Latest News

How Point-Of-Care Can Bridge The Gap Between Engineering And Medicine

May 22, 2019 | The point-of-care (POC) field should serve as a bridge between the engineering world and medicine world, says Ping Wang, Chief of Clinical Chemistry and Director of the Core Laboratory Hospital of the University of Pennsylvania. In her role, she focuses on the development and translation of novel technologies, with the end result being improvements in diagnosis and prognosis.

As a practicing clinical laboratory director and as a technology researcher and developer, Wang sees the implementation of these technologies from a unique perspective, using and leveraging the strengths in both fields for clinical applications. Collaborating with engineers and clinicians at various universities and institutions allows her to see the emerging trends in clinical care.

On behalf of Diagnostics World News, Mana Chandhok spoke with Wang about those current and emerging trends in clinical care, the lasting impact controversies such as Theranos' fraudulent claims have had on the point-of-care field, and where AI is making an impact today.

Editor's note: Mana Chandhok, a Conference Producer at Cambridge Healthtech Institute, is planning a track dedicated to Emerging Technologies at the Point-of-Care at the upcoming Next Generation Dx Summit in Washington, DC, August 20-22. Wang will be speaking on the program. Their conversation has been edited for length and clarity.

Diagnostics World News: You recently wrote a review article about current and emerging trends at the clinical care (DOI: 10.1373/clinchem.2018.287052). Can you briefly summarize the top points of the article? Are there new trends technologies that you are particularly excited about and why?

Ping Wang: Yes, so for the review article published in Clinical Chemistry, what we tried to do is summarize lots of the growth points and merging trends we have seen in fields of [POC] testing. The [POC] testing here is used in the broad sense, not limited to what we usually say as point-of-care testing, meaning in vitro diagnostics. We took a more broad approach by looking at in vitro diagnostics and some components of in vivo testing. For example, the continuous glucose monitors don't require you to take a sample; the analyte’s concentration is monitored continuously over a relatively long time. So we broadened our scope a little more to include those aspects as well.

The growth points and emerging trends are summarized in table 2 of the review article. We divide them into different areas such as specimens, testing innovation, analyzer, disposables, and data etc. So you can see there are a lot of new developments in this field.

Lots of the developments facilitate the current trends of telemedicine, which we have seen lots of growth in the last several years. The focus here is that the patients, especially those with chronic diseases, will be able to stay at home and be able to get consultations from their care providers over the internet. For example, through a secure connection, the care providers are able to give their consultations for a disease, and sometimes provide prescriptions in a remote way. That poses a question for the lab, which is, if the patient does not come to the hospitals or clinics anymore, how do we, as a diagnostic partner in the healthcare process, provide a new model of testing to support these needs of physicians and patients? I think a lot of trends we have seen in the past decade or so, such has mobile devices for smart phones, tablets, smart watches, and mobile phone applications for healthcare applications, and cloud usage for health data storage and transmission, will further facilitate the development of telemedicine.

At the same time, these technologies will allow the lab to gradually evolve from the current model of central lab testing models, in which the samples come to us, to more models of testing, allowing us to bring in more lab-on-a-chip devices or wearables or insertables into the field as well. That is something that I personally am very interested in and very excited about.

Even though it has been four years since Theranos was exposed as a fraud, point-of-care companies are still fighting against that association. How and when do you imagine the point-of-care field overcoming this reputation?

I haven't seen a lot of negative implications on the overall point of care testing fields from Theranos. The case of Theranos is broadly covered now, and  HBO and ABC have reported on it and have even made documentaries.

I don't know exactly if investors would be saying, "Well if Theranos didn't work we should not invest in [POC] testing anymore." I don't think that's the case. The vision to bring lab testing to more people, make it more user friendly , more affordable and with less blood  is still compelling. You need to have some solid technology to realize that vision.

Do you see any other major roadblocks to point-of-care taking over, or do you see anything standing in the way of success in the next five years?

I don’t think success should be defined as point-of-care testing taking over central lab testing. The field of point-of-care testing has existed for a long time, but we have not seen it take over the central lab testing. I don't think it will change in the near future, in the five or ten years or so. I think the central lab model is still a very strong one, as we have a lot of high volume testing, and we can turn around results relatively quickly for inpatients with a larger menu than what we have on [POC] devices right now. I don't think [POC] testing will replace central lab testing in a very short time frame.

In particular to individual technologies, I think there could be a lot of impediments for implementing them as real clinical devices. We have seen tests which were FDA approved but were not widely implemented, and we have technologies which had only stayed in the research phase but failed to move forward to FDA approval and the clinical implementation. I think there are still a lot of roadblocks in the way for a specific technology to proceed from the research setting to translational stage, clinical validation, and eventually to regulatory approval, and then to clinical implementation. We summarized some of those roadblocks in the review article, so some of the key questions that people can use as a checklist is summarized in table 4 and we also have a diagram illustrating all the key points along the development pathway in figure 1 of the review, so I hope those will be useful resources for anyone going down the path from technology development to clinical implementation.

What would be very beneficial to the developers is if they could find a compelling clinical application in the very beginning. Before they move down a pathway to focus on the technology itself, they should know what clinical questions they are trying to answer and the clinical pathway associated with those clinical questions, as well as what the future vision of the pathway after incorporating the novel technology is like. So those should be mapped out clearly, and that's why in the review we also mentioned that it is important that the technology development team incorporates some lab expertise and physician expertise fairly early on in their efforts, so that they know exactly what a good clinical target is and what the current clinical pathway is like.

Artificial Intelligence is being used all across healthcare and you can't really escape it at the moment. How do you envision AI being implemented in clinical care? Do you think it will help? Do you think it will hurt? What are your thoughts?

There is a lot of research and discussion around using AI in healthcare. And there have been some pilot studies and experiments out there. In the diagnostic field we have also seen research published in the last few years using primarily pattern recognition as one strength of AI. We are currently leveraging some of this strength in our own research as well, using AI with computer vision-based algorithms for image recognition.

There is also potential for AI to be used for pattern recognition in the data. In the clinical chemistry lab, we generate a large amount of data associated with clinical samples. So drilling down to those data and using AI to recognize the pattern within the data would be another interesting application. It may be very difficult for one individual to recognize the pattern, but it may be much easier for a computer to recognize once you train them with good data source.

Another interesting question is: how do we use AI to recognize particular disease patterns for an individual when you have multiple different test results from that patient all at once? A clinical scientist, pathologist or clinician might be able to do that on an individual case basis. But if we want to do it on a larger scale, I think AI can be very helpful. Importantly, everything is dependent on you having some good data sets and good images, to train your algorithms first.