Lab Soft News

The Joint Commission Takes a Giant Step Backwards Regarding Hospital Blood Draws

Robert Michel, who blogs over at the Dark Daily, calls our attention to a retrograde step on the part of the Joint Commission, which no longer requires phlebotomists to ask a patient to state his name prior to a blood draw (see: Joint Commission Changes Requirement for Patient ID during Blood Draws). Here is an excerpt from the article:

When The Joint Commission recently changed the patient identification requirement for drawing a blood sample, one national phlebotomy leader considered it a step backward in patient safety. “Per a revised policy issued by The Joint Commission, it is no longer required for the phlebotomist or person drawing the blood to actively involve patients by, for example, asking them to state their name,” commented Dennis Ernst, MT(ASCP), the long-serving Director of the Center for Phlebotomy Education.... “Having a patient state their name before a blood collection is a very important step in patient identification,” declared Ernst. “The Joint Commission now finds it acceptable for the phlebotomist to use the identification bracelet alone, which we all know can end up on the wrong patient. That’s disturbing to me because it’s taking a chance with a patient’s life. I think any time you dilute the requirements for patient identification, you create a possible scenario that is not favorable to anyone seeking healthcare. There doesn’t seem to be any logic that supports lessening the patient identification requirement,” said Ernst, who worries that not requiring phlebotomists to verify that the person about to be drawn matches the information on their wrist band will lead to unnecessary mistakes....“Given the frequency of these errors, it seems that asking someone to confirm their identity is a simple, positive requirement,” he added. Ernst queried The Joint Commission over this issue. He says that a representative from The Joint Commission informed him that their clients feel that it is “burdensome and unnecessary” to ask a patient to confirm his or her name.

Hospital phlebotomists are the unsung heroes of the clinical labs (see: A Tribute to the Phlebotomists). They function as the ambassadors from the labs to hospital patients. It is commonly one of the lowest paid jobs in the lab but yet one that requires great skill, personality, and tact. We send them out to inflict daily pain on patients and, appropriately, hold them to error-free work performance. Even a single error on their part without major clinical consequences puts their jobs in jeopardy. What the lab and hospital leadership owes them, in return, is a set of procedures relating to blood draws that helps to guarantee this perfect performance. I can say with some certainty, although it will be difficult to uncover the truth, that the lobbying for this change did not come from the clinical labs. My guess rather is that the pressure came from hospital nursing services.

When I was the director of the phlebotomy team in a major hospital, I was confronted with periodic struggles with nursing personnel regarding what lab personnel considered to be proper patient identification procedures. Back in those days, patients frequently had no wrist bands in place during phlebotomy rounds, having been cut off to start IV's and for other reasons. Our phlebotomists, based on our lab procedures, requiring two identity checks when possible (ask the patient to state his or her name, check the wrist band), would refuse to draw the blood absent the wrist band. This resulted in constant howls of indignation from the nurses. "Everyone in the patient unit knows that this is Mrs. Smith." 

At a time when lab professionals are seeking the goal of "six sigma" in terms of lab errors (i.e., 3.4 defects per million opportunities), hospital nursing services leave much to be desired, for example, when passing medications to patients. One 2006 study examined the error rate in a tertiary care hospital for this task (see: Frequency and Type of Medication Discrepancies in One Tertiary Care Hospital). Here are the results:

One thousand, four hundred twenty-four orders representing 197 patients from 13 nursing units were sampled for this study. Thirteen percent of the orders were discrepant and 61% of patients had at least one discrepancy. The most frequent types of discrepancies were drug omissions and unordered drugs.

Hopefully, lab personnel will be able to ignore this little "gift" from the Join Commission and continue to demand fail-safe hospital patient identification procedures. Oh, and be sure to check all of those pills in the paper cups before they are swallowed by your relative in the hospital.

U.S. Army Turns to Wiki Approach to Improve Field Manuals

I am currently engrossed in an excellent book entitled Emergence: The Connected Lives of Ants, Brains, Cities, and Software by Steven Johnson. Emergence is the term used to describe the process by which self-organizing systems use bottom-up information for guidance. Think of an ant colony as a prime example. The queen ant does not provide instructions to the members of the colony -- all she does is produce more ants. What, then, is the source of the information that is necessary to organize such a complex community? It appears that the ants in the colony, as they wander around, generate "local" messages via their pheromone trails such as "food this way" and "we are under attack." The Wikipedia is organized in a similar bottom-up fashion to generate content. The U.S. Army has now adopted a "wiki" approach for editing its critical field manuals, allowing input from all soldiers rather than just officers (see: Care to Write Army Doctrine? With ID, Log On). Below is an excerpt from the article:

In July, in a sharp break from tradition, the Army began encouraging its personnel — from the privates to the generals — to go online and collaboratively rewrite seven of the field manuals that give instructions on all aspects of Army life. The program uses the same software behind the online encyclopedia Wikipedia and could potentially lead to hundreds of Army guides being “wikified.” The goal, say the officers behind the effort, is to tap more experience and advice from battle-tested soldiers rather than relying on the specialists within the Army’s array of colleges and research centers who have traditionally written the manuals....In recent years, collaborative projects like the Firefox Internet browser or Wikipedia pages have flourished with the growth of the Internet, showing the power of thousands of contributors pulling together. Not surprisingly, top-down, centralized institutions have resisted such tools, fearing the loss of control that comes with empowering anyone along the chain of command to contribute. Yet the Army seems willing to accept some loss of control. Under the three-month pilot program, the current version of each guide can be edited by anyone around the world who has been issued the ID card that allows access to the Army Internet system....As is true with Wikipedia, those changes will appear immediately on the site, though there is a team assigned to each manual to review new edits. Unlike Wikipedia, however, there will be no anonymous contributors. Many in the Army have been suspicious about the idea, questioning if each soldier — specialist or not — should have an equal right to create doctrine, [an army officer] said.

First of all, let me say that I consider this to be an excellent move on the part of the Army -- to tap into the expertise of the entire range of personnel for guidance. It's also a tacit admission that high-ranking officers do not have a monopoly on new or creative ideas. In fact, the scales may tip in favor of enlisted personnel in this department. In addition, there is no question that this new process may run counter to the top-down command and control structure of the military. Note the last sentence above in which an officer questions whether "doctrine creation" should be the "right" of all ranks. For me, the answer is whether the goal is for the Army doctrine to be as good as possible.

I can't help but draw a parallel between Army field manuals and the procedure manuals that regulate the operations of clinical laboratories. Such manuals do not make light reading but are critical documents for assessing the quality of a lab at the time of inspection. These manuals are often written and revised by a small group of medical technologists and lab physicians within the lab. I was wondering whether some clinical labs have adopted a wiki process to keep their manuals up-to-date and modify them when necessary. If they can be modified using a wiki approach, the burdensome and concentrated review immediately prior to a lab inspection could be avoided.

Bioethicists Discuss Risks of Sharing Genetic Information on the Web

Consumer genomics offered on the web is a burgeoning enterprise. Genomic testing web sites fall roughly into two categories: those with a medical/scientific orientation such as DNADirect. Sites such as these use a vocabulary such as genomic medicine to indicate an interest in the diagnosis of genetic diseases in the population being tested and provide professional genetic counseling. Other web sites such as 23andMe provide disease discovery by genetic testing but use a vocabulary including words like ancestry, sharing, and community to indicate that one of their missions is to provide networking with other individuals who may be blood relatives. These latter sites have a social/networking orientation. It goes without saying that the business model of these latter sites raises some privacy issues (see: Risks of sharing personal genetic information online need more study, Stanford bioethicists say). Below is an excerpt from this article with boldface emphasis mine:

With just $399 and a bit of saliva in a cup, consumers can learn about their genetic risk for diseases from breast cancer to diabetes. Now, thanks to social networking sites set up by personal genomics companies, they can also share that information with family, friends and even strangers on the Internet....But according to bioethicists from the Stanford University School of Medicine, sharing genetic information online raises a host of ethical questions....Because genetic information applies to more than one person, issues of privacy and consent become complicated. "For example," [a bioethicist commented], "if you receive information on your breast cancer risk and share it with others, you might also be sharing information about your daughter's risk for breast cancer — even though she never consented to have that information shared."....In most cases, customers mail in a DNA sample for sequencing, and then get both raw data and an interpretation of their genetic profile. A few companies, including 23andMe, also let customers create a public profile and share their genetic data through a company-sponsored social networking site. For now, there aren't any laws that govern the exchange of genetic information online. But as genetic analysis becomes cheaper and more widespread, more and more people will have access to their DNA code — and experts fear that consumers may share genetic data without realizing the potential implications for themselves and their families. In addition, both consumers and their health providers may have trouble interpreting data provided by personal genetics companies....Estimates of disease risk are often based on small, unreplicated studies in the biomedical literature, but consumers may not understand how preliminary this data is. "Results depend on the number and type of markers that are used, as well as how robust their databases are," [the bioethicist] said.

As I understand the issues raised in this article plus prior knowledge, there are at least three problems raised by genomic testing web sites: (1) although the tested individual may provide informed consent for the test, the results may have a harmful effect on other blood-related individuals who have not provided such consent; (2) sharing personal genetic data with strangers may lead to mischief at some later time based on the knowledge that the tested individual is predisposed to develop a disease; and (3) estimates of disease risk may be based on faulty or inadequate data, which is to say that they may not be correct and the life of the recipient of the incorrect data may be colored or even ruined. Hospital clinical labs and reference labs are highly regulated by state and federal bodies due to the criticality of the test results being reported by them. It seem quite odd that there is no similar regulatory oversight for consumer genomics web sites. I suspect that the issue is not whether this regulation will occur but rather when it will take place.

The Challenge of Diagnosing Predisease in Our Healthcare Delivery System

New technology is inexorably driving medical diagnostics toward more frequent discovery of predisease. Medical diagnostics here should be taken to encompass both the clinical laboratories and medical imaging. Below is an article describing one further example of how a new nanotechnology technique is "thousands of times more sensitive than conventional methods to low levels of proteins or nucleic acids that show a person could develop a certain disease (see: Inventor to Receive MIT Prize). Below is an excerpt from it:

The Massachusetts Institute of Technology expects to name Chad Mirkin, creator of a range of sensitive diagnostic tests for diseases, as the winner of the $500,000 Lemelson-MIT Prize, a premier award recognizing invention....Mr. Mirkin's products use nanotechnology, which is based on the science of very small particles, to help researchers study diseases. His disease-detecting method, branded as Verigene, is thousands of times more sensitive than conventional methods to low levels of proteins or nucleic acids that show a person could develop a certain disease. The method can return results in an hour as compared with several days, in some cases....The scanning technique -- called dip-pen nanolithography -- creates patterns on several materials on an extremely small scale. Scientists can then use the patterns to learn more about the nature of certain cells and distinguish between normal and cancerous cells.

Our healthcare delivery system is ill-prepared for an influx of large numbers of patients with predisease (see: Predisposition to Disease and Pre-Disease on the Health Continuum; Wellness, Preventive Medicine, and the Classic Disease Model; Genomic Testing, Pre-Symptomatic Disease, and Health Insurance). Put another way, our capacity to diagnose disease is becoming less aligned with our ability to treat disease and reimburse physicians and hospitals for these treatments. Another unintended consequence of this fundamental shift in medical diagnostics is that many of our diagnosticians may be required to make therapeutic recommendations for which they may be unprepared. For now, perhaps the only solution is watchful waiting until we have a better understanding of this trend of increasing numbers of patients carrying predisease diagnoses.

Corrected Definition for a Pod Lab and a Look at In-Office Labs

In a previous note (see: Pathologist Satisfaction with "Pod Lab" Positions), I used the term pod lab in an incorrect way. My better understanding of the term comes after a conversation with Joe Plandowski who consults in this area. His web site is Twincrest. A pod lab is synonymous with a condo lab and both terms are used to refer to an off-site histology lab used to process tissue specimens for a medical practice.  Medicare billing for both types of labs are highly restricted under CMS anti-markup rules and thus unappealing. The term that I should have used in my note was an in-office lab which is the proper term for a lab that is built, owned, and operated by a group of clinicians such as gastroenterologists or urologists. Such a lab processes tissue specimens obtained for diagnostic purposes by the physicians working in the group that are then read and reported out to the group by an in-house pathologist.

Joe tells me also that CMS (Centers for Medicare and Medicaid Services) is now in the process of loosening its regulations such that an in-office histology lab no longer needs to be built inside of the suite of offices of the medical group but will need to be located inside the same building. Joe also tells me that, in his experience, the pathologists engaged to interpret the slides that are processed by in-office labs most commonly work on a fee-for-service, contracted basis for the medical group.

The installation of an in-office histology lab and the interpretation of the slides by a pathologist contractor to the group can be extremely lucrative for the group. The development of such a strategy is a substitute for sending the specimens to a commercial surgical pathology lab or to the local hospital pathologists. The commercial labs will be hit hardest by this emerging business model because office practices are the major source of specimens for such labs. Hospital pathology groups will still maintain control over inpatient specimens and those outpatient specimens generated by physicians working within the health system. The quickening pace of the emergence of large clinical groups and the financial appeal of in-office histology labs for them suggests that pathology groups and commercial labs may soon need to write off much of their small-specimen, office-generated business.

A New Resource for Information about the Medical Technologist Shortage

I have posted previous notes about the medical technologist shortage (see: Comments on the Medical Technologist Shortage; ARUP Labs Responds to the Lab Personnel Shortage). However, I have always been frustrated by the shortage of reliable information about this very important problem. I know that there is also widespread concern and interest about this issue because "medical technologist shortage" is one of the most common search terms used to conduct readers to this blog in spite of my scanty coverage of the issue. It seems that a new resource about this topic is now being made available by the Coordinating Council on the Clinical Laboratory Workforce, which is a task force formed by the American Society for Clinical Lab Science. This latter group is a professional society for clinical lab professionals. Here the vision statement for the CCCLW copied from its web site:

• Identify workforce issues and initiatives.
• Foster educational efforts at all clinical laboratory levels.
• Develop models for demonstration projects that are patient-focused and demonstrate the value of the clinical laboratory professional as it relates to positive patient outcomes.
• Promote partnerships with health care providers, industry,foundations, and organizations interested in the future of the clinical laboratory profession.
• Cultivate relationships with advocates external to the profession.
• Provide a balance between future vision and current clinical laboratory practice.

Here also are some Fast Facts from the CCCLW web site about the med tech shortage obtained from the October 2008 issue of Critical Values:

• 44% of laboratories currently report difficulty in hiring [medical technologists].
  
• About 150,000 new technologists will be needed by 2014: 81,000 to replace retirees; 68,000 for new positions.
• There has been a 67% drop in new CLS/MT graduates since 1977: 6,519 graduates in 1977; 2,141 in 2000.
• 40% of the current workers will retire in 10 years.
• There has been a 71% decline in the number of accredited training programs since 1975: 770 NACCLS accredited programs in 1975; 222 in 2007.
• Average age of current workforce is 49.2 years.

I fear that the med tech shortage problem will get worse before it gets better. I have heard rumors that various schools are currently reducing or eliminating their medical technologist programs due to cost pressures and difficulty in recruiting qualified faculty members. I will try to provide more information about this issue in the future.

My thanks to Paul Epner for allerting me to the work of the CCCLW.

Additional Comment on Results from Patient-Performed Home Lab Tests

In my note yesterday (see: Lab Test Results as Essential Components of PHRs), I wrote that the results from patient-performed home lab tests do not have the same standing as those performed by hospital or commercial clinical labs. I took special care in my choice of the work standing -- I thought that it was sufficiently ambiguous that it would not provoke any comment. I was wrong. William Shipley submitted the following:

Having just done my morning glucose finger stick and watched the numbers continue their steady track back to normality after the gastronomical delights of Las Vegas..., I read your comments on home glucometers not having the same standing as clinical tests. That's not particularly controversial, I've said it several times myself. But then I got to thinking about the assumption that a professional lab would do better than the 'free' glucometer in my hand. Well, clearly the technology is better. On the other hand, the drop of blood on my finger was most certainly mine -- no sample mis-identification. And while I don't run QC (and I even know better!), I nevertheless ask myself "does this make sense?" and repeat any outliers. My sample is also pretty fresh and has not suffered from handling, cooling, etc. Do I think I'm doing better than a professional lab? Probably not. But as we move toward personal health records and more self generated data, balancing that data with professionally gathered information will be an on-going issue. We should beware that we do not let our professionalism cause us to undervalue data gathered by the people who care about it the most.

I carefully chose the word standing because I did not want to disparage the use of patient-generated test results. After all, the patients themselves use these results to regulate their blood glucose levels. I have posted a number of previous notes about glucometers and believe that they are very useful extensions of hospital labs. I also favor the process of uploading and integrating these home results into hospital, office, and personal EMRs. Having said this, I also believe that the source of home glucometer results needs to be clearly identified in any electronic record.

It is a common practice for office or hospital-based physicians to repeat previously performed tests, particularly when they are important and the patient is new to them. In effect, they are questioning the standing of previous results. This is common practice and certainly relevant for home test results. Many test results are dynamic and it is necessary to assess the most current status of the patient. Physicians also tend to favor results from labs with which they have a long-standing relationship and therefore trust. This is partly a security-blanket and partly defensive medicine. The practice is also common in radiology and called over-reading. When a patient is admitted to the hospital, outside medical imaging studies are frequently reviewed by the in-house radiologists.

Finally, I want to refer to quality control (QC) of lab tests. I often refer to QC as the religion of lab professionals and they pursue it with a laser-like intensity. This bears on the accuracy of any test result not performed in an accredited lab. It is a mind-set not always understood or shared by other health professionals and probably not by healthcare consumers. I don't believe that patient-generated test results should be undervalued. I think that they are another set of data points available for interpretation. However, given that some of them may, in fact, be erroneous or misleading, they need to be placed into the proper clinical and information context, a phrase that I used in my previous note.

Who Should Not Receive Chemotherapy: The Next Step in Healthcare Reform

As we enter an era of healthcare reform with increased pressure to reduce costs, the issue of  "who not to treat" will begin to loom as large in our minds as the therapeutic imperative. I have addressed one aspect of this question in a recent post (see: A Look at Deloitte's Healthcare Reform Pyramid: A Strategy for Reducing Costs) with the references to the comparative effectiveness of diagnosis and treatment. Physicians will be increasingly asked whether a lower cost treatment will be just as effective as a higher cost one or whether treatment is even justified, as in the terminally ill. The increasing use of lower-cost generic drugs is a reflection of this same question.

I will be lecturing on this topic at Lab InfoTech Summit 2009. My lecture title will be the following: Biomarkers as a Key Enabler for Monitoring and Modifying Drug Therapy. Companion diagnostics will be increasingly used to select candidates for treatment with expensive biotech drugs. A recent article on the web (see: The Next Step in Cancer Drugs: Who Should NOT Get Them) raised the interesting and challenging topic of which cancer patients NOT to treat or which to treat differently. Below is an excerpt from the article with boldface emphasis mine:

Lots of patients with advanced colon cancer shouldn’t get some of the most advanced drugs used to treat the disease, according to a new recommendation from the American Society of Clinical Oncology (ASCO). It may seem odd that a group of cancer docs is recommending against a popular treatment for many patients. But figuring out which patients are unlikely to benefit from cancer drugs — which are not only very expensive but can also be toxic — is a high priority in cancer research these days....The new ASCO recommendation says patients whose tumors have certain mutations in a gene called KRAS should not receive a class of drugs that includes Erbitux, co-marketed in this country by Eli Lilly...and Bristol-Myers Squibb, as well as Amgen’s Vectibix. Studies have shown pretty clearly that patients with the mutations — which occurred in about 35% of cases this large study — don’t benefit from the drugs. Not using Erbitux as a first-line drug for these patients could save about $600 million a year, the WSJ reports. Amgen said the data warrant a label change for Vectibix, and a Bristol spokesman told the WSJ that Bristol and Lilly look forward to working with the FDA.

Interestingly enough, KRAS, mentioned in this excerpt, is the only biomarker that I know of with its own web site. It's very professionally done. A few clicks reveals that the sponsor or this web site is Merck, the developer of Erbitux.

In previous notes (see: Moving Resources from the Therapeutic to the Diagnostic Silo, Genetic Testing to Cull Out the "Un-Right Patient" as a Candidate for a Particular Drug Therapy), I commented on the feasibility and desirability of moving resources from the treatment silo to the diagnostic silo and sometimes making the decision of withholding treatment from the "un-right" patients. In the article cited above, we learn that not using Erbitux as a first-line drug for certain patients could save about $600 million a year. That's a fairly good start in moving resources, considering that the KRAS testing itself would consume only a tiny fraction of the resources that will be saved by not using the drug as the first-line agent. Look for more attention to drug label changes as directives from the FDA become more common requiring the use companion diagnostics to properly select patients for the use of certain chemotherapeutic agents.

LIS vs. LIMS: It's Time to Blend the Two Types of Lab Information Systems

Early in their developmental history, the two major types of laboratory information systems, LISs and LIMSs, began to diverge in their functionality. The differences between these two types of systems is not revealed by their names: LIS stands for laboratory information system and LIMS stands for laboratory information management system. LISs are installed in hospital labs and commercial reference labs. LIMSs are installed in industry settings, by which I mean facilities such as pharmaceutical research labs and other industrial labs such as those in food or chemical manufacturing. 

• LIMSs are designed to report results for batches of samples (e.g.,100 lab rats) to the responsible parties. They are required to satisfy good manufacturing practices (GMP), the FDA, and research scientists working in the pharmaceutical industry, as only one example.
• LISs are designed to report test results for individual patients back to the physicians caring for them. They need to satisfy the criteria of hospital accreditation agencies and HIPAA and also to avoid medical malpractice suits.

All of the above was brought home to me as I was listening to Dr. Jill Hagenkord's excellent lecture at the recently completed APIII conference in Pittsburgh (see: Array-Based Virtual Karyotyping: Ready for the Clinic But Waiting for a LIMS). During the course of her lecture, she made reference to LIMS as in the title of the lecture. However, she was clearly discussing clinical lab matters for which I felt that her use of the term LIMS was unexpected. When I discussed this with her afterwards, she was unsure of the differences between an LIS and a LIMS.

The source of her confusion quickly became apparent. She is representative of the new breed of pathologists who work in the area of genomic, proteomic, and genetic testing and for whom there is little gap between the clinical world and the world of medical research. For example, she uses sophisticated research tools but reports the results to clinicians caring for patients. The research and service worlds form a continuum for her and hence require a blending of the LIMS world with the LIS world.

One of the major goals of the translational research movement has been to bridge the gap from bench to bedside and any initiative to blend the LIS with the LIMS can be understood as an extension of this quest. The lesson from all of this is obvious. We now need to make an effort to understand the relative strengths of LISs and LIMSs and proceed to develop new systems that capture the best features of each of them.

OvaSure Conflict Between LabCorp and FDA Heats Up

I have posted a previous note about OvaSure, LabCorp's multiplex lab test for ovarian cancer  (see: Questions About the OvaSure Test for Ovarian Cancer). I have also posted a number of notes about IVDMIA lab testing, the term used by the FDA to refer to multiplexed lab tests which consist of multiple biomarkers plus a computer algorithm to help interpret the test results. OvaSure is an early example of such a lab test. The dispute between LabCorp and the FDA seems to be heating up based on a recent article (see: FDA Slams LabCorp for Selling Unapproved Ovarian Cancer Test). Below is an excerpt from it, with boldface emphasis mine, plus an anonymous comment in response to the article:

The FDA has warned clinical-test giant LabCorp that it has been marketing an ovarian cancer test without approval, vindicating skeptics of the assay who worried it wasn’t ready for prime time....LabCorp made the $220 test available in June, without prior FDA review, by taking advantage of a technicality. The regulator doesn’t require a premarket review for tests developed and performed by a single lab. But in a letter dated Sept. 29 and released today, the FDA determined that because the test was developed at Yale and that parts of it were manufactured elsewhere to the university’s specifications, it must meet the agency’s usual premarketing approval requirements, which could take up a year....In a study published last February, the OvaSure test correctly identified 95% of cancers in 224 samples, with a false positive rate of 0.6%. But cancer organizations were wary of it because those test were performed on women already suspected of cancer, not a general healthy population, and because OvaSure hadn’t gone through the usual FDA review....Researchers at Yale are continuing tests on OvaSure and expect to have more data later this year.

Comment to this article by "OvaSure for Sure?" Generally, a large number of archival specimens are batch processed together, within a very narrow time frame, by the same research team, so all the technical variables are minimized, which makes it much easier to get good results than in a real-world setting, where specimens are tested over a period of weeks, months, years, by different people, with different laboratory reagents, as occurs in the real-world.

There are three issues which jump out at me from this story that I think are worthy of further discussion:

• As noted above, premarket approval by the FDA is not required for "tests developed and performed by a single lab." These are the the so-called home-brew tests. Despite the name applied to such testing, I don't think that it is appropriate to refer to the performance of such tests as "taking advantage of a technicality." This has been a well-accepted strategy for the adoption of a large number of innovative new tests in the past. Obviously, there is some degree of ambiguity in the current definition about what is meant by the development of a test "by a single lab." I will leave this to the parties involved to hammer out.
• The second issue under discussion in the article above is the design of research studies of emerging biomarkers and their subsequent inclusion into multiplex assays. In retrospective studies using archived serum samples, various biases can be introduced relating to how the patients in the study came to be tested. In other words, what was the index of suspicion that the patients had cancer of the ovary and therefore came to be tested for the disease?
• The third issue, raised by the person commenting on it, is also very interesting. In a retrospective study using batched archival samples, there will always be a minimization of technical variables that is different from "real-world" testing in multiple labs. However, the same criticism can be leveled against a prospective study with all lab testing performed in a single lab or small set of research labs. It's difficult for a research study to closely approximate "real-world" conditions.