Transcript Intro_Labs

Laboratory Medicine :
Introduction and General
Principles
Laboratory Medicine Conference Series :
Goals and Objectives
ƒ Review principles of decision analysis as
they apply to lab tests
ƒ Review background physiology and
measurement methodology for lab tests
ƒ Present sensitivity, specificity, & general
clinical usefulness of specific lab tests
ƒ Determine cost effectiveness of common
lab tests
Laboratory Medicine Conference Series :
Lecture Topics List
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
ƒ
1. Introduction & Decision Analysis
2. Leucocyte, Erythrocyte, Platelet, & Clotting Studies
3. Electrolytes, Calcium, Glucose, & Osmolality
4. Renal & Liver Function Tests
5. Urinalysis
6. Cardiac Enzymes & Markers
7. Amylase, Lipase, Pregnancy Tests, Sedimentation Rate
8. Arterial Blood Gases
9. Sputum & Pleural Effusion Tests
10. Cerebrospinal Fluid Analysis
11.Toxicology Studies
12. Miscellaneous Microbiology & Serology Studies
Why Discuss the Rather Unexciting
Topic of Laboratory Medicine ?
ƒ Knowledge of test methodology helps in
analysis of test limitations
ƒ Many lab tests ordered actually have no impact
on clinical decision making
ƒ Potential to save large amounts of money by
more appropriate test ordering
ƒ Stimulate ideas for research projects to assess
lab tests' clinical usefulness & accuracy
Further Reasons that Discussion
of Lab Medicine is Important
ƒ 1/4 to 1/3 of current lab testing is estimated as
unnecessary
ƒ Overordering contributes to the false
impression that the E.D. is "too expensive"
ƒ Practicing cost-effective medicine represents
good medical practice
ƒ Clinically unimportant abnormal lab values can
"label" patients & cause them unnecessary
anxiety
Problems with Inappropriate or
Unnecessary Ordering of Lab Tests
ƒ Wastes money
ƒ May delay definitive care
ƒ Extra discomfort for patient if blood
draw otherwise unnecessary
ƒ Extra risk to patient of investigating or
treating false positive results
ƒ False assurance about disease
presence if test result is unrecognized
false negative
What are Some of the Reasons Behind the
Current Ordering of Too Many Lab Tests ?
ƒ "Habit is easier than thinking"
ƒ Anticipating expectations of
supervising physicians ; wanting to
appear complete & compulsive
ƒ False concept that this provides legal
protection
Sequential Steps in Evaluation of an
Individual Lab Test
ƒ Analytic analysis
–Precision, accuracy, technical sensitivity & specificity of
the test procedure
ƒ Diagnostic analysis
–Relation of test result to presence or absence of disease
ƒ Operational analysis
–Predictive clinical value of test result
ƒ Cost - benefit analysis
–Cost, safety, & relation of test to remainder of clinical
decision tree
Reasons to Order Laboratory Tests
ƒ Good reasons
–Screening
–Diagnosis
–Management
–Research
ƒ Crummy reasons
–Idle curiosity
–Placebo effect
–Medicolegal concerns
–Peer review concerns
Lab Tests Ordered for Screening
ƒ Screening = test not predicated on signs or symptoms
of disease
ƒ Requirements for a good screening test :
–Disease in question should be common enough in the
patient population to justify efforts to detect it
–Disease should have significant morbidity if
undetected
–Effective intervention for the disease is possible
–The test should be abnormal in a very high % of
patients with the disease
Four General Processes of
Diagnostic Reasoning
ƒ Formation of hypotheses
ƒ Exclusion of some hypotheses under
consideration
ƒ Inclusion of some hypotheses for
further consideration
ƒ Use of these hypotheses to guide the
search for further information
Definitions of Terms Referring
to Lab Tests
ƒ Accuracy
–Degree of agreement between test result & known standard
value
ƒ Precision
–Ability of particular lab testing method to produce same exact
results time after time
ƒ Sensitivity
–Probability that test is abnormal in a patient with the disease
ƒ Specificity
–Probability that test result is normal or negative in the absence
of disease
Test Predictive Value Matrix
Presence of Disease
Lab Test
Results
+
-
+
A
C
Sensitivity = A / (A + C)
Specificity = D / (B + D)
Positive Predictive Value = PPV = A / (A +B)
Negative Predictive Value = NPV = D / (D + C)
Prevalence = (A + C) / (A + B + C + D)
B
D
Definitions of Test Interpretation
Terms
ƒ Positive Predictive Value (PPV) :
–Estimate of % of time that a patient will actually
have the disease if the test is positive
–Or, the % of patients with a positive test who have
the disease
ƒ Negative Predictive Value (NPV) :
–Estimate of % of time that the patient will be free
of disease if the test is negative
–Or, the % of patients with a negative test result
who do not have the disease
Definitions of Test
Interpretation Terms (cont.)
ƒ Prevalence
–Probability that a patient in the group
under study already has the given disease
before the time of the study
–Or, the % of all individuals under study
who have the disease
ƒ As the prevalence of the disease
increases, the PPV increases & the
NPV decreases
Definitions of Test Interpretations
Test Value
Disease present
No Disease
Abnormal
True Positives (TP)
False Positive (FP)
Normal
False Negatives (FN)
True Negative (TN)
Sensitivity (SS)
TP / TP + FN
TP / all with disease
Specificity (SP)
TN / TN + FP
TN/ all without disease
Predictive Value of
Positive Test (PPV)
Predictive Value of
Negative Test (NPV)
TP / TP + FP
TP / all with positive test
TN / TN + FN
TN /all with negative test
Interpretations of PPV and NPV
ƒ As prevalence increases, value of a positive test
becomes more meaningful
ƒ At low prevalence, value of a positive test (even if its
sensitivity & specificity are high) is low
ƒ Negative test result for disease of low prevalence
virtually excludes disease (if sensitivity & specificity
are high)
ƒ Test with high PPV can be useful to "rule in" a
disease
ƒ Test with high NPV most useful to "rule out" a
disease
Interpretation of Test Predictions
ƒ A test can have 100 % sensitivity & specificity
only if the distribution of test results is
nonoverlapping between diseased & nondiseased patients
ƒ Prevalence of most diseases in unselected
populations is < 5 %
ƒ So use of tests to screen for disease in
asymptomatic or low incidence population
ends up that most positive results are false
positive (false negative rate also high)
Example of Effect of Prevalence on PPV :
*Prostatic acid phosphatase as predictor of prostate
cancer ( sensitivity 70 %, specificity 90 % ) :
Setting
Prevalence
PPV (%)
(cases / 100,000)
General population
Men age 75 or
more
Prostate nodule
present
35
0.4
500
5.6
50,000
93.0
* N.E.J.M. 1980 ; 303: 497-499.
Probability That a Healthy Person Will Have
Abnormal Results in a Biochemical Profile *
Number of Tests
1
Probability (%) That at Least
One Test Will Be Abnormal
5
6
26
12
46
20
64
100
99.4
*Ann Int Med 1987 ; 106 : 403-413
Benefit to Cost Ratio Analysis
ƒ Threshold probability (pT) = probability of disease at which
there is no difference in outcome between treatment or Dx
and lack of Rx or Dx
ƒ So when benefit to cost ratio is low, test needs a high pT
(high PPV) to be of use
ƒ When the benefit to cost ratio is high, the degree of
uncertainty can be greater
ƒ When benefit to cost ratio is low, test with high NPV not
required to exclude a Dx
ƒ When trying to exclude a Dx, & the benefit to cost ratio is
high, a test with a high NPV is needed
"Point of Service" (Bedside) Lab
Testing
ƒ
ƒ
ƒ
ƒ
Results available much quicker
No need for specimen transport
Less blood drawn
At present is 10 to 20 times more
expensive than "in-lab" tests
ƒ Can prevent use of lab tests as a
delaying or holding technique
"Clinical Laboratories Improvement Act
“ (CLIA)
ƒ Federal law passed to regulate all office &
hospital labs
ƒ Requires :
–Specific staff training & prerequisites
–lab Q / A program
–Periodic inspections
–Lots of paperwork & fees to register each lab
ƒ Regulations apply even to stool guiacs,
urine dipsticks, fingerstick glucoses, etc.
A.C.E.P. Lab Test Cost
Containment Project
ƒ Estimated that up to 30 % of E.D. lab
tests are unnecessary
ƒ Estimated that proper use of
guidelines could save $ 5.9 billion
annually (this # of $ would provide
health insurance for 1.6 million
people)
ƒ Overuse of even low charge items
adds up to a lot of $ in aggregate
Frequently Ordered Lab Tests for
Medicare Patients in 1987 (HCFA data)
Laboratory Test
Routine ECG
Routine venipuncture
CBC
CXR
Urinalysis
SMA-12
Glucose
Prothrombin time
Thyroxine level
Potassium
High density lipoprotein
Allowed Services
26,116,333
20,227,087
16,276,672
14,641,479
13,659,116
7,751,248
7,638,169
3,667,127
2,812,774
2,629,100
2,353,452
Allowed Charges ($)
534,714,028
60,652,091
150,480,379
318,617,533
73,651,635
135,974,256
49,762,334
24,963,037
28,176,841
18,465,803
28,469,571
Examples of Overuse of Lab
Tests in Emergency Medicine
ƒ "Routine" trauma or admission labs
not specific to the patient's illness
ƒ CBC to "uncover" bacterial illness
ƒ PT / PTT prior to anticoagulation
ƒ Electrolytes in simple gastroenteritis
or brief period of vomiting or diarrhea
Preop Labs Required at H.M.C.
ƒ Age 1 day to 44 years : hematocrit
ƒ Age 45 to 59 years :
–Hct, glucose, creatinine, ECG
ƒ Age > 60 : above tests plus CXR
ƒ Add other labs as indicated
Review of Preoperative Lab
Screening (J.A.M.A. 1985; 253 : 3576)
ƒ 2000 patients had CBC, lytes, glucose, PT /
PTT preop
ƒ only 40 % of these tests were indicated
ƒ There were only 10 unsuspected abnormal
results :
–6 were ignored
–one was incorrectly noted as "normal"
–Clinical impact of all the tests was zero
Sample Study on Lab Usefulness :
CBC for Bacterial Disease (Ann Emer Med 1986 ; 15 :
774)
ƒ
ƒ
ƒ
ƒ
ƒ
Retrospective of 172 E.D. patients
21 % sent by R.N. & not used by M.D.
PPV (WBC > 12.5) only 26 %
PPV of left shift only 33 %
Cost $ 10,100 per clinical impact
Sample Study of Lab Test Usefulness :
PTT Prior to Anticoagulation (Ann Emer Med 1993 ; 22 :
697)
ƒ Reviewed 199 E.D. admissions with
DVT or Pulm. embolus
ƒ Only 3 abnormals not predicted by
history (one congenital, 2 on warfarin)
ƒ Heparin was always started prior to
return of lab results
ƒ Caused no changes in heparin therapy
Sample Study of Lab Test Usefulness :
Electrolytes for Simple Seizure (Ann Emer Med 1986 ;
15 :33)
ƒ 180 E.D. patients with seizure
ƒ 27 (17 %) of cases had low bicarb
ƒ Other lytes were zero to 7 % abnormal
–Did not affect clinical management
ƒ One unsuspected hypoglycemia
Labs Recommended for New Onset
Seizure by Different References
ƒ Harrison :
–Na, Ca, Glucose, TFT's, Lead, Arsenic
ƒ Rosen :
–CBC, U/A, Glucose, Lytes, Ca, ABG
ƒ A.C.E.P. Clinical Policy for Seizure :
–No labs
–Consider CT or MRI
Study of Criteria Predictive of Clinically
Significant Electrolyte Abnormalities (CSEA)
(Lowe et al., Ann Emer Med 1991 ; 20 : 16)
ƒ Criteria Used :
–Poor oral intake, vomiting, chronic HBP, diuretic use, recent
seizure, muscle weakness, age > 65, alcoholism, abnormal
mental status, recent electrolyte abnormality
ƒ Prospectively validated these criteria in 984 E.D. patients
ƒ Sensitivity was 94 %
ƒ Eight "misses" :
–4 with low bicarb or Cl ; 2 had results circled in chart but not
acted on ; 2 had low K but empirically replaced
ƒ Following the criteria would have led to 24 % reduction in
obtaining lytes with no adverse clinical outcomes
Study of Clinical Guidelines Used to
Reduce Lab Testing (Gen Int Med 1990 ; 5 : 335)
ƒ
ƒ
ƒ
ƒ
Involved 79 M.D.'s & 1638 hospital patients
Used guidelines from consensus panel
Voluntary participation
Resulted in 21 % reduction in lab tests
ordered
ƒ No change in clinical outcomes as
measured by death or readmission within
90 days
Principles of Serologic Methods
ƒ Four general classifications of serologic
tests :
–Agglutination
–Precipitation
–Complement fixation
–Fluorescent antibody
ƒ All 4 utilize measurement of reaction of
antigen & antibody
Principles of Agglutination Tests
ƒ Can be done in test tube or on slide
–Slide proceure : mix patient's serum with
antigen on slide & observe for
agglutination
–Test tube procedure : serial dilutions of
serum mixed with constant amount of
antigen in successive tubes ; the highest
dilution showing agglutination is the
antibody titer
Principles of Indirect
Hemagglutination Tests
ƒ One type uses viral or parasitic
antigens that are adsorbed to
erythrocytes of animals (chickens,
rabbits, sheep or horses) or humans
ƒ Another type uses antigens absorbed
onto latex particles
–Allows longer storage & use of slide tests
Principles of Precipitation
(Precipitin) Tests
ƒ Show precipitation of antigen-antibody
complex
ƒ Can be done with capillary tubes where
precipitin forms at interface of antigen &
antibody
ƒ Or can be done with agar wells which allow
antigen & antibody to diffuse into each other to
form a line of precipitation
–If electric current applied (counterimmunoelectrophoresis),
this increases the speed of diffusion
Principles of Complement Fixation
Tests
ƒ Mix patient's serum with antigen &
complement, & then mix in red blood cells
coated with anti-RBC antibody
ƒ If any unbound complement is then available, it
will bind to the RBC's & lyse them
ƒ If there is no lysis, then all the complement has
been bound to the original complex & the
presence of antibody in the patient's serum is
indicated
Principles of Fluorescent Antibody
Tests
ƒ Uses a fluorescent dye (such as
fluorescein isothyocyanate) which
binds to antibody
ƒ The antigen-antibody complex is then
directly observed with a fluorescent
microscope
What Can You Do for More
Appropriate Lab Test Ordering ?
ƒ Only order tests that you would be
willing to pay for or have done on
yourself
ƒ Use clinical criteria for ordering when
these are available
ƒ Know charges for common tests
ƒ Educate your colleagues
ƒ Help develop clinical guidelines
Principles of Lab Test Ordering :
Summary
ƒ Perform test if additional information
required for diagnosis or management
ƒ Don't perform test if result will not alter
management
ƒ Select test based on need to rule in or
out particular disease
ƒ Review test methodology, sensitivity &
specificity, & cost - benefit ratio
ƒ Assess test result in terms of PPV or
NPV