Optum: Machine Learning at the center of Health Care

A brief analysis of Optum and its capabilities and challenges towards developing and implementing machine learning in the Health Care industry.

Health Care has always been a critical and sensitive industry for most countries. Trends such as the increase in life expectancy and the development of more sophisticated and expensive drugs are continually raising the stakes, particularly when met with inefficient health systems (where health results per dollar spent are sub-par): a Centers for Medicare & Medicaid Services study indicates that the USA will devote 20% of its GDP to Health Care by 2026 (in contrast to 18% in 2016)1.

At the center of the Health Care system is Optum. Part of the United Health Group, Optum is a leading information and technology-enabled health services business with over $91 billion dollars in yearly revenues2. The company’s critical task is to increase efficiency in this complex sector by providing actionable insights and connecting information and key players (e.g. by measuring the true impact of treatments or managing complex delivery and payment systems). Having access to an extremely large database of 100 million lives clinical data, 6.5 billion medical procedures, 18 billion lab results and 6 billion diagnoses3, the company is in a privileged place to tackle the Health Care efficiency issue.

At its core, machine learning is about using processing power to create algorithms that analyse and understand the relationship within large quantities of data in reasonable time4. Considering Optum’s business model and the amount of data available, it is no surprise that machine learning represents a major opportunity for the company. One could even argue that the future of Optum and machine learning are deeply intertwined, as Artificial Intelligence (AI) is seen as the probable saviour of the collapsing Health Care system. Moreover, the recent fast paced evolution of the core components that allow machine learning progress – more data, better algorithms and more powerful processing power – reveal that there is indeed opportunity for this technology today4.

As a company that mainly creates value through insights and implementation of technological solutions to corporate clients in the Health Care industry, Optum has taken several steps that enable the use – and development –  of machine learning. First, the company sees using data and intelligence to provide insights as a core competence. That means that the company’s organization and structures are designed around developing such competence, with a Business Unit dedicated to creating insights (OptumInsight) and positions that are empowered to drive value creation in this field (e.g. CIO at Optum and Chief Analytics Officer at Business Unit level). Second, Optum has also strived to set up initiatives and partnerships to advance the use of innovative solutions in the Health Care industry (included but not limited to machine learning). In 2013 the company launched Optum Labs, a collaborative research and innovation center co-founded with the Mayo Clinic5, but also featuring partnerships with important institutions in the Health Care sector, such as the Harvard Medical School6. Optum Labs CEO stated that they are “creating the infrastructure to use AI to make better decisions with data” and that “Ultimately, AI will give us better, faster insights”7. Third, Optum is also seeking innovation through an acquisition strategy. By the end of 2017, the company announced Optum Ventures, a $250 million fund to invest in small companies that use data and insights to help improve the Health Care system8. Finally, Optum – along with its parent company, United Health Group – is seeking to vertically integrate the industry9 (e.g. acquisition of DaVita Medical Group10), giving the company more access to data, critical for the machine learning development.

As seen above, Optum is taking important steps to lead the development and implementation of machine learning into Health Care. However, as the company grows bigger and more complex, the risk of losing focus in what matters also increases.  With competitive M&A movements (e.g. CVS and Aetna, creating an extremely valuable company is terms of reach and data11) and new players entering the market (e.g. Amazon positioning itself in several parts of the Health Care value chain12) the business is becoming more competitive and the line between the capabilities of data intensive companies of several industries (e.g. Amazon, in e-commerce) and Health Care data intensive companies is fading.  Given this changing landscape, Optum must be increasingly more positioned as a data and technology company to guarantee its competitive advantage. That means looking not only into Health Care, but also digging deeper into stand-alone prediction and machine learning technologies. An effective way of doing so would be to create a spin-off company focused on the development of machine learning and other AI abilities or acquiring players in such industry.

Going forward, besides creating advanced algorithms and having access to large data sets, what other competences will Optum need to develop to deliver value using machine learning?

Word count: 780.



  1. Centers for Medicare & Medicaid Services. (2018). CMS Office of the Actuary releases 2017-2026 Projections of National Health Expenditures. [online]. Available at: https://www.cms.gov/newsroom/press-releases/cms-office-actuary-releases-2017-2026-projections-national-health-expenditures [Accessed 11 Nov. 2018].
  2. Business Wire. (2018). UnitedHealth Group Reports 2017 Results Highlighted by Continued Strong and Diversified Growth. [online], Available at: https://www.businesswire.com/news/home/20180116005747/en/UnitedHealth-Group-Reports-2017-Results-Highlighted-Continued [Accessed 11 Nov. 2018].
  3. United Health Group. (2017). 2017 Annual Review. [online] Available at: https://www.unitedhealthgroup.com/content/dam/UHG/PDF/investors/2017/UNH-2017-Annual-Review.pdf [Accessed 11 Nov. 2018].
  4. Brynjolfsson and A. McAfee. (2017). What’s driving the machine learning explosion?.Harvard Business Review Digital Articles (July 18, 2017).
  5. Optum. (2013). Optum, Mayo Clinic Partner to Launch Optum Labs: An Open, Collaborative Research and Innovation Facility Focused on Better Care for Patients. [online]. Available at: https://www.optum.com/about/news/optum-labs.html [Accessed 11 Nov. 2018].
  6. Optum Labs. (2018). Our Partners. [online]. Available at: https://www.optumlabs.com/about/partners.html [Accessed 11 Nov. 2018].
  7. Optum Labs. (2018). Harnessing artificial intelligence (AI): Machine learning is key to the future of health care. [online]. Available at: https://www.optumlabs.com/work/artificial-intelligence.html [Accessed 11 Nov. 2018].
  8. United Health Group. (2017). Optum Announces $250 Million Fund to Invest in Next Generation of Health Care Innovation. [online]. Available at: https://www.unitedhealthgroup.com/newsroom/2017/1128optumventures.html [Accessed 11 Nov. 2018].
  9. Byers, Jeff. (2018). Optum a step ahead in vertical integration frenzy. [online] Healthcare Dive. Available at: https://www.healthcaredive.com/news/optum-unitedhealth-vertical-integration-walmart/520410/ [Accessed 10 Nov. 2018].
  10. Treichert, Erica. (2017). UnitedHealth’s Optum to buy DaVita Medical Group for $4.9 billion. [online] Modern Healthcare. Available at: https://www.modernhealthcare.com/article/20171206/NEWS/171209940 [Accessed 10 Nov. 2018].
  11. Abelson, Reed. (2018). CVS Health and Aetna $69 Billion Merger Is Approved With Conditions. [online] The New York Times. Available at: https://www.nytimes.com/2018/10/10/health/cvs-aetna-merger.html [Accessed 10 Nov. 2018].
  12. Farr, Christina. (2018). As Amazon moves into health care, here’s what we know — and what we suspect — about its plans. [online] CNBC. Available at: https://www.cnbc.com/2018/03/27/amazons-moves-into-health-what-we-know.html [Accessed 10 Nov. 2018].


Additive manufacturing – Will the world’s next mega city drop out of a 3D printer?


Crowdsourcing a moonshot: can a redesigned toilet help India win the battle against human waste?

Student comments on Optum: Machine Learning at the center of Health Care

  1. One of the most expensive parts of machine learning is compute power. The bigger the dataset and the more complex the algorithm, the more compute power it requires to run. Companies like Optum usually outsource this processing to cloud computing companies like Amazon (AWS) or Google (Google Cloud). I wonder if Optum should try to bring some of that processing into the company, especially if Amazon is making a play into the same space? Internal compute capabilities could also assuage data privacy concerns.

  2. Thanks Ricardo. The key here is their parent company relationship with UnitedHealth. Optum’s critical advantage is not their IP over machine learning techniques or engineering/data science talent, but rather their ability to access a breadth and depth of data across the health care system (from providers, pharma, and other payers) that they can use to compound the power of their predictive analytics offerings.

  3. While I would agree that Optum’s relationship to UHG is part of it’s competitive advantage in that it provides a wealth of data to plug into their system, it is also a weakness. Other private payers are much less likely to want to use Optum’s services for fear that their networks, reimbursement, and patient population data will be co-opted by UHC on the insurance sales side of the house. Given that all this is proprietary and important to making money in insurance markets that are increasingly regulated, I think there is a “ceiling” of sorts for Optum in the payer space.

    I’m also interested how much the insights derived from AI in Optum can drive change in the health care system, given how regionally based the economics tend to be.

Leave a comment