Process Optimization to Improve Immunosuppressant Drug Testing Turnaround Time

Transcription

1 Process Optimization to Improve Immunosuppressant Drug Testing Turnaround Time Vilte E. Barakauskas, PhD, 1,2 Tiffany A. Bradshaw, 3 Lonnie D. Smith, PharmD, 4 Christopher M. Lehman, MD, 5 and Kamisha L. Johnson-Davis, PhD 5,6 From the 1 Department of Pathology and Laboratory Medicine, Children s and Women s Health Centre of British Columbia, Vancouver, Canada; 2 Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, Canada; 3 ARUP Laboratories, Salt Lake City, UT; 4 University of Utah Health Care, Pharmacy Transplant Services, Salt Lake City; 5 Department of Pathology, University of Utah Health Sciences Center, Salt Lake City; and 6 ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT. Key Words: Process improvement; Process optimization; Turnaround time; Value stream map; Run chart; Immunosuppressants; Hospital laboratory Am J Clin Pathol August 216;146: DOI: 1.193/AJCP/AQW87 ABSTRACT Objectives: Timely reporting of immunosuppressant (ISP) drug level results is needed for transplant patient management. This study characterized the local ISP testing process, identified bottlenecks and implemented process improvements to meet turnaround time requirements. Methods: Laboratory information time stamps, direct observation and discussion with staff were used to construct a value stream map of the ISP testing process to identify process bottlenecks. Improvements were implemented to attain the required turnaround time. Results: Baseline performance of the existing ISP process (seven weeks, n ¼ 272 samples) indicated that only 28% of samples were reported by 2: PM. Major bottlenecks were identified to be the analytical run schedule, instrument delays, difficulty identifying ISP samples at intake, and difficulty collecting specimens. Process changes resulted in a median of 76% samples reported by 2: PM. Conclusions: Adjusting ISP collection and analysis processes improved the laboratory s ability to meet physician requested result reporting time of 2: PM. Immunosuppressant (ISP) medications are routinely used in the management of patients receiving solid-organ transplants. 1,2 Monitoring drug concentrations in the blood can guide adequate and safe dosing. ISP drug monitoring is performed using immunoassay or mass spectrometry methods. 1,2 It is known that immunoassay analysis is susceptible to cross-reactivity with inactive ISP drug metabolites and other interfering substances that may cause a significant positive bias between immunoassay and liquid chromatography-tandem mass spectrometry (LC-MS/MS) results. 3-6 However, immunoassay methods can be easily implemented in clinical laboratories using existing automated instrumentation, facilitating the timely analysis and reporting of ISP concentrations. 7 LC-MS/MS is considered the gold standard approach for ISP due to its high specificity and accuracy. 8-1 However, the complexity and cost of LC-MS/MS instrumentation may provide justification for laboratories to send-out ISP testing to reference laboratories that utilize LC-MS/MS. 7 While this is advantageous for many reasons, transport of samples to an off-site laboratory results in longer turn-around-times (TAT) for reportable results. Laboratories are thus challenged to balance the need for clinically accurate and economical ISP measurements with the need for timely results. At our institution, ISP monitoring is performed at a reference laboratory 1.5 miles away from the hospital laboratory by LC-MS/MS methods. ISP drugs are administered to inpatients in the morning. To obtain trough concentrations, sample collection is coordinated with dose administration. Generally, ISP drugs are administered every 12 hours, in the morning and evening. To obtain trough concentrations, 182 Am J Clin Pathol 216;146: DOI: 1.193/ajcp/aqw87 Downloaded from American Society for Clinical Pathology, 216. All rights reserved. For permissions, please journals.permissions@oup.com

2 samples must be collected predose (3-6 min prior at our institution). Nursing staff and phlebotomists collected venous samples for testing, while the laboratory staff was responsible for accessioning and packaging samples. Couriers are used to transport samples from the hospital to the testing laboratory. Once the specimens were received by the testing site, it took approximately 4.5 hours until results were available to the clinician. Results were transmitted electronically and were accessible in the hospital laboratory information system. Patient ISP concentrations were discussed during rounds to determine if dose adjustments were necessary. The TAT of patient results were impacted by the drug administration schedule, collection and sample transport scheduling, and analytical time needed for sample preparation and analysis. At the time of the study, care team physicians and pharmacists requested to have results reported earlier to facilitate dose adjustments, and therefore process improvement efforts were undertaken. This involved characterizing the total testing process to identify areas of improvement and metrics that could be used to monitor the process, implementing changes to the total testing process and finally, evaluating the effectiveness of these changes over time. Material and Methods Study Site The University of Utah Hospital (UH) is a 421-bed academic hospital in Salt Lake City. UH is a solid-organ transplant site, performing heart, kidney, liver, lung and pancreatic transplants. Both in- and out-patient ISP requests are encountered by the hospital laboratory. For the purposes of this study, the data set was limited to orders originating from in-patients. Tacrolimus constituted the majority of orders; Cyclosporine A and Everolimus orders were encountered much less frequently. Workflow ISP drug monitoring at UH is a multistep process involving sample collection by nursing or phlebotomy staff, specimen handling by the UH laboratory and analytical analysis performed by LC-MS/MS at a reference laboratory (ARUP) located 1.5 miles away. Daily ISP orders are placed for in-patients. Patients are followed by care teams consisting of nurses, physicians and clinical pharmacists. Medication dosing changes were made based on whole blood ISP concentrations, therefore it was necessary to have results available when the care team met for rounds. Historically the care team would round at 4: PM, but at the time of this study, rounds were changed to 2: PM. For the purposes of this study, analysis was limited to orders originating from four inpatient units: surgical specialty and transplant (SSTU), intermediate intensive care (IMCU), surgical intensive care (SICU) and the cardiovascular medical unit (CVMU); the units who most commonly care for transplant patients. Study Design This study involved three major phases with feedbackdriven iteration of some steps. First, the existing process of generating ISP results for UH inpatients was characterized and a process map assembled using the laboratory information system (LIS), direct observation, and discussion with laboratory and clinical staff. Second, bottlenecks in the process were identified and processes were modified to reduce the total process time to meet a result reporting time of 2: PM. Finally, the effect of these changes was evaluated prospectively with daily data review. Process Mapping The laboratory information system (LIS) was queried both retrospectively (August-October 212, 7 weeks; n ¼ 272 samples) and prospectively (October-December 212, 9 weeks; of which one week corresponded to initial process changes, n ¼ 269) for time stamp data. This included: scheduled collection time, actual sample collect time, the time a sample was received in the UH laboratory, time that sample was placed on a send-out list for transport to the reference laboratory, time received by reference laboratory and the time that final results were verified. The process of producing an ISP result was characterized using several approaches. Sample handling within the hospital and UH laboratory was characterized by direct observation of sample collection, transport and processing steps. Process sub-steps were identified, directly observed, and approximate time needed for these steps was estimated from LIS timestamps (7 weeks of retrospective data; median, min and max times between LIS timestamps were calculated). Medians were used to minimize the effect of nonconforming events, as these likely represent data entry errors, lost specimens, incorrect selection of dates or AM/PM times when placing orders. 11 Process boundaries were extended beyond the laboratory by speaking with nurses, physicians and pharmacists to identify potential causes of delays when samples were collected by floor staff. A process map was constructed to illustrate the various steps involved in generating an ISP result. The percentage of samples meeting turnaround time (TAT) requirements was evaluated weekly. Specimens that exceeded the required reporting time of 2: PM were investigated to identify the cause of the delay. Cause identification was limited to prospective data only. American Society for Clinical Pathology Am J Clin Pathol 216;146: Downloaded 183 from DOI: 1.193/ajcp/aqw87

3 Barakauskas et al /IMPROVING DRUG TESTING TURNAROUND TIME Process Improvement Bottlenecks in the process were identified using the process map. In addition, using the prospective data set, samples with late results were compared with those that were reported by 2: PM. The primary or root causes (most proximal to sample collection) of reporting delays, were tallied and evaluated using pareto charts. Process improvements were identified and implemented to alleviate bottlenecks in the testing process. Process Monitoring After implementing process changes, daily time stamps of ISP samples were monitored and late samples were followed-up to identify ongoing causes of delays. An investigation was conducted to determine whether the original cause of delays persisted, or if process changes revealed new causes of delays. The percentage of specimens that met the TAT goal was also monitored and summarized each month and distributed to the transplant care teams and laboratory staff. 11 to the testing site at the reference laboratory. To assist with the identification of specimens for ISP analysis, samples were sent to the laboratory in STAT testing bags. For transport to the testing site, samples were placed in a unique yellow bag, and transported by a dedicated courier. To expedite testing, the courier hand-delivered the yellow bags directly to the laboratory to bypass the specimen processing department in the reference laboratory. Median time to transport samples between the two laboratories was 43 minutes. The processes for sample analysis at the testing facility included the following: LIS entry to in-lab the specimens, set up for the sample run order, sample extraction and processing, then analysis by LC-MS/MS. Once the UH sample results were obtained, the quality controls (QC) and other quality assurance (QA) parameters were evaluated, if acceptable, patient results were reviewed, verified and released. Results were transmitted electronically to UH and were accessible in the hospital laboratory information system. On average, the TAT from in-lab status at the reference lab to the time of result reporting was about 4.5 hours, Figure 1. Results Process Characterization The process involved in generating an ISP result is illustrated in swim-lane format in Figure The process begins with the administration of ISP medication and a physician ordering a blood ISP concentration. This process involves six distinct groups of people/locations, denoted by vertical lanes in Figure 1. Sample collection time is determined by the time the medication dose was administered. Nursing staff either collected the blood specimens, or scheduled collection by laboratory phlebotomists. It appeared that samples were typically collected within 5 minutes of the scheduled collection time, Figure 1. The minimum and maximum time between scheduled collection and actual sample collection (Figure 1) appear to be outlier data points, and likely were the result of manual entry errors (for example, scheduling that a blood specimen be collected at 9: AM instead of 9: PM, or selecting yesterday s date instead of the current date). After sample collection, samples are transported to the UH laboratory either by hand, a pneumatic tube system or a dumb-waiter -type elevator. Median transport time was determined to be 17 minutes. Specimen processors must find the ISP samples among other samples arriving in the laboratory, enter them into the LIS (to show in-lab status), place them on a transport list and package them for transport Process Bottlenecks The process map identified several steps that had the potential to cause delays in result reporting (summarized in Table 1 ). First, several process steps were set up to meet a reporting time of 4: PM, as this was previously acceptable to clinical staff. This included a dedicated analytical run for UH specimens at 1: PM, with results usually reported by 4: PM. The process map shows that once a sample was received by the testing facility it took a median of 4 hours and 34 minutes to report a final ISP result. A run time of 1: PM therefore could not meet the new TAT expectation of 2: PM. The testing site processes multiple batched runs for ISP testing each day, such that if samples arrive earlier, they ccould be analyzed on a different batch at an earlier time of day. Courier pick-ups were also set for the 1: PM run time. Thus, scheduled run time was a potential process bottleneck. Another potential delay in the process was identified through direct observation the transport of samples from the hospital wards to the UH laboratory in STAT bags. Many non-isp samples were transported in STAT bags, especially in the morning, which reduced the effectiveness for using these bags to identify specimens for ISP analysis. Although specimen processors prioritize STAT bags, ISP samples were mixed in with these bags, and therefore not accurately prioritized. In addition, sample processing involved a number of subtasks (removing tubes from the bags, scanning barcodes, manual adjustment of collection 184 Am J Clin Pathol 216;146: American Society for Clinical Pathology 184 DOI: 1.193/ajcp/aqw87 Downloaded from

4 Physicians Nursing Phlebotomy UH Lab Courier Reference Lab Yes Prescribe ISP dose No Administer ISP Order ISP drug monitoring Schedule sample collection Collect blood sample Transport sample Collect blood sample Transport sample :5 [ 24:, 24:19] Receive sample bags :17 [:, 1:2] Sample processing Identify ISP samples :18 [:, 12:39] Prepare samples for transport Pick up samples Drive samples Hand-deliver samples Sample processing :43 [:19, 3:5] Build run Extraction Analysis QC and data evaluation 4:43 [2:18, 12:7] Result interpretation Data OK? Yes No Dose needs adjusting? and release Figure 1 Process map for immunosuppressant (ISP) result generation for University of Utah Hospital (UH) in-patients. Columns represent the major groups of personnel and locations involved in the process. Movement in the horizontal direction shows movement of the sample from one group to the other, illustrating hand-offs of the process. The sequence of events and steps involved is illustrated in the vertical direction, moving down the chart. Median, minimum and maximum times to complete groups of steps (delineated by shading) are noted when available. Legend: median hours:minutes [min, max]. QC, quality control. time, management of paperwork/requisitions), each of which could take a variable amount of time depending on whether samples were properly labeled, arrived with the correct paperwork and were associated with other exceptions (ie. incorrect sample type, tube or time). Discussions with the nursing staff identified blood draws as another potential bottleneck. Although nurses understood that specimens should be transported to the laboratory promptly, situations may occur that could result in specimen transport delay. In addition, discussions with the phlebotomy staff and review of internal QA notes indicated that difficulty in collecting samples could also contribute to delays. Collection delays may occur when patients were unavailable when the phlebotomists arrived, blood draws were American Society for Clinical Pathology Am J Clin Pathol 216;146: Downloaded 185 from DOI: 1.193/ajcp/aqw87

5 Barakauskas et al /IMPROVING DRUG TESTING TURNAROUND TIME scheduled at incorrect times or ordered too late to meet the requested TAT, or difficulty with drawing blood from the patient. Causes of Delays To add to the qualitative observations of bottlenecks identified, objective measures to link reporting delays to LIS timestamps were investigated. For samples with results reported after 2: PM, the cause of the delay was estimated from LIS timestamps and the cause most proximal to sample collection was tallied Table 2. Prior to major process changes, the most common causes attributed to result delays were the analytical run schedule, late collections and instrument problems Figure 2A. Process Improvement Based on identified bottlenecks and delay causes, several ISP sample process changes were implemented (Table 1). Phlebotomists were provided with continuing education to emphasize the need for appropriately timed specimen collections for ISP testing. Efforts were made to dispatch experienced phlebotomists for blood draws to reduce sample draw delays. To improve sample processing for ISP testing, samples were placed in STAT bags marked ISP to differentiate these bags from other STAT tests that arrived in the laboratory during the morning rush. Reeducation for phlebotomists and nursing staff was also undertaken to Table 1 Bottlenecks and Improvement Initiatives in the ISP Testing Process Bottlenecks Discovered Improvement Initiative Analytical run schedule Change analysis time for UH samples to 9:3 AM Dedicated staff for sample transport preparation Adjusted courier schedule Late sample collections Phlebotomist and nurse education Analytical/instrument delays Notification tree to inform team STAT bag usage ISP STAT bags Difficulty drawing from patient Assigning experienced phlebotomists ISP, immunosuppressant; UH, University of Utah Hospital. discourage batching of samples for pneumatic tube transport, so that an ISP sample would be transported immediately following collection. The reference laboratory added a 9:3 AM run to their daily run schedule to allow 4.5 hours of time for extraction, analysis and reporting of results. The courier schedule was also adjusted to allow the maximum time for sample draws and transport for samples to arrive at the testing laboratory in time for the 9:3 AM analytical run. Changes were implemented over the span of a week, although continued optimization occurred after initial process monitoring in response to staff feedback (see below). Process Monitoring A run chart was used to monitor the percentage of samples reported by 2: PM each week, and the median compared, pre and post-changes Figure 3A. Minor optimization occurred during the 9 weeks postimplementation (discussed below); however, because these were discovered on an ongoing basis, they were included in the post-implementation phase. Prior to process changes, only 27% of samples were reported by 2: PM. After initial changes this increased to a median of 76%. Sustained effects of process optimization were seen for months after, with 69 to 9% of samples meeting the reporting time goal Figure 3B. Reporting times continue to be monitored on a monthly basis. Initial process modifications were successful in mitigating the major cause of sample delays identified initially (analytical run schedule), and revealed other causes that remained (late sample collection, late receipt into the laboratory, collection was scheduled late; Figure 2B ) and could be the focus of future process improvement. Iterative Optimization During process monitoring, informal feedback was solicited from care teams, nurses, and the testing facility site. It was learned that not all floors had implemented the ISP -labeled STAT bags, and another request was made that all floors do so. To ensure that the testing facility did not miss the sample delivery, couriers were asked to announce the arrival of ISP specimens when they were Table 2 Time Stamps From Major Process Steps for Representative Samples Scheduled Collection Actual Collection Received at UH Lab Packaged for Transit In-Lab at Testing Site Result Reported Causes of Delay 1 8:3 8:25 8:39 8:52 9:31 13:4 2 8:3 9: 1:11 11:33 12:1 16: Late collection Long transit to lab 3 8:3 8:4 8:47 8:52 9:24 14:5 Instrument problems 4 9: 9:8 9:13 1:3 1:28 14:32 Collection scheduled late UH, University of Utah Hospital. 186 Am J Clin Pathol 216;146: American Society for Clinical Pathology 186 DOI: 1.193/ajcp/aqw87 Downloaded from

6 A 1 B Frequency (%) 6 4 Frequency (%) Analytical Run Scheduled Collected After 8:45 Instrument Problem Transit to Reference Lab Collection Scheduled Late Collected After 8:45 Received Late in UH Lab Scheduled Later Than 8:45 Missed Transfer List Other Figure 2 Causes of result delays identified through laboratory information system time stamps and internal quality assurance notes. A, Prior to and during process modification (n ¼ 35). B, After initial process modification (n ¼ 46). Where multiple causes contributed to delays, the cause most proximal to sample collection was tallied. Bars show the percentage of late samples; the black line indicates cumulative frequency. delivered, to provide an audible and visual cue that the analytical run could commence. In addition, three subsequent improvements have been implemented, which resulted in 9% compliance with 2: PM reporting rate. The hospital had implemented a purple ISP label to apply to the bags for better recognition of these specimens, which had been standardized for the phlebotomy and nursing units. The nursing staff had prelabeled bags near the pneumatic tube station to facilitate use. Lastly, the processing section now has a compliance graph posted in an area where the group has daily, 1 minute quality improvement discussions for ISP compliance and other quality initiatives. Discussion ISP drug monitoring is a complex process due to the need for timed specimens that are coordinated with drug dosing schedule the need for accurate analytical methods, timely result reporting and integration of many healthcare personnel to manage a single patient. Failure to meet test reporting requirements can be viewed as a nonconforming event. 13 Detecting and investigating nonconformities can provide objective data with which to target laboratory improvement efforts. 14 It has been recommended that laboratories investigate nonconforming events to uncover the reasons for delays, as the causes of delays can vary between tests, locations and individual laboratories. 14 Timely result reporting or TAT is a recognized laboratory Quality Indicator and expressing TAT as the rate of nonconformities is a means of evaluating the degree to which laboratory services meet the needs of those who use their services. 15 Causes of ISP result delays are scarce in the literature. The bottlenecks discovered here have been observed to delay TAT of other tests. Late sample collections, difficulty drawing specimens, instrument delays and specimen transit delay were all among the specific reasons for delayed TAT of STAT specimens originating from the ED and ICU in a CAP Q-Probes study of 496 laboratories. 14 Having identified causes of result delays, improvements require the coordination of many components, and information from all parties involved, as was demonstrated here. Several strategies for engaging personnel were used in this study. This included communication with all of the staff involved in the process, during various stages of the study, collecting empirical data rather than making assumptions American Society for Clinical Pathology Am J Clin Pathol 216;146: Downloaded 187 from DOI: 1.193/ajcp/aqw87

7 Barakauskas et al /IMPROVING DRUG TESTING TURNAROUND TIME A % of Samples Reported by 2 PM Aug Dec Week No. of Samples B % of Samples Reported by 2 PM Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Jan Feb Mar Apr May JuneJuly Aug Sep Oct Nov Dec Month No. of Samples Figure 3 Monitoring of immediate and sustained effects of process changes. A, Samples reported by 2: PM each week prior to process changes (week 6 to ) and after initial changes (week to 9). B, Sustained effects summarized by month. Total number of ISP orders each week is shown by the bars. Percentage of samples reported by 2: PM is shown by the triangles. Arrow indicates the time at which process changes began. Median is given by the dashed line, before and after implementation. about the process, and providing feedback and data to increase motivation and sustain improvements. Collaboration between laboratory and clinical personnel is key in process optimization and reduction of diagnostic errors. 16 Thorough process characterization and engaging staff at all levels may account for the success in TAT improvement, when compared to previously published strategies. For example, Melanson et al found that an IT-based prompt reminding nurses of the correct timing of vancomycin sample collection was ineffective in significantly improving sample timing. 17 The initiative described, however, was laboratory-initiated and may not have involved clinical staff from the development and onset. Changing the workflow in a hospital is a complex operation because a change in one area will affect other processes. In the present study, improving the analytical run schedule on its own was not sufficient, and required changes in courier schedules and sample order/collection time for patient dosing schedules. The laboratory could not have made all these changes on its own, illustrating that laboratory and clinical personnel collaboration is essential for successful laboratory process change. Continued monitoring of result reporting times (Figure 3B) indicated that improvements have been sustained. This suggests that process changes and regular feedback have been effective. Continuous monitoring of TATs of other laboratory tests was also associated with a decreased number of nonconforming events. 15 A drawback which can occur from process monitoring and routine feedback is that sustained process improvements may happen due to the Hawthorne effect 18 behavior modification that occurs because the process is being monitored. Unfortunately, it is not possible to separate the Hawthorne effect from the effects of process changes. In addition to achieving the desired reporting time goal, benefits of process mapping include discovery of previously unknown approaches to improve the process. In our case, this included unit-initiated prelabeling of ISP STAT-bags to make it easier for nurses to comply with this step of the process. During courier delivery of samples to the testing laboratory, it was discovered that a verbal announcement of sample arrival provided a signal for staff. Finally, it was recognized by both laboratory and clinical staff that there could never be a 1% guarantee of timely results because unexpected situations could arise (for example, instrument malfunction or inclement weather). In such cases it was identified that notifying the care teams of a potential delay was just as 188 Am J Clin Pathol 216;146: American Society for Clinical Pathology 188 DOI: 1.193/ajcp/aqw87 Downloaded from

8 valuable as timely results. This allowed care teams to adjust their meeting time, patient rounding order etc. to minimize impact on efficiency and patient care. Telephone notification provided a direct link between the laboratory staff and physician and/or pharmacist. These initiatives are examples of process improvements that were made after initial changes were implemented, as they only surfaced during process monitoring and feedback from personnel. Despite successful and sustained process changes, process improvement is a continuous activity, and future improvements are likely possible. Among sources of additional TAT gains, implementation of mass spectrometry-based methods for ISP testing in the hospital laboratory is an option. Also, as identified in the Pareto analysis, late sample collections remain among the top causes of late results (Figure 2B); further gains in TAT may manifest if additional means of ensuring on-time collections were sought. Several limitations in the present study reflect the realtime nature of this analysis. First, time stamp data was not available for all stop points in the process; it is limited to places in the process that sample barcodes were generated, scanned, or sample information was modified in the LIS. Exact times at which the sample was drawn could not be guaranteed because labels may be carried for some time after printing. Also, because samples were not scanned prior to transporting via the pneumatic tube system, transport time was not resolved from sample collection or processing times. Some process stages composed of smaller steps could only be characterized as one large step. Second, because ISP testing was already in place, process mapping and changes were undertaken on a live process. Pre- and post-change states could not be completely isolated from one another, as some changes occurred gradually, and other changes were added later on due to feedback and real-time monitoring of results. Third, the effect that ISP process improvement had on other laboratory tests and process was not assessed. Another limitation of this study was the effect of late laboratory results on patient care was not investigated. Novis et al, however, suggested that prolonged TAT could be used as an indicator of physician satisfaction. 15 Finally, the present study was not designed to assess whether ISP testing was utilized appropriately. Other therapeutic drug monitoring studies indicate that a significant portion of therapeutic drug monitoring orders were comprised of inappropriately timed specimens and identified collection time as an important factor of result and interpretation accuracy. 19 Results from study, could not characterize whether inappropriate timing of draws relative to dosing contributed to reporting delays. In summary, this study focused on optimizing the ISP testing process; however, the approach used here will likely be applicable for other tests. The approach may be particularly relevant to larger health care organizations operating with hub and spoke laboratory models, and rural or isolated laboratories that need to improve the workflow for send-out testing. Evaluation of the entire process, from patient care to reporting will provide empirical data to better focus process optimization efforts in many settings. The sustained results observed here demonstrate that it is well worth the effort. Corresponding author: Kamisha Johnson-Davis, 5 Chipeta Way, Mail Code 115, Salt Lake City, UT, ; kamisha.johnson-davis@aruplab.com. Acknowledgments: The authors thank the clinical and laboratory staff involved in transplant patient care and ISP testing, including Shawn Clark, JoD Fontenot, and Takara Blamires, ARUP Mass Spec II laboratory staff, UH laboratory staff, as well as nursing and other clinical staff working in transplant care at the UH. References 1. Tsunoda SM, Aweeka FT. Drug concentration monitoring of immunosuppressive agents: focus on tacrolimus, mycophenolate mofetil and sirolimus. BioDrugs. 2;14: Taylor PJ, Tai CH, Franklin ME, et al. The current role of liquid chromatography-tandem mass spectrometry in therapeutic drug monitoring of immunosuppressant and antiretroviral drugs. Clin Biochem. 211;44: Bolstad N, Warren DJ, Nustad K. Heterophilic antibody interference in immunometric assays. Best Prac Res Clin Endocrinol Metab. 213;27: Hoffer E, Kurnik D, Efrati E, et al. Comparison of everolimus QMS immunoassay on architect ci41 and liquid chromatography/mass spectrometry: lack of agreement in organtransplanted patients. Ther Drug Monit. 215;37: Lower DR, Cropcho L, Rosendorff A. Comparison of CEDIA FK56 assay with HPLC/MS/MS in a large cohort of pediatric patients. Am J Clin Pathol. 213;139: Saitman A, Metushi IG, Mason DS, et al. Evaluation of the waters masstrak LC-MS/MS assay for Tacrolimus and a comparison to the Abbott architect immunoassay. Ther Drug Monit. 216;38: Brandhorst G, Oellerich M, Maine G, et al. Liquid chromatography-tandem mass spectrometry or automated immunoassays: what are the future trends in therapeutic drug monitoring? Clin Chem. 212;58: Soldin SJ, Steele BW, Witte DL, et al. Lack of specificity of cyclosporine immunoassays: results of a College of American Pathologists study. Arch Pathol Lab Med. 23;127: Toraishi T, Takeuchi H, Nakamura Y, et al. Falsely abnormally elevated blood trough concentration of tacrolimus measured by antibody-conjugated magnetic immunoassay in a renal transplant recipient: a case report. Transplant Proc. 212;44: D Alessandro M, Mariani P, Mennini G, et al. Falsely elevated tacrolimus concentrations measured using the ACMIA method due to circulating endogenous antibodies in a kidney transplant recipient. Clin Chim Acta. 211;412: American Society for Clinical Pathology Am J Clin Pathol 216;146: Downloaded 189 from DOI: 1.193/ajcp/aqw87

9 Barakauskas et al /IMPROVING DRUG TESTING TURNAROUND TIME 11. Hawkins RC. Laboratory turnaround time. Clin Biochem Rev. 27;28: Lee Q, Snyder B. The Strategos Guide to Value Stream and Process Mapping: [Genesis of Manufacturing Strategy]. Bellingham, WA: Enna Products Corp; Clinical and Laboratory Standards Institute (CLSI) Nonconforming Event Management Document No. QMS11-ED Steindel SJ, Novis DA. Using outlier events to monitor test turnaround time. Arch Pathol Lab Med. 1999;123: Novis DA, Walsh MK, Dale JC, et al. Continuous monitoring of stat and routine outlier turnaround times: two College of American Pathologists Q-Tracks monitors in 291 hospitals. Arch Pathol Lab Med. 24;128: Thammasitboon S, Thammasitboon S, Singhal G. Systemrelated factors contributing to diagnostic errors. Curr Prob Pediatr Adolesc Health Care. 213;43: Melanson SE, Mijailovic AS, Wright AP, et al. An intervention to improve the timing of vancomycin levels. Am J Clin Pathol. 213;14: Adair JG. The Hawthorne effect: a reconsideration of the methodological artifact. J Appl Psychol. 1984;69: Morrison AP, Melanson SE, Carty MG, et al. What proportion of vancomycin trough levels are drawn too early? frequency and impact on clinical actions. Am J Clin Pathol. 212;137: Am J Clin Pathol 216;146: American Society for Clinical Pathology 19 DOI: 1.193/ajcp/aqw87 Downloaded from