Article Text

Download PDFPDF

Using Lean principles to optimise inpatient phlebotomy services
  1. Rachel D Le1,2,
  2. Stacy E F Melanson1,
  3. Katherine S Santos3,
  4. Jose D Paredes4,
  5. Jonathan M Baum1,
  6. Ellen M Goonan1,
  7. Joi N Torrence-Hill5,
  8. Michael L Gustafson6,
  9. Milenko J Tanasijevic1
  1. 1Department of Pathology, Clinical Laboratories Division, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA
  2. 2Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA
  3. 3Performance Improvement, Brigham and Women's Faulkner Hospital, Boston, Massachusetts, USA
  4. 4Continuous Improvement Solutions, UL (Underwriters Laboratories), Northbrook, Illinois, USA
  5. 5City of Hope National Medical Center, Duarte, California, USA
  6. 6Brigham and Women's Faulkner Hospital, Boston, Massachusetts, USA
  1. Correspondence to Milenko J Tanasijevic, Brigham and Women's Hospital, 75 Francis Street, Amory 2, Boston, MA 02115, USA; mtanasijevic{at}partners.org

Abstract

Background In the USA, inpatient phlebotomy services are under constant operational pressure to optimise workflow, improve timeliness of blood draws, and decrease error in the context of increasing patient volume and complexity of work. To date, the principles of Lean continuous process improvement have been rarely applied to inpatient phlebotomy.

Aims To optimise supply replenishment and cart standardisation, communication and workload management, blood draw process standardisation, and rounding schedules and assignments using Lean principles in inpatient phlebotomy services.

Methods We conducted four Lean process improvement events and implemented a number of interventions in inpatient phlebotomy over a 9-month period. We then assessed their impact using three primary metrics: (1) percentage of phlebotomists drawing their first patient by 05:30 for 05:00 rounds, (2) percentage of phlebotomists completing 08:00 rounds by 09:30, and (3) number of errors per 1000 draws.

Results We saw marked increases in the percentage of phlebotomists drawing their first patient by 05:30, and the percentage of phlebotomists completing rounds by 09:30 postprocess improvement. A decrease in the number of errors per 1000 draws was also observed.

Conclusions This study illustrates how continuous process improvement through Lean can optimise workflow, improve timeliness, and decrease error in inpatient phlebotomy. We believe this manuscript adds to the field of clinical pathology as it can be used as a guide for other laboratories with similar goals of optimising workflow, improving timeliness, and decreasing error, providing examples of interventions and metrics that can be tailored to specific laboratories with particular services and resources.

  • Quality Assurance
  • Laboratory Tests
  • Laboratory Management

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Phlebotomy services are under constant operational pressure to optimise service while carefully managing cost and resources in the context of increasing volume and complexity of work.1 A joint report from the College of American Pathologists and the Center for Disease Control Outcomes Working Group found that 41% (35 922/88 038) of errors including donor/recipient misidentification, wrong container used for specimen collection, and specimen mishandling, occurred in the preanalytical phase of transfusion medicine,2 highlighting the importance of upstream quality management. With errors observed in 0.60% (15 503/2 583 850) of inpatient tests versus only 0.039% (792/2 032 133) of outpatient tests in a single year at San Raffaele Hospital, Bonini et al3 recognised the significance of preanalytical errors in inpatients and speculated that such discrepancy between inpatients and outpatients is most likely due to the nature of inpatient testing marked by multiple specimen drawings and higher test complexity. Additionally, turnaround time (TAT) for inpatient laboratories has been shown to be critical for clinicians in order to improve inpatient management and decrease length of hospital stay.4 Collectively, these findings illustrate the importance of a streamlined phlebotomy service to minimise TAT and preanalytical errors in order to improve patient care management and satisfaction.

Since the late 1980s, clinical laboratories have employed Lean principles and tools derived from the Toyota Production System in order to identify and eliminate sources of waste and optimise workflow. Examples of these tools include kanban cards (stock supply indicators), spaghetti diagrams (visual depictions of actual flow), 5S (5 steps of workplace organisation: sort, set-in-order, shine, standardise, and sustain), Workouts (single-day, focused events dedicated to developing action plans) and Kaizens (multiday, rapid-improvement events involving frontline staff as experts in process brought together to identify issues and propose and iteratively test solutions in real time in order to improve overall performance and outcomes).5–10

In many pathology disciplines, the application of Lean tools has been shown to increase efficiency and productivity, and decrease TAT and error without additional resources.6–10 A study conducted by Zarbo et al7 found that the total number of specimen defects was reduced by 55% after standardisation of specimen collection and labelling processes in surgical pathology, while a study by Raab et al8 observed a 42 min decrease in TAT after implementation of 200 Lean interventions on the histopathology section of an anatomical pathology laboratory. Likewise, our group has optimised inpatient phlebotomy staffing using a data-derived staffing tool9 and reduced the number of mislabelled and unlabelled specimens from 5.5 to 3.2 per 10 000 draws through the implementation of an electronic handheld positive patient identification (PPID) device.10

In recent years, Brigham and Women's Hospital (BWH) clinical laboratory leadership identified inpatient phlebotomy as a target for process improvement in the context of increasing patient volume, test complexity, and clinician demands for shorter TAT, particularly during morning rounds when 60% of our inpatient specimen collections occur. In this study, we conducted three Kaizens and one Workout in inpatient phlebotomy focused on optimising supply replenishment and cart preparation and organisation, communication and workload management, blood draw process standardisation, and balanced rounding schedules and assignments. We assessed the impact of our interventions in conjunction with our efforts to optimise inpatient staffing9 using three primary outcome metrics: (1) percentage of phlebotomists drawing their first patient by 05:30 for 05:00 rounds, (2) percentage of phlebotomists completing 08:00 rounds by 09:30, and (3) number of errors per 1000 draws.

Methods

Study site

This study was performed at BWH, Boston, Massachusetts, USA, a large academic medical centre that houses 777 beds throughout 50 inpatient care units or ‘pods.’ The BWH pathology department employs 40 inpatient phlebotomists who perform approximately 50% of all inpatient draws (about 200 000 venipunctures annually or 600 draws daily) on all inpatient care units with the exception of the neonatal intensive care unit. Phlebotomy performs venipunctures on all patients without central lines (ie, approximately 50%), whereas nursing is responsible for collection in patients with central lines. Of the total inpatient phlebotomy draws, roughly 60% (ie, 360 draws daily) are performed throughout morning rounds (05:00 to 11:00), during which 16–18 inpatient phlebotomists are assigned to draw patients.

Phlebotomy workflow

The inpatient phlebotomy staff schedule consists of three shifts: 05:00 to 13:30, 13:30 to 24:00 and 24:00 to 08:30.9 Rounds are conducted at 04:00 (well newborns only), 05:00 (STAT and discharge rounds), 08:00 (routine morning rounds), 10:00 (additional morning rounds), 12:00, 14:00, 16:00, 19:00, 22:00 and 24:00. Phlebotomist duties from the start to end of a shift are depicted in figure 1. BWH does not have electronic order communication between physician order entry and the laboratory information system (LIS). Therefore, phlebotomists are unaware of their workload until they arrive on the pod, review the list of patients to be drawn or ‘pod sheet’, and acquire paper requisitions containing specific orders for each patient.

Figure 1

Typical inpatient phlebotomy workflow with titles, dates and steps targeted for each Kaizen and Workout event.

Study design

The study occurred in two phases: (1) process improvement (August 2009–April 2010): completion of three Kaizens (August 2009, October 2009 and February 2010, for supply replenishment and cart standardisation, communication and workload management, and blood draw process standardisation, respectively), a Workout (April 2010 for rounding schedules and assignments), and implementation of a new staffing model (April 2010)9 and (2) postprocess improvement (May 2010–July 2013): assessment of the cumulative impact of all interventions using three primary outcome metrics as described below.

Though the Workout addressed rounding assignments, we recognised that a thorough analysis of our current staffing model was necessary to maximise the benefits of all the other interventions. The details and impact of the new staffing model, which was also developed using Lean principles, are addressed in a previous publication.9 Because the staffing changes were in close proximity to our Lean improvements, we decided to include the new staffing model as a part of our process improvement period in this study.

Outcome metrics

The three primary outcome metrics used to gauge the cumulative effect of the Kaizens, Workout, Lean interventions, and staffing changes were (1) percentage of phlebotomists drawing their first patient by 05:30 for 05:00 rounds, (2) percentage of phlebotomists completing 08:00 rounds by 09:30, and (3) number of errors per 1000 draws, defined to include mislabelled specimens, unlabelled specimens, incorrect specimen received, insufficient quantity received, and no specimen received. These metrics were chosen to span across the entire phlebotomy process and evaluate quality and efficiency. Goals were set at 80% of phlebotomists to draw their first patient by 05:30 for 05:00 rounds, and at 90% of phlebotomists to complete 08:00 rounds by 09:30. Additionally, we sought to reduce the error rate to 4 per 1000 draws with no mislabelled and unlabelled specimens. Data for all three metrics were recorded between October 2009 and July 2013.

In order to capture first draw times for each phlebotomist, patient draw times for 05:00 rounds were collected once a week from our electronic handheld PPID device (n=195 weeks; 46 months).10 Round completion times were also collected once a week using manually recorded data by the team leader of when each phlebotomist checked-in upon completion of 08:00 rounds (n=195 weeks; 46 months). The median of the percentages for each week of the month for first draw and round completion data were determined to yield monthly data. For instance, recorded percentages of phlebotomists drawing their first patient by 05:30 for 05:00 rounds in November 2009 were 77% for week 1, 62% for week 2, 83% for week 3, and 54% for week 4. Thus, the median percentage for November was determined to be 69%. Lastly, the number of errors per 1000 draws was determined each month (n=46 months). Errors are generally identified and documented when laboratory staff compare specimens received with accompanying requisitions, and determine if the phlebotomist drew and labelled the appropriate specimens properly. Error data was compiled by a laboratory supervisor.

As a secondary metric, 5S scores for supply cabinet and cart organisation were determined by a phlebotomist using an audit tool with a scale from 0 to 15. Results were recorded weekly (n=195 weeks; 46 months) with the goal of achieving and maintaining a 5S score above 9.

Statistics

To evaluate the variability and assess stability of the outcome metrics during and after process improvement, process control charts were generated. Specifically, P charts were generated for percentage of phlebotomists drawing their first patient by 05:30 for 05:00 rounds, and completing 08:00 rounds by 09:30 as appropriate for proportional attribute data with variable subgroup sizes.11 By contrast, a C chart was created for number of errors per thousand draws, as the data reports count attribute data with invariable subgroup sizes.12 Upper and lower control limits were calculated based on appropriate formulas for the said process control charts.11 ,12

Results

Outcome metrics

The monthly median percentage of phlebotomists drawing their first patient by 05:30 for 05:00 rounds noticeably increased after process improvement (figure 2). We reached our goal of 80% of phlebotomists drawing their first patient by 05:30 for 05:00 rounds in 16 out of 39 months (41%) postprocess improvement. We also observed only 5 months’ postprocess improvement in which our percentage was less than 70%.

Figure 2

P chart of median percentage of phlebotomists drawing their first patient by 05:30 for 05:00 rounds. The black line represents the median percentage of phlebotomists drawing their first patient by 05:30. The solid horizontal grey line represents our goal of 80% of phlebotomists drawing their first patient by 05:30. The dashed vertical grey line separates the process and postprocess improvement periods. The dashed horizontal black and grey lines represent the mean of the median percentages and lower control limit, respectively (mean: 53, 78 percentage meeting criteria during process and postprocess improvement, respectively). The date of each Kaizen and Workout event and implementation of the new staffing model are also indicated.

Figure 3 illustrates a marked improvement in the monthly median percentage of phlebotomists completing 08:00 rounds by 09:30 postprocess improvement. We were able to reach our goal of 90% of phlebotomists completing 08:00 rounds by 09:30 in 36 out of the 39 months’ (92%) postprocess improvement.

Figure 3

P chart of median percentage of phlebotomists completing 08:00 rounds by 09:30. The black line represents the median percentage of phlebotomists completing 08:00 rounds by 09:30. The solid horizontal grey line represents our goal of 90% of phlebotomists completing 08:00 rounds by 09:30. The dashed vertical grey line separates the process and postprocess improvement periods. The dashed horizontal black and grey lines represent the mean of the median percentages and lower control limit, respectively (mean: 53, 97 percentage meeting criteria during process and postprocess improvement, respectively). The date of each Kaizen and Workout event and implementation of the new staffing model are also indicated.

Lastly, the monthly number of errors per 1000 draws decreased during the postprocess improvement phase (figure 4), though never reaching our goal of 4 errors per 1000 draws. However, we recorded 3 months’ postprocess improvement (February 2011, June 2011, February 2013) in which there were no mislabelled and unlabelled specimens (figure 5).

Figure 4

C chart of number of errors per 1000 draws. The black line represents the number of errors per 1000 draws per month. The solid horizontal grey line represents our goal of less than 4 errors per 1000 draws. The dashed vertical grey line separates the process and postprocess improvement periods. The dashed horizontal black and grey lines represent the mean of the errors per 1000 draws and upper and lower control limits, respectively (mean: 9.16, 6.64 errors per 1000 draws during process and postprocess improvement, respectively). The date of each Kaizen and Workout event and implementation of the new staffing model are also indicated.

Figure 5

Number of total mislabelled and unlabelled specimens. The black line represents the total mislabelled and unlabelled specimens per month. The grey dashed vertical line separates the process and postprocess improvement periods. The date of each Kaizen and Workout event and implementation of the new staffing model are also indicated.

Supply replenishment and cart standardisation (Kaizen #1, August 2009)

Prior to process improvements, frequently and infrequently used supplies were intermixed and stored in multiple locations (eg, cabinets, phlebotomy carts, and personal supply bags). Supplies also lacked labels or visual indicators resulting in inefficient supply management and expired items. Assessment of 5S scores prior to process improvement in supply storage areas showed a baseline score of 6. As a result of Kaizen #1:

  • 55 lbs (approximately 25 kg) of expired items were removed.

  • Supplies in cabinets were arranged based on frequency of use, given a single designated storage location, and clearly labelled.

  • A new inventory management system using kanban cards was implemented to ensure that supplies were always available, minimise instances of expired items, and visually know when to reorder without the need for counting supplies in inventory.

  • All phlebotomy carts were standardised to have the same layout and specific par levels for each item.

  • The process of restocking and cleaning the phlebotomy carts was moved to the end of the shift to increase efficiency at the start of the shift.

  • Individual scanners, printers, batteries and carts were assigned to phlebotomists for each shift to ensure accountability and help sustain improvements.

Postprocess improvement 5S scores were greater than or equal to 9, with the exception of 1 week, and were 15 in the last 25 consecutive weeks of the study (data not shown).

Communication and workload management (Kaizen #2, October 2009)

Issues identified related to communication and workload management included the absence of a phlebotomy sign-in/sign-out system, and the lack of a robust method for communicating STAT draws during shifts, and remaining draws between shifts. Such shortcomings resulted in missed draws that spanned shift change, a delay in STAT blood draws, and complaints from clinicians concerning TAT. Communication between phlebotomists within shifts was also hindered by the limitations of pagers which required phlebotomists to interrupt workflow to find a phone, page their colleague, and wait for a return call. After Kaizen #2:

  • A ‘team leader’ was assigned to manage each shift with the specific responsibilities of tracking, when phlebotomists completed their assigned rounds, in order to deploy them to assist other phlebotomists, and supervising STAT requests.

  • Designated check-in times were established for each round to monitor the progress of rounds.

  • A staff communication board was created to allow phlebotomists to sign-in/sign-out and indicate remaining draws for the next shift.

  • Direct connect phones replaced pagers as the primary form of communication between phlebotomists, and allowed for more efficient communication of workflow progress and complications including STAT draws.

Blood draw process standardisation (Kaizen #3, February 2010)

Prior to our process improvements, the blood draw process was inconsistent and varied by phlebotomist. Although broader steps were followed, substeps were often performed in various orders or omitted altogether, leading to inefficiencies and potential for error. Other observed habits included batching of requisitions resulting in bottlenecks in the laboratory specimen processing area. Spaghetti diagrams showed that phlebotomists had significant travel throughout the unit, often retracing their tracks as a result of the way in which they ordered patients for blood draws, to retrieve supplies, or to respond to pages. Additionally, phlebotomists were noted to revert to using a manual process of labelling specimens despite implementation of the electronic handheld PPID device. After Kaizen #3, certain steps of the blood draw process were eliminated, rearranged, and/or designated to a specific location.

  • Phlebotomists no longer initialled requisitions or highlighted pod sheets, as those steps were considered redundant.

  • Patients were drawn in order of room number and carts were locked and placed in or near patient rooms to reduce travel.

  • To assist with implementation and retention, a one-page visual diagram was created to highlight process safety checks and aid in remembering frequently missed steps.

  • A list of exceptions to the standardised blood draw process was created to ensure that each blood draw situation would be handled appropriately (eg, precaution rooms).

  • Labelling specimens using electronic PPID was standardised.

  • An audit tool was created to measure compliance with the new standardised blood draw process. All phlebotomists were systematically trained in the new process over a period of a month, retrained as necessary, and deemed competent at least annually.

Rounding schedules and assignments (workout, April 2010)

Prior to and during our process improvement phase, our inpatient phlebotomists complained about unequal rounding assignments. This was corroborated by the discrepancy in round completion times between phlebotomists assigned multiple high-volume pods, or pods with higher complexity patients versus phlebotomists assigned relatively low-volume pods, or pods with lower complexity patients. Several phlebotomists, including some with many years of experience at BWH, participated in the Workout. Based on their historical knowledge of patient volume (ie, number of patients without central lines) and complexity, the Workout team devised a modified rounding schedule in which assignments were more equally balanced. Pods were also grouped based on location to allow for efficient travel between assignments.

Discussion

In this study, we used Lean principles in inpatient phlebotomy to optimise workflow, increase the timeliness of blood draws, and decrease error by focusing on supply replenishment and cart standardisation, communication and workload management, blood draw process standardisation, and rounding schedules and assignments. We demonstrated improvement in all three primary metrics and made a positive impact on the efficiency and accuracy of inpatient phlebotomy services. We illustrate the benefits of Lean, and also provide examples of process improvements that can be implemented and metrics that can be used to monitor these changes.

All four process improvement events purposefully targeted different steps of the phlebotomy process (figure 1), and collectively helped us demonstrate an overall improvement in inpatient phlebotomy services. Kaizens #1 and #2 were particularly helpful in making communication between shifts more efficient, and allowing phlebotomists to begin and end their rounds and shifts on time, since supplies were organised and ready at the start of their shift. The Workout helped improve overall round completion times as redistributing workload between phlebotomists allowed for more timely and efficient rounding. Kaizen #3 was effective at reducing the number of errors and increasing efficiency through removal of unnecessary steps, creation of memory aids for critical error-prone steps, and implementation of an audit tool to monitor compliance. Though each Kaizen and Workout targeted a different step in the inpatient phlebotomy workflow, we believe that it is the cumulative effect of all these changes during the 9-month process improvement period that improved our metrics. We felt our definition of the process improvement period was the most accurate way to evaluate a constantly evolving process involving rotating personnel who needed time to adapt to workflow changes.

Development of robust metrics and attainable goals is important in order to continuously monitor progress and sustain improvement. Though Kaizen #1 was conducted in August 2009, we omitted our data from August 2009 to October 2009 because the metric initially developed was deemed unrealistic. Our original goal was for 70% of phlebotomists to draw their first patient by 05:20. However, early data collection showed that our goal was almost never met largely due to the amount of travel time required to arrive on the floor. Thus, the metric was modified during Kaizen #2 to our current goal of 80% of phlebotomists drawing their first patient by 05:30, allowing for a more reasonable standard.

The three primary metrics were obtained manually (ie, round completion times and errors per 1000 draws) and electronically (ie, first draw times). Though electronic data is ideal since it is often more accurate and easier to collect, we have shown in this study that it is possible to establish and sustain manual data collection. It should also be noted that the metrics used to gauge the impact of the new staffing model (ie, median time of morning blood draws and average number of safety reports filed for delayed draws)9 can be useful to other laboratories as well.

Based on qualitative feedback, the process improvement events also had a positive impact on phlebotomists and laboratory staff. Involved staff found the Kaizens and Workout to be helpful in optimising their own workflow and empowering them to create change. Additionally, they felt that their experience was valued by leadership, specifically concerning rounding schedules and assignments. Phlebotomists generally liked the new standardised blood draw process and felt that it was safer and would reduce collection errors. Overall, frontline staff noted increased satisfaction in the work environment particularly concerning interaction and communication with critical stakeholders for which they collaborated with on a regular basis.

There are a number of limitations to our study. With the exception of anecdotal complaints, we did not have extensive preprocess improvement metrics that allowed us to gauge the state of our inpatient phlebotomy services prior to the 9-month process improvement phase. Additionally, we acknowledge that statistical analysis would have been helpful to assess the impact of our interventions. Unfortunately, we were unable to perform statistical analysis due to the lack of preprocess improvement data. Lastly, we did not assess the downstream effects of our workflow improvements, such as resultant TAT, or clinician and patient satisfaction, which would have been helpful to further illustrate the impact of the interventions hospital-wide.

Despite our efforts, we were unable to meet our ultimate quantitative goal for number of errors per 1000 draws and are inconsistently meeting our goals for completion of first patient draw. We believe this is due to the current lack of order communication information systems and use of paper requisitions requiring phlebotomy to continuously perform rounds to determine which patients are scheduled for blood draws. BWH is in the process of implementing a new LIS and hospital information system (HIS) that will enable order communication and eliminate paper requisitions. The new LIS and HIS, including a seamless order communication module, should enhance our Lean efforts to date, and allow us to achieve our goals in the future.

Lean principles can be successfully used to improve processes in inpatient phlebotomy. Our study can be used as a guide for other clinical laboratories with similar goals, providing examples of interventions and metrics that can be tailored to specific laboratories.

Take home messages

  • Lean principles can be successfully implemented in inpatient phlebotomy services to optimise workflow, improve timeliness of blood draws, and decrease errors.

  • It is best to target and design interventions for all aspects of the inpatient phlebotomy service in order to achieve and sustain success.

  • Robust metrics should be developed and used to assess implemented Lean interventions.

Acknowledgments

We would like to acknowledge Tanika Patterson, Inpatient Phlebotomy Supervisor, for her contribution and support in this study. We would also like to thank all laboratory and phlebotomy staff involved in this study for their valuable insight, ideas, and commitment to Lean and continuous process improvement.

References

Footnotes

  • Contributors Study design: SEFM, KSS, JDP, JMB, EMG, JNT-H, MLG, MJT. Data gathering: KSS, JDP, JNT-H. Data review and final review: RDL, SEFM, KSS, JDP, JMB, EMG, JNT-H, MLG, MJT. Draft article: RDL, SEFM, MJT.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.