|Protocol Design Workshop --> A Business Approach|
10 December 08 -- hurry to register!
Start with the Claims
It helps to think in terms of the natural order of progression from want-to-have claims to substantialted claims.
Think like a marketer and write statements about your device that are persuasive and intended to make the physician want to use your product. Speak to the competition's weaknesses and how your device is better, faster, more accurate, easier to use, diagnostic of ..., an adjunct to..., or superior to....
If you're in regulatory, think in terms of intended use and indications for use. An intended use is a broad statement of purpose--e.g., the device is intended to maintain an open airway; the indications for use are specific and mention the disease, condition, or population in which you want to use the device.
Some companies write pages and pages of want-to-have claims. Fair enough, but you'll need to settle on one or two in order to focus your efforts.
In this workshop, we'll look at the relationship between claims, objectives, hypotheses, endpoints, methods, and sample sizes.
Item 1: Objectives
Think of objectives as a broad statement of business purpose. The ultimate purpose, of course, is to sell more devices and make money. The objectives are a polite way of expressing your ulterior motives for doing the study.
For example: you might want to conduct a pivotal study to support a 510(k) or PMA submission, demonstrate substantial equivalence, support marketing claims,
establish superiority to a competitor, give the physician an alternative, establish manufacturability, investigate consumer preference, market positioning (should you sell, lease, or rent?) or market position, acquire data for evidence-based medicine and reimbursement, conduct post-approval or registry studies, or to conduct confirmatory studies (i.e., US studies to 'confirm' OUS data.)
In this workshop, we'll look at common objectives for conducting a study, and review other sections that address the rationale, justification for the study, a contention of non-significant risk (if appropriate), and other elements of the protocol.
Item 2: Hypotheses
1) "allows physicians to extend pH collection to 48 hours -- 24 hours beyond the recording capability of conventional catheter systems."
The related testable hypothese may have been:
1a) "pH values can be collected for 48 hours, and 1b) competitor cannot collect data beyond 24 hours.
You should choose one primary hypothesis. This is the hypothesis on which you'll base sample size calcualtions and product success or failure.
You can have as many secondary hypotheses as you want, so long as they don't require different patient populations, treatments, or other design elements.
Item 3: Endpoints and Methods
A primary endpoint will correspond to your primary hypothesis, secondary endpoints will correspond to your secondary hypotheses.
Using the three products I randomaly selected from the web, the endpoints might have been:
1a) plots of continuous pH values in the gastrointestinal tract establish that readings are transmitted for 48 hours or longer, 1b) transmitters recovered naturally from the body are still able to accurately transmit pH values of known solutions.
2a) Audiometry tests show a 5% improvement in frequency perception over low ranges (from 150 to 1000 Hz) and a 15% improvement over high ranges (from 1000-8000 Hz) with our hearing aid over every other hearing aid on the market," 2b) "Patient diaries show at least 20% greater satisfaction with 'natural perception of sound' and 15% less annoyance with impact noises.
3a) "Photos of teeth with stained plaque after two-weeks brushing with Sonicare show a 50% decrease in staining measured by color density (number of pixels) and saturation (depth of hue) compared to photos of teeth after regular tooth brushing.
Notice that the endpoints must be measurable--a validated test must exist, are preferably objective-- although subjective data such as patient diaries may be necessary, and must be definitive--we must know for certain when/if the endpoint has been reached.
In this workshop, you'll practice writing hypotheses, endponts, and test methods.
Item 4: Sample Size
You can't just guess at the performance of the investigational and comparator devices: some type of prior testing or literature values are required for these calculations.
If the sample size is too large, work with the statistician to reword the hypotheses, or consider a different primary hypothesis altogether.
You should calculate a separate sample size for every hypothesis you can think of, then--if you are a start-up, go for the cheapest claims first; if you are an established firm, go for the claim with the most market value.
In this workshop, you'll learn some simple formulas and practice sample size calculations.
Item 5: Data Collection and Analysis
Your statician will use the data you've collected to substantiate or refute the hypothesis, hopefully allowing the you to meet your business objectives and make the marketing claims you need.
Poll du jour 7: What is the most commonly used database application for receiving data? Take the poll and find out.
Finally, we'll review the administrative sections of good protocols, such as responsibilities, monitoring plans, statistical analysis plans, glossaries, and more.
|You will receive:|
[x] ONE learner access per registration.
[x] A four-hour lecture.
[x] PowerPoint slides in pdf.
[x] Case studies and exercises.
[x] An example protocol.
[x] Two weeks to take the one-hour follow-up quiz.
[x] Two weeks to give course feed-back.
[x] An expert instructor, Nancy J Stark, PhD.
[x] 0.5 CEUs and Certificate of Attendance.
NOTE: I have no prior knowledge or information about the examples given in this promotion; aside from the information taken from the web, the information is purely conjecture. NJ Stark