Three New Technologies for 2004

February 1, 2004

Applied Clinical Trials

Applied Clinical Trials, Applied Clinical Trials-02-01-2004,

One new idea that?s ready for use; another that will be soon; and a third that?s not yet there.

Our relationship with technology is often love/hate. When it works well, it can lead us out of a morass of paperwork, into an organized, efficient digital world. Technology can be the beam of light from a lighthouse, guiding a boat safely home. The introduction of spreadsheets must have felt that way to accountants and business planners everywhere. However, technology can also be the Sirens of Homers Odysseybird-like creatures with womens heads whose song would lure sailors to their death on rocks. Lets analyze three compelling technologies to determine which, in 2004, will be the lighthouse and which the Sirens.

Before we start, let me tell you that the three technologies I am discussing are likely to be valuable, productivity-enhancing, and successfuleventually. The question is which one will be ready for prime time this year, and which will have to wait until next year, or maybe 2010? The three technologies I will discuss are instant messaging (IM), single sign-on (SSO), and handwriting recognition.

When thinking about the implementation of any new technology, we need to ask some probing questions. Does the technology solve a real business problem or one that has been created for the technology? Can it be integrated into my daily workflow? Is the technology mature? Does it require bringing together a number of different technologies/participants? Does it add work to my day? With any technology, the successful implementation requires a balance of people, process, and technology considerations.

Instant messaging
Instant messaging (IM) is a good place to start. This technology grew up alongside a whole series of (pre-) Internet technologies, including email, mailing lists, and bulletin board technologies. These communication technologies run along a spectrum of immediacy. The technology that is least immediate is the bulletin board, where people can post messages in threaded conversations that may or may not be read. It is a classic pull technology, where the reader must actively seek out the information, or it is never communicated. Next in immediacy is email and the mailing list. Both involve delivery of messages, but email is a warmer medium in that it often demands a response (excluding spam and corporate memos, that is). Finally, we come to IM, a hot medium that demands a more or less immediate response.

Since it is core to our discussion, I will briefly explain IM and its impact. I preface this explanation by telling you that I was recently explaining the impact of IM on teenagers and college students to a group of CRAs/CRCs in a training course and slowly caught on, to my embarrassment, that more than 90% of them were very familiar with IMI was the one out of touch. IM involves the installation of an application on the desktop that allows you to communicate with your friends and colleagues one at a time, in groups, or in many parallel discussions. The IM application announces to your correspondents that you are on your computer and available for a conversation, and displays to you the names of people who are available. Anytime you feel the need, you can start up a typed conversation with one of these people, and they will get the message instantly. Through IM you can have a brief discussion, or a prolonged one.

Instant messaging is becoming pervasive in home life in the United States. Older kids can sit by the computer having simultaneous IM conversations with 5 or 10 people. The first thing that new acquaintances do when they part is exchange IM screen names. Through IM, kids can keep in touch with distant relatives and old friends in real time. Parents are using IM to touch base with their kids at college in an ad hoc fashion. IM is being used by 50% of Internet-enabled households in the United States, a staggering number. IM has also begun to show up in major corporations where it is intended to enhance communications and increase productivity.

There are several limiting factors to IM adoption in the corporate workplace. To begin with, the lack of compatibility between the various IM offerings (AOL IM, Windows/MSN Messenger, and Yahoo Messenger) is an issue for intercorporate communications. As of this writing, there seems to be a definite difference in the users of the various services, with AOL IM appealing to young adults and families, and Yahoo appealing to a somewhat more sophisticated IT audience. As with many things, Microsofts offerings in this arena are doing the best job of capturing corporate interest, in part because IT buyers expect them to increasingly integrate with other Microsoft Office products. Furthermore, most corporate IM is within a single corporation, so the merging of different IM technologies is only an issue for people who want to communicate with their families outside corporate walls.

One issue with corporate IM is the noncorporate use of the technology. In a recently published survey, up to 80% of employees admitted to gossiping while at work using the IM technology, 64% used it to complain about management, and almost one-third used it to flirt. In fact, only 27% of people said that they confined their uses of IM to business purposes. This type of behavior is problematic for companies and certainly doesnt increase productivity. One behavior identified as negative is conspiring during conference calls. In my view, this is one of the more valuable uses of IM. When members of a company want to be on the same page when speaking with a customer or partner, but they are in different locations, how better than having a back-channel conversation on IM during the conference call?

Another issue is that standard IM conversations are not recorded to be available for future review, as emails are. In fact, corporate implementations of IM now typically include the capability to capture all IM conversations. This fact, if known to employees, should reduce the occurrence of gossiping/flirting.

The potential of IM in clinical trials is related to the instant communications that it enables. Specifically, those little questions that a site might ask a medical monitor can be answered without a phone call, and without a major disruption in the workday. Similarly, project-related questions can be addressed quickly during the normal process of work. When properly installed and supported, the IM conversation stream can be fully logged and recorded, enabling it to become part of the record of clinical trial correspondence.

IM is being tied into other tools that enhance collaboration. With one menu choice in an IM application, it is possible to start a voice conversation from the microphone that is part of modern laptops, or send a file. In addition, a whiteboard can be shared between two people in an IM conversation, allowing them to share ideas and concepts. Finally, it is fairly easy to invite someone to view an application that you have open on your desktop. Any application, not just Microsoft Office, can be shared so the person with whom you are conversing can actually control it. The major limiting factor in using these capabilities in a corporate setting is educating the user that they even exist.

Single sign-on
A second technology that is currently generating a lot of interest is single sign-on (SSO) technology. I am sure you are familiar with the problem that SSO addresses: the proliferation of passwords for all the Web sites and applications that you use. While you may get away with using the same simple password over and over again on commercial sites, most corporate applications have different requirements.

An electronic data capture system, for example, might require an eight-character password composed of small and capital letters, numbers, and symbols that gets generated randomly. Another application might require a passphrasea series of four or five words. Each may rotate every few months, and sooner or later, we are all tempted to write them all down and post them on the corner of our monitor on a Post-It note. Of course, the whole idea of a secret password breaks down at this point.

Single sign-on promises to answer this particular problem. A single sign-on solution might be implemented as follows: Each morning you would open up an application or Web browser to a particular page and sign in. Perhaps you would use a username/password, or maybe a SecureID card. You might even use a smart card, containing your digital certificate. In any case, once signed in, all your applications would become available to you. The portal application may securely store your applications passwords and make them available to each application on demand, or it may be linked into the security module of each application, allowing each to start without requiring a further password entry.

In fact, the technology required to sign up for Windows Messenger (Microsofts version of IM) includes a type of single sign-on unique to Microsoft, known as Microsoft.Net Passport. This Passport is, in fact, an attempt by Microsoft to provide SSO over the Internet. Once logged in, you can be automatically logged into any number of Web sites that have a trust relationship with Microsoft. If you sign up for an MSN Wallet and provide your credit card information, you can automatically buy merchandise from a variety of Web sites without entering any additional information. Microsoft is hoping that others will adopt this technology and build it into their applications, creating a universal single sign-on. This type of single sign-on is a centralized model where a central server provides authentication, and then communicates with the other applications/Web sites, passing the authentication information to them. Centralized models of SSO have some issues, especially for users outside of the corporate network, who may be forced to log onto the network to get passwords for their local applications.

Another type of single sign-on is client-centric SSO. In this scenario, the passwords are stored on the local machine. This approach requires the installation of some specialized software on each computer, which contains the passwords and provides them when needed. Except in highly customized solutions, this may be a manual process, whereby a user instructs local software to apply a password to a Web site or local application. The disadvantages are those of any local applicationlocal support, validation, interactions with other programs, and the security risk of having passwords available outside of a controlled server environment. There are several shareware programs available on the Internet that can perform this function.

Handwriting recognition
The third technology worth mentioning is handwriting recognition, one of the Holy Grails of computing. Ideally, a computer would be able to take handwritten text and accurately convert it to text. Handwriting recognition offers a lot of advantages to the user. First of all, the capture of information is more natural for many people. Writing can be focused on the task at hand, rather than typing errors. Furthermore, with handwritten documents it is easier to mix and match symbols, drawings, and other nontext items with text. Finally, with most modern handwriting recognition software, it is easy to capture individual key strokes as part of an audit trail and to authenticate the author based upon the style and tempo of their writing.

Handwriting recognition is best when whole words are captured, rather than individual letters. This is because errors in detection of a particular letter can be obviated by recognition of a whole series of letters together, making a word. This recognition is obviously dependent on a reference dictionary for which the match is performed. To optimize recognition, it is best to have a small dictionary and a small number of possible words. The very nature of clinical trials makes this requirement very difficult. When words are used that are not in the dictionary, there is no end to the amusement this can bring (see sidebar).

There are two different approaches that can be used separately or combined with the recognition of handwriting when handwriting is being recognized through the individual strokes of the pen. The first is asking people to adopt a particular writing style. The use of Graffiti in Palm Pilots is a good example of this technology. Unfortunately, it is difficult to get uncommitted people to change something they have been doing the same since they were in kindergarten. The next approach is to train the computer to recognize the strokes and shapes of letters that you use. This approach again requires some hours of writing to train. However, with some work it can currently give very good results.

Handwriting recognition has had a long and somewhat checkered history. The earliest commercial applications were in optical character recognition: typically the recognition of handwritten text after its completion. In fact, this technology was applied to clinical trials and is still in use today. Scanned or faxed CRFs can have some fields that are pretty easy for a computer to readare they checked or are they not checked? However, other fields may require numbers that can be confusing (7 and 1 can be confused in Europe; 4 and 9 worldwide), or, worse, words. Words and sentences are amongst the hardest for OCR technology to recognize, and they require a fair amount of manual interpretation. The problem for optical character recognition (OCR) and all other handwriting technologies in a clinical setting ishow large an error rate can we accept for the capture of primary data0.1%? 1%? 10%? The lower numbers are very difficult if not impossible to achieve. More to the point, identifying and correcting the errors is inefficient, time-consuming, and error-prone.

Handwriting recognition in real time has been around for several decades, but the technology entered the mainstream with the introduction of the Apple Newton, the predecessor of todays personal digital assistants (PDAs) that included a technology to learn native handwriting based on dictionary matches. Many swore by the device, while others swore at it. The fanfare associated with its introduction was deflated within the first week, when Garry Trudeau published a Doonesbury comic that lampooned the technology and its users. He had a character enter Catching on? and the Newton translated it as Egg Freckles. Some say that the strip was almost single-handedly responsible for sinking the technology. In fact, Apple continues to work with the underlying technology, incorporating it in their operating system.

Later attempts include the IBM/Cross Pad, which recorded pen strokes, and the Palm Pilot and similar technologies, which are dependent on modifying the users handwriting style. Microsoft has incorporated handwriting technology in its latest Tablet PC operating system, and its accuracy has predictably been the subject of some criticism.

The most recent entry into this field is the Anoto Pena quite different version of digital handwriting capture. Briefly, the pen contains a minute camera that captures the configuration of tiny, almost invisible dots on special paper. The dots are cleverly arranged so that the pen can identify and transmit the exact location on any piece of paper of a pen stroke. The pen captures individual pen strokes and can transmit them to a local or central computerproviding the equivalent of a photograph of the page with an audit trail for each pen stroke. In the end, though, the technology is dependent upon, and subject to, all of the foibles of the current state of the art of handwriting recognition.

Unfortunately, handwriting recognition has failed to reach a sufficient level for regular, reliable data recording. While recognition algorithms are improving, they still require manual review and frequent correction to achieve 100% capture. This is almost the same problem with OCR. While it is often possible to develop a confidence interval for recognition, this will only identify the low likelihood matches. In order to catch all errors, it would almost be necessary to look at every word. This would eliminate the advantages of handwriting recognition.

So, we have three technologies. Which fulfill the original criteria? Which give proper balance to people, process, and technology considerations. As mentioned, they all solve important problems, and they all can integrate well into a work environment. It is in maturity of the technology that they vary. IM technology is becoming mature and will only improve over time. SSO technology is still young, and many different versions are available. It is not yet clear which will prevail, or if this is even relevant. SSO may be the kind of technology that each company implements independently. Handwriting recognition is great for PDAs and other noncritical uses, but will be challenging if used in regulated data collection, such as clinical trial data management.

Handwriting is a people- and process-oriented technology, but fails on technology availability for mission critical applications. SSO works from a technology and process perspective, but will take significant re-education efforts when implemented. IM, though, has an established technology, a clear process, and is very strong on the people side; nontechnical people often enjoy using it. Of the three technologies discussed, IM is closest to being our lighthouse, while handwriting recognition currently threatens to dash us into the rocks.

SIDEBAR: Lewis Carrolls Jabberwocky, as recognized by the Apple Newton
(You can find the original Jabberwocky in Through the Looking-Glass and What Alice Found There, 1872 or at http://www.jabberwocky.com/carroll/jabber/jabberwocky.html)

TABLESPOONS
Teas Willis, and the sticky tours
Did gym and Gibbs in the wake.
All mimes were the borrowers,
And the moderate Belgrade.
Beware the tablespoon my son,
The teeth that bite, the Claus that catch.
Beware the Subjects bird, and shred
The serious Bandwidth!
He took his Verbal sword in hand:
Long time the monitors fog he sought,
So rested he by the Tumbled tree,
And stood a while in thought.
And as in selfish thought he stood,
The tablespoon, with eyes of Flame,
Came stifling through the trigger wood,
And troubled as it came!
One, two! One, two! And through and though,
The Verbal blade went thicker shade.
He left it dead, and with its head,
He went gambling back.
And host Thai slash the tablespoon?
Come to my arms my bearish boy.
Oh various day! Cartoon! Cathay!
He charted in his joy.
Teas Willis, and the sticky tours
Did gym and Gibbs in the wake.
All mimes were the borrowers,
And the moderate Belgrade.

It seems very pretty, Alice said when she had finished it, but its rather hard to understand! (You see she didnt like to confess even to herself, that she couldnt make it out at all.)
Lewis Carroll, Through The Looking-Glass

(c) 1993 Robert McNally.

Related Content:

eClinical