 |
|
Upcoming Events
September 9th 2010 07:30 AM
- Successful Office Kaizen Events
September 9th 2010 08:00 AM
- Sup 1: Fundamentals of Leadership for Manufacturers - St Paul
September 9th 2010 09:00 AM
- Sup 2: The Next Stage for Improving Performance for Manufacturers-Plymouth
September 9th 2010 01:00 PM
- Sup 1: Fundamentals of Leadership for Manufacturing - Plymouth
September 14th 2010 08:00 AM
- Results-Driven Sales Incentives
September 14th 2010 08:00 AM
- Fraud and Misconduct at Investigator Sites
September 14th 2010 01:00 PM
- Failure Mode and Effects Analysis (FMEA)
September 21st 2010 08:00 AM
- Project Monitoring- The Earned Value Method
September 21st 2010 08:00 AM
- Project Monitoring: The Earned Value Method & Project Management
September 21st 2010 01:00 PM
- Project Management for Product Development
Article Index
Medical Device CRO or Independent Consultant: How Do You Choose? Author Unknown The evolution of the clinical research department within the hallowed halls of medical device companies has been difficult and costly. In the early nineties, the addition of one to two technical experts to assist with the collection of data on the classic 150 - 200 patients on an Excel spreadsheet was considered exorbitant. Today, not only have clinical departments expanded, they are being fortified with physician medical directors, statisticians, preclinical and laboratory technicians. If you aren't a big manufacturer, though, where do you turn to support your trial? Your choices are varied and complex. National or local, large or small, pharmaceutical or device specific, or all independents? |
||
|
Can A Statistician Build A Better Mouse Trap - Or a Better Clinical Trial Article by: Chris Mullin Recently, interest has surged in new statistical methods that potentially reduce the time needed for product development and regulatory approval. Bayesian statistics and adaptive, or flexible-design, clinical trials are among the most discussed of these new methods. |
||
|
Risk Management, Regulation, and Relevant Steps: Thoughts on Optimizing Process Author Unknown Risk Management has received increased attention from regulatory agencies worldwide since publication of ISO 14971, "Medical devices - Risk Management - Application of Risk Management to Medical Devices." The U.S. Quality System Regulation only mentions "risk analysis" briefly in 820.30 (g) Design Validation. William Midgette, however, of the Office of Science and Technology at the Center for Devices and Radiological Health, has stated the regulation's preamble clarifies that "risk analysis" includes hazard identification, risk estimation, risk evaluation, risk control and post-production information (2/04 AdvaMed Risk Management Seminar). |
||
ADVERTISEMENT
|
||
|
Medical Device CRO or Independent Consultant: How Do You Choose? |
||
|
The evolution of the clinical research department within the hallowed halls of medical device companies has been difficult and costly. In the early nineties, the addition of one to two technical experts to assist with the collection of data on the classic 150 - 200 patients on an Excel spreadsheet was considered exorbitant. Today, not only have clinical departments expanded, they are being fortified with physician medical directors, statisticians, preclinical and laboratory technicians. If you aren't a big manufacturer, though, where do you turn to support your trial? Your choices are varied and complex. National or local, large or small, pharmaceutical or device specific, or all independents? Here's a top ten list of questions to help you make optimum choices: The challenge for many in the industry has been finding the CRO that allows the order-off-the-menu option: monitoring only, statistics when we need them, or regulatory strategy and FDA support ad hoc. However, the cost to provide these services on an as-needed basis has not allowed for any one organization to stand out as a leader. Many organizations flourish within the monitoring segment and have little or no support structure for the regulatory or statistical need. If you need to have all your investigational sites monitored and data cleaned, your core need is monitoring. Hence, the CRO with device-specific monitoring experience is your core criteria. Your most costly piece of outsourcing should be sourced from the CRO/consultant focused at that core. Core Competency Assessment How do you assess a CRO for core competence? After you have a shortlist of device-specific CRO support organizations, ask what portion of their business is Class 2 or Class 3 device specific. Consider these questions:
Support Staff Within all our organizations, we are only as good as our supports staff's ability to run interference competently and graciously with our enrolling investigational sites. This postulate holds true for the CRO and independent consulting staff you hire. If a site needs something quickly, your first line of defense may be that front desk, receptionist, or coordinator. His or her ability to calmly probe for as much detail about the site's needs, can make or break your relationship with the site. Know how the employees are trained, their supervisors, and how they will behave when confronted with "your" customer in stressful moments. Staffing Levels A common issue that arises is how many staff you expect to be dedicated to your project. Recall your motivation is to task the CRO or consultant with tasks you and your organization cannot accomplish in parallel. Will they pull personnel from other projects to meet your "event" schedules, or hire additional support? Most organizations pull support personnel to the "event" from other projects but unless you understand their billing level, you may end up with large overruns or out of scopes due to the number of people needed to meet the timetable. Expansion Services Both the CRO and the independent consultant have areas of coverage that aren't their core. When you plan to use any outsource organization, recall the adage of beginning with the end in mind, which has key strategic meaning. When you start, the need may be only one or two key items that can easily grow into additional tasking. Provisions for that expansion in human and budgetary resources need to be considered at the beginning. Check References It's also critical to check references. These organizations should provide you with a minimum of three good references. Be certain to obtain a reference that utilized the specific service for which you are contracting. Inquire about estimates and the organization's ability to come in on budget. Last, but certainly not least, when your executive team tells you the choice is fast, cheap, or right, pick the third choice–right. It is clearly the optimal choice in the long run.
|
||
ADVERTISEMENT
|
||
|
Can A Statistician Build A Better Mouse Trap - Or a Better Clinical Trial |
||
|
Recently, interest has surged in new statistical methods that potentially reduce the time needed for product development and regulatory approval. Bayesian statistics and adaptive, or flexible-design, clinical trials are among the most discussed of these new methods. Recently, FDA has issued a draft guidance document on the use of Bayesian statistics. Last July FDA held a conference on adaptive trial designs. Each has pros and cons to carefully weigh. Bayesian statistical methods are based on a simple mathematical formula derived by the British Reverend Thomas Bayes over 200 years ago. Methods using Bayes' theorem allow one to integrate past data within a current experiment. Incorporating prior information may lead to a shorter time to market, if a large and pertinent body of previous information is already available. For example, FDA draft guidance document references a breast cancer screening technology (T-Scan 2000, TranScan Medical, Inc.), in which Bayesian methods integrated data from several studies. This allowed the sponsor to demonstrate effectiveness with a smaller sample size. Bayesian methods have been successfully used in other therapeutic areas, including orthopedics, urology, and reproductive health. The "Summary of Safety and Efficacy" for these products is on FDA's website. A major difficulty in Bayesian methods is justifying prior information used. A convincing argument that the prior information is appropriate to combine with new data must be made. Another challenge is that the amount of statistical work required to design a sound Bayesian clinical trial may be far greater than that what is needed for a more traditional experimental design. Another statistical innovation receiving attention is the adaptive clinical trial. As opposed to the classical clinical trial, which has a pre-defined endpoint, treatment, population to be studied, and sample size, adaptive designs allow for the modification of one or more of these aspects during an ongoing trial. Both Bayesian and non-Bayesian versions of adaptive designs exist. Several methods exist for re-assessing the needed sample size based on initial observations in a clinical trial. This allows for an increase in the sample size if the initial assumptions of the trial were incorrect. These methods are useful when little to no information is available on what to expect in a trial prior to its start. The advantages of such flexibility are obvious. The downside is the logistical difficulty. If changes to a clinical trial are made mid-course, they can provide early indications of the study's results and bias behavior of investigators or patients. These designs require considerable pre-planning and pre-specification of what will be changed according to what is learned as the trial progresses. This may reduce the amount of flexibility in a trial design and thus questions the purpose for using such a design in the first place. Trial Design Method Pros, Cons, and When to Use  The figure below displays the advantage of a Bayesian analysis. The Bayesian calculation produces a more precise estimate of the treatment success rate in the form of the Bayesian posterior distribution. The solid black line above illustrates this. More precise estimates of treatment benefit allow smaller sample sizes to produce statistically significant results from clinical trials. Illustration of Bayesian Analysis  To summarize, there are many factors at play when designing a trial for your company. It is important to consider the pros and cons of each, and make your decisions accordingly.
|
||
ADVERTISEMENT
|
||
|
Risk Management, Regulation, and Relevant Steps: Thoughts on Optimizing Process |
||
|
Risk Management has received increased attention from regulatory agencies worldwide since publication of ISO 14971, "Medical devices - Risk Management - Application of Risk Management to Medical Devices." The U.S. Quality System Regulation only mentions "risk analysis" briefly in 820.30 (g) Design Validation. William Midgette, however, of the Office of Science and Technology at the Center for Devices and Radiological Health, has stated the regulation's preamble clarifies that "risk analysis" includes hazard identification, risk estimation, risk evaluation, risk control and post-production information (2/04 AdvaMed Risk Management Seminar). Many medical device companies have begun to implement risk management procedures, forms, and reports into their quality system. Implementation articles abound, but most contain the basic flow diagram of risk identification, analysis, evaluation, reduction, acceptance, communication, and few concrete examples. The Global Harmonization Task Force document, "Implementation of Risk Management Principles and Activities within a Quality Management System," or the GHTF Study Group 3, 5/05, online at http://www.ghtf.org/sg3/inventorysg3/sg3n15r82005.pdf, provides good practical examples and useful flow diagrams, including risks associated with manufacturing and other quality system processes. Another useful source is FDA's guidance document, Medical Device Use-Safety: "Incorporating Human Factors Engineering into Risk Management," online at: http://www.fda.gov/cdrh/humfac/1497.pdf. Even after understanding "what" should be done, attention is warranted on the process itself. Important process factors include the following three steps, named the "who," "when," and "how well" factors. WHO Are people with appropriate skills involved in hazard identification and later risk-management tasks? The process benefits from a cross-functional approach. This includes participants close to the user/customer, those from the clinical and marketing teams, as well as participants familiar with the technical aspects of the process. At design reviews, independent reviewers should be selected by ability to provide an unbiased perspective to examination of residual risk. WHEN Is hazard identification executed early enough in the process? Hazard identification is best executed before design concepts start to gel. This may seem obvious with novel devices, but many device projects involve components that have a track record. When a new intended use for an existing or assembled device is being evaluated, strong expectations for fast development may arise. Early review minimizes false expectations and validation surprises. Risks should then be re-evaluated periodically using inputs including testing, failure investigations, complaints, customer feedback. HOW WELL How well do you understand your device's intended use(s) and users? What are your information sources? Independent, in-depth research into an intended use provides insight into the probable risk profile. Relevant information is obtainable from physicians, published literature, standards, and regulatory guidance documents. Comparisons to competitor product are not necessarily sufficient. Without direct evidence, such as clinical data, benchmarking is banking on the quality of the competitor's risk management process. Even with well-established competitor product, independent understanding of the intended use allows focus on the most important design factors. Regulatory expectations of risk management continue to evolve. Careful planning and the following of the suggested process, and periodic examination and fine-tuning of a company's risk management process, may lead to a more effective regulatory process.
|
||
Chris Mullin is a Principal Statistical Consultant with The Integra Group, providing statistical expertise to the medical device and biotechnology industry. Contact him at 651 - 270 6442.