Making a design pattern library for mental health products

Sarah Drummond
9 min readOct 26, 2018

In April 2018, a group of practitioners from across the design, health, digital development and public policy professions came together to develop a public pattern library that aims to improve the design of digital products and services for the mental health and wellbeing sector. We held out first workshop in New York City at the beautiful Public Policy Lab offices.

We made this design pattern library

In just under two days, we produced 35 design patterns with a global network.

Our first principles on our pattern library

Today, we’re only releasing a snapshot of four, but with a view to build support to help us publish more.

Here’s how we did it.

Choosing a deadline

Public Policy Lab called Snook after seeing a blog we’d written about design patterns for mental health and wellbeing.

We chose a date, stuck to it, and the deadline held us accountable to running a two day workshop in NYC. We chose a short time period to prepare and it stopped us from over complicating the workshop design and frameworks for the patterns at an early stage.

Learning — deadlines are good to make a collaboration happen.

Choosing a focus

It was difficult to say ‘we will only look at this part’ of the landscape of mental health and wellbeing products and services.

What we essentially agreed on is that we wouldn’t use diagnosis versus non-diagnosis to curb our research and definition of patterns as we knew from an early stage that the same principles would apply to products used by anyone. This would be about mindfully designing products that are inclusive to all.

We chose ‘digital’ as our core point of focus. We’re aware, as service designers, that experiences are not limited to digital channels, but this starting point gave us a way to talk tangibly and more easily about what a pattern might look like in practice. In future iterations we will open this up further.

Researching and producing evidence

We both had existing published literature and projects out in the world so felt confident we had some learnings to mine that would help produce ideas on what this might start to look like.

However, we wanted to widen the insight net and include enough stimulus for others joining the workshop online and in person.

We undertook a series of online video interviews. These were mostly conversations with people who had shown interest in the project on twitter. These ranged from health professionals working in hospitals to investors into tech for good, mental health app designers to developers building CBT platforms.

We asked simple questions like: What makes a good service? What patterns do you see in your work? What learnings could be more widely shared when we focus on the digital mental health space.

We filmed these interviews and created a directory of the films for our workshops, plus ‘cheat sheets’ highlighting the common patterns of insight.

Setting out an early framework

We mapped out a high-level journey map — informed by our joint work — that helped us think about insights at specific stages of service use. This wasn’t to say, here is a linear journey of mental health, but it was to help categorise patterns and make it easier to navigate our findings.

Building an online community

We gave ourselves the hard challenge of running both a physical and virtual workshop. We had so much interest in the workshop that we created a slack community to hold conversations with participants.

We think it worked ‘ok’, in that, we did our best to hold online presentations streamed via Google Hangout and we invited people to comment and contribute where they could.

The workshop — day one

We ran across two half days. On day one, we ran an introductory presentation on what patterns are, and discussed our thinking on how they might evolve.

We split ourselves into three groups, and held initial conversations asking two questions:

  • What makes a good service in the mental health and wellbeing space?
  • What evidence do you have that backs up these sentiments?

This allowed us to map critical thoughts on patterns we thought should exist and supporting evidence for them.

We then reviewed the videos and held discussions on the patterns.

After our first session, we pinned these ideas to the relevant journey stages they may apply. It became clear, some of them were more general principles that could be exemplified across different stages of the journey map as features.

For example, the pattern No Surprises could be at the on boarding stage of a service, but also at the exit point.

Building our pattern language

These discussions were vitally useful in beginning to develop our pattern language.

We didn’t want to become ‘stuck’ on meta frameworks, but it was important we define ways of looking at communicating the data we were producing.

There are two distinct models we explored, first is our pattern language. We have been through three iterations of trying to define our pattern language, and its important to say, this is our first public iteration of this framework. We are convinced it will change as we dive deeper into surfacing more information around these.

We said that there are ‘Principles’. These are high-level values that should considered in every service, across all journey stages and delivery channels. We created six of them;

Accessibility: Users need access to suitable care regardless of their physical, emotional, or cognitive abilities; their age; or their immigration status.

Agency: Users need to be able to make informed choices about the beginning, end, and experience of care.

Confidentiality: Users need to know that they can seek care (or just information) without anyone else finding out.

Continuity: Users need a steady set of providers, with thoughtful hand-offs during care, and limited requests to re-tell their story.

Trust: Users need to feel confident that providers are reliable and able to care for them well.

Welcoming: Users need validation, kindness, encouragement, and respect from providers.

These are not publicly published yet, but helped guide our thinking in developing the patterns.

We said that there are service responsibility areas. These are overall guidelines for building service touchpoints. They can be exemplified at multiple stages across a user journey.

We felt that our patterns were the most useful, these are the example set of what others are doing online. At all times we kept in the front of our minds that our user might be a developer, or a designer, or someone architecting a service and this led us to ensuring these featured in our library. These are illustrations of how a pattern can be applied to design specific product features. They are non exhaustive and should not be viewed as inhibitive, rather a starting point and platform for further consideration.

Deeper patterns for no surprises and positive updates

We know we can take this further into a more detailed interaction pattern level (see above) which we did do during our workshop, but for now, we believe the service responsibility areas are a good start for anyone building a product or service in this space. We’re welcome to receiving specific patterns submitted by the public that are more detailed, which you can do here.

The goal of our project is to build the library — the information and digital framework to hold a large set of patterns, which will be submitted by designers and mental health experts from around the world, and find a way to sustain this network.

We also began to define the spaces for the types of services and products we might be discussing. We did not explore this further at the workshop, but will return to this as the catalogue becomes bigger in resource and needs further categorisation. One of our online participants had highlighted work by Huppert F. cited within the SHIFT Promoting Wellbeing paper. This captured conversations we’d had during the early part of the day on the different ‘spaces’ of where products and services fit.

It breaks down into four areas, and neatly helps us to identify where and how different patterns might be applied for future stages of the pattern library.

Mental Disorder — Prescribed products or services for diagnosed conditions (e.g A CBT platform)

Languishing — Recommended products or services through a health partner (e.g NHS app library)

Moderate mental health — General wellbeing apps (e.g An online meditation product)

Flourishing — All products and services (e.g Monzo bank account)

Producing pattern content

Our spreadsheet of early principles and patterns

We set up a rudimentary spreadsheet of the first patterns. We were looking for content to bring them to life that included:

  • Pattern name and simple description
  • Pattern contributors
  • Pattern evidence and examples
  • Pattern features and the stages they are relevant to
  • Stage relevance
  • Principle alignment

We began work on writing copy for each pattern and the hard work of looking and clarifying supporting evidence or further reading.

Developing a full pattern

Conjointly to the evidence searching, we researched existing patterns that align to the service responsibility areas. We used our own experience and the research videos we had created. We also incorporated the online submissions from Slack.

To develop a set of patterns let’s look at the service responsibility ‘Uphold user agency’:

Support users in arriving at their own understanding of their mental health, feeling a sense ownership over their own care/treatment, and defining their own wellness goals.

The key patterns for this service responsibility are:

Constructive Confirmation

Ensure that requests for user confirmation benefit the user, not just the service. Use confirmation requests to allow users to give feedback, change their treatment, and observe that they are in control of how they are being supported. Be careful not to over-ask.

Relevant at all stages.

Enable Self-Direction

Provide users with tips and advice on mental-health self care — ways users can support their wellbeing outside of and in addition to provider-based treatment. Suggested content includes information on how peers (in their community, age range, or situation) successfully manage their mental health. Preferably, this content should be provided in ‘public’ areas of the service, such as the home or sign-in page, so even non-users have access to it.

Relevant at the Awareness, Care/Treatment stages.

Stories not Symptoms

Provide plain language descriptions and stories about symptoms, rather than complex clinical descriptions. For example, using ‘Can’t sleep’ rather than ‘Insomnia’ might help users who wouldn’t identify as having a sleeping disorder know that they could seek support.

Relevant at the First Contact, Enrolment/Activation stages.

This is one of our current unreleased set of patterns but all can be viewed in our working pattern document.

Visualising the design patterns

We chose a simple visual language to highlight only key features. Again, the features are not inhibitive, they are there to inspire and allow users of our library to jump off from. At this early stage, the patterns and features are not all fully proven but do have supporting evidence to highlight where they may have been used before or hinted at academic research that provides backing for the feature. We’d like to take this further with a larger network and future iterations to bring people who have examples into the network to share research.

Building the library

Kacia from PPL undertook an amazing amount of hard work before the workshop to create a solid container for publishing the patterns. On our final day we grafted to try and finish at least 5 full pattern sets. We got them online and we ‘nearly’ launched but didn’t quite make it as some of the details of copy and descriptions for the overall site still needed final tweaks.

We decided at this stage to feature the service responsibility areas, not the pattern or principles up front and to keep the navigation simple. Over time, as we develop further content we will make the content smarter based on what people search for, utilise the stage names and go into a much deeper level of fidelity.

The workshop whilst producing content was also useful to develop our pattern language and frameworks which did take a good amount of our time into day two to get right.

What’s next?

We have published the first 4 high level patterns. It may not seem like much now, but we’ve done the hard work on planning out the framework for this.

We have 35 service responsibility areas with no patterns currently attached. Our intention is to set up workshops that bring experts together to build the pattern and evidence set for each and upload to the pattern library. This means we can refine our ask to experts to support the development of the pattern library.

For me, it was a personal delight to work with PPL and the other contributors on the day. We’re going to seek funding so we can undertake the research and evidencing of patterns more rigorously. If you are interested in talking to us about making this happen, you can get in touch. We’re looking for support to help us take this to the next level.

Join our mailing list

You can be kept up to date with all our developments by subscribing to our mailing list



Sarah Drummond

Founder @wearesnook @dearestscotland @cycle_hack @mypolice | Service Designer + Boss | GOOD Magazine’s Top 100 influencers 2016|Google Democracy Fellowship 2011