Rick Morrison, Comprehend Systems03.06.13
The life science industry is changing. Just 10 years ago, almost all trials were collecting data through paper-based systems and employed floors of programmers, but today all areas of clinical development are rapidly transforming. Companies are becoming more and more comfortable outsourcing to third-party vendors. They’re adopting new technologies at extremely quick rates, and these technologies are starting to have a serious impact on the entire drug development life cycle.
Every person involved in conducting clinical trials is impacted by these changes. CRAs are now being made instantly aware of what’s actionable at their sites and studies. Managers and directors can access and evaluate teams performance at all times. Data managers can clean data and find problems more effectively than ever before. Study managers and executives can accurately identify bottlenecks earlier in the process, allowing them to make adjustments that lead to cost savings and better outcomes.
We shall highlight several key technological areas that are enabling these rapid advances.
Going Mobile
It’s hard to overstate the impact mobile computing is having on all of our lives, from both a professional and personal perspective. There are now billions of people worldwide who are carrying what would be considered a supercomputer a decade ago. This mobile revolution has far-reaching impact, which will take years or decades for us to fully realize.
The ramifications of mobile devices on clinical trials are no less remarkable. We’re starting to see a wide range of different data collection systems for the patients themselves to use throughout the clinical trial to automatically collect data, log events and more. This trend will only expand. Devices like connected Holter monitors can continuously upload data to cloud servers, enabling constant surveillance by a centralized medical team. Alivecor’s mobile ECG sleeves automatically upload data as well; during a beta test, one unit revealed that a patient was unknowingly experiencing a heart attack! These devices automatically upload their data to a central repository, where doctors can immediately review and take action if necessary. These devices are changing the very nature of clinical trials, since they enabling monitoring and experimentation that previously would have been too costly, dangerous, or impossible.
Investigators and doctors are seeing the mass adoption of mobile data collection systems and eCRF devices, which drastically reduce the amount of time between collection and cursory review. They also increase correctness, auditability, and more. As users become more accustomed to mobile devices in their everyday lives and as the interfaces on existing technology become more seamless, this trend will continue. Current generation devices are much less clunky than those of previous generations, which enable easier deployment and better user capabilities. As speech recognition and hand recognition software gets better, it will only reduce delays further and speed up adoption.
Monitors and others involved in conducting clinical trials are also starting to take advantage of mobile devices. Monitors used to rely only on reviewing paper, completely separated from the outside world. Now, they’re able to review and communicate with clinicians and data managers who are taking advantage of centralized analysis to find problems earlier in the trials process. They can be instantly alerted in the field via e-mail and text message, enabling them to do their jobs faster and better, which in turn results in better execution across the entire study.
Mobile devices are making new processes and techniques possible, which are changing the fundamental structure of how clinical trials are designed and implemented.
Data Warehouses Get a Makeover
Different data collection systems and databases are now widely available and used. Even within a single study, there are often half a dozen or more data collection systems in use, each usually non-interoperable with the others. Some are cloud-based, some are onsite, some are commercial, some are custom. Most are created by different vendors.
This problem only gets exacerbated when sponsors and CROs start involving more than one study, or want to incorporate closed studies. While companies want to be able to explore all of their data and run analyses across the various systems, practically speaking it is all but impossible now. Interoperability is key.
For the past decade, data warehouses have been the preferred method for trying to achieve this goal. The objective is to hire a team of programmers to take all of the data and consolidate it into a single, well-structured database. Then, when anyone has a question, they can simply go to the data warehouse and find the answer.
There are several problems that prevent data warehouses from becoming a feasible reality. First of all, there are new data systems and studies coming online all the time, or changing mid-stream, which require the team of programmers to come back and restructure the entire data warehouse. Also, if users want to ask questions that were not initially pre-programmed into the data warehouse, programmers need to reprogram the entire system. In practice, these two problems result in data warehouses losing most of their anticipated value, because the questions users most want answered rely on the latest data, or on questions that weren’t necessarily preconceived.
In practice, data warehouses are extremely expensive and aren’t ever fully functional, and a lot of companies are abandoning plans after investments of tens of millions of dollars. However, we’re seeing a wave of technology that is fixing this problem. Think of these new technologies as a distributed analytics cube against all relevant databases. Even better, these tools are built specifically for clinical data, which means they handle some of the trickier aspects that data warehouses typically lack, including properly handling missing and null data, and making it easily accessible to all of the different user roles involved with the trial. These distributed systems work simultaneously across different vendors, types of systems and even custom systems, and can bring on new studies as soon as the data start being collected.
The old standard practice of hiring teams of programmers to constantly maintain a data warehouse that’s never really usable is fading away. Companies have wasted resources on these warehouses, and they’re no longer working at the level the industry demands. New technologies that are based on modern software techniques and are horizontally scalable will continue to replace antiquated systems.
Collaboration
Neither collecting data electronically nor analyzing it is an end in itself. Sponsors have yet to fully realize the ultimate end game; in order for data to be fully useful, the results need to be actionable and, most importantly, actions need to be taken.
The traditional methods of collaboration haven’t changes for decades and are still widely used. For example, if a clinical scientist needs a report looking at a couple dimensions of patient data, he typically has to ask a clinical programmer to create it, which can then take a week or two to develop. Remember, this often affects the critical path of getting the drug approved.
Similarly, if a data manager finds a problem that is actionable for a monitor in the field, he typically needs to log in to the specific query or discrepancy system for whichever study he’s currently looking at, and then log a query that he hopes will be picked up when the monitor next logs into the system. Or, they use e-mail, but messages are prone to getting lost and are not easily tracked. These are just a few of the current methods that are already used, but many new and exciting methods are becoming available.
Cross-study collaboration tools that are extremely prevalent and powerful allow users to access all data in a single location. But more than this, they actually allow teams to work more succinctly and efficiently, enabling all parties to get what they need, when they need it. When the clinical scientist needs something from the clinical programmer, he or she can just assign it in the same environment where the clinical programmer is already working. The programmer can then reuse or share that work when finished, allowing the clinical scientist to self-pull requests in the future.
Every person in the company is able to rely on push technologies, so that whenever a relevant data point is entered or changed, notification goes out automatically via e-mail or SMS message. Teams are able to share analysis and findings extremely easily, and then follow up on the results. Managers are able to see how effective their teams are actually performing. Every user in the organization can seamlessly get the answers needed to do their jobs well.
This type of efficient and next-generation collaboration has only recently become possible, with the advent of cloud technologies and the widespread distribution of mobile devices. CRAs in the field can just as easily collaborate on a project as a data scientist back home.
To the Cloud
Like mobile, it’s hard to overestimate the influence that cloud computing is having on enterprise data systems. Cloud computing is when data is stored on a large number of centralized servers “in the cloud,” meaning out in the Internet, typically hosted by an outside vendor, as opposed to being stored locally. What’s particularly powerful about cloud computing is that the customer doesn’t have to provision its own hardware or team of people to maintain the system, as it is handled by the outside vendor. Cloud systems are also priced so that customers pay for what they use. The more usage, the higher the price tag. Companies no longer need a multi-million-dollar upfront investment to start using a new technology – they can get started today for a relatively low amount, then pay as they grow.
During the past few years, there have been several phenomenally successful cloud vendors selling cloud software for life sciences, including Medidata Solutions and Veeva Systems, and this trend is only accelerating. In fact, there are now dozens of cloud EDC systems that are widely used throughout the industry, with new ones springing up all the time. Cloud allows vendors to scale each customer in its own controlled environment, maximizing performance and minimizing cost. It also makes it easy for other third-party vendors to integrate, for example, by doing so once and preventing the headaches that are associated with onsite deployments.
Cloud vendors offer their services cheaper because they can take advantage of the economies of scale. This allows customers to benefit as well because they can get started much cheaper than with an onsite deployment, and then scale up as needed.
Machine Learning
Systems are springing up that enable computers to do things that humans could never do, like finding the proverbial needle in a haystack. Not only that, but we’re starting see the proliferation of some particularly advanced forms of computing that weren’t possible just five years ago. For example, advanced signal detection and hypothesis generation, which allow computers to do what they do best – repetitive tasks on huge amounts of data.
Over the next several years we’re going to see this trend continue and combine with the other technologies outlined above to create some particularly amazing solutions. Companies that deploy these systems wisely will have a huge competitive advantage over their competition.
One example of this would be finding hidden drug-drug interactions on large sets of data, which were previously going undetected. Another example is helping to find ways to connect data sets that were previously disparate.
Hypothesis generation is another area that will become more prevalent as researchers have access to large heterogeneous datasets. They’ll be able to analyze large datasets using advanced computer-assisted tools to find interactions and answers that they didn’t even know to look for, which will significantly impact data analysis across studies. As a result, advanced discoveries, such as repurposing drugs and other significant trends, will be uncovered.
These trends are just some of the most significant advances in clinical IT that we’re expecting to see in the near future. As companies continue to embrace these technologies and adopt these best practices, they will continue realizing the value of these advances. It doesn’t even need to be challenging to begin adopting these technologies, as deployment techniques like cloud and Software-as-a-Service make it much easier than traditional onsite deployments. As the drug development market becomes more and more competitive, the companies that embrace change and innovate the most will reap the largest rewards.
Companies can start simple — adopting cloud technologies has a much lower barrier to entry than traditional enterprise software. Initiate change within smaller groups to experiment with technologies for a smaller upfront cost. Once study groups prove the value, sponsors can consider additional opportunities and start expanding from there.
The future of clinical IT innovation is limitless, and those who are receptive to adapting new technologies and practices will be enabled to do things previously unimaginable.
Rick Morrison is chief executive officer, Comprehend Systems. He can be reached at rmorrison@comprehend.com.
Every person involved in conducting clinical trials is impacted by these changes. CRAs are now being made instantly aware of what’s actionable at their sites and studies. Managers and directors can access and evaluate teams performance at all times. Data managers can clean data and find problems more effectively than ever before. Study managers and executives can accurately identify bottlenecks earlier in the process, allowing them to make adjustments that lead to cost savings and better outcomes.
We shall highlight several key technological areas that are enabling these rapid advances.
Going Mobile
It’s hard to overstate the impact mobile computing is having on all of our lives, from both a professional and personal perspective. There are now billions of people worldwide who are carrying what would be considered a supercomputer a decade ago. This mobile revolution has far-reaching impact, which will take years or decades for us to fully realize.
The ramifications of mobile devices on clinical trials are no less remarkable. We’re starting to see a wide range of different data collection systems for the patients themselves to use throughout the clinical trial to automatically collect data, log events and more. This trend will only expand. Devices like connected Holter monitors can continuously upload data to cloud servers, enabling constant surveillance by a centralized medical team. Alivecor’s mobile ECG sleeves automatically upload data as well; during a beta test, one unit revealed that a patient was unknowingly experiencing a heart attack! These devices automatically upload their data to a central repository, where doctors can immediately review and take action if necessary. These devices are changing the very nature of clinical trials, since they enabling monitoring and experimentation that previously would have been too costly, dangerous, or impossible.
Investigators and doctors are seeing the mass adoption of mobile data collection systems and eCRF devices, which drastically reduce the amount of time between collection and cursory review. They also increase correctness, auditability, and more. As users become more accustomed to mobile devices in their everyday lives and as the interfaces on existing technology become more seamless, this trend will continue. Current generation devices are much less clunky than those of previous generations, which enable easier deployment and better user capabilities. As speech recognition and hand recognition software gets better, it will only reduce delays further and speed up adoption.
Monitors and others involved in conducting clinical trials are also starting to take advantage of mobile devices. Monitors used to rely only on reviewing paper, completely separated from the outside world. Now, they’re able to review and communicate with clinicians and data managers who are taking advantage of centralized analysis to find problems earlier in the trials process. They can be instantly alerted in the field via e-mail and text message, enabling them to do their jobs faster and better, which in turn results in better execution across the entire study.
Mobile devices are making new processes and techniques possible, which are changing the fundamental structure of how clinical trials are designed and implemented.
Data Warehouses Get a Makeover
Different data collection systems and databases are now widely available and used. Even within a single study, there are often half a dozen or more data collection systems in use, each usually non-interoperable with the others. Some are cloud-based, some are onsite, some are commercial, some are custom. Most are created by different vendors.
This problem only gets exacerbated when sponsors and CROs start involving more than one study, or want to incorporate closed studies. While companies want to be able to explore all of their data and run analyses across the various systems, practically speaking it is all but impossible now. Interoperability is key.
For the past decade, data warehouses have been the preferred method for trying to achieve this goal. The objective is to hire a team of programmers to take all of the data and consolidate it into a single, well-structured database. Then, when anyone has a question, they can simply go to the data warehouse and find the answer.
There are several problems that prevent data warehouses from becoming a feasible reality. First of all, there are new data systems and studies coming online all the time, or changing mid-stream, which require the team of programmers to come back and restructure the entire data warehouse. Also, if users want to ask questions that were not initially pre-programmed into the data warehouse, programmers need to reprogram the entire system. In practice, these two problems result in data warehouses losing most of their anticipated value, because the questions users most want answered rely on the latest data, or on questions that weren’t necessarily preconceived.
In practice, data warehouses are extremely expensive and aren’t ever fully functional, and a lot of companies are abandoning plans after investments of tens of millions of dollars. However, we’re seeing a wave of technology that is fixing this problem. Think of these new technologies as a distributed analytics cube against all relevant databases. Even better, these tools are built specifically for clinical data, which means they handle some of the trickier aspects that data warehouses typically lack, including properly handling missing and null data, and making it easily accessible to all of the different user roles involved with the trial. These distributed systems work simultaneously across different vendors, types of systems and even custom systems, and can bring on new studies as soon as the data start being collected.
The old standard practice of hiring teams of programmers to constantly maintain a data warehouse that’s never really usable is fading away. Companies have wasted resources on these warehouses, and they’re no longer working at the level the industry demands. New technologies that are based on modern software techniques and are horizontally scalable will continue to replace antiquated systems.
Collaboration
Neither collecting data electronically nor analyzing it is an end in itself. Sponsors have yet to fully realize the ultimate end game; in order for data to be fully useful, the results need to be actionable and, most importantly, actions need to be taken.
The traditional methods of collaboration haven’t changes for decades and are still widely used. For example, if a clinical scientist needs a report looking at a couple dimensions of patient data, he typically has to ask a clinical programmer to create it, which can then take a week or two to develop. Remember, this often affects the critical path of getting the drug approved.
Similarly, if a data manager finds a problem that is actionable for a monitor in the field, he typically needs to log in to the specific query or discrepancy system for whichever study he’s currently looking at, and then log a query that he hopes will be picked up when the monitor next logs into the system. Or, they use e-mail, but messages are prone to getting lost and are not easily tracked. These are just a few of the current methods that are already used, but many new and exciting methods are becoming available.
Cross-study collaboration tools that are extremely prevalent and powerful allow users to access all data in a single location. But more than this, they actually allow teams to work more succinctly and efficiently, enabling all parties to get what they need, when they need it. When the clinical scientist needs something from the clinical programmer, he or she can just assign it in the same environment where the clinical programmer is already working. The programmer can then reuse or share that work when finished, allowing the clinical scientist to self-pull requests in the future.
Every person in the company is able to rely on push technologies, so that whenever a relevant data point is entered or changed, notification goes out automatically via e-mail or SMS message. Teams are able to share analysis and findings extremely easily, and then follow up on the results. Managers are able to see how effective their teams are actually performing. Every user in the organization can seamlessly get the answers needed to do their jobs well.
This type of efficient and next-generation collaboration has only recently become possible, with the advent of cloud technologies and the widespread distribution of mobile devices. CRAs in the field can just as easily collaborate on a project as a data scientist back home.
To the Cloud
Like mobile, it’s hard to overestimate the influence that cloud computing is having on enterprise data systems. Cloud computing is when data is stored on a large number of centralized servers “in the cloud,” meaning out in the Internet, typically hosted by an outside vendor, as opposed to being stored locally. What’s particularly powerful about cloud computing is that the customer doesn’t have to provision its own hardware or team of people to maintain the system, as it is handled by the outside vendor. Cloud systems are also priced so that customers pay for what they use. The more usage, the higher the price tag. Companies no longer need a multi-million-dollar upfront investment to start using a new technology – they can get started today for a relatively low amount, then pay as they grow.
During the past few years, there have been several phenomenally successful cloud vendors selling cloud software for life sciences, including Medidata Solutions and Veeva Systems, and this trend is only accelerating. In fact, there are now dozens of cloud EDC systems that are widely used throughout the industry, with new ones springing up all the time. Cloud allows vendors to scale each customer in its own controlled environment, maximizing performance and minimizing cost. It also makes it easy for other third-party vendors to integrate, for example, by doing so once and preventing the headaches that are associated with onsite deployments.
Cloud vendors offer their services cheaper because they can take advantage of the economies of scale. This allows customers to benefit as well because they can get started much cheaper than with an onsite deployment, and then scale up as needed.
Machine Learning
Systems are springing up that enable computers to do things that humans could never do, like finding the proverbial needle in a haystack. Not only that, but we’re starting see the proliferation of some particularly advanced forms of computing that weren’t possible just five years ago. For example, advanced signal detection and hypothesis generation, which allow computers to do what they do best – repetitive tasks on huge amounts of data.
Over the next several years we’re going to see this trend continue and combine with the other technologies outlined above to create some particularly amazing solutions. Companies that deploy these systems wisely will have a huge competitive advantage over their competition.
One example of this would be finding hidden drug-drug interactions on large sets of data, which were previously going undetected. Another example is helping to find ways to connect data sets that were previously disparate.
Hypothesis generation is another area that will become more prevalent as researchers have access to large heterogeneous datasets. They’ll be able to analyze large datasets using advanced computer-assisted tools to find interactions and answers that they didn’t even know to look for, which will significantly impact data analysis across studies. As a result, advanced discoveries, such as repurposing drugs and other significant trends, will be uncovered.
These trends are just some of the most significant advances in clinical IT that we’re expecting to see in the near future. As companies continue to embrace these technologies and adopt these best practices, they will continue realizing the value of these advances. It doesn’t even need to be challenging to begin adopting these technologies, as deployment techniques like cloud and Software-as-a-Service make it much easier than traditional onsite deployments. As the drug development market becomes more and more competitive, the companies that embrace change and innovate the most will reap the largest rewards.
Companies can start simple — adopting cloud technologies has a much lower barrier to entry than traditional enterprise software. Initiate change within smaller groups to experiment with technologies for a smaller upfront cost. Once study groups prove the value, sponsors can consider additional opportunities and start expanding from there.
The future of clinical IT innovation is limitless, and those who are receptive to adapting new technologies and practices will be enabled to do things previously unimaginable.
Rick Morrison is chief executive officer, Comprehend Systems. He can be reached at rmorrison@comprehend.com.