Can Artificial Intelligence Demystify Medical Underwriting?

Industry transparency and signposting

In fact, in 2020, the UK’s financial regulator ruled that all firms that sell travel insurance must signpost (openly direct) customers in questionable health to a directory of firms specialising in covering pre-existing conditions. And there are a good number: some covering any or all pre-existing conditions (even ones under treatment), others offering limited cover even for terminal conditions, some requiring back up from GPs, but all of them dependent on precise, complete, verifiable medical information provided by the applicants, frequently with the client’s GP’s verification, and if necessary, a full health record. In effect, full medical underwriting.

In the US, there is less generosity for the afflicted. According to one of the leading comparison sites, Squaremouth, 65 of the 94 policies on its website can offer pre-existing coverage. But such coverage relies on those conditions having been quiescent (no symptoms, treatment, testing, change of medication) over the previous few months or years – this varies. After that, the pre-exes may be covered. But in the US, medically underwritten policies for travel insurance (as opposed to long-term expat cover) are relatively uncommon.

In Canada, medical underwriting is the mainstream for any but the young ‘invincibles’. But this entails completing highly detailed medical questionnaires that require yes or no answers to two or three pages of queries such as: “In the last 24 months have you been diagnosed with or had an investigation, medical consultation, been prescribed medication or received treatment for: a) Digestive tract disorder?; b) Internal conditions? And for a definition of Internal conditions, look to page…etc.” Sometimes complicated, often challenging for a layman, but always with an advisory to ‘check with your physician’ if they are defeated by the question; an advisory rarely followed.

Technology allows for accurate assessment

A considerable advance in assessing the risk of pre-exes has been the travel risk rating tool developed by Verisk Analytics to help customers self-declare their medical conditions over a more individualised, almost proactive platform. According to Lara Suttie, Director of Business Development at Verisk: “Before the pandemic, when there were no travel restrictions, Verisk was responsible for supporting the successful completion of more than 21 million risk assessments per year with its global travel insurance partners, including brokers, distributors, banks and price comparison websites.”

She adds: “While Verisk does not currently have a similar presence in the US travel market as in the UK, we are currently evaluating needs and potential solutions relative to the US market. The US has typically been focused on cancellation cover rather than medical expenses.”

The tool is primarily used during an online sales journey via a direct insurer website or a price comparison website. In this scenario, the applicant would select high blood pressure and diabetes within the risk rating tool. This would present the applicant with a series of dynamic questions designed to ascertain the relative risk of those pre-existing conditions. Once answered, the responses will generate risk scores that can be applied to the insurer’s own underwriting rules, presenting the applicant with coverage and pricing options. And though the tool ‘does not factor in variables such as age or gender of an applicant, where appropriate, the insurer could include this in their own underwriting matrix’, added Suttie.

Initiatives such as these certainly can add a layer of objectivity to the use of a client’s private health profile and do it fairly and virtually instantaneously. But relying on an applicant to be objective in assessing one’s own frailties engenders a huge leap of faith and puts the patient/client in an untenable conflict of interest knowing that every yes response is going to run up the fee. And though honesty may be the best policy, it can be costly.

Can travel insurers take heart from the fact that automated medical underwriting is already working, and working effectively for other lines of insurance? Munich Re Automation Solutions Ltd announced recently that Storebrand Livsforsikring AS has selected ALLFINANZ NOVA to automate its life insurance underwriting process. NOVA, a cloud-based platform from Munich Re Automation Solutions, will help make the application process easier and quicker. “As well as reducing the application processing time by an estimated 65 per cent, ALLFINANZ NOVA has allowed Storebrand to simplify how they ask for health information, helping increase the rate at which applicants finish an application. Storebrand has also implemented the Underwriter Workbench and Insight Data Analytics services from Munich Re, analysing the questions that are fed through NOVA to identify ways to increase the straight-through-processing rate,” said Munich Re. Heidi Skaaret, Executive Vice President, Head of Retail Markets at Storebrand, added: “A key driver for our partnership with Munich Re is improved customer satisfaction and ensuring they have a quick and easy experience when applying for life insurance.  Since incorporating NOVA into our systems, we have seen a solid increase in customer satisfaction, and we have already doubled the rate of automation in our underwriting process.” 

Artificial intelligence could revolutionise the insurance industry

Enter the engineers of artificial intelligence and data analytics providers, who are reshaping the information landscape available to underwriters, claims managers, assessors, actuaries, any who are  tasked with balancing the risk of a given applicant’s health profile, with a price that justifies a sale.

At an international innovations symposium held in Toronto in 2019, Melissa Carruthers, Senior Manager for Strategy Consulting Practice for Deloitte, emphasised that life and health insurers (including travel insurers) have for too long been encumbered by corporate legacy systems and they needed to accelerate their efforts to leverage advances in artificial intelligence, digitalisation, data analytics and the output of emerging insurtechs to better understand the need and expectations of their customers.

In a report predicting the impact of AI on the future of insurance, global financial services firm McKinsey & Company emphasised that ‘artificial intelligence (AI) has the potential to live up to its promise of mimicking the perception, reasoning, learning and problem-solving of the human mind’. “[And] in this evolution, insurance will shift from its current state of ‘detect and repair’ to ‘predict and prevent’, transforming every aspect of the industry in the process.” The report goes on to warn: “The pace of change will also accelerate as brokers, consumers, financial intermediaries, insurers and suppliers become more adept at using advanced technologies to enhance decision-making and productivity, lower costs, and optimise the customer experience.”  (From Insurance 2030 – The impact of AI on the future of Insurance, authored by Balasubramanian, Ari Ramnath Libarikian, and Doug McElhaney.)

In respect to insurance underwriting, Kate Brown, Senior Vice President of Swiss Re Corporate Solutions, pointed out: “Human beings just don’t have the capability to deal with the massive amounts of data we are getting now. That’s the competitive advantage of AI, being able to use this data deluge to make better decisions.”

SCOR, one of the world’s largest reinsurers, has teamed up with Verisk in developing EHR Triage Engine Plus (launched in 2020), a system that can scan and distill thousands of pages of a patient’s electronic health records and deliver an easy-to-read one-to-two-page summary of key risk elements along with a recommended underwriting classification. And it does it in a matter of minutes. This is not to be confused with Verisk’s travel risk rating tool, which is used for a different purpose, but it is nonetheless a proof of concept of which travel and international health insurers should take note.

Paul Donnelly, Vice President EMEA of Munich Re Automation Solutions, told ITIJ that increasing adoption of automated underwriting is bringing a host of benefits to insurers and customers alike: “For consumers, automated underwriting guarantees a smooth application process by reducing unnecessary questions and leading to almost instantaneous results. For insurers, it reduces the amount of time spent sifting through copious amounts of data by identifying specific patterns, which reduces friction for the customer.

“Globally, we also notice a particular rise in adoption in industry sectors where the underwriting has traditionally been more complex and onerous. A clear example would be international private medical insurance (IPMI) which requires a rigorous underwriting process, similar to that of life insurance. For IPMI insurers specifically, automated underwriting will greatly speed up the underwriting process, and enhance the customer experience.”

Has data harvesting of health records gone too far?

How far has this harvesting of health data gone and how much confidence do medical providers have in the technologies?  The Wall Street Journal reported in 2020 that several major US hospitals have teamed up with Microsoft, IBM and Amazon to scan millions of patient health records purportedly to help them identify clues and develop algorithms to improve diagnostics via AI and gain clues to quicker, more accurate, and cost-effective treatment modalities.

This kind of attention by Big Data has predictably prompted health privacy concerns among regulators globally. The California Consumer Privacy Act was passed in 2020 primarily for the purpose of expanding the rights of its residents to know exactly who has received their health information and how it’s being used. The European Union’s (EU’s) General Data Protection Regulation (2018), designed to increase data privacy for EU citizens, has a policy of levying fines of up to four per cent of revenue or £17 million, whichever is the greater, for entities not meeting the regulations.

And as recently as August 2021, the UK Council for Science and Technology released a white paper calling for the adoption of technology in the UK health system to enable ‘transformative change that will benefit the health and wellbeing of the UK and promote economic growth’, but also cautioned that technology alone cannot overcome the inequalities that lead to disparities in health outcomes and that digital tools for health should be accessible to all, or risk exacerbating health inequalities as a result of a ‘digital divide’. 

The use of electronic health records and other deep pools of personal information once thought to be impregnably private, is going to be argued long and hard as AI developers extend their beachheads into industries and activities far and wide. Back in 1999, Scott McNealy, Chairman of Sun Microsystems fatefully intoned the warning: “You have zero privacy anyway. Get over it.”

If only it were that simple.