HBS 3: Using Predictive Analytics to Reinvent Healthy Cities

The Public Talk on “Using Predictive Analytics to Reinvent Healthy Cities: A New Business Model in Consumer Health” was successfully held at ESSEC Asia-Pacific on 16th July. Gathering experts in analytics from academia, public sectors, and industry, the talk provided a facilitative platform for discussion with regard to the opportunities that predictive analytics brings for the health services research and outcome improvement at both in-hospital and community settings. The Public Talk was the third in the Health, Business and Society series, which was initiated with the aim of facilitating and disseminating insights to relevant stakeholders in the industry, so as to deliver meaningful actionable policy recommendations to both public and private decision makers.

The talk was moderated by Dr. Allen Lai, Director of Institute of Health Economics and Management at ESSEC Business School Asia Pacific. The panel consisted of Mr. Callum Bir, Director, Health and Social Services, Microsoft Asia-Pacific, Prof. Ashwin Malshe, Assistant Professor of Marketing, ESSEC Asia-Pacific, and Mr. Wu Dan, Head of Healthcare Analytics Unit, Khoo Teck Puat Hospital, Alexandra Health System.

The Public Talk was opened by Prof. Ashwin Malshe, who gave an introduction to the state of analytics in the healthcare industry. To Prof. Malshe, analytics refers to the study of data to generate actionable insights for individuals, managers and policymakers. He emphasised that the insights which we generate need to be able to be used to improve decision making. Prof. Malshe framed healthcare analytics by introducing the 6Ps as follows:

• Patients
• Physicians
• Providers
• Payers
• Pharmaceuticals
• Policymakers

The greatest challenge for healthcare analytics, as Prof. Ashwin argued, lies in getting patients to release the data and allow the data to be used for such purposes due to problems such as privacy issues as well as data exploitation. To provide the audience with a brief overview of analytics, Prof. Malshe introduced 3 different categories of analytics:

• Descriptive analytics: the business intelligence aspect of analytics
• Predictive analytics: looking at past data and predicting what is going to happen in the future
• Prescriptive analytics: analytics that enables decision makers such as patients or hospital managers to take actions

Mr. Callum Bir introduced the topic through addressing the current trend of an ageing population across the region, and shared the challenges brought by ageing populations. He posed the question of how we can use technology, such as home-based post-stroke rehabilitation gaming programs, to solve some of the problems contributing to an ageing population such as life-style related and preventable illnesses. 

To put things in perspective, Mr. Bir shared various factors which drive chronic diseases, such as diabetes, alcoholism, tobacco, poor physical activity, high blood pressure and obesity. Such chronic diseases have considerable economic impacts on cities, taking a toll on economies through:

• Decreased wages, taxes and savings
• Increased absenteeism, disability and early retirement 
• Decreased worker productivity

Mr. Bir explained that cities, being the economic growth engines of countries, have much to gain from taking action to solve these problems. Following this, Mr. Bir spoke about Citizen Generated data, and the problems faced in obtaining such data. The challenge lies in getting people to record their data, as obtaining accurate self-reported figures is not as easy as it looks. Raising the example of having individuals providing data about their diet, Mr. Bir proposed using technology to get people to share details about what they eat in a more innovative way, such as encouraging them to share these photos on social media, where analysts can then deduce the nutritional content of their food based on the photos. However, the audience raised concerns about privacy issues with such applications. Other than Citizen Generated data, Mr. Bir added that medical devices such as weighing scales and glucometers are able to capture data and allow users to extract the data for analysis. Lastly, wearable devices such as the Fitbit records actionable data which allows individuals to better understand their life patterns or habits. 

Mr. Bir also addressed the question of whether we can really drive behavioural changes with the use of predictive analytics through providing real-time feedback and insights to call for actions by individuals. 

A group discussion on “Who should pay for the technology?” then ensued. Members of the audience were encouraged to discuss this topic, and suggestions such as segmenting different levels of consumers and charging them differently were raised. Mr. Bir also pointed out that even though many innovations have been developed to help improve the current situation, the adoption rate is very low in reality, due to issues such as affordability of these new innovations. He also added that his team is constantly thinking about business model innovations that can allow them to address the problem of who is going to pay for the system. Mr. Wu Dan shared his perspective on this issue, elaborating that if people do not pay for the service themselves, they are unlikely to feel pressure to use it. Focusing on the unique nature of healthcare, Prof. Malshe added that the key issue lies in whether a negative externality is produced when an individual is facing health problems, and if the people around this individual are affected by this negative externality so much that others would pay for services to improve this individual’s situation. He elaborated that this could possibly be the case for sicknesses, but with regard to the issue of whether an individual is fit or unfit, people around an unfit individual are unlikely to feel seriously affected by this individual. A member of the audience also added that the government is usually the payer for such healthcare services, and assessments on which service or solutions to use are often made based on demonstrated clinical outcomes of these solutions. 

The Group Discussion was followed by Mr. Wu Dan’s presentation, where he spoke about the conception of the analytics unit at Khoo Teck Puat Hospital, projects which the team has partaken in, as well as the challenges which the team has faced and the solutions which they developed through the use of analytics. 

During the Q&A session, a member of the audience highlighted the fact that predictive analytics is based on assumptions, but assumptions can also be wrong and randomness could affect the predictions which we make. Thus, care needs to be taken when assumptions are made, as well as when determining to what extent these assumptions are true. A question on whether it is more important to know about models for analytics, or about asking the right questions to get the right insights was raised. Prof. Malshe explained that a gap exists in that those doing the analytics might not be able to ask the right questions, while those who are able to ask the right questions might not know how to carry out analytics. 

The Public Talk was an engaging one, with speakers sharing their views from different perspectives, allowing the audience to develop a better idea of the challenges faced in using predictive analytics in organisations, possible solutions to overcome these challenges and the opportunities brought to us by predictive analytics.