Science has come so far since I was in college; it’s difficult to paint a picture without sounding like one of those “I walked 5 miles to school in a blinding snowstorm” stories.

While I didn’t have to walk that far to school, the Mesozoic era of chemistry that I worked in is laughable when I look back at it. Let’s see; we wrote programs for physical chemistry labs by using a line editor, typing on teletypes, linked to a distant computer via an acoustic coupler modem. I don’t even want to tell you how slow these things were.

How about the x-ray diffraction equipment? We took crystals, mounted them in a small container, surrounded the container with a strip of undeveloped x-ray film, set it in the center of a room, then ducked behind a barrier to blast the thing with x-rays from a projector. We’d develop the film, and then the fun began with interpreting the lines on the film.

Ah, the very word “interpretation” brings on the shivers. In qualitative organic analysis class, we’d be given unknowns, and told to identify them. This was actually a fun class, but using today’s equipment, the lab would have taken about ten minutes. Back then, derivatives frequently had to be made as the unknown compound may not have lent itself to conventional spectral analysis. Don’t ask me about my little thionyl chloride adventure whereby I was working with a compound upon which I was trying to run a chlorination. I had a small beaker of the compound in an aqueous solution, placed in another beaker of ice. I closed the hood down over my hands and, using a Pasteur pipette, shot a healthy amount of thionyl chloride into the solution. My lab book basically instructed us to “add five ml of thionyl chloride…” which is what I read when I unloaded the pipette. The problem was, I hadn’t turned the page to see, in bold print, “slowly, one drop at a time, and wait for the reaction to slow down between drops!”

Ooops. My haste was rewarded with an instantaneous explosion that pulverized all of the glassware I’d been working with, driving some of it right through my lab coat, and scaring me so badly that I took the skin off the top of my wrists as I yanked them out of the hood in about a millionth of a second.

After we ran infrared spectra on our unknowns, we’d have to go to the library and start looking at the Sadtler wavelength charts to compare the peaks of our unknowns to those in the literature. Typically, we’d have to do further library research using the Beilstein series of books. The problem was; as good as Beilstein was, it was all in German. Nonetheless, we all developed a working knowledge of the language to be able to navigate the series.

Perhaps the most humorous inconvenience we had was working with the mass spectrometer. This vile and temperamental instrument was assigned a permanent graduate student babysitter, and no one else could approach it. I can recall few times where I peered into the room and didn’t see the technician buried in it up to his waste. The machine easily had a working:not working ratio of about 25:1. Now, when I walk into a lab and see GC/MS instruments, I want to grab the analyst and tell them about the hard times, but I know they wouldn’t care.

These days, the instrumentation is far more sophisticated, as is the topic of informatics platform unification.

One of my favorite books by the late, but phenomenally gifted author, James Clavell, is Shogun. Set in feudal Japan in 1600, just one of the complex plot lines revolves around the existence and propagation of fiefdoms. These are revenue-producing properties that were fiercely defended by their rulers, often with brutal force.

Fiefdoms also describe the common culture in large multidisciplinary scientific enterprises whereby turf protection superseded the greater good of the corporation. It used to be more important to fight to the death over matters of local jurisdiction and policies than to worry about the IT department’s quest to control and manage the computing infrastructure for the corporation.

To praise another site’s methodologies was a sign of weakness, and to capitulate by adopting those policies resulted in a career-limiting stagnation by virtue of some peer being perceived as being an innovator, while another was perceived to be a follower. Now, more than ever, the scientific enterprise is far less receptive to fiefdoms, and the word “harmonization” is spreading throughout industries as a call to battle. Indeed, it’s a buzzword with a plethora of interpretations.

In the pure pharmaceutical/biotechnology manufacturing world, “harmonization” means that the enterprise desires to consolidate operations, analysis methods, naming nomenclature, review/approval workflow, test code definition, results limits, rounding algorithms, etc.

But now, let’s look at the progressive scientific enterprise that desires to fully utilize the functionality in a product such as STARLIMS, and they intend to address a myriad of issues attendant to the world of eClinical data management and statistical analysis.

I encourage you to read a detailed analysis on the subject of harmonization in the clinical world that STARLIMS’s Joe Tehan, Senior Account Manager in our Clinical Operations group, has posted. You can find Joe’s excellent piece at:

Starlims Clinical ICT

Remember, clinical environments add multiple dimensions to an informatics approach. There is the need for sophisticated specimen tracking and possibly biobanking. There are issues attendant to matching collected specimens to the subject they were obtained from. And let’s not forget that the clinical world presents entirely unique challenges for data privacy.

Another topic Joe’s article mentions is STARLIMS excellent language support feature. The STARLIMS interface can easily be configured to reflect the native language of the users, but there are some interesting validation debates bandied about by our customers on this point.

Bear in mind that STARLIMS is used in a wide variety of industries, and some of those are regulated. In the world of FDA-regulated companies, and we’ll specifically refer to pharmaceutical and biotechnology enterprises, New Drug Applications (NDA) and other submission documentation must be submitted in English. This isn’t a problem in the STARLIMS world, but some customers take an extra precautionary step in how they present the user interface to their users. Some corporations want to be absolutely certain that all users are interpreting the user interface in precisely the same way, so they have adopted a policy whereby all screens in their LIMS must be presented in English. In fact, there are major pharmaceutical companies who, for this and related reasons, mandate that all technical employees are fluent in spoken and written English.

This approach may be considered overkill by some corporations, but essential to others. It comes down to the fact that FDA policies are really guidelines in many cases, and not spelled out to the minutest degree. So, there is a certain amount of risk assumption that regulated companies must consider. As a general rule of thumb, less risk = more stringent (and, hence, costly) validation policies. Conversely, more risk = less stringent (and less expensive) approaches to validation.

Which is the “right” approach? There is no answer to that question. Better stated; it is a moving target, and really open to interpretation (no pun intended) of the company’s compliance organizations.

We frequently are asked by FDA-regulated scientific enterprises: “how much will validation cost?”

This is an impossible question to answer without asking quite a few questions in return. So, sadly, the response often follows the lines of “it depends”.

There are companies who follow almost Draconian validation policies, and each project they must produce budget estimates for requires a healthy amount to be set aside for validation. Other companies may bypass “conventional” validation approaches and use their own internal policies and procedures to attain their own interpretation of a compliant environment. As a general rule, validation consultants will start with an assumption that stringent compliance is desired, unless a company explicitly details past approaches/templates that suggest a different (read: less stringent) approach.

In all cases, it is not the validation consultant who assumes the risk; it is the company, so the final approach is really a matter of corporate risk assumption and a constant monitoring of evolving FDA requirements.

Regardless, STARLIMS can allow strict control in terms of language support, or present a multi-faceted, flexible interface for users. STARLIMS will not limit the scientific enterprise’s ability to invoke their compliance policies; that is a matter for the enterprise’s compliance organization.