Apart from being the biggest search engine, Google is also a key researcher in the tech space. In its recent posts, Google has put forward the research work it has been performing in various fields, including health technology.
Google’s research in health has earned Alphabet a spot among the five most impactful corporate research institutions on the Nature Impact Index from 2019 to 2022. With its focus on artificial intelligence (AI) in healthcare and applied machine learning (ML) programs, Google Research is making a significant impact in health and life sciences publications.
What’s more? Google is taking a collaborative approach to innovation in health and is committed to earning the trust of healthcare professionals and patients alike. In this article, we’ll explore the three main themes that emerged from Google’s health research in the past year:
- The criticality of technology partnerships
- The shift towards mobile health
- The use of generative ML in health applications.
Read along to discover Google’s latest breakthroughs in health tech.
Criticality of technology partnerships
In healthcare, responsible innovation takes time and investment to make a real impact. But what if we could accelerate the journey from research to human impact? That’s where technology partnerships come in.
By partnering with established med-tech companies, ML technologies can be safely deployed to patients with less delay. These partnerships have yielded Google impressive results, like:
- Google’s collaboration with iCAD to integrate ML technology into mammography software to improve breast cancer screening.
- A similar partnership with RadNet’s Aidence has led to advancements in low-dose CT scans for lung cancer screening.
- In genomics, Google’s partnership with Pacific Biosciences resulted in a new genome sequencing tool powered by ML techniques.
- It’s collaboration with Stanford University to diagnose genetic diseases by combining sequencing technologies and ML.
A combination of cross-industry partnerships and scientific openness plays a crucial role in achieving the benefits of human-centred AI in healthcare and medicine. These partnerships should be seen as a complement to, rather than a replacement for, open research and open software that drives progress in the health field.
Shift towards mobile health
Google’s healthcare research is taking a new approach to patient care by shifting from centralised clinics to mobile ML-powered solutions that can be brought to patients. By utilising smartphone sensors, it has already made strides in measuring heart rate and identifying skin conditions.
However, new sensing technologies can expand the potential for mobile health capabilities. Google’s one of the most exciting projects in this area is the development of mobile maternal ultrasound for under-resourced communities.
With complications from pregnancy and childbirth disproportionately affecting low-income populations, ML-powered mobile ultrasounds can bring obstetric ultrasound to those in need. With a low-cost ultrasound device and smartphone, healthcare workers can detect problems before they become critical.
Google’s research has already shown promising results in estimating gestational age and foetal malpresentation with accuracy comparable to professional sonographers. In the future, this technology could allow for on-site screening and triage of various medical issues, extending care to millions with minimal training.
This is just one example of the exciting possibilities that lay ahead in the shift towards mobile health.
Generative ML in health
Generative ML is an emerging technology poised to complement pattern recognition systems in healthcare. It has shown promising results in creating synthetic data for privacy preservation and simulating clinical trials.
It also has the potential to personalise medicine. For example, Google has
- Explored the use of generative image models in data augmentation
- Generated realistic electronic medical records for research purposes
However, Google must proceed with caution to develop generative ML systems, as it cannot underestimate the gravity of impacting real people with real health issues.
Recently, Google published a pre-print on large language models (LLMs) and the encoding of clinical knowledge, which introduced their own research-grade medical question answering LLM, Med-PaLM.
Med-PaLM has outperformed other systems on medical exam questions, but receiving a ‘passing’ mark falls short of the safety and accuracy required for real-world medical question answering. Further research and partnership are necessary to fully develop the potential of generative ML in healthcare.
Google’s research in Health: What’s the way forward?
Google’s research in health has shown great potential for improving healthcare access and outcomes. While mobile ML solutions and new sensing technologies can greatly benefit under-resourced communities and extend access to care for millions, there are concerns about privacy breaches and bias in AI decision-making.
Google’s plans for health research include increased collaboration and partnership between academic, med-tech, AI-tech and healthcare organisations. It will allow for more efficient mapping of the arc from innovation to impact in AI for healthcare and potentially expand the role of smartphones and mobile sensors in providing care beyond what we currently imagine telehealth to be.
As technology and our understanding of human health continue to evolve, let’s see what new possibilities and challenges can arise.
What do you think? Let us know in the comments.
