Three Risks of a Startup

In a conversation with Duane Knight the CFO of Denver Biomedical, recently acquired by Cardinal Health, he mentioned the three main risks of a startup company. Although you may be saying "just three?" I thought these summed up the topic, I like lists of three and no one would read this blog if I listed fourteen. So here goes:

1. Management Risk
2. Technical Risk
3. Market Risk

Management risk encompasses all the risks involved with the management team: do they have experience, can they execute the business plan, can they work together and can you even round up the right people. Startups are especially difficult because highly qualified people may not be available or may not want to quit more stable jobs for the volatile environment of a startup.

Technical risk is the risk involved with the product or service the startup intends to sell. A startup by its very nature is usually trying to do something that hasn't been done before whether that's a new drug compound or new food product. New drugs have especially high technical risks due to the uncertainties when administered to humans. In comparison, software is relatively low risk. That may come as a surprise to those organizations that have pumped millions into IT systems only to have them scrapped. But software is more adaptable and, when managed properly, most IT projects eventually hit their targets.

Market risk is the risk associated with all those things concerned with selling the product/service: does the market need such a product, are there competing products, will the market buy enough for the company to eventually make a profit?

The startup company must deal with each risk and develop ways to reduce these risks in order to build value in the company; the lower the risk, the greater the value. Eventually, with enough work, patience, money and luck, the risk can be managed to the point where the company can actually forecast their revenues and profits a month, quarter or even a year in advance. A sure sign that the company's ready to go public.

Gateway to the FDA

The FDA launched a new initiative in May 2006 called the Electronics Submission Gateway (ESG) which lets pharmaceutical companies make regulatory submissions electronically. Although a modest step into the digital era, by all indications it appears to be a success. Before the ESG, drug companies were required to deliver the information in some kind of physical form, be it paper, CD-ROM, or digital tape. In the case of paper, the deliverable could be truckloads. The ESG allows for huge amounts of information, up to 100 Gb, but now it takes only minutes to upload, not days to ship.

Not only is transit time reduced, but the sender also receives a reliable acknowledgement of receipt; confirmation that the submission is in process. With the physical media there was always some question as to whether the documents made it to the right place or not.

Currently the FDA accepts submissions to CBER, CDER and CDRH. They also support receipt of Adverse Event Reporting System (AERS) reports. The next phase of the ESG will support submissions to the Center for Veterinary Medicine (CVM), Center for Food Safety and Nutrition (CFSN) and the Office of Orphan Product Designations (OOPD).

AERS reports topped 12,000 for March 2007, by far the largest percentage of submissions. These tend to be much shorter than the other submissions which have been growing steadily since launch. Users appear satisfied with the system and there are now between 150 and 200 registered users of the system.

Each of the above centers has its own set of requirements for submission, however, documents are free text which requires manual processing. Although delivery time of the documents has been reduced from days to minutes, the bulk of the regulatory work will proceed at the same pace as it had before. This can certainly be streamlined, as it has been in many other agencies, but this will require quite a significant investment, one which carries far more risk than what has been done so far.

The Future of the Web

BusinessWeek has run a special report on the future of the web in their April 9 issue and seems to be pointing the direction of the Semantic Web technologies. As an example, BW sights the pharmaceutical giant Eli Lilly as an early adopter of SW technologies for use in drug development. In an attempt to slash a third of the cost off the development of a new drug, they are implementing SW technologies to better manage their data.

Tim O'Reilly is quoted as saying "Web 2.0 is the messy way that the Semantic Web is actually happening." Messy in the fact that it's a bottom-up folksonomy rather than the top-down ontology driven structure required by SW specifications.

Does this exposure in BW mean SW is ready for prime time? "I'm still trying to figure out what Web 2.0 is" remarked one VC at last night's MIT Enterprise Forum in Boulder. But while the entrepreneurs try to sort out the business model, Lilly's IT department goes it alone, "There aren't a lot of people we can turn to with experience." Seems like there's got to be a business model in saving a pharmaceutical company $400 million off the cost of developing one drug.

Perverse Incentives

In a recent piece in the Wall Street Journal entitled Perverse Incentives in Health Care, John Goodman describes the environment in today's healthcare system that disincentivizes efficiency and quality in healthcare and rewards the status quo. Goodman places the blame for this on a system in which the payer calls the shots; Medicare or the insurance companies dictate what procedures they will authorize and how much they will pay for them. It's the healthcare providers' job is to provide the best care for their patients while staying within the boundaries set by the payers.

This has the effect of reducing and discouraging any innovations in the field because the provider may not be reimbursed for them. If a healthcare provider, medical device or pharmaceutical company has a better, more efficient or effective way of treating a patient this must be reviewed and approved by the payers before a healthcare provider can get paid. Although it is in the long term interest of the payers to reduce costs and improve quality, they are large bureaucracies and do not respond quickly to changes in the market. As a result such innovations as email and electronic medical record systems are not as easily implemented in healthcare as they are in virtually every other sector of the economy.

Goodman points out that those sectors of healthcare that are not reimbursable but rather paid for directly by the consumer, such as cosmetic or laser eye surgery, are both competitive and entrepreneurial. These sectors offer competitive and dropping prices, rapid gains in technology and constantly improving quality. In these sectors the coupling between payer and provider is looser; the consumer is simply interested in a smaller nose, better eyesight or fewer wrinkles and does not want to sweat the details of how much a syringe, anesthetic or some sub-task costs.

Some healthcare providers, such as the Mayo Clinic and Intermountain Healthcare, have bucked the system and have implemented efficient practices resulting in dramatically reduced healthcare costs. This requires not only vision, but a large investment of resources to implement. Most healthcare providers are too constrained by their current financial situation to take such a risk. Goodman suggests a way for Medicare to reward innovations that will lead lower costs in the long run. It may be a start, but the whole idea of a government agency managing innovation strikes me as equally perverse and ultimately futile.

Terminology vs Knowledge

In our perpetual climb up the value chain beginning with the bits and bytes of data, then to recognizable information and finally knowledge, terminology fits right in the middle. Terms, or words, are recognizable slices of information that may be organized into a hierarchy or system. In fields such as medicine or biology with long histories of terminology development, many different systems of terms have cropped up due to the various needs of sub-groups of users and their isolation from each other. As a result, there is much duplication, overlap and confusion in terminology use even within the knowledge domain.

The need to interface between various networks of healthcare payers and providers has driven the demand for organization in this confusion, resulting in systems such as Current Procedural Terminology (CPT), SNOMED Clinical Terms, and International Classification of Diseases all developed to assist healthcare providers in finding a common vocabulary to describe the services they provide.

Selecting a standard terminology provides a common framework and is a significant step forward, but these are merely words; text strings in a matrix of other text strings without the ability to transfer significant quantities of knowledge. Humans naturally make the connection between a term and the object it represents and in fact expect that term to be imbued with the requisite information, but such is not the case with computers. To a computer "Joe Smith" is just a nine character text string.

Object Oriented Programming (OOP) sought to change all that, providing the knowledge underpinnings to turn simple terms into full-fledged objects that behave as do the terms they represent, or at least to the degree necessary to fulfill the requirements of the software. And there you have it: what exactly are the requirements? For a terminology management system it is to standardize and organize the diverse terminologies and there it stops. Most knowledge management systems aspire to loftier goals such as supporting decision making processes.

And thus we have a grey area: terminology management systems that aspire to be knowledge management systems. Or users who want them to be. A successful terminology management system is one which includes and classifies as many terms in the domain as possible whereas a successful knowledge management system includes as many functions in the domain as possible.

Let's take for example an anatomical terminology management system. If this includes a complete catalog of anatomical parts, including synonyms and locations, this will fulfill the requirements of users who wish to know what term to use and in what context. However, even if we know that the femur is attached to the hip and is in the leg, the terminology management system may not indicate that it is a bone or that it could suffer a fracture. This is the type of information that would be contained within a knowledge management system.

Perhaps the most effective knowledge management systems would be ones that incorporate a terminology catalog at the front end; a place where every term could be found and then once found could capture the information needed to model that object. As terminology management systems finally catch on and fulfill the needs of the industry, knowledge management systems will not be far behind.