All opinions are my own.

Monday, August 31, 2015

Could Software Technologies be the “Creative Destructionist” for Pharmaceuticals and Chemicals?

Revolution happens in every industry but some revolutions are widely noted and others are not. Nucor Steel revolutionized the steel industry through its ”mini-mills”. They are the largest steel producer in the United States. Their electric furnaces overtook the traditional blast furnaces. It was a perturbation. Many established biggies have been dwarfed.

A quiet but fast paced evolution is taking place in the food industry. Berkshire Hathaway, Warren Buffett, and 3G Capital (Brazil) have started a revolution in the food industry by acquiring Heinz, Kraft, Burger King and Tim Horton (Canada). NestlĂ© Chairman Mr. Peter Brabeck-Letmathe commented, “3G and Buffett have pulverized the food industry market, particularly in America with serial acquisitions” (1). Outsiders are causing changes.

Pharmaceuticals are going through their mergers and acquisitions but they are very unlike in other industries. These are not based on manufacturing technology innovations. Most are related to potential new drug or increasing market share. 

Japanese and European automobile companies led their evolution through product quality and designs. On the other hand fine/specialty chemical industry, pharma’s older cousin due to lack of technology innovation, has seen its shift from developed countries to the developing countries.

Software’s influence:

Advances in software have led evolutions all over. In recent years software has progressively added conveniences for our comfort.

In retailing, through on-line stores and social media, it has changed our purchasing behavior. Google’s venture in driverless car is well known. Apple is rumored to be looking at similar opportunities. Theranos, 23andMe, Nanobiosym are using software technologies to simplify healthcare diagnostics. Google (2) and 23andMe (3) are also involved in drug discoveries. There are other companies like Verseon who are using software to reduce time for new drug discoveries (4). I am sure NIH and pharma companies have capabilities that are used to reduce drug development time. I expect that industry outsiders might have higher success because they do not have many preconceived notions and have a fresh look using latest software modeling methods.

I conjecture that combination of software and miniaturization advances might soon be available at doctor’s office. My speculation is that a routine blood analysis would be done using a handheld device that has a sensor to read the blood not requiring a blood sample. Results would be analyzed instantaneously and patient given the most accurate treatment for her/his ailment.

We are seeing significant and increasing influence of software’s prowess in our lives. Can software’s influence percolate to innovate pharmaceutical and chemical manufacturing technologies and software could be called pharma’s “creative destructionist”? Answer would be positively yes. Pharma does need a major evolution (5) and a spring.

We need to recognize that every API and its formulation is a fine/specialty chemical that has disease-curing value and has been formulated with inert excipients and packaged for easy dispensation.

Current practices:

Fine/specialty chemical manufacturing processes can be improved and changed after commercialization. However, such changes do not happen easily for pharmaceuticals. FDA regulation 21CFR314.70 has to be followed. Current practices produce the desired drug but most of the processes are inefficient and unsustainable. Speed to market is the major cause. There is significant room for improvements. Pharma’s current manufacturing practices are about 30-40 years behind times especially from fine/specialty chemicals, its older cousin. Pharma has not had the need to change because of its assured profitability.

At present, drug patented life is limited after the regulators have approved it for commercial use. Most of the active pharma ingredients and their formulations produced use methods and processes that are laboratory work based. Laboratory practices are essentially duplicated as commercial manufacturing practices even down to how the quality is checked of the intermediates. They add to inefficiency. Generics follow similar practices. Whole business process becomes and stays inefficient.

Future practices:

There is considerable discussion about change/improvement of pharma’s manufacturing practices. For it to happen paradigm shift in process development and their translation to commercial practices is needed. Change has to be initiated and has to happen within the companies. No regulatory body can force change. In addition, it has to happen before clinical trials because change later is challenging, can be expensive and has to follow regulations e.g. FDA 21CFR314.70.

Manufacturing technology innovation change has to happen at the fundamental level and it starts when any paper process chemistry is being reviewed and is going to be tested in the laboratory for feasibility and commercial viability.

From my perspective, before we start exploring/playing with the chemistry in the laboratory, it is necessary to understand the sociochemicology of all of the chemicals and that includes solvents, intermediates and produced products. Sociochemicology is the relative behavior of individual and collective behavior of chemicals. It is dependent on their physical and chemical properties (6). Many can and will ask/say why we need to know individual and mutual behavior of chemicals.

Answer to this question is simple. Individual properties and behavior of many reactants are available but mutual behaviors are not readily available. Physical properties and behavior give us many processing clues. By knowing and understanding individual and mutual behavior, we can develop great to excellent processes. We can also manipulate and exploit sociochemicological behavior to create simpler processes. Chemical engineers and chemists understand values mentioned above. Safe handling information is also important.

Having all of the above information before we go to the laboratory can significantly cut short the laboratory process development and commercialization time. However, there is big issue and it is “we do not have all of the above postulated information about sociochemicology of every chemical involved and produced”.

Information about many chemicals is available from different sites (7,8,9,10,11,12,13,14). There are additional sources. Physical property Information is covered in e.g. Exxon’s blue book and Hydrocarbon Processing (about 1960-1965) etc. and can be of considerable value. If information about chemicals and their sociochemicalogy (their mutual behavior) is not available, it will have to be generated in the laboratory. Chemists and chemical engineers end up generating information what they think is relevant but they may review all available options.  

We do exploit sociochemicological behavior of chemicals but to limited extent. Current and ongoing advances in software technology, through modeling, could also be tapped to create/predict sociochemicological behavior of chemicals. Modeled information about behavior of chemicals would be of great value. It could give us many different process design options that we had not considered. With this information at hand, human genius using his/her creativity and imagination along with training in chemistry and chemical engineering will facilitate development and design of the best manufacturing processes in much shorter time. Such processes will consistently produce quality products at the lowest cost. Well-designed processes could significantly lower regulatory oversight from the current levels. We could also leapfrog the process of “continuous improvement” before the potential drug/s go to clinical trials. At some time in future 21CFR314.70 might not be necessary.

Some benefits of understanding and using behavioral information of chemicals are discussed. Physical state of reactants and intermediates can be used to improve laboratory process development. We may be able to limit number of solvents used to desired two and not to exceed three. Process productivity and conversion yield will also be improved. Total number of conversion steps may be minimized. Intermediate and final product solubility information can be used to develop a better commercial process. Separation processes can be greatly improved. Reaction process kinetics can be well understood and exploited to our advantage. Unit operations and processes can be simplified. Supply chain and operations scheduling can be significantly improved. Asset utilization and total business process will also improve. Capital investment can be significantly reduced.

I may be overoptimistic but having broader scope of information could catapult pharmaceutical manufacturing processes (API and their formulations) from QbA (quality by analysis) to QbD (quality by design) from inception and eliminate “B” bureaucracy (government and internal) and “C” consternation (should we think about continuous improvement) from pharma vocabulary. Processes thus developed would have many of the following benefits.

  1. Longer patent life for new molecules due to shorter drug discovery and development time
  2. Lowest product cost for brand and generic drugs
  3. Consistent and uniform product quality
  4. Use of best technologies
  5. Reduced time to market
  6. Highest profits
  7. Highly sustainable processes with minimum environmental impact
  8. Much higher customer base allowing economies of scale advantage
  9. Potential of continuous API manufacture
  10. Highest possibility of continuous formulations

Outlined benefits clearly suggest that it is time to accelerate change in pharma’s process development and commercialization practices. However, if we are content with the existing practices nothing will change.

Software powerhouses can create the needed information if it has not been created and/or is in private domains. Created information will definitely improve API and chemical manufacturing and their formulation methodologies. Success here would have widespread benefits for billions and pharma landscape would be changed forever. In addition, every industry that uses chemicals will also benefit.

I firmly believe that software enhancing our understanding of chemical behavior can be a REAL “creative destructionist”. Process of continuous innovation will become way of life and change many landscapes.

Girish Malhotra
EPCOT International

  1. NESTLE CEO: Warren Buffett just 'pulverized the food industry market http://www.businessinsider.com/r-nestle-says-taking-action-to-keep-top-slot-in-food-industry-2015-4 accessed August 25, 2015

  2. Large-Scale Machine Learning for Drug Discovery, http://googleresearch.blogspot.com/2015/03/large-scale-machine-learning-for-drug.html accessed August 26, 2015

  3. In Big Shift, 23andMe Will Invent Drugs Using Customer Data http://www.forbes.com/sites/matthewherper/2015/03/12/23andme-enters-the-drug-business-just-as-apple-changes-it/ accessed August 26, 2015

  4. Malhotra, Girish: Are Software Technologies Going to be Pharma’s Creative Destructionists? Profitability through Simplicity, http://pharmachemicalscoatings.blogspot.com/2015/07/are-software-technologies-going-to-be.html accessed August 25, 2015

  5. Malhotra, Girish: Does the Pharmaceutical Industry Need A Steve Jobs? Profitability through Simplicity. http://pharmachemicalscoatings.blogspot.com/2011/11/do-pharmaceuticals-need-steve-jobs.html November 8, 2011 
  6. Malhotra, Girish: Focus on Physical Properties To Improve Processes: Chemical Engineering, Vol. 119 No. 4 April 2012, pgs 63-66
7.     Chemblink

11.  Merck Index

Friday, August 7, 2015

Impact of 21CFR 314.70 on The Life of a Pharmaceutical Manufacturing Process

Regulations are necessary for quality assurance of drugs. FDA established 21CFR314.70. It is a very important rule. It assures that there is no “by manufacturer’s choice” deviation from the manufacturing methods and practices that have been filed for the active pharmaceutical ingredient (API) and their formulations, labeling and packing etc.  

My interpretation of 21CFR314.70:

21CFR 314.70 calls for reporting any changes that have been made in anything to do with production of a saleable pharmaceutical product and that includes API. Changes include process method or operating conditions, operating parameters, equipment or test methods etc. Their impact on the drug efficacy and performance has to be assessed by the drug producer. Basically the idea is the producer has to assure that the efficacy and performance of the drug does not change from the approved product.  

I call this regulation an opportunity if the pharma companies want to have the best of the technologies to maximize their profits and increase their patient base. To achieve an effort or lets call a different thinking would be needed. Since it would be different from the current thinking, many would say, “Can’t be done.” That would be considered normal.

21 CFR314.70 encourages “continuous improvements” in the processes that will create the best product for clinical trials and that’s the way it should be. QbD (quality by design) becomes natural part of the process development before a process is commercialized. However, in my estimation all this has to be done prior to going to clinical trials. Benefits will be infectious.  

The Simplest and the Difficult Part:

Reporting part of the regulation is the simplest of the change exercise. However, section 314.70 (2) specifically asks “holder of an approved application…. must assess the effects of the change before distributing a drug” and it is the difficult part. 

This CFR asks for assurance that the process changes have not altered the performance and efficacy of the drug that has been approved by the regulatory bodies. In order to prove this aspect, company may have to re-do all of the necessary tests that were done including clinical trials to get the necessary approval. This would be like starting from scratch and can be an expensive and time consuming process.

Since changes can have a significant financial impact, unless there are compelling reasons and cost justification, no company would go through the change process and that includes continuous improvements. This is especially true for a brand drug due to its limited patent life after the drug has been approved. Generics could incorporate the best of the technologies and methods but due to competitive pressures to be the first to market, not sufficient time is spent to perfect what all is involved.

Thus, any process change after the drug has been approved is rather difficult. This inability has significant ramifications. The biggest is no process improvements can be made on majority of the existing approved drugs for the fear of different performance, impurities and altered efficacy. Costs related to process inefficiencies are passed on to the patients, a sad and sorry consequence of the good intentioned regulation. Additional unfortunate part is that many of the drugs become out of reach of needy.

Other Side of the Coin:

Is there a solution to include the best methods? Yes there is and it is simple. As stated earlier normal reaction would be “can’t be done”.

Companies manufacturing APIs and their formulations will have to think and act differently. They have to give away their deep-rooted and imbedded traditions. Process development chemists and chemical engineers who would scale-up and eventually commercialize the products would have to get involved in the development process at an early stage. Every one involved would have to consider and imagine how the commercial processes would operate. In these initial stages the developers would need to review safety and environmental aspects keeping in mind how cGMP requirements etc. will impact asset utilization. There will be lot of learning and the knowledge gained would facilitate efficient process development.

It is normal in the laboratory to isolate intermediates and that is fine for a lab process. However, for a commercial process it would be necessary create a path where isolation is to be avoided. This is not normal in pharma manufacturing. Different thinking would be needed.

Manufacturing recipes exploiting sociochemicology of chemicals will have to be included to create the best processes. Chemical industry, pharma’s older cousin, has practiced these methods and they can be easily translated to produce API and their formulations.

FDA and EMA are issuing various guidances to cajole the industry to improve their manufacturing practices and technologies. Till the industry takes the steps to modernize their methods, not much will change. By the way using better analytical equipment is not process technology improvement, at least not in my book. To me it is like replacing a dull knife with a sharper knife.


Finally 21CFR314.70 latently suggests creating and having the best processes before clinical trials are the best option. Pharmaceuticals can vigorously incorporate process of “continuous improvements” in their API manufacturing and formulation processes before they go to clinical trials. They can very easily exploit “what if” scenarios to serve different levels of population. Such an analysis would allow them to best consider and incorporate batch vs. continuous (1, 2, 3) processes for API manufacture and/or formulations, an unprecedented opportunity.

In my analysis pre-21CFR314.70 thoughtful process development can be a tremendous win. Changes made after drug approval will increase workload for the companies and regulatory bodies and this makes changes difficult to incorporate at best. 

Girish Malhotra, PE
EPCOT International

  1. Malhotra, Girish: Why Fitting a Square Plug in a Round hole is Profitable for Pharma and Most Likely Will Stay? Profitability through Simplicity, http://pharmachemicalscoatings.blogspot.com/2014/08/why-fitting-square-plug-in-round-hole.html
2.     Malhotra, Girish: A Blueprint for Improved Pharma Competitiveness, Contract Pharma, Vol. 16, 7, Pg. 46-49, September 2014, http://www.contractpharma.com/issues/2014-09-01/view_features/a-blueprint-for-improved-pharma-competitiveness/   

  1. Malhotra, Girish: Continuous Process in Pharmaceutical Manufacturing: Considerations, Nuances and Challenges, Contact Pharma, June 2, 2015, http://www.contractpharma.com/issues/2015-06-01/view_features/continuous-process-in-pharmaceutical-manufacturing-considerations-nuances-and-challenges