Middleware and the Clinical Hematology Laboratory

Balance? Clinical Needs vs. Laboratory Reality TAT Accuracy Resources Costs

It doesn t have to be a contest between efficiency and patient care!

Today s Outline Background of CBC and diff Verification of the differential count Practice modeling: pre-middleware 5 phases of middleware utilization Automate Rules for Manual Differentials Focus on Patient Safety Issues Enhance Laboratory Effectiveness Support Management Goals Unify Multi-site Practices

Evolution of the Complete Blood Count (CBC) Early 1900 s: Counting chambers Colorimetric methods for hemoglobin 1940 s and 1950 s: Wintrobe and the hematocrit tube, MCV, MCH, MCHC 1956: Wallace Coulter and aperture-impedence cell counting method 1970 s: Light scatter cell counting 1980 s & 1990 s: Air- cooled lasers, computers, etc. 1990 s: Application of immunologic methods 2000 s: Middleware and other software approaches

Evolution of the Leukocyte Differential Count (Diff) Early 1900 s: Evolution of manual techniques Early 1970 s: Image capture techniques Hematrak,, IMI Micro 21, LAFIA system, Difffmaster, etc. 1981: Technicon Hemalog D First instrument to perform automated differentials 1983: Coulter Counter Plus IV 3-part differentials 1980 s/90 s: Coulter STKS, Sysmex 9000, Cell Dyne 3500 5-part differentials 2000 s: Image recognition software programs Still in its infancy

Why Focus on the CBC and Diff? CBC is the most frequently ordered single laboratory test Highest reimbursed laboratory assay by CMS Used in outpatient and inpatient settings Need 24/7 availability Used as a screening tool for health Opens the door to the patient - looking for possible hematologic, infectious, inflammatory, and oncologic diseases Used specifically to evaluate for a variety of acute and chronic hematologic diseases

The Manual Differential The manual differential Is labor intensive Requires both highly skilled and experienced technologists and back-up and availability by pathologists Errors with the manual differential Observer errors Slide distribution errors Sampling errors with a 100 cell count Data entry errors

The Automated Differential The automated differential Is geared toward not missing anything Must recognize a wide spectrum of white cell diff abnormalities Lymphoid and myeloid; acute and chronic; pre- and post- therapeutic No single cell type perfectly represents each disease process Can t consistently distinguish between reactive and disease Every instrument technology has its own inherent idiosyncrasies and will over-capture and under- capture particular morphologic abnormalities

Manual vs. Automated Diff? The Challenges? Need to identify important abnormal findings while minimizing the time required for either normals or minimal abnormals In other words, maximize the use of automated differentials while minimizing the number of manual differentials The Technology vs. The Art

Verification of the Diff: Where? Instrument Auto-verify rules In the Lab Technologist Middleware LIS Auto-verify rules

Verification of the Diff: How? Instrument Is geared toward not missing anything Quantitative flags (RBC & WBC) Set by the laboratory Should not be based on normal values Qualitative flags Established by the manufacturer Autoverification at the instrument is dependent on the local patient population Mayo Health System Clinics: 60% auto-release rate U of Michigan and Mayo: 40% auto-release rate

Verification of the Diff: How? LIS Dependent on vendor/system Many/Most have rule-writing capabilities Easier for the lab to meet regulatory requirements IT support and back-up Often dependent on non-lab people to write the rules Responsiveness can be at a bureaucratic, glacial pace; non-flexible

Verification of the Diff: How? Middleware: Usually purchased from instrument manufacturer Other non-instrument vendors are available Need information from both the instrument and LIS Need to bring middleware vendor, instrument manufacturer, and LIS resources together to understand the interface and what limitations the interface will impose Back-up and IT support is critical for a laboratory. Does the lab have to develop their own expertise?

Verification of the Diff: How? Middleware and Mayo Laboratory Hematology Hematology Instrument Interface System (HIIS; Coulter): November 1996 Aqueduct (Orchard Software): July 1999

Differential? Instrument auto-verify Scan and release Manual LIS / EMR Minimal abnormal or normal Significant abnormal

Practice Modeling Instrument auto-verify Scan and release Manual: Normal/minimal Abnormal (signif) How your practice cases are distributed in these buckets will determine your middleware value. Is your instrument auto-verify at: 80% or 40%? Is your scan and release at: 5% or 30%? Manual differentials: how many have true value added by morphologic review? 5% or 50%?

Practice Modeling - Mayo Instrument auto-verify Scan and release Manual: Normal / minimal Manual: Abnormal (significant) Pre-Middleware (1995): 40% 20% 30% 10% Current (2008): 82% 8% 2% 8%

Practice Modeling: The Value Equation Middleware Value = Quality (satisfaction, outcomes, safety, and service) Cost Value must drive the process of whether and how middleware should be used in a hematology laboratory.

Today s Outline Background of CBC and diff Verification of the differential count Practice modeling: pre-middleware 5 phases of middleware utilization Automate Rules for Manual Differentials Focus on Patient Safety Issues Enhance Laboratory Effectiveness Support Management Goals Unify Multi-site Practices

Five Phases of Middleware Support management goals Enhance laboratory effectiveness Focus on patient safety issues Automate rules for manual differentials Unify multi-site practices

Automate Rules for Manual Differentials Simply write rules for what the technologist does If they look up results in the computer If they compare results to previous studies If they base a decision on where the patient is or who their doctor is If they minimize or ignore certain results or certain instrument flags What are their algorithms for deciding whether a diff gets performed or not?

Automate Rules for Manual Differentials Quantitative numbers Ranges sensitive to age and gender Instrument flags Delta check to hold Delta check to release Delta check on instrument flags By physician By type of CBC By lab location By patient location By patient Elapsed time between specimen testing

ISLH Consensus Guidelines AGE Parameter Primary and/or Secondary and/or Tertiary Action 1 Neonate First sample Slide review CBC Parameter Primary and/or Secondary and/or Tertiary Action 1 WBC <4.0 or > 30.0 and Delta failed and < 3 days Slide review PLT <100 or >1000 and First time Slide review HGB <7g/dl or > 2g/dl above upper reference range for age and sex and First time Slide review RDW >22 and First time Slide review

ISLH Consensus Guidelines DIFFERENTIAL Parameter Primary and/or Secondary Action 1 No diff or incomplete diff Slide review Neut # <1.0 or > 20.0 and First time Slide review Lymph # >5.0 (adult) or >7.0 (<12 yrs old) and First time Slide review Mono # >1.5 (Adult) or >3.0 (<12 yrs old) and First time Slide review NRBC # any value and First time Slide review

ISLH Consensus Guidelines INSTRUMENT FLAGS Parameter Primary and/or Secondary and/or Tertiary Action 1 Suspect flag (except Imm G/B) Flag + and First time and Adult Slide review Dimorphic RBC Flag + and First time Slide review PLT clump flag Any count Check sample for clots Immature gran flag Flag + and Previous confirmed result and Positive delta fail for WBC Slide review Left shift flag Flag + Follow lab SOP Atypical lymphs Flag + and First time Slide review Blast flag Flag + and Previous confirmed result and Positive delta fail for WBC Slide review NRBC flag Flag + Slide review

Mayo Aqueduct Rules QUANTITATIVE FLAGS Tests Action Lo Hi Value Tests Held/Trapped Tests Pre-Transmitted Assigned Panels Hgb < X 6.1 RBC, Hgb, Hct, MCV, MCH, MCHC WBC, 5-pt diff, RDW, Plt CBC, CBCC, CBCN, CBC4/NC Plt < X 40.1 WBC, Plt WBC < X 1.0 WBC 5-pt diff, RBC, Hgb, Hct, MCV, MCH, MCHC, RDW 5-pt diff, RBC, Hgb, Hct, MCV, MCH, MCHC, RDW, Plt CBC, CBCN, CBC4/NC CBC, CBCN, CBC4/NC Plt Contains > R 39 WBC, Plt 5-pt diff, RBC, Hgb, Hct, MCV, MCH, MCHC, RDW CBC, CBCN, CBC4/NC

Mayo Aqueduct Rules INSTRUMENT FLAGS Value Tests Held/Trapped Tests Pre-Transmitted Assigned Panels Imm. Ne 2 5-pt diff CBC CBC, CBC4/NC Low Event # 5-pt diff CBC CBC, CBC4/NC Ly Blasts 5-pt diff CBC CBC, CBC4/NC Platelet Clumps CBC & diff None RBC Interference CBC WBC, 5-pt diff CBC, CBCN, CBC4/NC CBC, CBCC, CBCN, CBC4/NC, CBCCG Variant LY 5-pt diff CBC CBC, CBC4/NC

Mayo Aqueduct Rules DELTA CHECK Tests Action Lo Hi Value Tests Held/Trapped Tests Pre- Transmitted Assigned Panels Plt Exceeds Hold Delta (Current is both + & - 80%. New release You can choose one or the other or both) 80%/ 4 days CBC & diff None CBC, CBCC, CBCN, CBBCCG, CBC4/NC RDW Exceeds Release Delta > Location is not X 10%/ 7 days 17.5 Spec. Sta. RDW WBC, 5-pt diff, RBC, Hgb, Hct, MCV, MCH, MCHC, Plt CBC, CBC4/NC

Lab NE St. Marys (SMH) SMH Lab (ER, OR, ICUs) Lab NW Lab Kasson Methodist (MH) MH Lab (OR, ICUs) Lab Kenyon Lab Bldg Lab Bldg Lab Bldg Outpatient Clinics Gonda Lab (Chemo) Blood Draw Central Auto. Lab Heme Lab

Central Auto. Lab n=1719 March 19, 2008 Middleware auto-verify n=1404 82% LIS / EMR Trapped by Middleware: Quantitative CBC: 195 Quantitative Diff: 168 Instrument Flags: 302 Individual Patients: 6 Heme Lab n=315 18% 2% 6% 10% Instrument auto-verify n=42 Scan and release n=105 Manual diff n=168 13% 33% 54%

Middleware - Impact Instrument auto-verify Scan and release Manual: Normal / minimal Manual: Abnormal (significant) Pre-Middleware: Current: 40% 20% ~30% ~10% 82% 8% 2% 8% Within 3 months of implementation: Decreased 3.0 FTE TAT of Priority 1 CBC s: went from 50% to 85% released < 30 min. Median TAT of all CBC s went from 90 min. to 30 min. Average CBC time in central automated lab: <20 min.

Automate Rules for Manual Differentials CBC Chemotherapy Applied different rules for patients receiving chemotherapy for myeloma and lymphoma and certain leukemias The entire clinical question is: Are there enough platelets and neutrophils so that the hematologist can continue to give the next dose of chemotherapy?

Automate Rules for Manual Differentials CBC Chemotherapy (CBC-C) C) Created a new type of orderable CBC available only to the hematologists: CBC-C Clinical needs: The CBC is no longer a screening assay it is a therapeutic monitoring assay. An accurate diff is not a concern, just: are there PMN s and platelets? A rare blast or a slight left shift is not important Rapid turn-around around-time is the driving force

Automate Rules for Manual Differentials Separate auto-verify rules for the CBC-C WBC>99.9; Hgb<6.1; MCHC>37; BA%>5 5-part R flag; Interference flags; non-results Plt delta check Clinical awareness is important Cannot be used for other clinical scenarios Need to call the lab to convert the CBC to a real CBC if there are any clinical findings that raise other questions such as infection, etc.

Automate Rules for Manual Differentials Pre CBC-C ~150 CBC s per week from chemo unit Autoverify through Aqueduct 35% Average TAT: 20 min. vs. 2.5 hours Not a happy clinical group! CBC-C C implemented in 2001 Post CBC-C Autoverify: 88% TAT = 20 min from blood draw Audits: Neutropenia Thrombocytopenia Left shift RBC poik (non-specific) Rare blast (LEBR) Rare lymphoma or plasma cell ~1 CBC-C C converted to routine CBC per month

Focus on Patient Safety Issues Why call it patient safety? In the hematology laboratory, true primary sentinel events are blessedly rare BUT: : Revised reports from labs do contribute to unnecessary clinical steps, which may lead to unnecessary patient events Every lab needs to look at all their revised reports and near misses. Why did they happen? What was wrong with the process? Until you collect the info, you will underestimate the severity of the issue

Focus on Patient Safety Issues Revised reports and near misses Revised diff counts (instrument or tech) After pathologist review of a bone marrow After tech review of a subsequent PB smear After a call from a clinician Revised platelet count due to clotted specimens or EDTA-associated clumping After tech review of a subsequent PB smear After a call from a clinician Dependency on sticky notes and memory

Focus on Patient Safety Issues Revised reports and near misses (con t con t.) Didn t follow through per SOP (memory) Did a scan and release instead of a manual diff Forgot to have a second review by senior tech or MD Etc. Manual reporting / clerical error (direct interface) Mixing up slides and paperwork (barcodes)

Focus on Patient Safety Issues Individual patients act like individuals they don t always follow the rules for the group. Let s call our patient: Mrs. Johnson Blasts may be missed on follow-ups Lymphoma cells may be missed on follow-ups Morphologic features may be difficult to interpret EDTA platelet clumpers create problems Red cell agglutination create problems Clinician or pathologist need specific follow-up on Mrs. Johnson Protocol/clinical trial requirements

Focus on Patient Safety Issues Middleware can help Immediately in the lab can create a rule to trap the next sample that comes through from Mrs. Johnson A time line is added to the rule; e.g., expires in one month or three months Doesn t matter where in the system Mrs. Johnson walks into main clinic, ER, outlying clinic, etc.: same outcome for the CBC

Focus on Patient Safety Issues Pre- Mrs Mrs.. Johnson Revised differentials 2.5 A events / mo. Communicate by sticky notes and emails Post- Mrs Mrs.. Johnson Revised differentials <1 A event / mo. 5 to 10 Mrs. Johnson rules in place at any given time < 2 hrs. from request to being live

Enhance Laboratory Effectiveness Use middleware to streamline the presentation and flow of data from instrument to workstation regardless of where data was acquired Minimize the number of manual steps between instrument and slide review Blends with patient safety issues

Enhance Laboratory Effectiveness Use middleware to make the lab paperless At each workstation: Transmit scatterplots to any workstation Make all CBC/diff data easily visible Apply color coding to aid visual recognition Visually see trapped vs. released data Display invoked rules for trapped data Multiple ways to review old data By CBC date By each analyte

Enhance Laboratory Effectiveness Lab continues to run if LIS is down Use rules as SOP reminders E.g.: new employee can t report out schistocytes without 2 nd review All new adult lymphocytoses needs MD review All blasts >1% need initial MD review Disease processes Set frequency of review after initial diagnosis of a disease: e.g., CLL, MDS, etc.

Enhance Laboratory Effectiveness The instrument s QC package (1( 2s N=2 or 3) is not optimal for each of the CBC s components. A high false rejection rate breeds complacency: LH750 Sigma Metric QC Rule Hgb 8.75 1 3.5s N=3 WBC 8.82 1 3.5s N=3 RBC 7.50 1 3.5s N=3 Plt 7.57 1 3.5s N=3 Can you use middleware as your QC manager?

Support Management Goals Is there laboratory information that cannot be gathered through routine reports? Technologist productivity Type of cases released by technologist Number and type of cases trapped by each rule Monitor utilization of rules Identify new clinical scenarios for potential new rules

Technologist Productivity Avg. PB cases / shift 100 90 80 70 60 50 40 30 20 10 0 A C E G I K M O Q S U W Y AA CC

Technologist Productivity Avg. PB cases / shift 80 70 60 50 40 30 20 10 0 E R S W X CC DD Scan Manual

Future Needs: Unify Multi-site Practices Multiple sites within a hospital/clinic Multiple hospitals and clinics within an organization local, regional, national May have different instrumentation Different clinical needs at each site Different patient characteristics at each site Not all sites may do manual differentials May or may not have a common EMR

Future Needs: Unify Multi-site Practices Middleware offers you a common language through which laboratories can communicate Middleware offers you a starting point for standardization / integration of your medical practice

Middleware Conclusions Before you purchase / implement: Model your lab practice to understand where you might benefit from middleware and to what extent Keep in mind the Value Equation both quality and costs can be improved with middleware Envision the coordinated roles that your instrument, LIS, and potential middleware system will play in your lab practice Instrument and LIS might be sufficient in some labs Work with your IT people to understand their needs and their willingness to help provide support in the lab

Middleware Conclusions Automate rules for differentials this is the easy part! Benefits include: Reduce costs / FTE s Improve TAT Improve employee satisfaction Improve accuracy of differential results Challenge your concept of the CBC and diff and think differently about specific patient populations and their needs

Middleware Conclusions Focus on patient safety issues. Every lab needs to look at all their revised reports and near misses. Why did they happen? What was wrong with the process? Use middleware to fill in gaps in your patient safety processes: The Mrs. Johnson Rules SOP and disease-specific specific Pop-Up reminders for the techs

Middleware Conclusions Enhance laboratory effectiveness Create the paperless lab Ensure the ideal screen presentations of data Minimize the steps between CBC analyzer and the microscope/work station Improve your QC approach in the lab Support management goals Middleware is another source of information about your laboratory Unify multi-site practices A common language to cross-communicate communicate with other labs A starting point for standardization / integration of your multi- site laboratory practice

Balance Balance? Achieved! Clinical Needs and vs. Laboratory Reality Outcomes Satisfaction TAT Accuracy Management Patient safety Resources Costs Middleware!

Thank You!