Hyping of 99%
Noninvasive prenatal screening (NIPS) tests have been hyped for being “99% accurate.” This promotion of NIPS’ accuracy comes first from the NIPS laboratories themselves:
- Sequenom, maker of MaterniT21: “Results from a method validation study demonstrated high performance of the MaterniT21 PLUS test (>99 percent specificity) in accurately detecting these autosomal trisomies and select microdeletions.”
- Sequenom on its targeted test for Down syndrome and Trisomy 18, VisibiliT: “demonstrated a greater than 99% sensitivity and specificity for trisomies 21 and 18.”
- Ariosa, maker of Harmony: “the Harmony test can be performed as early as 10 weeks of pregnancy and offers a greater than 99 percent accuracy rate for risk assessment of trisomy 21, which causes Down syndrome.”
- Natera, maker of Panorama: “Across multiple clinical trials, Panorama has been validated globally for detection of trisomy 21, trisomy 18, trisomy 13, monosomy X, and now triploidy, with a sensitivity of greater than 99 percent for trisomy 21, trisomy 18, trisomy 13, and triploidy, 92 percent for monosomy X, and a less than 0.1% false positive rate for all syndromes tested.”
- From the website of Illumina, maker of verifi (screenshot taken June 7, 2015):
- And the same representations are made by Illumina’s licensees of its verifi technology, Progenity and LabCorp’s InformaSeq.
Due to the laboratories’ marketing, patients report being told by their obstetrician or their genetic counselor that “the test results are 99% accurate” as though those are the odds of that patient having a child with Down syndrome. Indeed, that is the headline of an article written the week of this blog post (see image at top).
This is a confusion between what is called the test’s “sensitivity” and the test’s “positive predictive value” or “PPV.”
Sensitivity, Specificity, & PPV
Sensitivity is what percentage of pregnancies actually carrying a child with Down syndrome are detected by the test. For the claim of 99%, the laboratories are saying that out of 100 pregnancies carrying a child with Down syndrome, their test detected 99 of them. This also means their test did not detect 1 of the pregnancies actually carrying a child with Down syndrome, who received what is called a “false negative” result.
There is also a test’s “specificity.” Specificity is the percentage of pregnancies not carrying a child with Down syndrome that the test accurately reports are not affected. Referring to Illumina’s claim (above) of 99.94% specificity for its test verifi, that means of 10,000 pregnancies not carrying a child with Down syndrome, verifi accurately reported that 9994 were unaffected.
But that means verifi incorrectly reported that 6 pregnancies were carrying a child with Down syndrome, i.e. 6 “false positives.”
Positive predictive value is the computation taken of a mother’s chance for having a child with Down syndrome, based on her age or a first-level screen result, and then applying NIPS sensitivity and specificity rates to that chance.
For instance, a 30-year old mom has approximately a 1-in-a-1,000 chance for having a child with Down syndrome. If you had 100,000 30-year old expectant moms, then 100 of them would be pregnant with a child with Down syndrome and 99,900 would not.
Applying the sensitivity rate, 99 of the 100 actually pregnant with a child with Down syndrome would be detected. Applying the specificity rate, 60 out of the 99,900 moms not carrying a child with Down syndrome would instead be given a false positive result. For each 30-year old mom, then, when she receives a “positive” verifi result, she could be one of the 99 actually carrying a child with Down syndrome or her test result could be one of the 60 false positives because she is not carrying a child with Down syndrome.
This breakdown means a verifi result for a 30-year old woman has a positive predictive value (PPV) of 62%, i.e. she has a 62% chance that she is actually carrying a child with Down syndrome and a 38% chance that she is not … but not a 99% chance that she is carrying a child with Down syndrome.
Standard of Care
If a provider tells a patient her NIPS test came back “positive” so she has a 99% chance of having a child with Down syndrome, the provider likely has committed malpractice. Not just because the provider is giving his or her patient incorrect information, but it is against the standards of care set by professional medical guidelines:
- ACOG, the American College of Obstetricians and Gynecologists, advised when reporting screening results: ““it is preferable to provide patients with their numerical risk determined by the screening test, rather than a positive versus negative screening result using an arbitrary cutoff.”
- The Society for Maternal-Fetal Medicine (SMFM) instructs: “counseling should include a discussion of the predictive value of cfDNA as a screening test (including the possibility that the result is a false positive)”.
- The National Society for Genetic Counselors (NSGC) states in its 2015 Fact Sheet for NIPS:
A >99% risk score does not mean there is a greater than 99% chance that the pregnancy is affected with a condition. * * * In order to determine the chance for a high risk result to be a true positive, one must calculate the positive predictive value. * * * Positive predictive value (PPV) is the proportion of positive results that are true positives. In other words, PPV answers the question: “What is the chance an abnormal NIPT means the fetus has this condition?”
I do not throw around a term like “malpractice” lightly. It immediately puts professionals on the defensive. But women keep being told that based on their MaterniT21/verifi/Harmony/Panorama test result they have a “99% chance” of having a child with Down syndrome, when almost always that is a confusion by the professional of the test’s sensitivity rate and not the test result’s PPV.
If you were told you have a 99% chance that is based on the sensitivity rate and not the PPV, your provider committed malpractice. He or she gave you inaccurate medical information contrary to the standards of care for interpreting NIPS results.