Canadian medical anthropologist Dr. Leigh Hayden and UK-based service designer Kat Matfield describe how techniques from behavioural psychology can help practitioners work more effectively with patients with complex needs.

Introduction

Patients with complex needs have a confluence of physical and mental health issues that are often compounded with psychosocial challenges. They range from “high users” of health care to the “silently sick”, but overall account for a large proportion of health spending with relatively poor outcomes and experience of care. To improve engagement and experience of this patient population, the health care industry would benefit from new approaches to designing services currently organised around single diseases or organ systems. This paper focuses design exercises using concepts from behavioural psychology to engage patients with complex needs.

Women’s College Hospital is a health care organization focused on providing high quality outpatient care for patients with complex needs. Adaptive Lab is a digital products and services innovation agency that draws on the best practices of the technology sector which successfully embraces user-centricity in its design. The two organizations joined forces to tackle one of the most ‘wicked’ problems in health care today.

The three design methods we describe in this paper use concepts from behavioural psychology, which can be applied to health care to engage patients with complex needs to:
1. Build rewards into a service by prompting new ideas on how to incentivise health behaviours
2. Estimate cognitive load to identify moments of low “cognitive load” which could impede successful engagement of patients with a health service
3. Identify low and high points experienced, and brainstorm ways of minimising memory of lows and maximising memory of highs.

Context
Health care’s challenges in responding to complex needs

Across the globe, but most acutely in developed countries, we are seeing a growth of people with multiple chronic (or long-term) conditions, often compounded with psychosocial challenges. This population is commonly referred to as patients with complex needs.[1]

The growing prevalence of patients with complex needs poses new challenges to medicine and health care which are currently organised by disease or organ systems. Challenges for patients with complex needs include the following:
Care at various clinics that is often disconnected and sometimes competing.[2]
Motivation, and social, financial, and cognitive resources required to engage in self care and follow medical instructions.[3]
Burden imposed by care that requires engagement in several activities to achieve optimal health,[4] which manifests itself as poor patient experience in patients with multiple chronic conditions reporting lower satisfaction and quality of care.[5]
Deprived social conditions (e.g., poverty, housing issues, under-employment) that increase chronic disease risk, are barriers and disruptions to medical care and adequate self care, and threaten available resources for health care and health promotion.[6]
The expense and threat to economic livelihoods imposed by a chronic condition.[7]
These challenges can also result in poor patient compliance, worsening disease outcomes, increased morbidity and added cost. Overall, health care would benefit from re-design to provide a more collaborative, streamlined and engaging approach for people with complex needs.
Greater focus on “high users”

“High users” are defined as the 1-5% of a population that uses the greatest amount of services, whether measured by utilisation rates or cost.[8] Those at risk include patients with HIV/AIDS, a serious and persistent mental health condition, or multiple chronic conditions.[9] Contributing social issues include lack of home ownership, low incomes, and food insecurity.[6]

Developed countries are focusing on containing the utilisation and/or costs of high users. In Ontario, Canada, the top 5% account for 66% of health care spending.[10] In New York state, the top 5% account for 48% of health care spending.[9]

The hard to reach “silently sick”

At the other end of the spectrum are the “silently sick”, who choose not to, or are unable to seek care from the mainstream health system, despite poor health. This disenfranchised group live with morbidity that could be treated.

Reasons for this potential disenfranchisement include:
Mistrust of the system for historical or cultural reasons
Health-seeking as an activity is not valued
Difficulty in accommodating health appointments and self-management requirements
Lack of awareness of available health and social services.

When the “silently sick” do seek care, they are often confronted with previously identified deterrents, which then perpetuate tension, and may cause confusion and inappropriate care.

User-centricity in technology

Behavioural psychology has had a marked influence on how the technology sector approaches product and service design. Fogg’s 2009 influential behavioural design theory has three key elements: motivation, ability and trigger.[11] These must intersect simultaneously in order for a behaviour to occur. A user must have both the ability and motivation to engage in that behaviour which must be prompted or triggered until it becomes a habit. The more difficult the behaviour (ability), the higher the motivation required. Motivation is difficult to sustain, so Fogg proposes “motivation waves” to capitalise on periods where users are naturally more highly motivated. Ability is typically easier to target than motivation.

Design theory and methods often use these approaches to produce a desired behaviour, which may be as equally influential on consumer behaviour, (e.g., downloading an app), as on facilitating a health-related behaviour, (e.g., improving sleep patterns). To complement such approaches, anthropological insights are needed to appreciate the cultural and socio-economic context in which decisions are made, and the power and class matrices in which certain behaviours, like eating vegetables, are preferred over others, like drinking alcohol.

Objective

Our objective was to develop tailored design techniques to specifically engage patients with complex needs by using research-based insights into the struggles of people with complex needs and theoretical direction from behavioural psychology.
Design Methods

We tested three newly-developed design methodologies in order to tailor them to people with complex needs, in whom illness may result in lower physical and cognitive endurance, lower concentration, and lower tolerance for services that do not produce an immediate desired result.

For each exercise, we defined the flow, desired outcomes, participant criteria (number and make-up), and materials required. Participants recruited included designers and web developers at Adaptive Lab. Exercise review and feedback through retrospectives enabled us to identify gaps, possible modifications, utility, appropriate timing, and guidance on re-design. This iterative design approach enabled us to implement modifications to each exercise.

Design Method 1 – Reward Card Game

This exercise was designed to build “rewards” into a service. It is a card game intended to prompt new ideas about how to incentivise health behaviours.

Health care is a participatory practice, often involving new practices (e.g., appointment attendance, fasting for blood tests, dietary changes), which are inherently difficult to integrate into a routine. The promise of uncertain future benefit competes with certain present enjoyment. In health care, the information deficit model is often used; if patients understand what they were asked to, they will conduct themselves accordingly. Empirical evidence suggests this model often fails.[12] Providers, without other tools, default to guilt and shaming to encourage patients to behave according to the medical model. Behavioural psychology offers approaches which could be built into the service, increasing the effectiveness of the patient-provider partnership, and avoiding negative tension and resentment in patients.

Nir Eyal’s influential book “Hooked: how to build habit-forming products” outlines his techniques to help designers build products that users habituate into their daily lives.[13] These so-called “habit-forming products” have inbuilt cognitive rewards which promote use and habituation. His concepts can be applied to service design by exploring how different “types” of rewards can be used to create more satisfying and habit-forming products and services.

Eyal groups rewards into three categories demonstrated in Table I.

Table 1: Eyal’s Cognitive Rewards [13]

Rewards
Relation
Types of reward
Rewards of the Tribe
Social acceptance and belonging
Valued social rewards by social beings:
Importance
Acceptance
Inclusion
Attractiveness
Rewards of the Hunt
Need to acquire resources to aid survival, context dependent
Things of value (e.g., food)
Money
Information
Rewards of Self
Personal gratification, and sense of pride and purpose
Mastery
Consistency
Completion

Goal
Develop a number of possible methods to reward desired user behaviour to enable a product or service to become “stickier”.
Scenario
The “Virtual Ward” programme for patients with complex needs provides close monitoring after hospital discharge, but engages patients in a number of self care activities which are often encouraged using guilt and not rewarded using good design.
Participant criteria
The exercise was conducted with one moderator and three participants.
Preparatory work
Preparing for the exercise took one person one hour. Two sets (colours) of cards were developed: 1) User behaviours that the product or service require to work optimally; one behaviour per card, 2) All of the reward categories (listed in Table 1); one reward per card.

Findings
The exercise took 30 minutes from explanation (including the behavioural psychology concepts) to completion. Budget for 45 minutes for a group unfamiliar with design exercises and concepts from behavioural psychology. Participants enjoyed the exercise, however the concepts, although interesting, required a large assimilation of new knowledge in a short timeframe. Creation of a figure to demonstrate the types of rewards with a brief description (or examples) of each will facilitate improved understanding at the onset of the game.

To maintain the flow of the exercise, if a participant is not able to think of a reward of the type drawn, the turn should move onto the next participant after 10-15 seconds. This occurrence is a reality as some behaviour-reward combinations may not make sense.
Discussion
This exercise, although helpful for brainstorming ideas for rewarding programme-specific behaviours, can also be used to initiate a design exercise which aims to form habits. The exercise can be modified for larger groups by splitting them in two groups, each doing the same exercise. This increases competitive engagement and produces more ideas. Reconvening after the exercise allows for comparison of answers and voting to select the best.

Overall, this exercise has promise. Seeking and receiving health care typically has multiple disincentives and barriers, such as long wait times, long travel distances, or recommended activities that are confusing and/or difficult to incorporate into daily routines. Incorporating rewards into a health service, to keep patients engaged, give them additional incentive and motivation to withstand inevitable difficulties, is novel and has promise for patient engagement, patient activation, and patient-provider relationships.

Design Method 2 – Assessing Cognitive Load

This exercise involves estimating cognitive load to find patients’ most receptive moments. It has been designed to identify moments of low “cognitive load”, to help choose the appropriate moment to onboard a patient into a new programme or service.

Cognitive load refers to the amount of information being processed by the working memory at any given time. It is a sum of the intrinsic and extraneous loads. Intrinsic load is inherent to the contents of the material being learned; more difficult or complex concepts have a higher intrinsic cognitive load. Extraneous cognitive load is anything that consumes working memory or cognitive processing which is not concerned with the material at hand. This includes unrelated information or material, or other cognitive inputs such as noise or other distractions, which compete for attention and working memory.

Cognitive load is influenced by:
Novelty – whether the learner has been exposed to it before; if they have, it consumes less working memory
Other distractions, e.g., multi-tasking
Learner’s current “mode” – this could be “habit” mode (e.g., early morning when getting ready for the day) or “learning” mode (e.g., evening when more relaxed)
Complexity of the material presented, which can be reduced through simplification and avoidance of jargon and new vocabulary
Learner’s emotional state – if they are in an excited state, they have less capacity for decision-making.
Seeking and receiving health care and self care requires patients to learn and integrate these activities, and the rationale for conducting them, into their lives. When introduced at a period of high cognitive load, integration of new ideas and activities is more challenging, possibly reducing ability to engage in the care or service. This creates a poor patient experience which could potentially be avoided or reduced.

Goal
Determine the best “moment” for engagement and consider ways of reducing cognitive load.
Scenario
The “Virtual Ward” programme for engagement was used
Three engagement timeframes were used – at the hospital before discharge; 48 hours after discharge at the patient’s home; and 1 week after discharge at the patient’s home.

Participant criteria
The exercise was conducted with four people (three designers and one programmer). Two people moderated the exercise.
Preparatory Work
Preparing for the exercise took two people one hour. Cognitive load mapping “cards” (Figure 1) were developed for participants to score the cognitive load dimension for each scenario. Once complete, the cognitive load can be assessed at a glance to quickly identify moments of low cognitive load. Questions cover the following concepts that influence cognitive load: background, environment, complexity, impact and novelty. Pointers to note are as follows:
The first cognitive load question should focus on anxiety/excitement.
The description of the concepts should enable clearer questions.
A familiar topic/situation should be selected to improve the utility of the exercise.
The cognitive load questions should be grouped by topic (context, current environment, decision, emotion).

Figure 1: Cognitive Load Mapping Card
Are they in an excited state?
mild
moderate
extreme
What time of day is it?
morning
afternoon
evening
Has their day been busy?
not at all
a little
very
Is their current environment distracting?
not at all
a little
very
Are they currently multi-tasking?
not at all
a little
very
Do they have to absorb lots of information?
a little
some
lots
Do they have to grasp any new concepts, or words?
none
some
lots
Have they done something like this before?
done this before
done something similar
never done anything like this
Are there a lot of options of what to do?
just one
a few
lots
Is it hard to choose between the options?
not at all
a little
very
Will the impact of the choice/task be serious and/or long lasting?
not at all
a little
very

Findings
The group completed the exercise for a problem (engagement with the Virtual Ward programme) with which they were unfamiliar. As such, they struggled with the exercise and required explanation from the moderator regarding the user journey and circumstances.
In a second iteration of this exercise participants focused on a relevant problem, the engagement of patients with a project they were working on. Using the project’s user journey map, the cognitive load at three distinct moments was observed, to give some empathy into the state of the user during the various steps of the journey. The exercise took 30 minutes. About half way through the exercise, the group led itself.
Discussion
Assessment of the usability and specific utility of this exercise highlighted the following insights:

In health care, programme engagement, patient education and other communication-rich activities are typically completed at times convenient to the care provider or system. This exercise is valuable because it forces programme designers to consider how receptive patients may be and when best to undertake these activities. This level of empathy is uncommon in health care design and could be especially helpful for patients with complex needs.
It is ideal for the participants to be familiar with the problem at hand, ideally patients/users themselves.
The exercise could be combined with a user journey map, to explore cognitive load at each step of the user’s journey.
It is beneficial conducting the design method exercise early in a project design process, but not right at the beginning. This is allows for insight provided by this exercise to be more easily incorporated into a service design.
To focus discussion after completion of the exercise, the following question should be posed to participants: are there ways of removing cognitive load (e.g., by using default options), spreading the load out or shifting it, or simply removing elements to simplify the user experience?
This exercise will capture hypotheses about users – specifically about their environment and context. The group may need to make a number of assumptions about the user’s context and environment. These assumptions should be tested to determine their validity. Alternatively, this exercise could be done by researchers who have developed a deep understanding of the patient experience.

Design Method 3 – Designing for the remembering self

Daniel Kahneman’s book “Thinking Fast and Slow” has heavily influenced behavioural economics and popular understandings of human behaviour.[14] He describes the “two selves” – the experiencing self and the remembering self.

He argues that happiness is a function of our experiencing self, that assesses our current life, and our remembering self, that assesses our memories of prior experiences. His research suggests that our remembering self has a greater influence on our happiness. We live according to the stories we create of events (which may not fully reflect the experience we had at the time). Moreover, the endings of our stories are especially important and frame the entire event.

Social science also recognises the importance of personal narratives – the constructions of memory that provide meaning and coherence to our experiences and sense of self. The differences between actual events and our memory of them and the role of narrative in linking discrete data to enhance memory and comprehension are of interest.

The importance of memory can be applied to the patient experience. For patients, health care can have unavoidable negative experiences, e.g., pain, discomfort, embarrassment, and/or confusion. In addition, health care can have avoidable negative experiences, such as inefficiency, ineffectiveness, poor treatment, lack of treatment.

To improve patient experience (one of the triple aims of health care, identified by the Institute for Healthcare Improvement), work can be done to remove the avoidable potentially negative experiences. However, even if this work succeeds, not all negative experiences are avoidable. In these cases, an attempt can be made to influence how patients remember their experience, and thus improve their overall experience and wellbeing. In addition, by improving patients’ positive memories of the experience of a health service, we can also increase the likelihood that they will continue to use the service and be engaged in it.

Goal
Identify the high and low points of an experience, and brainstorm how to minimise the memory of the lows, and maximise the memory of the highs.
Scenario
This exercise could be conducted during the pilot phase of a new programme, or during the design phase, by exploring patient experiences of similar programmes. It could also be used to improve existing programmes.
Participant criteria
The exercise can be conducted with one moderator and two to four participants.
Preparatory Work
Budget one person 30 minutes to prepare for the exercise. They must develop two sets of cards: 1) for reducing memory; red in colour, 2) for increasing memory; green in colour. Each set consists of three of cards stating – “repetition”, “attention”, “emotional impact”.

Findings
The exercise took 45 minutes including explanation, the exercise, and a retrospective. Participants conducted the exercise using both service and product experiences. Examples included: signing up for a web service and ordering something for home delivery. Participants struggled with remembering a “poor” or “exceptional” experience, and subsequently many of the details around that experience.

The exercise is best done for real problems encountered, because we found that we needed to probe participants to be as granular and specific as possible about what went well and poorly. It worked better for service experiences than product experiences, as these had multiple different steps.
Discussion
This method is best used for experience design, and paired with a user journey map to map out the experience then chart the good or bad qualities of each step.[15] From this map, it is possible to focus in on particularly good or bad experiences, or clusters. This initial map is best conducted with real users who can describe their experience in detail, although it may not be appropriate to involve these users in discussing how the memory of the experience can be improved.

It is optimal if a bad experience can be avoided or designed out. Merely decreasing its memorability, when you can eliminate it, is suboptimal. In health care, there is a need to deeply interrogate whether bad patient experiences can be avoided, and avoid presuming that these experiences are inevitable. However, some of the discomforts of “poking and prodding” in health care are unavoidable, and dampening their impact and memory may be both compassionate and beneficial.
General Findings

Our work has demonstrated that behavioural psychology concepts can be used to develop exercises to explore health care design problems. Although the exercises above were designed with health care in mind, it is possible that they may be applied to other design problems. It would be worthwhile to test their broader applicability.
Our work also demonstrates that participants can successfully learn how to apply behavioural psychology concepts to assess situations and provide design insight. An important caveat is that the exercises were not tested with “patients with complex needs”, but rather, with design lab employees who are skilled at learning and applying new concepts. Using internal staff and other industry people to test exercises first to identify major problems is helpful and cost effective. The next step is to test the tested and revised exercises with real users focusing on their health care design problems or possible experiences.
The success of these exercises in the design process is a product of several factors, including identifying the:
appropriate design problem
relevant user population
most receptive moment for implementation.
Further exploration and testing of the exercises’ situational utility in relevant populations will promote further user-centric improvements.
Conclusion

We have identified, trialled and adapted user-centric design methods for health care in order to inform design of technology solutions, to enable patients to better engage with health care services and health improving behaviours. We have taken these approaches from the tech industry one step further and applied novel design methods to better address the health service needs of patients with complex needs. Using concepts from behavioural psychology, we developed these design methods to specifically engage patients with complex needs who are arguably underserved in the current system, and to improve user experience (and potentially) outcomes. Patients with complex needs often live very different lives compared to those who design and provide health care, and their care is in itself very difficult to do well, so soliciting their input could be a valuable way to improve their experiences and outcomes. Our approaches, informed by behavioural psychology, are intended to solicit their views and explore their lives.
Our work represents a novel approach to engaging complex and disenfranchised users. The exercises used in co-design with this population are typical exercises (consumer mapping, developing user profiles, card sorting), or modifications of these common exercises. To our knowledge, no others have developed exercises based on behavioural psychology, which target patients with complex needs. Theory-driven approaches to design exercise development are uncommon and have real potential.
Further application in health care to address real problems with real patients is needed. However, this research has enabled a crucial first step. The exercises above are both novel, adding to the arsenal of design exercises commonly used, and specific to use in instances in which low acceptability and low satisfaction are probable, instances that are not uncommon to health care.
The next phase is to implement the design methodologies in the community by partnering with service organizations (such as health care organizations and patient groups). This can also draw on previous experience with re-designing services and systems for other complex and disenfranchised populations, such as vulnerable families,[16] and homeless people [17]. Designing services with end-users may identify wider system barriers. While designers are process-driven, providers are solution-driven and have intimate knowledge of patients’ struggles and needs, and can help overcome these barriers.[16]
Behavioural psychology concepts, although not originally designed for application in services for patients with complex needs, can be used to facilitate increased acceptance of health services by focusing on the “how” of service design, and less on the “what” of service design.

Dr. Leigh Hayden is a medical anthropologist working at Women’s College Hospital, in Toronto, Canada and Kat Matfield is a Service designer working at Adaptive Lab, London, UK.

References

1. Schaink AK, Kuluski K, Lyons RE, et al. A scoping review and thematic classification of patient complexity: offering a unifying framework. Journal of Comorbidity. 2012; 2(1):1-9.
2. Zulman DM, Asch SM, Martins SB, et al. Quality of Care for Patients with Multiple Chronic Conditions: The Role of Cormorbidity Interrelatedness. J Gen Intern Med. 2014 Mar; 29(3):529-537.
3. Eton DT, Ridgeway JL, Egginton JS, et al. Finalizing a measurement framework for the burden of treatment in complex patients with chronic conditions. Patient Relat Outcome Meas. 2015 Mar 27; 6: 117-26.
4. May C, Montori VM, Mair FS. We need minimally disruptive medicine. BMJ 2009 Aug 11; 339:b2803.
5. Burgers JS, Voerman GE, Grol R, et al. Quality and coordination of care for patients with multiple conditions: results from an international survey of patient experience. Eval Health Prof. 2010 Sep; 33(3):343-64.
6. Fitzpatrick T, Rosella LC, Calzavara A, et al. Looking Beyond Income and Education: Socioeconomic Status Gradients Among Future High-Cost Users of Health Care. Am J Prev Med. 2015 Aug; 29(2):161-71.
7. Greene R, Dasso E, Ho S, et al. Patterns and expenditures of multi-morbidity in an insured working population in the United States: insights for a sustainable health care system and building healthier lives. Popul Health Manag. 2013 Dec; 16(6):381-9.
8. Veillard J, Denny K. Transformation through Clinical and Social Integration: Meeting the Needs of High Users in Healthcare. Healthcare Papers. 2015 Apr;14(2):4-7.
9. Zucker HA. Tackling the Issue of High Users of Healthcare. Healthcare Papers. 2015 Apr; 14(2):48-51.
10. Woodchis W, Austin P, Newman A, et al. The Concentration of Health Care Spending: Little Ado (yet) About Much (money). Canadian Association for Health Services and Policy Annual Research Conference, Montreal. May 30 2012.
11. Fogg BJ. A Behavior Model for Persuasive Design. Persuasive Technology Lab, Stanford University. 2009 Apr. Available at: http://bjfogg.com/fbm_files/page4_1.pdf
12. Brown JP, Clark AM, Dalal H, Welch K, Taylor RS. Effect of patient education on the management of coronary heart disease: a systematic review and meta-analysis of randomized controlled trials. Eur J Prev Cardiol 2013 Aug; 20(4):701-14.
13. Eyal N. Hooked: How to Build Habit Forming Products. 2014. Penguin Books.
14. Kahneman D. Thinking Fast and Slow. 2011. Farrar, Strauss and Giroux, New York
15. Service Design Tools. Communication Methods Supporting Design Processes – Customer Journey Map. Available at: http://www.servicedesigntools.org/tools/8
16. Body J, Forrester S. “Synthesizing Policy and Practice, the case of co-designing better outcomes for vulnerable families” Design for Policy, Edited by Christian Bason, 2013
17. Simm W, Southern J, Whittle J, Ferrario MA, Dillon R. Co-Designing With and For Chaotic Lives: #Patchworks, a Catalyst Case Study. 2013. Available at: http://openlab.ncl.ac.uk/vulnerability/files/2013/02/Simm_DFWVP2013.pdf