Utilization of Arterial Blood Gas Measurements in a ... - Semantic Scholar

Clinical Chemistry / ARTERIAL BLOOD GAS TEST UTILIZATION

Utilization of Arterial Blood Gas Measurements in a Large Tertiary Care Hospital Stacy E.F. Melanson, MD, PhD,1 Trevor Szymanski,2 Selwyn O. Rogers Jr, MD, MPH,3 Petr Jarolim, MD, PhD,1 Gyorgy Frendl, MD, PhD,2 James D. Rawn, MD,3 Zara Cooper, MD,3 and Massimo Ferrigno, MD2 Key Words: Arterial blood gases; Utilization; Laboratory tests DOI: 10.1309/ELH5BPQ0T17RRK0M

Abstract We describe the patterns of utilization of arterial blood gas (ABG) tests in a large tertiary care hospital. To our knowledge, no hospital-wide analysis of ABG test utilization has been published. We analyzed 491 ABG tests performed during 24 two-hour intervals, representative of different staff shifts throughout the 7day week. The clinician ordering each ABG test was asked to fill out a utilization survey. The most common reasons for requesting an ABG test were changes in ventilator settings (27.6%), respiratory events (26.4%), and routine (25.7%). Of the results, approximately 79% were expected, and a change in patient management (eg, a change in ventilator settings) occurred in 42% of cases. Many ABG tests were ordered as part of a clinical routine or to monitor parameters that can be assessed clinically or through less invasive testing. Implementation of practice guidelines may prove useful in controlling test utilization and in decreasing costs.

604 604

Am J Clin Pathol 2007;127:604-609 DOI: 10.1309/ELH5BPQ0T17RRK0M

Analysis of laboratory test utilization has gained increasing importance for health care facilities attempting to contain costs and identify process inefficiencies. Inappropriate utilization and blood collection not only escalate cost but may also create unnecessary patient discomfort, cause vascular and nerve injuries, lead to blood loss and anemia, and introduce or spread infection. Laboratory tests should be ordered with a plan to use the information obtained to make accurate diagnoses, change management, monitor the response to management, or predict prognosis.1,2 However, many tests may be ordered for other reasons, such as availability of the patient, access to an arterial line, ease of performance, medicolegal reasons, or simply habit.1,3 Arterial blood gas (ABG) tests are widely utilized, particularly in intensive care units (ICUs). A total of 99,456 ABG tests were ordered in our 722-bed hospital in 2004. Some authors have developed guidelines for the appropriate utilization of ABG tests4-7 in ICUs, including significant changes in minute ventilation, significant decline in oxygen saturation, and a significant change in the clinical condition. In ventilated patients in stable condition, Wang et al6 and Roberts et al5 suggest ordering an ABG test every 12 hours, and Pilon et al 4 recommend routine ABG tests once daily. However, not all changes in the fraction of inspired oxygen (FIO2) or in ventilator settings should require an ABG test. Because of the large volume and cost of ABG tests at our institution, we decided to examine ABG test utilization patterns throughout the hospital in an attempt to identify opportunities for improved utilization of and to determine who is ordering ABG tests, for which patients, and for what reasons. In addition, we studied whether or how the ABG test results affected patient management. © American Society for Clinical Pathology

Clinical Chemistry / ORIGINAL ARTICLE

Materials and Methods

Results

The clinical laboratory at Brigham and Women’s Hospital, Boston, MA, performs blood gas measurements on the Rapidlab 865 (Siemens Medical Solutions, Tarrytown, NY). The blood gas analyzers are validated to report arterial pH, total CO2, PCO2, PO2, oxygen saturation, hemoglobin, hematocrit, sodium, potassium, ionized calcium, and whole blood glucose values. ABG tests performed outside the laboratory at point-of-care sites (ie, cardiac operating room and neonatal ICU) were not included in this study. Approval for the study was obtained from the Partners Human Research Committee, which waived the requirement for informed consent. The 24-hour day was divided into 12 twohour intervals (eg, 12:00 AM-2:00 AM, 2:00 AM-4:00 AM, 4:00 AM-6:00 AM). To produce a representative sample with 2 days’ worth of data, covering all 24 hours of the day with samples from every day of the week, data were gathered from each interval twice on different days of the week (eg, 12:00 AM-2:00 AM on Wednesday and Saturday, 2:00 AM-4:00 AM on Monday and Thursday, 4:00 AM-6:00 AM on Tuesday and Sunday). At the end of a given 2-hour interval on the predetermined day of the week, a list of all the ABG tests ordered during that interval was obtained from the laboratory information system. This list included the patient medical record number and location within the hospital. Only ABG tests on arterial samples were included. An ABG survey form ❚Appendix 1❚ was taken to the patient’s location, and the physician who ordered the ABG test or, if not available, the nurse attending to the patient was asked but not required to complete the survey used in this study. The data collection took approximately 3 weeks during the months of July and August 2005. The utilization survey inquired about the level of training of the ordering clinician (defined as physician or nurse), the reason for ordering the ABG tests, the most important parameters used, and the effect of the results on patient management. As indicated by “see attached” in Appendix 1, there were different options to check as the reason for ordering the ABG test, within the following 3 broad categories: (1) respiratory event, eg, tachypnea, hypoxia, or asthma; (2) cardiovascular event, eg, myocardial infarction, congestive heart failure, or arrhythmias; and (3) metabolic event, eg, metabolic acidosis or respiratory acidosis. The survey also queried about the influence of the ABG test on patient management. A clinician who responded that the ABG test resulted in a change in management was asked to specify how management was changed. For patients in whom a change of ventilator settings was performed (as indicated by the clinician on the form), the medical records and ICU nursing flow sheets were reviewed to determine the specific change in settings. All ventilator setting changes 2 hours before and 2 hours after the ABG measurements were assessed.

We analyzed 530 ABG tests performed in the laboratory during 24 two-hour intervals, representative of different staff shifts throughout the 7-day week. Of these 530 samples, 39 were excluded because they were venous or mixed-venous blood gases, leaving a total of 491 ABG tests. Physicians or nurses who were intimately involved in the care of the patient at the time of the ABG test responded to the ABG survey. Of 491 surveys, 405 were completed, yielding a response rate of 82.4%. Unavailability of the clinician (eg, clinician in the operating room or clinician off the floor owing to staff shift change) at the time of the survey was the major reason for unfilled surveys. ❚Figure 1❚ illustrates the average number of ABG test orders per hour for 10 consecutive days during the study period. A peak occurred between 4:00 AM (mean ± SD, 18.3 ± 4.3 tests) and 5:00 AM (mean ± SD, 22.6 ± 5.3 tests), before morning rounds. The volumes reached a minimum at 8:00 AM (mean ± SD, 4.3 ± 2.5 tests) and, at other times, averaged 7 to 15 tests per hour. Of the ABG tests, approximately 87% were ordered in ICUs with the following breakdown: cardiac surgical ICU, 18%; medical ICU, 14%; thoracic ICU, 14%; surgical ICU, 14%; neurological ICU, 12%; burn trauma ICU, 10%; cardiac ICU, 4%; and bone marrow ICU, 1%. The remainder were ordered in the operating rooms (9%), postanesthesia care unit (1%), and non-ICU settings (4%). Of the patients, approximately 97% were intubated, and the majority had arterial lines. Although physicians ordered the ABG tests, the majority (68%) were suggested by nurses.

30

ABG Orders

25 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Time (hr)

❚Figure 1❚ Number of arterial blood gas (ABG) tests ordered per hour. The data were obtained by averaging the hourly volume from 10 consecutive days during the study period and plotting the average (gray bars) ± SD (error bars). The peak at 5:00 AM represents 22.6 ± 5.3 ABG tests.

Am J Clin Pathol 2007;127:604-609

© American Society for Clinical Pathology 605

DOI: 10.1309/ELH5BPQ0T17RRK0M

605 605

Melanson et al / ARTERIAL BLOOD GAS TEST UTILIZATION

Clinicians were asked to mark their rationale for ordering the ABG test using a list of common indications (Appendix 1). The most common reasons for requesting an ABG test were a change in ventilator settings (27.6%), a respiratory event (26.4%), and routine (25.7%) ❚Table 1❚. Of the 107 respiratory events, 56 (52.3%) were identified as hypoxia and 18 (16.8%) as tachypnea. Of the 36 metabolic events, 13 (36%) were described as respiratory acidosis and 10 (28%) as metabolic acidosis. Cardiovascular events were noted only twice as a reason for an ABG test. Some reasons for ordering an ABG test as noted in Table 1 may be considered routine or they may lack a clear indication. For example, if the percentages in the routine, transfer to a different unit, and postprocedure laboratory categories are combined, they make up 30% of the total ABG volume. In addition, the category termed “other” in Table 1 (5.4% of total) included the following reasons for ordering an ABG: “arterial line placed,” “poor chest x-ray,” and “postcode.” The clinicians considered PO2, pH, and PCO2 the most important parameters in the ABG test results ❚Table 2❚. A minority (≤20.2%) selected oxygen saturation, base deficit, hematocrit, or total carbon dioxide as the most important parameter. Clinicians were instructed to rank the parameters in order of importance; however, many simply checked all parameters that applied. In 67 patients (16.5%), PO2 and/or oxygen saturation were marked as the only important parameters. Of the results, approximately 79% were expected. A change in patient management occurred in 172 cases (42.5%). There was a change in ventilator settings in 64.5% (111/172) of patients. Among the most frequent changes in ventilator settings were a change in FIO2 (27%) and a change in pressure support (13%) ❚Table 3❚. Less common reasons for a change in management in response to an ABG test result included a blood transfusion or extubation. Some of the changes classified as other in Table 3 included fluid status changes, delaying surgery, changing or increasing medications, and administering electrolytes. When specifying how management was changed, some clinicians indicated the specific change in ventilator settings (eg, change in pressure support level), whereas others (56/111 [50.5%]) simply wrote “change in ventilator settings.” For the cases in which the specific change in settings was not mentioned, the flow sheet was reviewed to obtain more details. In 6% of cases, there was no change 2 hours before or 2 hours after the ABG analysis (Table 3). In 3% of cases, the flow sheets were not available to determine what change was made.

Discussion This hospital-wide study examining ABG tests reveals an opportunity to improve ABG test utilization and decrease hospital costs. We confirmed our hypothesis and found that many 606 606


❚Table 1❚ Reason for Performing Arterial Blood Gas Measurement in 405 Cases No.* (%) of Cases Change ventilator settings Respiratory event Routine Metabolic event Postextubation Postintubation Preextubation Low hematocrit value† Follow-up on abnormal results Transferred to a different unit Postprocedure laboratory testing Unreliable pulse oximeter data Altered mental status Establish brain death Cardiac event Preintubation Other * †

112 (27.6) 107 (26.4) 104 (25.7) 36 (8.9) 16 (4.0) 12 (3.0) 15 (3.7) 14 (3.5) 9 (2.2) 7 (1.7) 7 (1.7) 7 (1.7) 4 (1.0) 4 (1.0) 2 (0.5) 2 (0.5) 22 (5.4)

Clinicians were allowed to mark more than 1 reason for ordering the test. Most of the tests ordered for low hematocrit values were for patients in the operating room.

❚Table 2❚ Parameters of Arterial Blood Gas Measurement Considered Most Important by Clinicians in 405 Cases No. (%) of Users Considering Parameter Most Useful*

PO2 pH PCO2 Oxygen saturation Base deficit Hematocrit Total carbon dioxide Other *

267 (65.9) 256 (63.2) 250 (61.7) 82 (20.2) 50 (12.3) 39 (9.6) 37 (9.1) 27 (6.6)

Clinicians were allowed to indicate more than 1 parameter.

❚Table 3❚ Change in Management in Response to Arterial Blood Gas Test Results No. (%) Change ventilator settings Change FIO2 Change pressure support level Change bipap settings Change in respiratory rate Change PEEP settings Change mode No change Other Chart not available Transfuse blood Extubate Multiple interventions Intubate Other

46 (26.7) 13 (7.6) 8 (4.7) 7 (4.1) 6 (3.5) 4 (2.3) 11 (6.4) 10 (5.8) 5 (2.9) 10 (5.8) 9 (5.2) 6 (3.5) 2 (1.2) 32 (18.6)

Bipap, bilevel positive airway pressure; FIO2, fraction of inspired oxygen; PEEP, positive end-expiratory pressure.

© American Society for Clinical Pathology


ABG tests are ordered as part of a clinical routine or to monitor parameters that can be assessed clinically or through less invasive testing. Our results may be beneficial to other academic centers struggling with ABG test utilization. Many techniques to reduce laboratory utilization have been reported in the literature. Interventions, including education, guideline development, and computer models, have produced varying degrees of success. van Walraven et al8 showed that guideline development, laboratory requisition modification, and a change in funding policy led to a significant decrease in utilization. Computer-based interventions, including redundant laboratory test alerts or specific order template designs, have also been shown to be effective.6,9 Educational efforts are an important aspect of laboratory utilization and should be targeted at physicians, especially new residents, and nurses. Solomon et al10 developed the PRECEDE model of behavior change, which categorizes interventions into predisposing, reinforcing, or enabling. A combination of several interventions is usually more effective than a single intervention for improving inappropriate utilization. Only a few studies have looked at the utilization of ABG tests, and they are limited in scope.4-7,11-14 For example, they examine ABG test utilization in a specific patient population or a single ICU, and they focus on reduction in utilization. Furthermore, they do not describe the rationale for ordering the ABG test, the relative importance of specific ABG parameters, the significance of the results, or their effect on patient management. After interventions including guideline development, computerized order template design, and educational effort, most of these studies showed a mild to moderate reduction in ABG test utilization, with no apparent change in clinical outcomes.4-6 Muakkassa et al13 studied ABG test utilization in a surgical ICU and concluded that the presence of an arterial line was the most powerful predictor of the number of ABG tests per patient. Unlike previous studies, our study not only describes hospital-wide ABG test utilization patterns but also attempts to assess the usefulness of ABG tests and their effect on patient management. In our hospital, there are no guidelines for ABG test use, and clinicians are not limited as to the number of ABG tests they can order. In fact, the order entry system allows a physician to place a standing order for an ABG test every 6 hours for the first 24 hours, regardless of the patients’ clinical status, and does not limit additional orders within the 6-hour windows. We found that many ABG tests are ordered as part of a clinical routine. Of the clinicians, 25.7% selected routine as their reason for ordering an ABG test. Several of them included other reasons for ordering ABG tests, such as the transfer of a patient to a different unit, arterial line placement, or a postprocedure test. We can speculate that some of these cases may be part of a clinical routine. The ABG test ordering pattern follows a consistent daily routine with a peak at 4:00 AM to 5:00 AM. In some cases, residents may be ordering an ABG

test to have the results available for morning rounds, even though testing at the time may not be indicated. No national organization explicitly states a policy on the utilization of ABG tests to assess whether a patient is ready for extubation. Two recent studies examined the usefulness of ABG tests in conjunction with a spontaneous breathing trial to predict successful extubation.15,16 Both studies concluded that ABG values affected extubation decisions in relatively few patients. These results corroborate our clinical opinion that an ABG test is not necessary before and after each intubation or extubation or for every change in ventilator settings. The aforementioned reasons for ordering ABG tests were indicated by approximately 40% of the clinicians. In all of those cases, an ABG test may not have been necessary to supplement or justify a clinical assessment. Laboratory tests should be performed if their results can assist to establish a diagnosis, alter management, or monitor therapy. We examined the specific change in ventilator settings that occurred as a “change in management” to determine their appropriateness. Of the changes in ventilator settings, 27% consisted of a change in FIO2. In some settings, pulse oximetry can substitute for an ABG test to determine if a change in FIO2 is necessary. It is interesting that for 6% of patients, a change in ventilator setting was recorded on the form based on the results of the ABG test; however, according to the medical record and flow chart, no action was taken 2 hours before or 2 hours after the ABG results. This suggests that an ABG test was not necessary in this 4-hour window, that the change was not properly documented in the flow chart, or that the clinician incorrectly recalled the details of the patient’s management. By using data from Table 1, as outlined in the “Results” section, and the preceding arguments, we estimate that at least 30% of ABG tests may be unnecessary. The laboratory would save the cost of reagents, the lease payment on 1 blood gas analyzer, and 0.75 full-time equivalents if the ABG test volume was reduced by 30%. This translates into a savings of approximately $100,000 per year. There are several limitations to our study. This was an observational study. We can only speculate about the number of ABG tests that were ordered unnecessarily, and we cannot determine whether not performing an ABG test in these situations would affect the patient’s outcome. This study was performed at a large academic medical center with residents and fellows, and the results may not be applicable to small or medium-sized hospitals. In addition, we used a survey to obtain some of the information on ABG test utilization. Surveys have limitations, including interpretation and recall bias. ABG tests should be ordered to assess a significant physiologic change in the patient. Our data suggest opportunities to improve arterial blood gas utilization and decrease hospital costs. These results may be beneficial to other large academic medical centers. Education about the usefulness of ABG tests Am J Clin Pathol 2007;127:604-609



607 607

Melanson et al / ARTERIAL BLOOD GAS TEST UTILIZATION

and the implementation of practice guidelines could prove helpful. Removal of arterial lines according to developed guidelines may prevent overutilization of ABG tests. Furthermore, computer order entry may be used to guide test ordering by providing ABG test cost or reminding clinicians about redundant or inappropriate orders.

From the Departments of 1Pathology, 2Anesthesiology, Perioperative and Pain Medicine, and 3Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA. Address reprint requests to Dr Melanson: Dept of Pathology, Brigham and Women’s Hospital, 75 Francis St, Amory 2-217F, Boston, MA 02115.

❚Appendix 1❚ ABG Survey Patient's MR No. _________________

Primary Hospital Diagnosis: _____________________________________________________

Date / Location ICU

Time __________

Inpatient

Outpatient

OR

PACU

Floor

/

ED

Pulmonary Lab or

Other Lab 1. Who ordered ABG? (Do not mention name.) PGY (please mark years of training) Attending Fellow Nurse-initiated Part of a lab panel (if yes, which panel ________________) 2. Reason for ordering ABG (Mark all that apply.) Respiratory Event (See Attached) Cardiac Event (See Attached) Metabolic Event (See Attached) Preintubation Preextubation Other (specify below) ___________________________________________ Postintubation Postextubation 3. What were most important parameters you were looking for in the ABG? (Rank all that apply.) PO2 Total CO2 O2 Sat PCO2 Base Deficit Hematocrit pH Other specify by circling it/them (K, Na, lonized Ca, Glu, CO Sat, MET Hgb) 4. Were results expected or unexpected? Expected Unexpected 5. Did results of ABG influence patient's management? Yes No In case of Yes, please mark all that apply to the diagnosis/management for which the ABG was sent. Change in diagnosis Change in management: How ________________________________________________ Diagnosis confirmed Present management confirmed 6. Value at time of ABG Pulse oximeter (mm Hg) End-tidal PCO2 (if available) 7. Intubated Yes No 8. Mental Status Sedated Unconscious Awake Disoriented Drowsy ABG, arterial blood gas; ED, emergency department; Glu, glucose; Hgb, hemoglobin; ICU, intensive care unit; MET, methemoglobin; MR, medical record; OR, operating room; PACU, postanesthesia care unit; PGY, postgraduate year; Sat, saturation.

References 1. Lundberg GD. The need for an outcomes research agenda for clinical laboratory testing [laboratory]. JAMA. 1998;280:565-566. 2. van Walraven C, Naylor CD. Do we know what inappropriate laboratory utilization is? A systematic review of laboratory clinical audits. JAMA. 1998;280:550-558. 3. Axt-Adam P, van der Wouden JC, van der Does E. Influencing behavior of physicians ordering laboratory tests: a literature study. Med Care. 1993;31:784-794. 4. Pilon CS, Leathley M, London R, et al. Practice guideline for arterial blood gas measurement in the intensive care unit decreases numbers and increases appropriateness of tests. Crit Care Med. 1997;25:1308-1313. 5. Roberts D, Ostryzniuk P, Loewen E, et al. Control of blood gas measurements in intensive-care units. Lancet. 1991;337:15801582.

608 608


6. Wang TJ, Mort EA, Nordberg P, et al. A utilization management intervention to reduce unnecessary testing in the coronary care unit. Arch Intern Med. 2002;162:1885-1890. 7. Merlani P, Garnerin P, Diby M, et al. Quality improvement report: linking guideline to regular feedback to increase appropriate requests for clinical tests: blood gas analysis in intensive care. BMJ. 2001;323:620-624. 8. van Walraven C, Goel V, Chan B. Effect of population-based interventions on laboratory utilization: a time-series analysis. JAMA. 1998;280:2028-2033. 9. Bates DW, Kuperman GJ, Rittenberg E, et al. A randomized trial of a computer-based intervention to reduce utilization of redundant laboratory tests. Am J Med. 1999;106:144-150. 10. Solomon DH, Hashimoto H, Daltroy L, et al. Techniques to improve physicians’ use of diagnostic tests: a new conceptual framework. JAMA. 1998;280:2020-2027.

© American Society for Clinical Pathology


11. Inman KJ, Sibbald WJ, Rutledge FS, et al. Does implementing pulse oximetry in a critical care unit result in substantial arterial blood gas savings? Chest. 1993;104:542-546. 12. Kellerman AL, Cofer CA, Joseph S, et al. Impact of portable pulse oximetry on arterial blood gas test ordering in an urban emergency department. Ann Emerg Med. 1991;20:130-134. 13. Muakkassa FF, Rutledge R, Fakhry SM, et al. ABGs and arterial lines: the relationship to unnecessarily drawn arterial blood gas samples. J Trauma. 1990;30:1087-1093.

14. Diby M, Merlani P, Garnerin P, et al. Harmonization of practice among different groups of caregivers: a guideline on arterial blood gas utilization. J Nurs Care Qual. 2005;20:327-334. 15. Pawson SR, DePriest JL. Are blood gases necessary in mechanically ventilated patients who have successfully completed a spontaneous breathing trial? Respir Care. 2004;49:1316-1319. 16. Salam A, Smina M, Gada P, et al. The effect of arterial blood gas values on extubation decisions. Respir Care. 2003;48:1033-1037.

Am J Clin Pathol 2007;127:604-609



609 609