Method Comparison and Medical Decision Limits

In her lecture on Friday, Dr. Flatland talked about (among other things) analytical error, breaking it down into random error (imprecision) and systematic error (bias).  She also discussed the difference between constant and proportional bias when comparing two different methods of measuring the same analyte.  These are certainly important issues to consider when interpreting laboratory data.  Another way to compare methods is error grid analysis, which focuses on the clinical relevance of error.  Sometimes differences in results of two methods have no practical effect on clinical decision-making; on the other end of the spectrum, differences in lab results can alter clinical decision-making with dangerous consequences; and there are lots of scenarios that fall between the two extremes.  If you're interested in learning more about a specific application of error grid analysis in veterinary medicine, see this article (especially Figure 3):  http://avmajournals.avma.org/doi/pdfplus/10.2460/javma.235.11.1309

Sensitivity, Specificity, and Predictive Value

When we covered this topic in class, I emphasized how the prevalence of a disease influences the predictive value of a test for the disease:  i.e, PPV increases as prevalence increases, and NPV increases as prevalence decreases.  More recently, I had a good question from a student about the relationship between a test's sensitivity and its predictive value:

It would seem to me that as sensitivity increases, your PPV would decrease since the likelihood of your test getting a false positive increases as sensitivity increases.  Is this a true assumption or does the relationship even exist like that?

This was from my reply:

The flaw in your thinking is that the likelihood of getting a false positive test result increases as specificity decreases, but not necessarily as sensitivity increases.  If you do the math, you’ll see this is so.  For example, try working through 2 scenarios that both have 20% prevalence and 90% specificity, but that have respective sensitivities of 75% and 95% -- you’ll find that the PPV actually goes UP in the scenario with higher sensitivity.

If specificity is 100%, then PPV = 100%.  Why?  Because all unaffected individuals test neg, so any pos results are true positives (an especially desirable trait in a confirmatory test).

If sensitivity is 100%, then NPV = 100%.  Why?  Because all affected individuals test pos (an especially desirable trait in a screening test), so any neg results are true negatives.

Then it occurred to me why the student might have equated an increase in sensitivity with a decrease in specificity, so I added to my response:

The reasoning I gave you earlier refers to a scenario where sensitivity varies but specificity and prevalence remain constant.  However, as I talked about in class, if you are using a "cut point" value as the threshold for determining whether a test result is positive or negative, then sensitivity and specificity are inversely related – that is, as sensitivity goes up, specificity goes down, and vice versa.
   
If you're in the VM888 class and you follow the logic of this post, then you'll probably do fine on any questions related to this topic on Friday's exam!

RBC shape abnormalities

There was a good question in class today about abnormal erythrocyte shape.  The catch-all term for RBC shape abnormalities is poikilocytosis.  I mentioned that the types of "poik" that you should know for the purpose of the course are spherocytes (which usually indicate extravascular hemolysis), schistocytes (RBC fragments, also called schizocytes, which are often seen in patients with DIC or other types of "microangiopathies"), and eccentrocytes (which, like Heinz bodies, are a manifestation of oxidative damage to RBCs).  Your textbook has a nice section on poikilocytosis, summarized in Table 3.7, and some good images in Plates 5 and 6.  The poikilocytosis section of Cornell's e-ClinPath website is also a good resource.

Hope y'all have a great Labor Day weekend.  See you on Tuesday.

1st v. 2nd edition of required textbook

I had a good question about the required text:  "Does it matter if we buy the 1st or 2nd edition of the required textbook?  The second edition is obviously much more expensive, and I was wondering if you knew of any significant differences in the newer edition?"

This was my response:  The differences in the content of the 1st and 2nd editions are relatively small, but not irrelevant.  I think the library has copies of both, so I suggest you take a look at them yourself before you decide.  Remember, you have free access to the e-book of the 2nd ed.

Plasma protein terminology: "globulins" v. "immunoglobulins"

As Dr. Flatland indicated, the value for globulin concentration on a serum or plasma chemistry panel is determined mathematically:

[Total protein] - [Albumin] = [Globulins]

So, in the context of clinical chemistry, "globulins" means any proteins detected by the biuret method, except albumin.

Sometimes people say "globulins" when they really mean "immunoglobulins", which means antibodies.  An immunoglobulin (Ig) is a type of globulin, but the terms are not synonymous because the globulin fraction also comprises many non-Ig proteins, including coagulation factors, acute phase proteins, transport proteins, enzymes, hormones, cytokines ... and I'm probably leaving some out.

Measured and calculated erythrogram values

To follow up on what we covered in class today, this is how erythrogram values are determined using an automated hematology analyzer: 

RBCs (x 10⁶/µL) – the analyzer counts the # of RBCs in a given volume of blood, and also determines the size (volume – see MCV) of each RBC. 

Hgb (g/dL) – the analyzer lyses RBCs and measures [Hgb] by spectrophotometry. 

MCV (fL) – the analyzer software calculates an average value based the sizes of all RBCs measured. 

Hct (%) – the analyzer software calculates a value based on RBCs and MCV.  Working through the math on a hypothetical blood sample with an RBC concentration of 6.5 x 10⁶/µL and an MCV of 72 fL:

     RBC               = 6.5 x 10⁶/µL = 6.5 x 10¹²/L  

     MCV               = 72 fL = 72 x 10^-15 L

     Hct                  = RBC x MCV

                             = (6.5 x 10¹²/L) x (72 x 10^-15 L)

                             = 468 x 10^-3 = 0.468 = 46.8%

If you just used the reported values for RBC and MCV without converting the volumes to liters (instead of µL or fL), then it would work out like this: 

     Hct                  = RBC x MCV
                             = 6.5 x 72 = 468 (decimal points are off w/out unit conversions) 

MCH (pg) – the analyzer software calculates a value based on Hgb and RBCs.  Working through the math on a hypothetical blood sample with a Hgb concentration of 15.4 g/dL and an RBC concentration of 6.5 x 10⁶/µL: 

     Hgb                 = 15.4 g/dL = 15.4 g/10^-1 L

     RBC               = 6.5 x 10⁶/µL = 6.5 x 10¹²/L

     MCH               = Hgb ÷ RBCs
                             = 15.4 g/10^-1 L ÷ 6.5 x 10¹²/L
                             = 2.37 g x 10^-11 g = 23.7 g x 10^-12 g = 23.7 pg 

If you just used the reported values for Hgb and RBC without converting the volumes to liters (instead of dL or µL), then it would work out like this: 

     MCH               = Hgb ÷ RBCs
                            = 15.4 ÷ 6.5 = 2.37 (decimal points are off w/out unit conversions)

MCHC (g/dL) – the analyzer software calculates a value based on MCH and MCV.  Working through the math on a hypothetical blood sample with a MCH of 23.7 g/dL and an MCV of 72 fL:

     MCH               = 23.7 pg = 23.7 g x 10^-12 g

     MCV               = 72 fL = 72 x 10^-15 L

     MCHC            = MCH ÷ MCV
                             = 23.7 g x 10^-12 g ÷ 72 x 10^-15 L
                             = 0.329 g/10^-3 L = 32.9 g/10^-1 L = 32.9 g/dL 

If you just used the reported values for Hgb and RBC without converting the picograms to grams and the femtoliters to liters, then it would work out like this: 

     MCHC            = MCH ÷ MCV
                             = 23.7 ÷ 72 = 0.329 (decimal points are off w/out unit conversions) 

RDW (%) – the analyzer software calculates a value (SD/mean) based the sizes of all RBCs measured.