Audits are a critical component of quality systems, but are they guided by formal assessments of risk to your products? In this world of ICH Q9, can you offer even a semi-quantitative justification for your audit priorities? We have spoken to many people in the industry, and almost all mention a risk assessment being undertaken prior to an audit. But we have not found many people that formalize that risk assessment, or keep it updated from audit to audit. Even fewer communicate their scoring of risk to either their internal clients or the vendor that has been audited.
A new trend in auditing is to use a form of risk assessment both before and after the audit. A popular form is the Failure Modes and Effects Assessment, or FMEA (see, for example, http://www.sre.org/pubs/Mil-Std-1629A.pdf). In a traditional FMEA, risks of failure are identified in a detailed fashion, and scored in three categories related to the failure’s probability, detectability and severity. Scoring is done on a semi-quantitative or relative basis using an arbitrary scale such as 1-10. For an audit, you might use the same categories as they relate to a particular vendor's (or department's) ability to deliver a product or service, failure free. You could organize your FMEA according to the critical quality attributes of the product or service being delivered or according to a list of requirements from a guideline or the CFR's. Your FMEA should receive input from affected departments, and should be used for prioritization of audit items. You should have the FMEA in mind as you conduct your audit, and remember why various items received high prioritization. You may change ratings for probability or detectability based on what you observe. If instead, you confirm your evaluation, you should probe remediations that decrease your firm's primary concern. A remediation that addresses detectability, when the issue was probability, likely won’t mitigate the risk of failure.
When you return from your audit, rescore the FMEA with assessments based on your observations and data that you collected. Make sure that you share your analysis with the stakeholders. And monitor the performance of the vendor until the next audit; the data will help inform your next FMEA.
Do you already use FMEA's in audit preparation and reporting? Let us know your practices.
Tuesday, August 4, 2009
Tuesday, June 23, 2009
Why is Quality by Design so heavy with statistics?
Why is the literature on Quality by Design so laden with statistics and experimental design space jargon? After all, the definition of the term “design” doesn't seem to include the analysis of messy data leading to rough correlations with results that are valid only over a limited range. So what gives?
The idea behind QbD was to use mathematical, predictive models to predict process outcomes. This concept can be applied directly to simple unit operations, such as drying, distilling, heating and cooling. However, unlike in the petrochemical business, the thermodynamic properties of most active pharmaceutical ingredients are not known and are difficult to measure. The unit operations used to manufacture common biotechnology products, such as cell culture, chromatography and fermentation have been modeled, but the models are very sensitive to unknown or unmeasureable adjustable parameters. The batch nature of these operations also makes their control difficult, as classical control theory relies on the measurement of an output to make an adjustment to an input to correct the output back towards the design specification.
Since there does not appear to be a clear path to using models, an approach has been chosen that emphasizes getting as much phenomenological information from as few experiments as possible. This is the Design of Experiments approach, where input conditions or operating parameters are systematically varied over a range and the process outputs measured, with statistics used to deconvolute the results. The combined ranges tested become the “design space”, and the process performance outputs with the variations closest to the process failure limit become the critical performance parameters. The results are useful, but only within the design space, and only with the certainty that the statistics report. Also, since the results are phenomenological, the effect of scale is often unknown.
The statistical approach is acceptable, and for the immediate future it's probably the best that we can expect. But the focus on this approach seems to drown out the more pressing need for good process models and physical properties data. These are the elements that allowed the petrochemical and commodity chemicals industries to scale up processes with assurance that quality specifications would be met. There are countless models available for bioprocessing's more complicated unit operations, but they have parameters that we don't know and can't calculate from first principles. There is no question that we need to find ways to collect this data, and a commitment to publish or share it. There are also simpler unit operations that we can model, scale up and scale down with complete assurance. These include operations such as mixing and storing solutions, filtration, diafiltration, centrifugation and some reactions. We shouldn't let the more complicated operations that still require statistical DOE approaches prevent us from applying the true principles of QbD to our simpler unit operations.
The idea behind QbD was to use mathematical, predictive models to predict process outcomes. This concept can be applied directly to simple unit operations, such as drying, distilling, heating and cooling. However, unlike in the petrochemical business, the thermodynamic properties of most active pharmaceutical ingredients are not known and are difficult to measure. The unit operations used to manufacture common biotechnology products, such as cell culture, chromatography and fermentation have been modeled, but the models are very sensitive to unknown or unmeasureable adjustable parameters. The batch nature of these operations also makes their control difficult, as classical control theory relies on the measurement of an output to make an adjustment to an input to correct the output back towards the design specification.
Since there does not appear to be a clear path to using models, an approach has been chosen that emphasizes getting as much phenomenological information from as few experiments as possible. This is the Design of Experiments approach, where input conditions or operating parameters are systematically varied over a range and the process outputs measured, with statistics used to deconvolute the results. The combined ranges tested become the “design space”, and the process performance outputs with the variations closest to the process failure limit become the critical performance parameters. The results are useful, but only within the design space, and only with the certainty that the statistics report. Also, since the results are phenomenological, the effect of scale is often unknown.
The statistical approach is acceptable, and for the immediate future it's probably the best that we can expect. But the focus on this approach seems to drown out the more pressing need for good process models and physical properties data. These are the elements that allowed the petrochemical and commodity chemicals industries to scale up processes with assurance that quality specifications would be met. There are countless models available for bioprocessing's more complicated unit operations, but they have parameters that we don't know and can't calculate from first principles. There is no question that we need to find ways to collect this data, and a commitment to publish or share it. There are also simpler unit operations that we can model, scale up and scale down with complete assurance. These include operations such as mixing and storing solutions, filtration, diafiltration, centrifugation and some reactions. We shouldn't let the more complicated operations that still require statistical DOE approaches prevent us from applying the true principles of QbD to our simpler unit operations.
Wednesday, June 3, 2009
Ten Steps to Choosing your Contract Manufacturer
For many young companies, choosing a contract manufacturer, or CMO, for their lead pharmaceutical candidate is critical. Choose the wrong contractor, and you could be faced with delays and cost overruns with which your investors and patients won’t be very sympathetic. While there is no guarantee that you will always make the right decision, here are some tips that can help you make your choice in an organized, thoughtful, meaningful and objective way:
1) Make a list of all the possible suppliers. Such lists may be purchased, but they are also easily assembled from internet searches. In fact, you can do a little pre-screening with your own internet search.
2) Screen potential suppliers with a phone call. You will probably speak with a business development or sales person representing the contractor, but usually these people are quite knowledgeable about their company’s capabilities, and common problems encountered in the industry. We recommend you not reveal too many details about your project, and be prepared mostly to listen. However, you should have three or four key capabilities or proficiencies that you are looking for in all of the potential vendors. If possible, try to rule out vendors who do not meet your “must-have” requirements at this stage. Stay tuned for a blog on how to establish your showstopper list, it’s a critical exercise, and may extend beyond key capabilities and proficiencies.
3) Keep a matrix, and record the date you first contacted the vendor, when they responded to you, the status of any confidentiality agreements, and all the contact information that you can gather (email addresses, cell phone numbers, main switchboards and extension numbers). Also note the responses that each contact had relative to your three or four show-stopper criteria.
4) Meet with your team, and select three to five potential vendors to request a proposal from. We don’t recommend more than five: getting good, comparable proposals is a lot of work, like 2n times the work, where n = the number of proposers. Not to mention the work that contract manufacturers go through to read your RFP and prepare a proposal. Your three or four showstopper criteria should help you limit the number of proposers; if necessary you can begin to narrow down based on “nice-to-have” criteria as well. You may also deliberately choose to look at a range of vendors that represent different strengths/weaknesses (for example, do you prefer a “one-stop shop” that is convenient, but maybe not the best at everything, or a “specialty” vendor that provides higher levels of expertise, but will require you to select and manage multiple vendors?).
5) Solicit proposals. Most contract manufacturing seekers have a Request for Proposals (RFP) process that includes a document. These RFP documents vary from one page requirements descriptions, to lengthy, legalistic documents that require a team, and a month, to respond to. You should do what is comfortable for your organization. There needs to be enough information so that the vendor is able to respond with a meaningful proposal. There is some legal danger, particularly with intellectual property, so it’s not a bad idea to get your RFP reviewed by your legal counsel. And you should only send an RFP to a vendor after they have signed a mutually agreed confidentiality agreement.
6) Score your proposals. Find some basis to make apples to apples comparisons. RMC uses a modified Kepner Tregoe analysis, but many forms of analysis will work. You should have determined how you will assess and weight qualitative data before you begin. And in doing so, you should not under-estimate intangible factors: the ability to communicate, time zone differences, good references (you’ve checked, right?) are some examples. At this stage you should be ranking and scoring on “nice-to-have” criteria as well as comparing cost/timeline, capabilities, capacity, and viability of the business. You might form your opinion of the viability of the business by reading annual reports and press releases, and by assessing how busy the manufacturing area and support labs look during your visit.
7) Visit the top two or three vendors in person. Vendors may not allow a formal quality audit prior to signing a contract, but be sure to bring your quality representatives even for an informal “technical visit”. If the project is large, you may take the resistance to a pre-use quality audit as a red flag. Again, spend as much time as possible listening, rather than talking. Get a tour, and copies of all presentations. Ensure that you have a meeting between the decision makers for both sides as well as aligning discussions between key technical and quality personnel. If there are disputes or further work to do, your decision makers must have a good working relationship.
Your visit is also your best opportunity to break past the business development group and take a true measure of the business. Chat with the people in the lab or production area if you can. Look over the state of the equipment, the cleanliness of the facility and the stock in the warehouse. Check their flexibility-- what can they make happen for you, vs. what will have to be run past someone in another building, or another city? Ultimately, you should think about hiring a contract manufacturer similarly to how you hire an employee, by hiring for expertise as well as fit with your team.
8) Consider entering contract negotiations with at least two vendors. Things can go wrong in negotiations, and your position is stronger if you can legitimately walk away. We typically don’t let any of the final three candidates off the hook until the ink is dry on the final contract. If your budget can justify it, having a second contractor performing development work and verifying the primary contractor’s results is an excellent idea. It may ultimately spread your risk in supply chain, and give you leverage in negotiating commercial supply agreements later on.
9) Revisit your analysis tool. You may learn new things in your contract negotiations that cause you to adjust your evaluation. Don’t be afraid to be frank if you feel like terms have been changed since the selection was made. This is a good reason to keep a back up vendor.
10) You must now manage the project according to the criteria by which you selected your supplier. Hold your supplier and yourself accountable to these criteria. For example, after selecting a vendor because they can meet a very aggressive timeline, do not put the project timeline at risk by failing to order back up critical path supplies, in case the primary order doesn’t arrive, or fails to meet specifications on arrival. We will have more to say about managing a contract manufacturer in a future blog.
Choosing the right manufacturing partner is critical for your success as a drug developer. Spend the time required to make a good decision. This time should be spent gleaning as much tangible and non-tangible information on all your options as possible, and then objectively comparing it. You should have an idea about how you’re going to evaluate and weight non-tangible factors into your decision. And once you have made the choice, manage according to your criteria. Although everyone has their own methods for vendor selection, these are some suggestions that have worked well for us and our clients. If you have questions or comments, please visit http://www.rmcpharma.com/ or email us at info@rmcpharma.com.
1) Make a list of all the possible suppliers. Such lists may be purchased, but they are also easily assembled from internet searches. In fact, you can do a little pre-screening with your own internet search.
2) Screen potential suppliers with a phone call. You will probably speak with a business development or sales person representing the contractor, but usually these people are quite knowledgeable about their company’s capabilities, and common problems encountered in the industry. We recommend you not reveal too many details about your project, and be prepared mostly to listen. However, you should have three or four key capabilities or proficiencies that you are looking for in all of the potential vendors. If possible, try to rule out vendors who do not meet your “must-have” requirements at this stage. Stay tuned for a blog on how to establish your showstopper list, it’s a critical exercise, and may extend beyond key capabilities and proficiencies.
3) Keep a matrix, and record the date you first contacted the vendor, when they responded to you, the status of any confidentiality agreements, and all the contact information that you can gather (email addresses, cell phone numbers, main switchboards and extension numbers). Also note the responses that each contact had relative to your three or four show-stopper criteria.
4) Meet with your team, and select three to five potential vendors to request a proposal from. We don’t recommend more than five: getting good, comparable proposals is a lot of work, like 2n times the work, where n = the number of proposers. Not to mention the work that contract manufacturers go through to read your RFP and prepare a proposal. Your three or four showstopper criteria should help you limit the number of proposers; if necessary you can begin to narrow down based on “nice-to-have” criteria as well. You may also deliberately choose to look at a range of vendors that represent different strengths/weaknesses (for example, do you prefer a “one-stop shop” that is convenient, but maybe not the best at everything, or a “specialty” vendor that provides higher levels of expertise, but will require you to select and manage multiple vendors?).
5) Solicit proposals. Most contract manufacturing seekers have a Request for Proposals (RFP) process that includes a document. These RFP documents vary from one page requirements descriptions, to lengthy, legalistic documents that require a team, and a month, to respond to. You should do what is comfortable for your organization. There needs to be enough information so that the vendor is able to respond with a meaningful proposal. There is some legal danger, particularly with intellectual property, so it’s not a bad idea to get your RFP reviewed by your legal counsel. And you should only send an RFP to a vendor after they have signed a mutually agreed confidentiality agreement.
6) Score your proposals. Find some basis to make apples to apples comparisons. RMC uses a modified Kepner Tregoe analysis, but many forms of analysis will work. You should have determined how you will assess and weight qualitative data before you begin. And in doing so, you should not under-estimate intangible factors: the ability to communicate, time zone differences, good references (you’ve checked, right?) are some examples. At this stage you should be ranking and scoring on “nice-to-have” criteria as well as comparing cost/timeline, capabilities, capacity, and viability of the business. You might form your opinion of the viability of the business by reading annual reports and press releases, and by assessing how busy the manufacturing area and support labs look during your visit.
7) Visit the top two or three vendors in person. Vendors may not allow a formal quality audit prior to signing a contract, but be sure to bring your quality representatives even for an informal “technical visit”. If the project is large, you may take the resistance to a pre-use quality audit as a red flag. Again, spend as much time as possible listening, rather than talking. Get a tour, and copies of all presentations. Ensure that you have a meeting between the decision makers for both sides as well as aligning discussions between key technical and quality personnel. If there are disputes or further work to do, your decision makers must have a good working relationship.
Your visit is also your best opportunity to break past the business development group and take a true measure of the business. Chat with the people in the lab or production area if you can. Look over the state of the equipment, the cleanliness of the facility and the stock in the warehouse. Check their flexibility-- what can they make happen for you, vs. what will have to be run past someone in another building, or another city? Ultimately, you should think about hiring a contract manufacturer similarly to how you hire an employee, by hiring for expertise as well as fit with your team.
8) Consider entering contract negotiations with at least two vendors. Things can go wrong in negotiations, and your position is stronger if you can legitimately walk away. We typically don’t let any of the final three candidates off the hook until the ink is dry on the final contract. If your budget can justify it, having a second contractor performing development work and verifying the primary contractor’s results is an excellent idea. It may ultimately spread your risk in supply chain, and give you leverage in negotiating commercial supply agreements later on.
9) Revisit your analysis tool. You may learn new things in your contract negotiations that cause you to adjust your evaluation. Don’t be afraid to be frank if you feel like terms have been changed since the selection was made. This is a good reason to keep a back up vendor.
10) You must now manage the project according to the criteria by which you selected your supplier. Hold your supplier and yourself accountable to these criteria. For example, after selecting a vendor because they can meet a very aggressive timeline, do not put the project timeline at risk by failing to order back up critical path supplies, in case the primary order doesn’t arrive, or fails to meet specifications on arrival. We will have more to say about managing a contract manufacturer in a future blog.
Choosing the right manufacturing partner is critical for your success as a drug developer. Spend the time required to make a good decision. This time should be spent gleaning as much tangible and non-tangible information on all your options as possible, and then objectively comparing it. You should have an idea about how you’re going to evaluate and weight non-tangible factors into your decision. And once you have made the choice, manage according to your criteria. Although everyone has their own methods for vendor selection, these are some suggestions that have worked well for us and our clients. If you have questions or comments, please visit http://www.rmcpharma.com/ or email us at info@rmcpharma.com.
Thursday, May 21, 2009
How do you establish a reference standard for release of your first GMP lot?
There are some different approaches depending on situation, but here is one general approach that many people use:
* Preclinical material that was used to support IND-enabling tox studies is characterized and used as an interim reference standard. This doesn't have to be manufactured by GMP; could even be research grade material. Characterization tests should be performed and include methods to determine content, primary structure, bioactivity, tentative release tests, etc. Important to have a good handle on content from this standard (i.e., for a protein drug you might look at total protein content by elemental analysis.) Potency units could be defined here as well.
* Use this interim reference standard as a comparator to release your initial GMP material.
* Use material from the first GMP campaign to create a primary reference standard. Again, this material will need to be subjected to a panel of characterization tests in addition to the release tests (typically at least or more in-depth than the characterization done on initial reference standard).
* The primary reference standard supercedes the interim reference standard after it is created and qualified. This typically happens during or immediately after the first GMP campaign.
From an assay validation standpoint, you will always have to start with some initial standard (typically, the interim reference standard, which is all you may have available before GMP manufacture), and you can get into some circularity with accuracy (i.e., assaying the standard against itself). So for your initial standard and assay validation you rely somewhat on orthogonal methods of analysis. Also, as per ICH Q2, accuracy can be inferred from precision, linearity and specificity.
By Lori Nixon, RMC Pharmaceutical Solutions, Inc.
Find out more at www.rmcpharma.com
* Preclinical material that was used to support IND-enabling tox studies is characterized and used as an interim reference standard. This doesn't have to be manufactured by GMP; could even be research grade material. Characterization tests should be performed and include methods to determine content, primary structure, bioactivity, tentative release tests, etc. Important to have a good handle on content from this standard (i.e., for a protein drug you might look at total protein content by elemental analysis.) Potency units could be defined here as well.
* Use this interim reference standard as a comparator to release your initial GMP material.
* Use material from the first GMP campaign to create a primary reference standard. Again, this material will need to be subjected to a panel of characterization tests in addition to the release tests (typically at least or more in-depth than the characterization done on initial reference standard).
* The primary reference standard supercedes the interim reference standard after it is created and qualified. This typically happens during or immediately after the first GMP campaign.
From an assay validation standpoint, you will always have to start with some initial standard (typically, the interim reference standard, which is all you may have available before GMP manufacture), and you can get into some circularity with accuracy (i.e., assaying the standard against itself). So for your initial standard and assay validation you rely somewhat on orthogonal methods of analysis. Also, as per ICH Q2, accuracy can be inferred from precision, linearity and specificity.
By Lori Nixon, RMC Pharmaceutical Solutions, Inc.
Find out more at www.rmcpharma.com
Friday, May 15, 2009
RMC Pharma is 5!
RMC Pharmaceutical Solutions celebrated its 5 year anniversary today. We have grown from a couple people looking for their first consulting opportunity each in our spare bedrooms or kitchen tables to an integrated team capable of taking almost any health care product from conception to commercialization. Today, we are 10 regular employees, with 6000 square feet of office and lab space in Longmont, CO. We've helped 40 clients on three continents with projects ranging from producing pre-clinical proof of concept materials to overseeing pre-approval inspections and quality audits from the biggest pharmaceutical companies in the world. Our capabilities have expanded beyond those of a typical consulting group, to include specialized software such as Slimstat, Minitab, Omnisign, SuperPro and DPL decision modeling software, to laboratory capabilities, including zetasizing, HPLC and Akta based process development. Still, our most valuable resource is our people and the relationships that we've built in our five years. The experience has been very rewarding for me personally, and I hope that my partners, associates and clients have benefited from it as I have.
Subscribe to:
Posts (Atom)