Why It Matters
A unique approach to segmentation may be the key to providing personalized care for thousands of people at a time.
SIGN UP FOR IHI EMAILS
Processing ...

Scale Down and Scale Up: How to Mass Customize Personalized Care

By Helen Bevan | Tuesday, May 14, 2019


Photo by Gerhard Bögner | Pixabay

Helen Bevan is Chief Transformation Officer, NHS Horizons England, and known around the globe for her efforts to build learning networks to energize large-scale change in health and care. In the following interview, she addresses the challenge of providing personalized care for individuals (scaling down) while also improving care or health for a large population (scaling up). Bevan will share more on this topic at the IHI/BMJ International Forum on Quality and Safety in Healthcare: Asia Pacific in Taipei (September 18–20, 2019). 

What does it mean to scale down and scale up in health care improvement?

I’ll start with an example from my own country in the National Health Service in England. It’s a massive health and care system that provides support for 54 million people. We’ve just launched a long-term plan for how we’re going to transform our entire health and care system over the next 10 years.

There’s a tension at the heart of the plan because, on the one hand, it talks about how we’ll create personalized and individualized care and support for people that is based on what matters to patients. That’s the scaling down part. At the same time, we want to do this on a massive scale. That’s the scaling up part.

What’s an example of how to personalize care on a large scale?

When we think about “scaling” in health care improvement, it’s usually about taking a change that makes a positive difference for a group of patients and seeking to make similar changes for a much bigger group of patients. But if we are talking about personalized care at scale, it can’t be about giving everyone the same change. People have different wants and needs. So, to practically scale up and scale down at the same time, we need to group or segment our population of patients or service users.

There’s a countrywide initiative in Sweden called the “patient compact.” It means that anybody that uses health services in Sweden has an agreement with their providers around the kind of highly personalized care they can expect. To do this, they have come up with a unique way of segmenting patients. What we very often do [in the health care systems in the UK and elsewhere] is segment people by their clinical needs. If you imagine a triangle, the people with the most needs are at the top of the triangle and those with the least need are at the bottom of the triangle. 

What they’re doing in Sweden, led by the Swedish Association of Regions and County Councils, is segmenting people by their behaviors and attitudes to the health and care system. They talked to thousands of people around what matters to them and, based on this information, they’ve segmented people into four categories:

  1. Independent and engaged — People who are independent and engaged trust the [health and care] system and when they need health information they’ll seek it out themselves. They’re the kind of people who want to take part in health care decisions.
  2. Worried and engaged — This second group is also engaged, but they trust the health and care and system less. They have a higher level of anxiety regarding their encounters with the system.
  3. Traditional and concerned — These are people who trust the system and they prefer providers to make their decisions for them. They don’t proactively seek out health information.
  4. Vulnerable and concerned — These people don’t trust the health care system. They don’t proactively seek out health information and they experience high levels of anxiety when they encounter their health and care system.

If we’re going to scale down and personalize care, but do it at a big enough scale, we need to design the system to address different groups by their different needs. For instance, somebody who is independent and engaged can make their own appointments digitally. They’re also willing to monitor their own health through self-checking blood pressure, sleep patterns, exercise, blood sugars, etc. In contrast, someone who is vulnerable and concerned wants to be telephoned before their appointment. They like having an assigned contact person and personal follow-up calls afterwards. If we can understand what segment people are in, we can then create something at scale that meets people where they are. They’re at the start of a journey, but I think the Swedish work is groundbreaking.

Are there keys to successfully scaling down and scalping up that you’ve found so far?

This has to be about co-design and co-creation. We can’t go in with an expert mindset that assumes we know what people want. We have to ask people what matters to them. We also need to start with a strengths-based approach. Let’s focus on what people can do and want to do rather than use a deficit illness model.

We need to have a hypothesis when we’re thinking about scaling down and scaling up. In terms of scaling up, in addition to segmentation, we need a clear theory of change. We need to theorize about what we think the cause and effect is going to be. That might be a driver diagram. It might be a logic model.

Some of the underlying principles have been around for a long time. Twenty years ago, in the National Health Service in England, we were talking about mass customization and using principles from production management to segment patients. I think those ideas have come around again. What we’re seeing across the world is people trying to understand and design for both personalization and scale at the same time. What can we learn from the Swedes? What can we learn from other industries? Learning is at the heart of this.

Editor’s note: This interview has been edited for length and clarity.

You may also be interested in:

Do These 4 Things When You’re Scaling Up Improvement

7 Questions to Guide Your Large-Scale Improvement

first last

Average Content Rating
(0 user)
Please login to rate or comment on this content.
User Comments

​​