Learning the basics of blockchain

What is the scope for blockchain technologies?

“There is technology becoming available … I don’t claim to be an expert on it but the most obvious technology is blockchain” boasted Chancellor Philip Hammond at the Conservative Party Conference, when discussing the Irish border challenges of Brexit.

What problems can blockchain tackle?

Now, I also do not claim to be an expert in blockchain technology, but from what I understand, I did not think blockchain technology was going to help solve the Brexit Irish border issue. Nevertheless, what this revealed to me was that I was poorly informed as to what blockchain technology truly represented beyond the hype of cryptocurrency, and what problems blockchain technology was reasonably equipped to tackle. Certainly, within healthcare, reports often appear stating that medical records or genetic test results will be “put on the blockchain”, but what this actually represents remained rather opaque to me.

Figure 1: My prior understanding of blockchain technology (source)

Training in blockchain fundamentals

To up-skill myself — and help determine whether the hype associated with blockchain technology is justified — I enrolled on Coursera’s blockchain foundations course, created by ConsenSys Academy.

Having recently completed the course, I certainly feel better informed regarding blockchain technology and can now sensibly engage in discussion regarding its implementation, beyond the superficial knowledge I have regarding Bitcoin and Ethereum. The course does a good job at outlining the basic concepts of what a blockchain is, how and why it was created, provides an overview of the technical components underpinning the technology and, most importantly for me, provides a framework to establish whether or not a blockchain technology should be considered as a possible solution, for a given challenge.

Summarising blockchain utility

The first thing to appreciate is that a blockchain is only useful when a database is required. Furthermore, blockchains really only have an advantage over traditional databases when multiple users require “write” access to that database, and those users are either unknown, untrusted or in possession of conflicting interests. Even then, databases can accommodate these scenarios, typically through a trusted 3rd party. But when such an entity cannot be relied upon, a blockchain solution might be a possible option.

Figure 2: Framework outlined by Coursera

If a blockchain solution is required, there are a range of additional considerations that need to be addressed to determine whether a public or private blockchain is most suitable. Jeremy Millar, Chief of Staff at ConsenSys, has outlined an approach for this here.

Conclusion

Whilst the content of the course might not stimulate blockchain aficionados, it provided the necessary foundations to appreciate how blockchain technology can be applied and what limitations are currently present.

Assisted Intelligence: reflections from Nvidia’s GPU Technology Conference (GTC) 2018

This past week I attended Nvidia’s GPU Technology Conference (GTC) where I learnt more about the disruptive force of artificial intelligence within society.

Whilst it is easy to get distracted by the many exhibitions highlighting broad applications of artificial intelligence, I focussed on the intersection between artificial intelligence and healthcare, an area I have previously discussed at the Royal Society for Medicine’s meeting series (2015, 2016, 2017).

One thing I hadn’t appreciated before attending the event, was that Nvidia have been innovating within the healthcare domain for over 10 years, including partnerships with Siemens and GE. Having now experienced GTC I can now better appreciate why healthcare is a key vertical for Nvidia, given that the FDA has approved over 11 indications within 2018 that relate to artificial intelligence.

Reflecting on my GTC experience, I came away with a few key messages:

  1. Assisted intelligence: an increasing number of peer-reviewed articles have been published outlining how artificial intelligence compares against clinicians for the diagnosis of disease. Whilst there was reference to many of these studies throughout GTC, it was more common to hear how artificial technology can be used to assist the clinical decision process, rather than replacing it — this was discussed with reference to diabetic retinopathy.
  2. More-from-less: within the GTC keynote Jensen Huang, CEO of Nvidia, described how with the latest ‘real time rendering technology’, conspiracy theories regarding the space landing can be debunked. This was fascinating, and it appears that similar approaches are being applied within medical imaging, as outlined within the project CLARA update. This included the possibility that adverse reactions to contrast agents could be minimised by enhancing images that are generated using a far lower concentration of contrast media, and limiting radiation exposure through shorter capture times.
  3. DNA sequencing: there is potential to improve the DNA sequencing pipeline with deep learning, but we are not there yet. For example, Oxford Nanopore gave an impressive demonstration that explained how GPU technology can be applied to accurately read the A,T,C and G bases inferred from an electrical signal that is generated as a strand of DNA passes through a nanopore — the process that enables the DNA sequence to be read by the DNA sequencer. This is a remarkable use of deep learning technology, and has resulted in dramatic improvements in their technology. Beyond this specific application, Oxford Nanopore also outlined their vision to generate a device that consumers can afford, potentially enabling anyone to sequence anything within the world around them. This vision stretches well beyond the exciting opportunities reliable long-read sequencing affords to researchers studying the human genome, such as the ability to detect large structural re-arrangements, and capture areas of homology and highly repetitive regions that were challenging using established short-read methods.
    However, after attending one of the Deep Learning Institute’s hands-on sessions on variant classification for genomic variants, it was apparent that beyond base-calling, there are plenty of opportunities throughout the sequencing pipeline where deep learning may be applied to improve on current performance.
  4. Organisational benefit: beyond these clinical applications, GTC provided me with an opportunity to appreciate what artificial intelligence is, and what it is not. Understanding the scope of this (and any) technology is critical. Several speakers discussed the challenges data scientists encounter when trying to deploy artificial intelligence within an organisation, and often it related to a inappropriate expectations regarding the technology and the role of the data scientist within the process.

Recent Developments in Digital Health 2017

Yesterday I hosted the 4th annual Recent Developments in Digital Health conference at the Royal Society of Medicine. This was the 3rd year I organised and chaired the event.

It is a real pleasure to run these events and provide members of the healthcare community with insights into how healthcare can potentially be improved through digital solutions.

In previous years (2015 and 2016) we have had some incredible speakers, including: Mustafa Suleyman (Google DeepMind), Prof Euan Ashley (Stanford University), Prof Tony Young (NHS England), Martin Kelly (Health XL), Ali Parsa (Babylon Healthcare), Dr David Clifton (University of Oxford), Dr Kazem Rahimi (University of Oxford) and Alan Barrell (Judge Business School).

This year we heard from a series of experts within digital healthcare. This included: Sir Mark Walport (Chief Scientific Advisor to UK Government) who discussed blockchain technology within healthcare; Professor Martin Cowie (Imperial College London and E-health Lead for the European Society of Cardiology) regarding monitoring in heart failure; Dr Charlie Davie (Managing Director, UCLPartners and Digital.London) and Mr Shafi Ahmed (Royal London Hospital).

My talk focussed on the current digital health landscape. The slides to this talk are available on slideshare and here.

A Handheld Convergence

Predictions suggest 50 billion Internet enabled mobile devices (IEMDs) – including smartphones, tablets and wearable devices – will be in use globally by 90% of individuals over the age of 6 years old by 2020. Already 61% of Americans own a smartphone.

An icon of our generation, and emblematic of the unrelenting advances that have occurred in transistor technology, IEMDs have the potential to upgrade current healthcare surveillance and delivery strategies worldwide. Furthermore, these initiatives are estimated to generate $6 billion in savings for companies in the United States, and partly explain why the mobile healthcare (mHealth) industry has been supported with financial investment despite a paucity of clinical evidence to date.

Although not a recent development, telemedicine is only now beginning to be recognized as an appropriate vehicle for healthcare surveillance and delivery. Benefitting from the increased penetrance of mobile phones throughout society, public health organizations have been capable of closely tracking communicable disease outbreaks through geographically tagged text-messages. Such strategies were successfully implemented in Haiti for cholera during 2012 and throughout Western Africa for ebola in 2014. However, it was not until 2015 that telemedicine was authenticated for healthcare delivery when the largest insurance company within the United States, UnitedHealth, announced that it would reimburse physicians for virtual-doctor appointments. This came in response to the growing popularity of numerous online portals, such as Doctor on Demand, which connect patients with thousands of medically qualified staff whenever required. It is notable however that there remain strict regulations restricting physicians from providing advice to individuals outside the federal state in which they hold medical licensure. Such policies impinge on telemedicine’s potential within the United States, and mean successful telemedicine strategies observed throughout the developing world are unlikely to be replicated in the USA. Take for instance, Narayana Hrudayalaya, a specialist cardiac unit based in Bangalore, India. The telemedicine program in Narayana Hrudayalaya extends across 150 telemedicine centers across Asia and Africa and facilitates specialist input into the care of hundreds of thousands of patients with cardiovascular diseases, irrespective of geographical location.

The current generation of IEMDs go far beyond the ability to simply text-message and video-network on demand. Vastly improved computational power and the ability to effortlessly connect with additional devices, has opened up a large armamentarium of possible mHealth interventions for patients, doctors and researchers.

Despite a poorly defined regulatory framework, a surfeit of medically-related apps have been developed. Providing an alternative to traditional disease prevention and chronic disease management strategies, IEMDs have enabled patients to take control of their health, should they wish. From tracking medication compliance, to the recording of an endless stream of physiological data – such as blood pressure, heart rate and blood glucose — transmitted in real-time via wearable sensors, a whole range of possibilities exist. And consequently those patients who commit to this approach appear more empowered, by generating their own data and questioning the paternalistic dogma regarding ownership of medical records. For doctors, IEMDs become the go-to-guide for hard-to-reach diagnoses, guidance or support. Furthermore, by combining IEMDs with adjunctive hardware a smorgasbord of diagnostic tools emerge, digitally recreating routine tools such as otoscopes and opthalmoscopes, and providing handheld equivalents of expensive equipment such as ultrasound. Point-of-care testing for biochemical and microbiological information, traditionally generated in laboratories, can also be derived following advancements in microfluidigm technology. Such versatility has enabled life-saving procedures to be performed in resource-deplete settings, from screening for obstetric and neonatal risks during pregnancy to diagnosing deadly parasites in blood. The current generation of IEMDs represents the swiss-army knife of clinical medicine, with the greatest value being seen within the developing world.

But ironically the individuals most likely to benefit from mHealth strategies often cannot access this technology. Affordability of IEMD is improving, with devices harboring most features present on top end models now available for ~$35, making acquisition of an IEMD possible, if not for patients for healthcare workers. However, the far larger issue is bringing over 58% of the world’s population, ~4.2 billion people, online. Efforts from Google and Facebook are attempting to resolve this issue, providing Internet access to previously neglected geographical areas through innovative solutions. These efforts so far have been remarkably successful, with estimates suggesting that ~10% of individuals are being lifted out of poverty as a direct result of these interventions.

Beyond the direct benefits for patients and doctors, IEMDs also provide exciting potential for the way clinical trials are performed. Escaping from the contrived one-off data points that are convention within clinical research, Apple’s ResearchKit enables users record their own data seemlessly, either manually or via a sensor. Over 11,000 individuals signed up for Stanford’s MyHeart Counts within the first 24 hours of launching, a cohort population that would usually take over a year to acquire from 50 medical centers. Now ResearchKit is open-source and having overcome early ethical concerns regarding patient recruitment, provides a platform for investigators to recruit from the hundreds of millions of iPhone users worldwide and silently monitor their activities accurately, overcoming challenges previously felt in traditional clinical trials. And whilst the average iPhone user may not necessarily represent the typical patient a clinical trial would recruit, the sheer volume of data available through ResearchKit makes this approach a game changer for clinical research. With such rapid advancements it is foreseeable that smartphone based genome sequencing will soon be available, a venture that could yield further insight into the composition of human health, but will no doubt face multiple legal and ethical obstacles.

With so many developments within telemedicine and mHealth it is easy to get overwhelmed by the hype. Importantly, we need to remain guarded until the efficacy of the latest generation of mHealth interventions is formally evaluated in well-constructed clinical trials. Undoubtedly there is a huge opportunity to revolutionize healthcare surveillance and delivery globally.