S., Martin, B., & Weaver, C. (2007). Evaluation of a new pulse oximeter sensor [Electronic version]. American Journal of Critical Care, 16(2), 146-152.
GE Healthcare. (2009). Tuffsat handheld oximeter. Retrieved from http://www. gehealthcare.com/euen/patient_monitor ing/products/imm-monitoring/pulseoximeters/monitoring_tuffsat.html
Haynes, J.M. (2007). The ear as an alternative site for a pulse oximeter finger clip sensor [Electronic version]. Respiratory Care, 52(6), 727-729.
Hill, E., & Stoneham, M.D. (2000). Practical applications of pulse oximetry. Retrieved from http://www.nda.ox.ac.uk/wfsa/html/ u11/u1104_01.htm
Howell, M. (2002). Pulse oximetry: An audit of nursing and medical staff understanding [Electronic version]. British Journal of Nursing, 11(3), 191-197.
Jensen, L.A., Onyskiw, J.E., & Prasad, N.G. (1998). Meta-analysis of arterial oxygen saturation monitoring by pulse oximetry in adults [Electronic version]. Heart Lung, 27(6), 387-408.
Noridian Administrative Services. (2011). Local coverage determination (LCD) for oxygen and oxygen equipment (L11457). Retrieved from https://www. noridianmedicare.com/dme/coverage/do cs/Icds/current_lcds/oxygen_and_oxygen_equipment.html
Oximetry.org. (2002). Principles of pulse oximetry. Retrieved from http://www. oximetry.org/pulseox/principles.htm
Polit, D.F., & Beck, C.T. (2008). Nursing research: Generating and assessing evidence for nursing practice. Philadelphia, PA: Wolters Kluwer Health.
Severinghaus, J.W., Naifeh, K.H., & Koh, S.O. (1989). Errors in 14 pulse oximeters during profound hypoxia. Journal of Clinical Monitoring, 5(2), 72-81.
Stoneham, M.D., Saville, G.M., & Wilson, I.H. (1994). Knowledge about pulse oximetry among medical and nursing staff [Electronic version]. Lancet, 344(8933), 1339-1342.
Sund-Levander, M., Forsberg, C., & Wahren, L.K. (2002). Normal oral, rectal, tympanic and axillary body temperature in adult men and women: A systematic literature review. Scandinavian Journal of Caring Sciences, 16(2), 122-128. doi:10.1046/j. 1471-6712.2002.00069.x
Cynthia L. Johnson, BSN, RN, is a Graduate Student, University of Illinois at Chicago, Moline, IL.
Mary Ann Anderson, PhD, RN, is Associate Professor, University of Illinois at Chicago, Moline, IL.
Pamela D. Hill, PhD, RN, FAAN, is Professor, University of Illinois at Chicago, Moline, IL.
Note: The authors and all MEDSURG Nursing Editorial Board members reported no actual or potential conflict of interest in relation to this continuing nursing education article.
TABLE 1. Ambient and Tympanic Temperatures Number of Ambient Temperature Tympanic Temperature Location Participants (Degrees Fahrenheit) (Degrees Fahrenheit) 1 38 71.1 -72.1 94.1 - 99 2 23 68.4 - 73 95.8 - 99.1 3 16 66.7 - 72.7 94.9 - 98.6 4 27 71.1 -76.3 95.4 - 98.6
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||CNE SERIES|
|Author:||Johnson, Cynthia L.; Anderson, Mary Ann; Hill, Pamela D.|
|Date:||Mar 1, 2012|
|Previous Article:||Maintaining placement of temporary enteral feeding tubes in adults: a critical appraisal of the evidence.|
|Next Article:||Patients with cancer more likely to face higher cost burdens than those with other chronic conditions.|