Abstract Objectives To determine the need and impact of repeating critical values in hematology and coagulation. Methods We prospectively evaluated the need for repeating critical values. The cost of this practice was estimated using a workflow analysis. Retrospective chart review before and after removal of this process was performed to assess the clinical impact of removing this practice. Results Over 95% of the repeated values remained critical and all but one of the repeats were within the expected analytical precision of the assays. The practice of repeating critical values delayed turnaround time for these results and wasted resources, most notably manpower. The delay associated with repeating hematology critical values resulted in delayed administration of blood product (RBC units). Conclusions Repeating critical hematology and coagulation results was found to be an unnecessary process that wasted laboratory resources and lengthened turnaround time, delaying clinical intervention. Critical values, Resource utilization, Laboratory management Critical values, also known as alert or panic values, are laboratory results that indicate a possible life-threatening situation for the patient that requires rapid clinical intervention to avert significant patient morbidity and mortality.1 Given the importance of these values, there is a pressing need to report them accurately and rapidly. Although there are no regulatory requirements to verify critical values by repeat analysis, the practice of repeating all critical values is longstanding and common in laboratories.2 Multiple studies have called into question the necessity of this practice, with most concluding it is unnecessary for accurate result reporting.2-6 Some reports have shown limited utility of repeat testing in specific instances, such as when the results fall outside the analytic measurement range.7 In addition to being unnecessary, the process of repeating critical values in chemistry was shown to delay reporting of the results. Studies have shown delayed turnaround times (TATs) ranging from 5 minutes for blood gases to 42 minutes for calcium.4,6 Most of the studies assessing TAT delays have focused on clinical chemistry assays; far fewer studies have looked at delays caused by repeating hematology or coagulation critical values. The one study looking at hematology values reported median delays of approximately 8 minutes for all laboratories, with 10% of laboratories in the Q-probe having median delays of 17 to 21 minutes.2 Even more worrisome, in the Q-probe study on repeating critical values, 20% of laboratories reported at least one adverse event related to delayed reporting of laboratory values.2 While the Q-probe did not further investigate this, it raises the possibility that delayed reporting of critical values caused by repeat testing is delaying urgent clinical interventions and putting patients at risk. Motivated by the existing literature, we undertook a prospective evaluation of our laboratory’s policy of repeating critical hematology and coagulation results. This evaluation included an assessment of the necessity of performing these repeats, the impact to the laboratory from a resource perspective, and the potential impact of removing this practice on patient care. Materials and Methods Coagulation and Hematology Testing For coagulation, we evaluated the necessity of repeating activated partial thromboplastin time (APTT), international normalized ratio (INR), and fibrinogen (FIB) on an ACL-TOP 500 CTS (Instrumentation Laboratory, Bedford, MA). For hematology, we evaluated the necessity of repeating hemoglobin (Hb), hematocrit (HCT), platelet (PLT) count, and WBC count on a UniCel DxH 800 (Beckman Coulter, Miami, FL). Adult critical value limits for these analytes and the defined acceptable precision are listed in Table 1. Pediatric critical values differ for some of these analytes but were not considered in this analysis. All repeat critical value testing had to be manually programmed after retrieval of the sample; no critical values repeats were autoprogrammed for hematology or coagulation in our laboratory. Before reporting any critically low hematology results, all samples were manually inspected for the presence of clots. In addition, critically low PLT values on new patients were confirmed by manual review of a smear before reporting. These quality checks remained in place after removal of repeating critical values. For APTT and INR, the samples were all recentrifuged prior to repeat testing. When the process of repeating was removed, this step was also eliminated. All instruments were operated by licensed medical technologists and used in accordance with laboratory procedures, which included regular quality control and proficiency testing. Table 1 Adult Critical Value Limits for Coagulation and Hematology Testsa Analyte Precision Lower Limit Upper Limit Hemoglobin, g/dL 2%/0.3 6 22 Hematocrit, % 1.7 18 65 WBC count, × 103 15%/0.2 — 100 Platelets, × 103 7%/5,000 10 1,000 International normalized ratio 0.3 — 5.0 Activated partial thromboplastin time, s 10% — 150 Fibrinogen, mg/dL 15%/5 100 — Analyte Precision Lower Limit Upper Limit Hemoglobin, g/dL 2%/0.3 6 22 Hematocrit, % 1.7 18 65 WBC count, × 103 15%/0.2 — 100 Platelets, × 103 7%/5,000 10 1,000 International normalized ratio 0.3 — 5.0 Activated partial thromboplastin time, s 10% — 150 Fibrinogen, mg/dL 15%/5 100 — aA critical value was defined as any value at or below the lower limit or at or above the upper limit. Dashes indicate that there is no upper/lower limit set for this analyte. View Large Sample Sets Five separate cohorts were included in this study. The first cohort was prospectively collected to determine the necessity of repeating critical values. This cohort included 896 consecutive critical values during a 1-month period. In this cohort, technologists recorded initial and repeat critical value results and the time that each result was completed. The time the sample was received was also noted. The second and third cohorts were retrospectively obtained from the laboratory information system (LIS) to assess the impact of removing critical value repeats on TAT. The second cohort included randomly selected critical results (100 for all analytes except WBC and FIB, for which only 50 could be obtained) from the month prior to the study period. The third cohort was a comparable random selection of critical results but this time obtained from the month after cessation of the repeat policy. The fourth and fifth cohorts were studied to examine the clinical impact of removing critical value repeats. These cohorts were obtained by taking randomly selected patients from the third and fourth cohorts who had critically low Hb values and had received RBC transfusions. Statistics TAT was calculated from the time the sample was received in the laboratory to the time a result was verified in the patient’s medical record. Data analysis was performed in R.8 Differences in turnaround time before and after removal of the critical value repeat practice were assessed using the unequal variances t test (Welch t test). Results Necessity of Repeating Critical Values During the evaluation period, 896 coagulation and hematology critical values were repeated. The most frequent critical repeated was PLTs (349 [39%]), and the least was WBCs (44 [5%]) Table 2 . As shown in Table 2, most critical value repeats remained critical and within the defined precision (863 [96.3%]). For those tests where the repeat was noncritical, the repeat was within the defined precision for all but one (a critically low PLT). For the low PLT that fell outside the expected precision, the original value was 9 and the repeat was 15. In this instance, supervisory review resulted in posting of the original critical value. Thus, in no instance did repeating critical values alter the reported result. Table 2 The Impact of Repeating Critical Values for Coagulation and Hematology Tests Analyte Critical Value Repeated, No. Repeat Still Critical, No. (%) Repeat Within Precision, No. (%) Hemoglobin 117 112 (95.7) 117 (100) Hematocrit 99 96 (97.0) 99 (100) WBC count 44 44 (100) 44 (100) Platelets 349 332 (95.1) 348 (99.7) International normalized ratio 115 111 (96.5) 111 (100) Activated partial thromboplastin time 127 125 (98.4) 127 (100) Fibrinogen 45 43 (95.6) 45 (100) Analyte Critical Value Repeated, No. Repeat Still Critical, No. (%) Repeat Within Precision, No. (%) Hemoglobin 117 112 (95.7) 117 (100) Hematocrit 99 96 (97.0) 99 (100) WBC count 44 44 (100) 44 (100) Platelets 349 332 (95.1) 348 (99.7) International normalized ratio 115 111 (96.5) 111 (100) Activated partial thromboplastin time 127 125 (98.4) 127 (100) Fibrinogen 45 43 (95.6) 45 (100) View Large Clinical Impact of Repeating Critical Values Given the lack of benefit from repeating critical values, the decision was made to discontinue this practice and explore the impact this had on critical value TAT. Data from before and after discontinuation of the practice were assessed for TAT from sample receipt in the laboratory to result verification. Notched boxplots of the TAT before and after discontinuation of repeats are shown in Figure 1 and Figure 2. A number of samples in both data sets had times from received in laboratory to resulted that were outside the realm of possibility (ie, faster TAT than the analytic time of the assay); such values occurred in the before and after data sets where they comprised less than 1% of the total data. These values likely related to missed sample receipt in the LIS, causing them to be scanned as received after testing was complete. The median, mean, and maximum TATs for all analytes are given in Table 3. All analytes showed a statistically significant (P < .05) decrease in the TAT. The largest impact was realized for PLTs and WBCs. In addition to the decrease in TAT, the TAT for critical values showed improved consistency, as evidenced by reduced interquartile ranges evident in the notched boxplots. Along with this improved consistency, there was a reduction in the maximum values encountered (Figures 1 and 2). For instance, the longest observed TAT for WBC with repeats was 196 minutes, while the longest observed TAT for WBC without repeats was 128 minutes. This reduction in the longest TAT for reporting critical values was observed for all analytes. Table 3 Turnaround Time (TAT) Before and After Discontinuing the Practice of Repeating Critical Values Analyte Median TAT Before, min Median TAT After, min Mean TAT Before, min Mean TAT After, min Hemoglobin 29 16 37 19 Hematocrit 26 16 35 22 WBC count 71 29 97 42 Platelets 96 32 107 46 International normalized ratio 45 25 59 29 Activated partial thromboplastin time 46 26 66 30 Fibrinogen 32 23 45 27 Analyte Median TAT Before, min Median TAT After, min Mean TAT Before, min Mean TAT After, min Hemoglobin 29 16 37 19 Hematocrit 26 16 35 22 WBC count 71 29 97 42 Platelets 96 32 107 46 International normalized ratio 45 25 59 29 Activated partial thromboplastin time 46 26 66 30 Fibrinogen 32 23 45 27 View Large Figure 1 View largeDownload slide Turnaround time for hemoglobin (Hg), hematocrit (HCT), fibrinogen (FIB), activated partial thromboplastin time (APTT), and international normalized ratio (INR) before (gray) and after (white) discontinuing the practice of repeating critical values. The notches indicate the 95% confidence interval about the median. The indicated P values were calculated using Welch t test. Figure 1 View largeDownload slide Turnaround time for hemoglobin (Hg), hematocrit (HCT), fibrinogen (FIB), activated partial thromboplastin time (APTT), and international normalized ratio (INR) before (gray) and after (white) discontinuing the practice of repeating critical values. The notches indicate the 95% confidence interval about the median. The indicated P values were calculated using Welch t test. Figure 2 View largeDownload slide Turnaround time for WBC count and platelet count (PLT) before (gray) and after (white) discontinuing the practice of repeating critical values. The notches indicate the 95% confidence interval about the median. The indicated P values were calculated using Welch t test. Figure 2 View largeDownload slide Turnaround time for WBC count and platelet count (PLT) before (gray) and after (white) discontinuing the practice of repeating critical values. The notches indicate the 95% confidence interval about the median. The indicated P values were calculated using Welch t test. Retrospective evaluation of the impact of removing the repeat practice was then performed—in particular, what if any impact delaying reporting of the critical values to perform the repeats had on our medical center. To answer this question, the time to clinical action was assessed. This involved selecting 15 patients with critically low hemoglobin who received transfusion with packed RBCs (pRBCs).Table 4 shows relevant timelines before and after removal of repeat testing. There was a nonstatistically significant decrease in the median time from laboratory order to pRBC order following removal of repeat testing. The bulk of this time savings was from the reduction in time from laboratory order to result verification. This time savings (11 minutes) was consistent with the time savings realized in Hg TAT (Table 3); the time from result verification to pRBC order, on the other hand, stayed relatively consistent. Resources Used in Repeating Critical Values The financial impact of repeating critical values was assessed using workflow modeling. This analysis considered the costs to the laboratory associated with repeating critical values: in particular, the cost of reagents/supplies and the labor. A small portion of this cost was in the actual reagent usage, at an average of $0.47 per test for the analytes considered. The much larger cost of this practice was in the labor associated with performing the repeats. This labor involves retrieving the sample, reprogramming the analysis, and assessing the repeat to determine which value to report. The amount of time spent performing these tasks varied widely among employees and at various times throughout the day; a workflow assessment yielded an average time per repeat of 28 minutes. The percentage of time that the employee was actively involved in the repeat process was also highly variable. Some technologists effectively multitasked while waiting for the critical repeat, while others cited concerns they would get distracted and delay reporting the critical values as reasons why they did not engage in multitasking. This variability makes it challenging to calculate the true labor cost. A “worst-case scenario” cost accounting was therefore done assuming that technologists devoted 100% of their time to repeating the critical values. As shown in Table 5 , the annual cost of this practice varied based on the number of critical values encountered in any given month. The worst-case scenario average annual cost of this practice to our laboratory was therefore estimated at $116,159. The actual cost incurred was something less than this depending on the amount of multitasking performed while repeating these critical values. Table 5 Estimated Financial Cost of Repeating Critical Valuesa Volume No. of Critical Values per Year No. of Annual Hours Annual Labor Cost, $ Annual Supply Cost, $ Total Cost, $ Low volume 4,416 2,061 72,135 2,076 74,211 High volume 9,408 4,390 153,650 4,422 158,072 Average 6,912 3,226 112,910 3,249 116,159 Volume No. of Critical Values per Year No. of Annual Hours Annual Labor Cost, $ Annual Supply Cost, $ Total Cost, $ Low volume 4,416 2,061 72,135 2,076 74,211 High volume 9,408 4,390 153,650 4,422 158,072 Average 6,912 3,226 112,910 3,249 116,159 aThe time to perform repeats was estimated at 28 minutes based on a workflow assessment. The labor cost was set at $35 per hour for a full-time, licensed medical technologist. The cost varied with the number of critical values and is therefore shown for low- and high-volume critical values as well as the average encountered in our medical center. View Large Discussion The practice of repeating critical values has been shown to add little value in chemistry but has been less well studied with regard to hematology and coagulation. The results shown here support the conclusions of others that this practice is not necessary for accurate reporting of hematology and coagulation critical values. In addition, we found this process to waste a substantial amount of resources and cause delays in clinical action. To our knowledge, this is the first report to explore the impact that removing repeat critical values has on the timeliness of clinical interventions. The prospective assessment of repeating critical values showed that 949 of the 950 repeats all fell within the expected analytical precision. The one sample that did not was close to the analytical precision and was reported as a critical value using the original critical result. While the repeats did not affect the results reported, it should be noted that other quality control measures were found to be important for preventing the release of erroneous critical values. The manual assessment of clots in samples with critically low hematology values was found to be an essential quality check to prevent erroneously reporting inaccurate results. However, the actual analytic repeating of the values was not necessary. While this practice has been known to use resources, the workflow assessment shown here provides an estimate of the worst-case scenario cost of this practice to clinical laboratories. In a time when clinical laboratories are constantly being asked to do more with less, it is helpful to evaluate the financial cost of practices that add little to no value to patient care. Our assessment shows that this practice could consume almost two full-time equivalents in our medical center. The actual cost incurred would depend on how much time technologists were able to multitask while repeating the critical values. In addition, this number will decrease for centers with fewer critical values. Still, it could represent a substantial amount of resources devoted to a process that does not improve patient care. Beyond wasting resources in the clinical laboratory, we found that the practice of repeating critical values was impeding patient care by delaying the TAT of these life-threatening laboratory values. For all tests, there was a statistically significant decrease in the median and mean TAT (Table 3 as well as Figures 1 and 2). The median delay in reporting was most significant for PLTs (64 minutes) and WBCs (42 minutes). This represents a much larger delay in TAT than has been reported for chemistry analytes such as glucose (17 minutes) or calcium (42 minutes).4,6 It is also substantially longer than has been reported for these two analytes.2 The cause of this is not entirely clear. One contributor could be the lack of automatic repeat options in hematology testing in our laboratory. Another contributor could be the volume of samples our laboratory processes. In the College of American Pathologists Q-probe study, only four of 73 hospitals studied had over 400 beds.2 Our medical center has 862 beds in addition to a substantial outreach program (total hematology volume of over 5.5 million annually). This high workload creates more situations where technologists are distracted with other tasks while trying to retrieve a sample to repeat a critical value. Whether or not this is the source of the longer time to repeat in our laboratory compared with that published in the Q-probe study is unclear. In addition to faster TAT, removing repeats was found to yield a more consistent process in the laboratory. This is evidenced by the shrinking of the interquartile ranges in the notched boxplots (Figures 1 and 2). This consistency is likely caused by simplifying the workflow and having less room for the technologist to get interrupted in the midst of confirming/reporting the critical result. The high variability in time required to perform the repeat analysis was observed in the workflow assessment. An additional benefit of simplifying the workflow and minimizing room for technologists to get interrupted was that a reduction in the maximum TAT was observed (eg, 196 minutes for WBCs with repeats compared to 128 minutes without). Most of the extremely long TATs were found to occur during shift change where a repeat could be forgotten in the handoff; removing the repeat process mitigated this issue. Perhaps one of the most intriguing observations was the possible reduction in time to order pRBCs in patients with critically low Hg (Table 4). While these results were not statistically significant, this could be related to the small numbers studied. A larger assessment to determine if reporting critical values faster leads to faster clinical intervention is needed. Still, the results in Table 4 suggest it is possible that the delays associated with performing repeat critical analysis actually translate to delays in clinical action. This would be consistent with the Q-probes study that found a high frequency of adverse events associated with delayed result reporting.2 Furthermore, this suggests that the process of repeating critical values is not only an unnecessary and wasteful process but also detrimental to patient care. Table 4 Median Time From Order of Hemoglobin Test to Order of Packed RBCs (pRBCs)a Group Laboratory Order to Result, min Result to pRBC Order, min Total Laboratory Order to pRBC Order, min With repeats 22 30 52 Without repeats 11 28 39 Group Laboratory Order to Result, min Result to pRBC Order, min Total Laboratory Order to pRBC Order, min With repeats 22 30 52 Without repeats 11 28 39 aRandomly selected patients with critically low hemoglobin values before and after removal of repeat testing (15 patients were chosen for each group). View Large There are three major limitations of this work. First, the operational savings are based on our laboratory, in which repeating critical values required manual intervention by a medical technologist. It is this labor that contributed the most to the delays in reporting critical values when performing repeats and also comprised most of the resources used for this practice. Thus, in a laboratory where repeats are programmed to occur automatically (either on the instrument or in the middleware), the improvement in TAT and cost savings will be minimal. However, it should be noted that while the impact to laboratory operations is less with automatic repeats, the necessity of this practice is still highly questionable. The second limitation is that only adult samples were evaluated; pediatric samples introduce additional preanalytic variables that were not considered in this study. However, our clinical laboratory does process pediatric samples, and we have ceased repeating critical values for these patients as well. So far, we have not been made aware of issues related to inaccurate critical values for these patients. We have, however, noticed that strict sample integrity checking processes are essential for accurate results on pediatric patient samples for critical and noncritical values alike. With proper quality control of preanalytic errors, it seems highly likely that analytic repeating of critical values in pediatric samples will also prove to be unnecessary. A final limitation is that this study only looked at the hematology and coagulation instruments used in our laboratory (ACL-TOP 500 CTS and DxH 800). The applicability of these findings to other instruments is thus unclear. It is likely that other instruments would perform comparably, and this was the finding of the CAP Q-probe, which showed no association between instrument make or model and the agreement between repeat critical values.2 Overall, this work shows that repeating critical hematology and coagulation results is an unnecessary practice that delays reporting of critical test results, wastes laboratory resources, and may contribute to delays in prompt clinical interventions. Any laboratory still performing critical value repeats should carefully evaluate this process in its facility. Our own laboratory has now completely discontinued the practice of repeating critical values. The major barriers we had to overcome in implementing this change were (1) longstanding beliefs that this was important for ensuring accurate results and (2) beliefs that advantages of changing this practice were too insignificant to warrant the effort. Other institutions might face comparable barriers. For the first, the approach described here (and in the cited literature) can hopefully serve as a template for how to conduct a similar study and use data to show the necessity (or lack thereof) of repeating critical values. For the second barrier, it is hoped that the data shown here are convincing that repeating critical values incurs more cost than some may think and has a price in terms of reduced quality of service, especially delayed reporting of life-threatening results. References 1. Lundberg GD. Critical (panic) value notification: an established laboratory practice policy (parameter). JAMA . 1990; 263: 709. Google Scholar CrossRef Search ADS PubMed 2. Lehman CM, Howanitz PJ, Souers Ret al. Utility of repeat testing of critical values: a Q-probes analysis of 86 clinical laboratories. Arch Pathol Lab Med . 2014; 138: 788- 793. Google Scholar CrossRef Search ADS PubMed 3. Chima HS, Ramarajan V, Bhansali D. Is it necessary to repeat critical values in the laboratory? Today’s technology may have the answers. Lab Med 2009; 40: 453. Google Scholar CrossRef Search ADS 4. Deetz CO, Nolan DK, Scott MG. An examination of the usefulness of repeat testing practices in a large hospital clinical chemistry laboratory. Am J Clin Pathol . 2012; 137: 20- 25. Google Scholar CrossRef Search ADS PubMed 5. Toll AD, Liu JM, Gulati Get al. Does routine repeat testing of critical values offer any advantage over single testing? Arch Pathol Lab Med . 2011; 135: 440- 444. Google Scholar CrossRef Search ADS PubMed 6. Onyenekwu CP, Hudson CL, Zemlin AEet al. The impact of repeat-testing of common chemistry analytes at critical concentrations. Clin Chem Lab Med . 2014; 52: 1739- 1745. Google Scholar CrossRef Search ADS PubMed 7. Niu A, Yan X, Wang Let al. Utility and necessity of repeat testing of critical values in the clinical chemistry laboratory. PLoS One . 2013; 8: e80663. Google Scholar CrossRef Search ADS PubMed 8. R Core Team. R: A Language and Environment for Statistical Computing . Vienna, Austria: R Foundation for Statistical Computing; 2013. © American Society for Clinical Pathology, 2018. All rights reserved. For permissions, please e-mail: firstname.lastname@example.org
American Journal of Clinical Pathology – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud