Disruptive technology in business insights – for better or worse?
Technology in the medical world has seen huge advances over the past decade, as Professor Shafi Ahmed, Harvard Medical School and Bart’s Medical School, demonstrated at the BHBIA Annual Conference 2023. From 3D printed replacement joints and bionic arms to AI assisting with clinical trial data gaps and participant coaching to robotics in surgery.i,i,iii,iv
Professor Ahmed inspired attendees with examples of augmented reality (AR) and
virtual reality (VR) already in use for patient support, training and surgery,
including his personal experience of leading VR training for 55,000 surgeons in
one day. Attendees saw a glimpse into the futuristic plans for NHS digital
transformation with the potential for full remote monitoring for patients with
chronic illness in their own home - the ‘Housepital’. We live our lives
digitally with the exponential adoption of new technology.
In business insights (BI), much of the focus of this new technology has been on the role of Artificial Intelligence (AI) in the analysis of complex datasets and how access to new social media platforms, personal-use apps and gamification techniques can help us derive deeper insights into attitudes and behaviours. More recently, the topics of large language models (LLM) and how the pharma industry might use ChatGPT have come to the fore.
The recent pace of change has been radical and accelerated during the Covid pandemic, making it hard for some to keep up with.v, vi With this in mind, the BHBIA raised questions about how we as a BI community are keeping pace and how we can use innovation to deliver high-value insights that can positively impact our nation’s health. Conference presentations and workshops probed technology disruption with varying degrees of positivity and caution. Is the rapid evolution seen in other areas of healthcare or the technology disruption in other BI industries being mirrored in healthcare BI? Or will a slower revolution of existing ideas and methods make the biggest difference?
The adoption of digital technology in BI
Discussions at conference concluded that pharma is behind the curve in adopting new technologies. We have compliance requirements to adhere to and project-by-project, siloed adoption of technology can hamper enterprise-wide adoption. In pharma, we don’t have the same pressures as consumer brands which are adopting new and innovative approaches to on-demand consumer insight.
Despite the risks and fear of introduction, we have a burning platform to overcome. It’s important to remember that our customers and their patients interact digitally with consumer brands and life lives outside of medicine and healthcare. As they experience the benefits of digital technology in other aspects of life, they are demanding we improve our online relationship with them too.
The adoption of technology in BI must be viewed in two parts: Firstly, understanding the business need and collecting the data to address it, and secondly, aggregating data and eliciting insight into the data to tell a story. Technology can help advance both parts of the equation and as such is being embraced by the BI industry, expanding the vision of fieldwork with the availability of bigger and potentially better data, for example. It has notable applications in reaching more people and different cohorts of the population through innovative platforms and in measuring performance and aiding statistical analysis, but it’s the analysis of this data that requires the most caution.
Big data
In personal healthcare,
we have seen a boon in data availability and collection vi, with individual data originating primarily
from apps and wearables.vii,viii,ix The development of the personal health record has also contributed to an
explosion of individual data which has the potential to be mined into large
data sets and analysed using AI, and could provide insights into disease risk
and development.x The compliance and regulation around the storage, use and transfer of this
personal data means more stringent guidance needs to be met across borders to
ensure maximum availability and use.xi
The Collection of large physiological data omits the
important psychological and behavioural data which informs many of the
decisions made by pharmaceutical companies for their brands. It’s arguable
whether large data sets can contribute much to this, as physicians’ decision-making
tendencies can be questionable, according to one study conducted in 2010. A
study of 10 physicians found that of their 1,188 decisions over the course of a
week related to prescribing specific brands, 15% were down to ‘instinct’, 37%
due to anecdotal experience and 33% for ‘other reasons’.xii This data set of over 1,000 decisions, gave little insight into how a brand
team could tap into their decision-making process to focus their marketed brand
attributes.
New digital technologies, customer experience software and online communities have allowed faster data collection. Coupled with compliant cloud software and computational tools, this has enabled the collection and storage of larger data sets and data collection from market research participants.
The outstanding questions concern the keys to leveraging the amassed data sets. The caution (which we’ll come onto later) is focused on AI’s ability to dig into data to provide meaningful, non-biased insight, the skills for which currently lie very much with humans.
Automation and Data Quality
On the flip side of
access to more data and the ability to house larger data sets is the quality of
the digitally produced data, which for some, has proven questionable.vi AI and LLM have opened a potential panacea
for the analysis of big data to assist in providing a narrative and insight
more rapidly than any human could do. Yet there are critical watch outs if we
omit human interpretation from analysis.
Online data collection,
for example, through social media throws a wide net out to an audience. There
is inherent caution required in assessing both the quality of the data
retrieved and the safety of the data.xii
Biases are widely accepted to be inherent in AI as we rely on humans to input
the code, with biases favouring the white male.xiv,xv,xv,xvii This leads to claims that even with digitally created data there is an innate
gender and ethnic bias in algorithm modelling, training and usage. These biases
need to be viewed with a socio-cultural perspective as data could reinforce
existing biases rather than challenging them.xvii
Market research is an art as well as a science and we need to embrace connection with respondents and engage with the technology to enhance statistical analysis.vi Another way to put this is that without human interaction, thought and context, data is just data and sophisticated analysis to provide insight requires human intervention to avoid the potential for erroneous data collection and misuse of platforms and data.xix
Access to wider audiences
The acceleration of
digital platforms for online interviews and focus groups during Covid-19,
allowing participants with health conditions or those who are time-poor equal
involvement, has been one of the positive changes in recent years. Adopting
newer digital platforms and methodologies has enabled access to larger and more
varied participant audiences.
Uday Bose, Head of
Human Pharma Global Go-To-Market and Business Steering – Boehringer Ingelheim
Ltd, used his keynote speech to remind us that we need to focus on the Customer
Experience, with people being the most important asset in healthcare BI. This
was in reference to customer insight, where it is important to meet with people
and learn their experience first-hand, and the real need for human intelligence
to layer on top of data to overcome data bias and realise the ongoing issues
with bias in the pharmaceutical industry.
As shown at conference, notable successes include improved speed of completion by utilising software and apps and improved access to hard-to-reach audiences. This has been achieved through the implementation of tools which aim to reduce completion time and make involvement in market research more enjoyable. Speeding through market research questions on apps including Fast Open Ends or Rapid Sort, a ‘Tinder-like’ preference-based questionnaire also helps gain accurate ‘gut reaction’ responses from participants. Gamifying the entire process of fieldwork with participants via online worlds can help engage hard-to-reach audiences and keep them engaged in market research, thereby eliciting important insights.
Rate of change
With technology advancing at such an astounding speed, an important question is, whether transformation must be rapid to be considered true innovation. ChatGPT has changed the face of AI and amplified the promise of new digital technology in the short time since its launch in Autumn 2022. However, technology can be quickly utilised by society while also potentially taking a generation to embed into our consciousness.xx
The exponential change leaves many users struggling to keep up, for example, with frequent technology platform updates leading to a near-constant need for learning and adaptation. Compliance and medical teams’ involvement in ensuring advancing technology remains compliant means it’s more than just the teams directly using the technology that need to keep pace, but also those who are approve and monitor its use. It could be argued that in healthcare, these additional layers contribute to the slower uptake of new technology.
Does this pace of change automatically mean exciting innovations? It’s important to note that agencies pursuing new and exciting technology in BI will leave aside as many innovations as they adopt as many are rejected following testing. Rigour is paramount to the successful advancement of technology when we are dealing with people’s health. It is worthwhile, therefore, to take the time to think about what we are trying to achieve with technology in BI, what the potential impact of our work is and how we want the industry to change. We must always follow the ‘Why?’ behind innovating and then we can pursue technology that has a measurable positive impact on the world.
Managing change and digital transformation
As we’ve seen, the speed of change and a perceived need to keep up to date is often overwhelming. When planning for technological change, we learned at conference that organisations can unwittingly set unnecessarily harsh KPIs, forgetting that technological change often needs existing methods and protocols to be adapted or replaced before the ‘new’ can be embraced.
Progress can happen in steps and does not need to happen all at once. Appreciating that incremental change can achieve great results over time and will motivate the same number of people to change is important. Taking the long view and working with champions to plan and ensure adoption, acceptance and action all take place could achieve better results than attempting to adopt every innovation.
As with aspects of digital advances, technological change can only be successful if the human brain and communication are integrated into the process. Anecdotal feedback leads to the premise that technological change is unlikely to be successful without strong leadership and communication.
What next?
Digital transformation is continuing to disrupt the status quo. However, the acceleration seen during Covid has been dampened somewhat as people returned to the office and face-to-face interactions commenced again. Digital solutions can offer new and exciting avenues for introducing and applying digital approaches to fieldwork and accessing market research participants. But when it comes to analysis, caution is the watchword. It is imperative for the continued success of healthcare BI that innovation is not automatically assumed to mean improvement.
Transformation does not need to happen overnight. We do not need to rush into the next promise of innovation or adopt the next ‘shiny new’ platform, but can transform with a keen eye on compliance, risk reduction and acceptance for adoption while benefiting from the advances newer technologies bring to add value to our industry.
The BHBIA is working to open ways to share experiences and best-practice from thought leaders in digital transformation, including relevant learning and development opportunities in line with our new Skills Framework. The BHBIA Ethics & Compliance Committee is working on two new sets of guidance, one on social media and one on AI. These will cover some of the compliance challenges that are inherent in adopting new technologies, including guidance on working with Digital Opinion Leaders and advice from regulators and ChatGPT examples. These will be available to members in autumn 2023.
References
i University of Bath. (2021) Personalised 3D printed knee implant could bring relief to thousands of arthritis sufferers. Bath.ac.uk. https://www.bath.ac.uk/announcements/personalised-3d-printed-knee-implant-could-bring-relief-to-thousands-of-arthritis-sufferers/
ii NHS England. (2022, November 5). NHS England» NHS offers life-changing bionic arms to all amputees. England.nhs.uk. https://www.england.nhs.uk/2022/11/nhs-offers-life-changing-bionic-arms-to-all-amputees/
iii Harrer, S., Shah, P., Antony, B. J., & Hu, J. (2019). Artificial Intelligence for Clinical Trial Design. 40(8), 577–591. https://doi.org/10.1016/j.tips.2019.05.005
iv Leprince-Ringuet, D. (2018, September 4). Meet Versius, the surgical robot about to take aim at your organs. WIRED UK; WIRED UK. https://www.wired.co.uk/article/surgical-robot-uk-versius
v Dean. E., Haney, C., & Khatri, L (2022) Market research is the foundation for experience transformation in 2022. 2022 Market Research Trends Report. Qualtrics. https://www.qualtrics.com/uk/ebooks-guides/2022-market-research-trends/
vi Furtner, D., Salil Prakash Shinde, Singh, M., Chew Ming Wong, & Sajita Setia. (2021). Digital Transformation in Medical Affairs Sparked by the Pandemic: Insights and Learnings from COVID-19 Era and Beyond. 36(1), 1–10. https://doi.org/10.1007/s40290-021-00412-w
vii Apple (United Kingdom). (2017). Healthcare - Apple Watch. Apple (United Kingdom). https://www.apple.com/uk/healthcare/apple-watch/
viii Oura Ring. (2023). Oura Ring: Accurate Health Information Accessible to Everyone. (2023). https://ouraring.com
ix Whoop. (2023, January 31). WHOOP 4.0 | Next Generation Fitness and Health Monitoring. https://www.whoop.com/en-gb/membership/strap/
x Hiba Asri, Hajar Mousannif, Hassan Al Moatassime, & Zahir, J. (2020). Big Data and Reality Mining in Healthcare: Promise and Potential. 122–129. https://doi.org/10.1007/978-3-030-51935-3_13
xi European Council. (2023) Data act: member states agree common position on fair access to and use of data. Europa.eu. https://www.consilium.europa.eu/en/press/press-releases/2023/03/24/data-act-member-states-agree-common-position-on-fair-access-to-and-use-of-data/
xii Darst, J. R., Newburger, J. W., Resch, S. C., Rathod, R. H., & Lock, J. E. (2010). Deciding without Data. 5(4), 339–342. https://doi.org/10.1111/j.1747-0803.2010.00433.
xiii Ventola, C. L. (2014). Social media and health care professionals:
benefits, risks, and best practices. P & T: A Peer-Reviewed Journal
for Formulary Management, 39(7), 491–520.
https://www.ncbi.nlm.nih.gov/
pmc/articles/PMC4103576/
xiv Alba, D. (2022, December 8). OpenAI Chatbot Spits Out Biased Musings, Despite Guardrails. Bloomberg.com; Bloomberg.
xv Bloomberg. (2022) ChatGPT Open AI Chatbot is Spitting Out Biased Sexist Results https://www.bloomberg.com/news/newsletters/2022-12-08/chatgpt-open-ai-s-chatbot-is-spitting-out-biased-sexist-results
xvi Stack Overflow. (2022, December 5). Temporary policy: ChatGPT is banned. Meta Stack Overflow. https://meta.stackoverflow.com/questions/421831/temporary-policy-chatgpt-is-banned
xvii Vock, I. (2022, December 9). ChatGPT proves that AI still has a racism problem. New Statesman; New Statesman. https://www.newstatesman.com/quickfire/2022/12/chatgpt-shows-ai-racism-problem
xvii Ferrer, X., Van Nuenen, T., Such, J., Coté, M., & Criado, N. (2020). Bias and Discrimination in AI: a cross-disciplinary perspective. https://arxiv.org/pdf/2008.07309.pdf
xix Carmel, O. (2023, June 22). Council Post: Combining And Leveraging
The Strengths Of AI And Human Consultants. Forbes.
https://www.forbes.com/sites/forbesbusinesscouncil/2023/06/20/
combining-and-leveraging-the-strengths-of-ai-and-human-consultants/
xx Palandrani, P. (2020, February 10). A Decade of Change: How Tech Evolved in the 2010s and What’s In Store for the 2020s. Global X ETFs. https://www.globalxetfs.com/a-decade-of-change-how-tech-evolved-in-the-2010s-and-whats-in-store-for-the-2020s/