PAT in Perspective: Estimate PAT’s Real Costs

If you base the purchase of any instrument on cost, and the project takes longer or, worse yet, fails because the instrument is inadequate, you could easily waste $1,000,000. Photo courtesy of Dichroma.com.

In light of a PAT or even a simple analytical method development project, that hourly figure takes on new meaning. Under the old paradigm of analytical method development, the cost of instrument we purchased, for example an HPLC, was assumed to be amortized over numerous projects. In order to keep within a (compartmentalized) budget, the best price was considered before ordering. When you add the 15th or 20th LC to your lab, the true costs of development are blurred.However, when you begin a stand-alone PAT program, costs are amplified in the eyes of managers, who often assume that the first PAT project will be a single-bullet type. That is, we choose a process and proceed to optimize it. The analytical instrument, e.g., a NIR spectrometer, will be used (it is assumed) for this one job. We proceed to apply the lab-based paradigm of cost effectiveness and look for a bargain. This is the “penny-wise” approach. What we do not account for in this “price” are the other costs. I will use as for an example a project I worked on to assay tablets via NIR transmission. After we purchased the analyzer for “X” dollars, we began the project.The lab team consisted of a group leader, a lead chemist, and a “second” chemist, and an outside consultant (not me). After a month or so, the team decided that it needed specialty tablets of a larger range; this involved two people from Product Development and their equipment. We needed to analyze the tablets using a reference HPLC method. This involved the “second” chemist and another chemist from a second group (with his robotic system). As the work proceeded, we involved the Quality Assurance department with the validation report; an investigator and a technical writer were also assigned to the project.But that doesn’t count the input from the Information Technology group and the time spent helping with our software validation. The assistance from the instrument company was gratis. Our director didn’t work for free, either; he needed to oversee all our progress. Had we been doing an actual at-line or on-line method, Production personnel would have been involved along with people from the engineering department. Of course, the purchasing department was a big help, too. So, the project took well over a year from concept to final approval from QA and Management.Now, let’s look at the true cost of developing a simple NIR assay method. Since we all worked on other projects and “put out fires” throughout the year, we cannot count 100% of the time, but I did some conservative costing for the people involved. The total number of people who “touched the ball” included:

• Three from the Methods Group
• Two from Product Development
• Two from Quality Assurance
• One from Analytical
• One from Information Technology
• One manager
• One consultant
• One from Purchasing
• Freebies from the instrument company

We must have spent nearly one million dollars on the method and its validation, and that’s a conservative estimate.Now, let’s examine the cost of the instrument used. We went with one that cost $60,000. Until this analysis, that would have sounded expensive, but it is only 6% of the total cost! What is my point? If you base the purchase of any instrument on cost, and the project takes longer or, worse yet, fails because the instrument is inadequate, you have wasted $1,000,000! In a process method, where you add in engineering and production personnel costs, the instrument cost becomes even less important.In short, DO NOT BUY EQUIPMENT SIMPLY BECAUSE IT IS CHEAPER! In any analysis, lab or process, the monitor will be the least expensive part of the project. If you delay the NDA because of an inferior instrument, well, just consider that a blockbuster drug can deliver between $5-6 million per manufacturing day to the producer. Would you want to weigh a month or more delay of such a product against saving $10,000-20,000?