Over 10 years experience of Traceability Solutions

drug discovery pharma medical health care research shutterstock
17 Jan

By pharmatrax

Category: Technoloy

Machine Learning, Clouds & Pills: Biopharma AI’s Big Impact No Comments

Machine Learning, Clouds & Pills: Biopharma AI’s Big Impact

Share This Post

Advances in AI and cloud technologies are helping health and pharmaceutical companies improve services, but they come with regulatory challenges. Let’s look at how smart companies are leveraging smart tech to gain a competitive advantage, and stay on the right side of the law, in 2020.

We tend to think of technological advancements arising in response to needs, such as customers’ constant demand for faster shipping driving Amazon’s supply-chain innovations. But in reality, the relationship between technology and use-conditions is more reciprocal. In fact, innovation often precedes a full understanding of its potential impacts.

Today’s pharmaceutical industry is in some ways entering just such a “catch-up” phase with AI and cloud technology. Recent MarketsandMarkets research indicates the market for biopharma AI will increase from US $198.3 million in 2018 to a whopping $3.88 billion by 2025—good for a 52.9 percent CAGR.

Growth Factors

This massive growth is attributable to three main technological factors:

  1. Skyrocketing Data Volumes: Big Data philosophy, and a proliferation of data capturing and analytics tools, means pharmas have more information to mine for insights and optimization potential. Examples include tracking data from health apps, electronic health records, medical imaging data and more.
  2. Next Generation Computing Power: Manually processing these mammoth data archives would take many humans many years to accomplish, but machine learning AIs can be trained to complete these tasks with incredible quickness and precision. These tools are resource-intensive, and as recently as a decade ago there were only a few labs in the world that could handle projects at the scale today’s business regularly demands. As computing power advances, so too do the ambitions of tech-savvy companies.
  3. Low Computing Costs: Happily, these improvements in processing muscle have arrived hand in hand with a dramatic reduction in the cost of high-volume computing services. As demand for AI chipsets has increased, more have come to market, and per unit costs have fallen. Moreover, cloud computing means more companies can outsource their processing and storage needs to third parties — a formula for big savings on infrastructure investments.

AI for Optimization

Google parent Alphabet recently tasked its intelligence lab DeepMind with exploring the world of drug discovery and development, a US $9.4 billion industry in 2018. By “feeding” its AI massive volumes of biological research – data collected from clinical studies and aggregated patient medical records – DeepMind’s developers have generated speculative new drugs showing promise in laboratory tests combined with faster development cycles.

Most pharma companies still use AI in this “brute force” fashion to crunch as much data as possible and then parse the speculative results for ideas that might work. But it’s likely that the real innovations will come as AI is more thoroughly integrated into every stage of the drug discovery process. The idea is to generate data that is tailor-made from the moment of its creation to be compatible with ML algorithms.

Cloud: A Game Changer

Let’s go back to considering those massive volumes of data AIs need. Storing petabytes upon petabytes of data is a challenge for all but the most robust on-site data storage facilities, taking into account the infrastructure and upkeep costs such monoliths suggest. It probably won’t be news to you if you’re reading this blog that most companies have switched to enterprise cloud computing options.

In fact, Deloitte Global predicted that in 2019 “all industries will accelerate their use of cloud-based AI software and services” (emphasis ours). Fully 70 percent of these are expected to obtain their capabilities from cloud-based enterprise software.

Advantages

  • Remote Collaboration: Many pharma operations facilities and partners are spread across the country and globe. The cloud lets them collaborate like cubicle-mates.
  • Lower Upkeep, Fewer Updates: Downloading data management tasks to cloud providers means letting someone else handle infrastructure updates and maintenance. This is a boon for smaller companies trying to keep pace with giants.
  • Data Security: Making wise choices with one’s cloud providers, and instilling good in-house security practices, can make private data stored off-site highly secure. Particularly given the complex regulatory environment healthcare and pharma companies operate in, using encryption, passwords and firewalls is simply good business.

Regulatory Risks

Getting back to DeepMind, it has astonishing technology to be sure, but that technology is hungry for data, and sometimes the way that data is acquired is less than ethical. In November 2019, a whistleblower leaked information that Alphabet had secretly acquired the medical information of 50 million Americans without their consent. While it’s common practice for medical information to be bought and sold at scale, this transaction was different: the data was unencrypted and included the names and personal details of the patients affected.

While the fallout from this disclosure is ongoing, it’s unclear if Alphabet will be able to fully brush off regulatory ramifications — fines for knowingly violating HIPAA were recently reduced to $250,000 per infraction, or less than the company makes in the time it took you to read the number in your head.

For smaller companies, however, as well as those who take seriously their ethical responsibilities to their clients, regulations like HIPAA and Europe’s GDPR still bare teeth. Most notably, failure to meet data security obligations can create entanglements when it comes to drug approval.

The best approach is to lean into the benefits that the cloud and AI offer vis a vis regulatory submission. For example, using AI makes it easy to record exactly what actions were performed throughout the research process. It also allows for pinpoint tracking of how and where sensitive data is accessed. This streamlines the auditing process and ensures the replicability of test results. Remember: machines and regulators alike adore the scientific method — and the smartest companies recognize this means they need not be at odds.

Advances in AI and cloud technologies are helping health and pharmaceutical companies improve services, but they come with regulatory challenges. Let’s look at how smart companies are leveraging smart tech to gain a competitive advantage, and stay on the right side of the law, in 2020.

We tend to think of technological advancements arising in response to needs, such as customers’ constant demand for faster shipping driving Amazon’s supply-chain innovations. But in reality, the relationship between technology and use-conditions is more reciprocal. In fact, innovation often precedes a full understanding of its potential impacts.

Today’s pharmaceutical industry is in some ways entering just such a “catch-up” phase with AI and cloud technology. Recent MarketsandMarkets research indicates the market for biopharma AI will increase from US $198.3 million in 2018 to a whopping $3.88 billion by 2025—good for a 52.9 percent CAGR.

Growth Factors

This massive growth is attributable to three main technological factors:

  1. Skyrocketing Data Volumes: Big Data philosophy, and a proliferation of data capturing and analytics tools, means pharmas have more information to mine for insights and optimization potential. Examples include tracking data from health apps, electronic health records, medical imaging data and more.
  2. Next Generation Computing Power: Manually processing these mammoth data archives would take many humans many years to accomplish, but machine learning AIs can be trained to complete these tasks with incredible quickness and precision. These tools are resource-intensive, and as recently as a decade ago there were only a few labs in the world that could handle projects at the scale today’s business regularly demands. As computing power advances, so too do the ambitions of tech-savvy companies.
  3. Low Computing Costs: Happily, these improvements in processing muscle have arrived hand in hand with a dramatic reduction in the cost of high-volume computing services. As demand for AI chipsets has increased, more have come to market, and per unit costs have fallen. Moreover, cloud computing means more companies can outsource their processing and storage needs to third parties — a formula for big savings on infrastructure investments.

AI for Optimization

Google parent Alphabet recently tasked its intelligence lab DeepMind with exploring the world of drug discovery and development, a US $9.4 billion industry in 2018. By “feeding” its AI massive volumes of biological research – data collected from clinical studies and aggregated patient medical records – DeepMind’s developers have generated speculative new drugs showing promise in laboratory tests combined with faster development cycles.

Most pharma companies still use AI in this “brute force” fashion to crunch as much data as possible and then parse the speculative results for ideas that might work. But it’s likely that the real innovations will come as AI is more thoroughly integrated into every stage of the drug discovery process. The idea is to generate data that is tailor-made from the moment of its creation to be compatible with ML algorithms.

Cloud: A Game Changer

Let’s go back to considering those massive volumes of data AIs need. Storing petabytes upon petabytes of data is a challenge for all but the most robust on-site data storage facilities, taking into account the infrastructure and upkeep costs such monoliths suggest. It probably won’t be news to you if you’re reading this blog that most companies have switched to enterprise cloud computing options.

In fact, Deloitte Global predicted that in 2019 “all industries will accelerate their use of cloud-based AI software and services” (emphasis ours). Fully 70 percent of these are expected to obtain their capabilities from cloud-based enterprise software.

Advantages

  • Remote Collaboration: Many pharma operations facilities and partners are spread across the country and globe. The cloud lets them collaborate like cubicle-mates.
  • Lower Upkeep, Fewer Updates: Downloading data management tasks to cloud providers means letting someone else handle infrastructure updates and maintenance. This is a boon for smaller companies trying to keep pace with giants.
  • Data Security: Making wise choices with one’s cloud providers, and instilling good in-house security practices, can make private data stored off-site highly secure. Particularly given the complex regulatory environment healthcare and pharma companies operate in, using encryption, passwords and firewalls is simply good business.

Regulatory Risks

Getting back to DeepMind, it has astonishing technology to be sure, but that technology is hungry for data, and sometimes the way that data is acquired is less than ethical. In November 2019, a whistleblower leaked information that Alphabet had secretly acquired the medical information of 50 million Americans without their consent. While it’s common practice for medical information to be bought and sold at scale, this transaction was different: the data was unencrypted and included the names and personal details of the patients affected.

While the fallout from this disclosure is ongoing, it’s unclear if Alphabet will be able to fully brush off regulatory ramifications — fines for knowingly violating HIPAA were recently reduced to $250,000 per infraction, or less than the company makes in the time it took you to read the number in your head.

For smaller companies, however, as well as those who take seriously their ethical responsibilities to their clients, regulations like HIPAA and Europe’s GDPR still bare teeth. Most notably, failure to meet data security obligations can create entanglements when it comes to drug approval.

The best approach is to lean into the benefits that the cloud and AI offer vis a vis regulatory submission. For example, using AI makes it easy to record exactly what actions were performed throughout the research process. It also allows for pinpoint tracking of how and where sensitive data is accessed. This streamlines the auditing process and ensures the replicability of test results. Remember: machines and regulators alike adore the scientific method — and the smartest companies recognize this means they need not be at odds.

Source:https://www.enterpriseai.news/2020/01/08/machine-learning-clouds-pills-biopharma-ais-big-impact/

Share this Post!

Leave a Reply


The reCAPTCHA verification period has expired. Please reload the page.

Categories