How do we know this works?

Standard

“You can’t use an old map to explore a new world.” Albert Einstein

Why do we need evidence?:

Whilst the majority of colleagues intuitively understand LfE, a minority of colleagues reserve judgement and request to see “hard evidence” to support its implementation.

“…show me the interventions for all this touchy-feely stuff; I need hard facts and tools”  NHS senior leader (anonymous)

Initially, I ignored these comments as the benefits of LfE were obvious to me, and the rapid uptake of the initiative was evidence that it was perceived as useful by many staff in the NHS.  However, over time, I became curious and set about looking for evidence to support LfE. 

This blog post is a summary of the evidence I presented at the #LfEconference18 in November 2018 at Birmingham Town Hall.  This is by no means an exhaustive presentation and I hope that, over time, our Community of Practice will collectively build on the evidence presented here.  The library section on this website is updated regularly; please contact me if you wish to suggest articles / data / other evidence for our site.

How to articulate the message:

Whilst researching the evidence base for LfE, I also started thinking about how best to articulate the message to those who don’t intuitively understand it.  It seems that we do not all speak a common language.

A recently published report from the Carnegie Institute describes two different languages (or “lexicons”) used by policy makers: the rational lexicon and the relational lexicon.  This report (written by Julia Unwin) is primarily about kindness in public policy.  Whilst not directly about healthcare, the content of the report was useful in highlighting different ways to articulate the potential benefits of LfE to policy-makers and leaders in healthcare. 

Occasionally I hear about a colleague who wishes to implement LfE in their organisation, but their manager puts a block on the implementation due to a lack of evidence or “rational” metrics (see the quote above).  According to the Carnegie report, the manager in this example is fluent in the “rational lexicon”, but may have difficulty listening to or understanding a rationale presented in the relational lexicon.  In the words of Julia Unwin, framing arguments in the relational lexicon (e.g. about kindness) may be perceived as “unhelpfully interrupting the adult flow of conversation”.

Taken from the Carnegie report, here are the characteristics of the rational and relational lexicons:

From Kindness, emotions and human relationships. Carnegie Report.

Both lexicons have their risks and benefits if they are used exclusively, and I believe it is possible to recognise both types of language when communicating about LfE (or any other initiative or activity).  In reality the distinction between the two lexicons is not absolute, and it shouldn’t be taken as a fixed dichotomy.  Nevertheless, I have found the model very useful when discussing the potential benefits of LfE and the evidence to support it.

To the evidence:

The evidence collected in this blog post is divided into the following sections:

  1. Stories / anecdotes
  2. Staff survey data
  3. Data from the PRAISe study
  4. Related science and research from outside the LfE movement

Stories and anecdotes:

Although anecdotal evidence does not stand up well to academic critique, story-telling can be a very compelling way to articulate a message.  Each LfE report is a short story: a story describing an episode of excellence.  Simply reading a report, and discussing it with the staff members involved, is often enough to demonstrate worth of the initiative.  The excellence described in the reports is typically a type of non-technical skill, and recipients of reports frequently report “going about their work differently” as a result of receiving recognition.  Often, they were not aware that there actions had had such a positive impact until they received the report.

Within the Community of Practice, there is a growing collection of stories of impact of the initiative, within various different healthcare settings. One example is captured in this excellent podcast describing the implementation of LfE in Derby.

We have also uploaded a few videos of staff describing their personal stories and insights related to LfE on our youtube channel.  Stories describing impact of LfE are often shared informally, and we would like to start compiling some on our library page of the website.  Please contact us if you wish to share stories of impact of LfE.

Staff surveys:

Many staff surveys have conducted across the LfE Community of Practice, some of which have been displayed at our conferences (see our conference poster compilation).   At BCH, we have conducted two large surveys of our PICU workforce.  These surveys were conducted in 2015 (1 year after implementation) and 2018.

The BCH staff surveys show highly positive perceptions of the value and impact of LfE within our department.  The 2018 survey is still under analysis, and is being replicated in other centres currently, but here is a summary of a selection of questions from the surveys along with the proportion of positive responses:

Year of survey Question % of positive responses
2015 By reporting excellence, I am helping to improve patient care 86%
2015 I learn best from studying examples of good practice 87%
2015 Excellence reporting can improve team morale 93%
2015 Excellence reporting can boost my motivation 87%
2018 Receiving excellence reports increases the likelihood of me practising in a similar way in the future 89%
2018 Excellence reporting leads to improvements in the safetyof care we provide 85%

We will be carrying out further analysis of the 2018 survey, and to replicate it in other centres over the next few months.  Please contact us if you wish to use the questionnaire.

PRAISe project data:

The PRAISe project was designed as a proof of concept project to measure the impact of LfE interventions (positive reporting and appreciative inquiry) on an area of clinical interest.  We chose antibiotic stewardship / sepsis as a test area due to a connection with my clinical interests, but the PRAISe methodology could be applied to almost any Quality Improvement situation. 

The project is summarised in full in the Health Foundation report, and we are currently writing a manuscript for submission to a peer-review journal, but some of the results are shared below.

We found that positive reinforcement via excellence reporting and appreciative inquiry interviews was associated with an increase in quality in some of our measured processes (all of which were orientated around behaviours of healthcare professionals). There seemed to be a dose-response effect, in the sense that the largest improvements were seen in the processes receiving the highest concentration of positive feedback. 

The best example was seen in improvements in antibiotic prescribing: the rate of gold standard prescriptions increased during (and after) a period of positive feedback (the red arrows show the start and end of the intervention period):

Our summary interpretation of the PRAISe project is that “if you show staff what they are doing well, they will do more of it”.

Related science and research from outside the LfE movement.

Theoretical evidence from neuroscience:

“Failure is the best teacher”

Some scepticism about LfE emerges from the assumption that we learn best from failure, and therefore we should concentrate our efforts on recognising and interrogating episodes of failure (e.g. error and harm).  This is compelling to anyone who has ever made an error (i.e. all of us).  However, it ignores the fact that success is actually highly instructive – in fact, it is probably more instructive than failure.

Experimental evidence from neuroscience shows that success leads to faster and stronger memory formation (i.e. learning) than failure.  However, this learning can only take place if feedback has been recevied – i.e. we need to know if we have achieved success or failure.  Without feedback we are not aware of the outcome.  The outcome of many of our interactions in healthcare is not always made known to us – this is especialy true for interactions relying on non-technical skills.

This is where LfE reporting has a unique advantage: very commonly, recipients of LfE reports state that they were unaware of the positive impact of their intervention or behaviour. 

LfE therefore provides the feedback to allow learning from success – in an environment where the prevailing approach to learning is to highlight failure.

Staff engagement:

One of the key functions of LfE is to provide recognition to staff (individuals or teams) who have performed excellently.  We know from the thousands of reports we have received, that many of these episodes of excellence are everyday activities – e.g. lending a hand; supporting a peer; going above and beyond;  showing kindness to a patient or colleague…

Thus, the LfE report serves as a means of recognising a colleague’s excellent work.  Recognition is a key factor in staff engagement: staff who feel recognised are more engaged in their work.  This has been demonstrated in several large studies; a compelling example comes from the WorkTrends (TM) survey from IBM:  – a survey of >19,000 workers in 26 countries, from a cross-section of industries.

The survey results demonstrate the positive relationship between recognition (measured as  agreement with the statement, “I receive recognition when I do a good job”) and staff engagement.

The importance of staff recognition is also shown very clearly in this report from Gallup:

Employee Recognition: Low Cost, High Impact.

But, does staff engagement correlate to any meaningful (and measurable) outcome?  In short, yes.  West and Dawson’s 2012 report for the King’s Fund (Employee Engagement and NHS Performance) clearly demonstrates the positive relationship between staff engagement in the NHS, and multiple  important outcomes including quality of services and quality of financial performance:

West, Dawson. King’s Fund 2012.

Thus, LfE provides recognition for staff members; staff recognition is related to staff engagement, and staff engagement is related to multiple important, measurable outcomes.

In conclusion:

I have described evidence to support LfE from within the initiative, and from sources outside the LfE movement.  But, this is not an exhaustive report:  I have scratched the surface, and there is a lot more to discover and explore.  I have also not looked at the potential risks of LfE.  There may be a shadow side of which I am not aware.  Whilst I continue to look for evidence and data from this initiative I welcome comments below and via our forums.

 

Adrian

There is no error?

Standard

I recently attended a FRAM workshop. FRAM is a form of resilience engineering, which allows us to create a model of a complex system. It stands for Functional Resonance Analysis Model; but don’t let that put you off! I spent the first half of the workshop feeling rather confused, but ended up with a basic understanding and an enthusiasm to try the method to create models to better understand my work.

FRAM is underpinned by several principles, the most striking and interesting is: the equivalence of success and failures.

This may not seem intuitive, as we are conditioned to evaluate failure as more significant than success. (E.g. see loss aversion, from Kahneman). But actually, I believe this principle is true for work in complex systems. And it is profoundly important as it provides us with a way to understand our work (and improve it) without having to apportion blame for error. The tendency to apportion blame is, in my opinion, a major hindrance to progress in safety and healthcare in general.

If you make a serious attempt to make rules for every aspect of your work, you will soon realise that is impossible to explain every single action for every possible environment and situation. Thus you will see that in order to go about your work successfully, you will need to continuously make small adjustments. (In FRAM this is called ‘approximate adjustments’). These adjustments are usually successful but occasionally they lead to failure. But whether or not they lead to success or failure, they are essentially the same adjustments.

If we are able to understand this part of our work we will start to have a method for removing blame from error. In fact, this approach removes the idea of error. There is no error; there is only adjustment which may lead to success or failure.

 

Adrian

15s30m

Standard

 


This guest blog is from Rachel Pilling and Dan Wadsworth – founders of a fantastic social movement in healthcare called 15seconds 30 minutes (@15s30m).

So firstly Dan and I are really excited to be writing a blog for Learning for Excellence.  We’ve admired from afar and can see the huge impact this sort of social movement can have on staff and patients alike.

It connects with our social movement really neatly – we are all about Joy in Work – the third leg of Quality Improvement, alongside Patient Safety and Cost Effectiveness.  The role of Joy in Work in underpinning staff engagement, performance, sickness and workforce retention has been championed by the IHI, and the newly published NHS Improvement Health and WellBeing (http://www.nhsemployers.org/your-workforce/retain-and-improve/staff-experience/health-work-and-wellbeing/health-and-wellbeing-framework) brings it under the spotlight for the NHS.

 

 

But lets go back a few steps…what is our social movement?

It’s an idea called 15 seconds, 30 minutes or 15s30m for short.  It asks any member of staff to think of a small 15 second task they can undertake which will save someone else 30 minutes later on, and in doing so reduce frustration and increase joy for themselves, colleagues and patients.  We call these tasks 15s30m Missions and anyone, from chief executive to porter, can get involved.

Here’s an example:  a few weeks ago, we had a power cut at 815am, just as the eye clinic reception was about to open.  The shutters were stuck down and only one PC was working.

Our brilliant receptionist Carol decided that she would stand at the front door and spend 15 seconds greeting patients as they arrived, taking down their name and date of birth on a note pad, and passing it round to the receptionist behind the shutter to start booking patients in – this not only meant that the nursing staff could begin to check visions and put in drops, but when the power came back on we knew who was in the department and there wasn’t a big queue of patients waiting to check in…..but more than that, patients weren’t left “in the dark” (literally!), unsure if clinic was open, worrying that someone might not know they are there. The impact of her 15 seconds was felt across the clinic – what could have been a frustrating day for everyone, was instead a really smooth run session.  Carol is one of our 15s30m Heroes – she sees when something needs doing to make the experience better for the patient, doesn’t wait for permission,  and other people follow her example.

 

We know that the people who know how a department, ward or office is best run is the people on the front line.  By standing on a hospital corridor asking staff, we have collected a list of ideas – 100 ideas in 100 days in fact – and we think some of them will work for any organisation.  We’ve made some little videos on YouTube – go and have a look.  We also invited staff to come to our workshops, where we help them reconnect with the Joy of work, why they enjoy being in healthcare, how they know when they’ve made a difference, what makes a “good day”.  We help them express an idea they have to improve the way work is done – a mission – and empower them to launch it!

What started as a silly conversation about a WashBasket (you’ll have to check out the website, launching July 2018 www.15s30m.co.uk) has grown out from our trust, into our CCG, community nursing homes and other trusts. But this isn’t “our” social movement – its yours. This isn’t a “Bradford” idea, it’s a global one.  We know every hospital has heroes like Carol – people who can make those small changes which have a big impact on someone else – the NHS is built on them.  So we want to help make it easier for people to make the changes, to have the confidence to tell someone how we can make it better. 

There’s lots of ways you can get involved.We’d love you to follow us on Twitter – @15s30m – there’s lots of ideas we post. We are running a 15s30m Festival on 3 October in Bradford – a longer version of the workshop sessions, with a “headline set” from Helen Bevan – and some tools and hints you can take away to start your own @15s30m movement in your local organisation.

Or just drop us an email to 15s.30m@bthft.nhs.uk – we’d love to chat about your idea, help you get going – or just tell us what you’ve done so we can start planting 15s30m Heroes in trusts up and down the country.

Rachel Pilling  is a consultant Ophthalmologist and Dan Wadsworth is a Transformation Manager for Bradford Teaching Hospitals NHS Trust.  Their social movement 15seconds 30minutes was the winner of the NHS Improvement  Sir Peter Carr Award in 2017, recognising a clinician-manager partnership and offering personal development opportunities for them to improve their leadership and managerial skills.

Choice architecture

Standard

It’s very easy to focus on the negative aspects of one’s experiences. It’s the path of least resistance – it actually takes more effort to refocus one’s attention onto the vast amount of excellence which occurs everyday. Much has been written on the reasons for this negativity bias, and I won’t explain it here. And I do concede, very willingly, that much of our progress in healthcare (and in society in general) has resulted from our ability to notice the negatives; so I would never say that we should stop doing it.

But there are also many benefits from noticing the good. When you notice what is working, you learn something new. When you show appreciation or gratitude for the good, you improve your mood and that of those around you. Unfortunately, it takes effort to use the lens of positivity. Can we make it easier to access our positive worldview?

I contend that we can use choice architecture to make it easier for us and our colleagues to CHOOSE to see the positive aspects of their work. Providing easy access to an excellence reporting system, and making sure that the system works (i.e. positive feedback is forwarded in a timely manner) are key steps to making it easier to recognise excellence and to show appreciation. Linking the excellence reporting system with the adverse incident reporting system also provides an easy opportunity for staff to balance their observations of their workplace.

We know from 4 years of experience (and thousands of excellence reports) that this does not reduce the amount of adverse incident reports- it simply adds more intelligence to the reported data.  It also  makes it easier for staff to chose to notice what is working, and to show appreciation to colleagues.

 

Adrian

Work as done and work as imagined: a role for LfE?

Standard

Much has been written on the concept of work as done (WAD) vs. work as imagined (WAI). Essentially, the idea is that the work done at the sharp-end / on the shop floor (i.e. WAD) differs significantly from that which is documented in standard operating procedures (WAI). Understanding this difference is key to understanding why adverse events occur, yet this approach is often overlooked: the prevailing approach to adverse incident analysis is often based on an assumption that WAI is the reality.

Understanding WAD is not easy and probably requires a completely different approach to commonly used methods like Root Cause Analysis (RCA). WAD depends on variability of performance including improvisation and work-arounds. This variability is essential for socio-technical systems (such as health-care) to function, but this variability can also be the source of failures. Unfortunately, adverse incident analyses tend only to highlight the negative side of human variability, so efforts to make systems safer often result in the imposition of more and more constraints.

How can we understand WAD better? In particular, is it possible to capture the positive side of variable performance?

Capturing WAD necessarily requires real workers describing how real work is done. Hence, the understanding must come from the ‘sharp-end’. Various methodologies exist, but I would advocate the value of excellence reporting. The vast majority of LfE reports describe non-technical skills whereby success has occurred despite difficult conditions.  These non-technical skills (e.g. generosity, kindness, going the extra mile) are not featured in WAI, yet they are assumed.  It is my contention that positive human interactions are a core component of WAD and should be actively noted and appreciated. LfE is designed to do just that.

 

Adrian

Hawks and Doves

Standard

I just spoke about LfE to some laboratory staff. There was an excellent question from the audience (paraphrased):

“I perceive excellence as rare and exceptional. As a ‘hawk’ this makes sense to me. Is LfE about capturing this stuff, or more about the everyday ‘good’ activities?”

In my answer I tried to articulate that excellence is subjective. It doesn’t come with a priori definitions. It also doesn’t have to be intimidating. We have thousands of LfE reports from our institution and the vast majority describe a simple episode of non-technical activity involving one or more colleagues. On the face of it these reports often seem to be descriptions of ‘people just doing their jobs. But in every case, there was something excellent about the episode, in the eyes of the reporter.

That is all that is required to trigger an excellence report. The positive feedback is powerful and informs the recipient about the impact of their actions, the extent of which is often not known by the recipient until the report arrives.

Hawk or dove, you can use LfE to show appreciation to colleagues, based on your own definition of excellence.

 

Adrian

Story of implementation in County Durham and Darlington NHS Trust

Standard

Here is a guest blog from Dr Richard Hixson, describing the rapid implementation of excellence reporting in his trust – County Durham and Darlington NHS Trust.

Excellence Reporting; conception, implementation and experience.

Working as a Deputy Medical Director in Patient Safety was eye opening and rewarding but at times confusing. We focussed on the incidents, the actions and learning but however we dressed it up, it always felt as though we couldnt shake the negative. An averted incident could only be reported as a near-miss and there was simply no mechanism to understand how somebody’s positive actions had prevented harm from occurring. The machine that processed incidents responded rapidly whilst individuals cited in patient experience documents remained blissfully unaware of the praise they were receiving. We simply had an imbalance in our processes.

Wrapped up in the world of Serious Incidents, complaints and mortality, I was totally unaware of concepts such as ‘Safety 2’ and ‘Learning from Excellence’. I repeatedly expressed frustration, tagging a slide onto every presentation I gave stating we needed a smaller stick, a bigger carrot, a focus on learning from the positives and a mechanism to recognise the good stuff. One stairwell rant was overhead by a dietician, Jennie Winnard who was aware of the work taking place in BCH. We therefore decided to team up and pay the Executive Body a visit. Our vision was presented in June 2016 and we were immediately challenged with creating a fully functional excellence reporting platform for the largest trust in the Northeast of England. We had just over 2 months to deliver over the summer holidays working to a budget of £0.00.

Utilising our Patient Safety colleagues and Ulysses, the company behind Safeguard, we succeeded in creating a module that sat alongside incident reporting and whilst being similar in aesthetics was much, much simpler to complete. Due to the time pressure, there was little fanfare accompanying the September 1st launch as we relied upon simple communications: emails, trust bulletins, screen savers and word-of-mouth. We sat back and waited to see whether anyone else ‘got it’ eager to see how our initiative was received by colleagues.

Fast forward 16 months and the results can only be described as staggering. In the first year alone, 1131 reports were filed naming 1634 members of staff with 75% for ‘going the extra mile’ and ‘team work/peer support’. Summaries were being provided to Care Groups and integrated into governance meetings, bulletins ran short stories on ‘the good stuff’ whilst surveys revealed recipients felt more positive about themselves, their colleagues, their job and even the Trust for weeks or months after receiving a report.

What started as a pure ICT portal has now extended to ‘ER cards’ which can be used for staff such as domestics who do not access email. Even without re-marketing, excellence reporting continues to gather pace with increasing numbers of reports filed every week. Positivity is addictive with recipients looking out for and recognising the excellence in others which simply manifests as high-quality care being provided by ‘ordinary staff’ who feel they are ‘just doing their job’.

As one of our recipients reported – “Of all the changes in the Trust, the little addition of excellence reporting has made a tangible difference to the working lives of many. Most of us, including me would not like to tell the world how good we are at what we do. Excellence should be perceived by others. Excellence in patient care should remain our motto and inspiration”.

On reflection, it just seems so obvious that this is just what the staff and Trust needed. I just cannot believe it took us so long to appreciate and implement.

Richard Hixson, Consultant in Anaesthesia and Critical Care.

County Durham and Darlington NHS Foundation Trust.

richardhixson@nhs.net

 

Is Learning from Excellence “Safety-II”?

Standard

Learning from Excellence is often described as a “Safety-II” initiative.  I can see why this is the case, but the truth is that I implemented LfE before I had even heard of Safety-II.  That’s not to say the Safety-II was an underground movement; I was just a bit slow to find out about it.

The principle aims of LfE are to improve quality (through gaining insight by looking at hitherto under-studied parts of our system) and to improve morale (through formal positive feedback).  Safety-II is a concept based on the idea that safety can be considered a condition where as many things as possible go right; rather than the prevailing approach to safety – Safety-I – where we consider safety a condition where as few things as possible go wrong.  In the history of ideas, Safety-II is a very new one.  Whilst its theoretical principles are increasingly well defined, there is a distinct lack of practical application – particularly in healthcare.  How does one actually “do” Safety-II?  While we wait for the answer to this question, which may take years to come, we are tempted to “fit” initiatives, like LfE, into the Safety-II label.

Does LfE fit into Safety-II?  I think the answer is yes and no.  But a bit more yes than no.  LfE is about identifying success, and viewing it through a learning lens.  The name of the initiative suggests that it tends to identify extremely good (i.e. excellent) examples of work.  Indeed, this was the original idea of the initiative.  However, after looking at over 2000 reports, I have concluded that we are not capturing rare episodes of excellence – we are actually capturing “everyday excellence”.  The vast majority of the reports are a description (or short story) of a small work-around, improvisation, or a generous human touch which allowed success to occur in difficult circumstances.  It turns out that neither difficult circumstances, nor generous human touches are hard to find in healthcare. The success of the initiative is due to the fact that these have been happening since the start of organised healthcare, but have been unrecognised through formal reporting systems.

In the Safety-II construct, we could make our systems safer by understanding day to day work better.  Since success happens most of the time, we should be studying and understanding what happens most of the time, in order to recognise the work-arounds / adaptations / improvisations which create (and underpin) the conditions which allow success to happen.

If most of our LfE reports are about everyday work, I would argue that we have essentially created a system for “doing Safety-II”; at least in part.  What we haven’t created is a system which understands every element of everyday work.  This is someway off.

The fact that LfE is not a perfect practical solution for Safety-II is not a reason to change it. On the contrary, it continues to grow and spread positivity and positive change in healthcare (and beyond), so there is no need to make it fit into anything.

Show me the “learning”

Standard

Questions from the sceptics: 2. Show me the “learning”.  This is the second in a series of short blogs about common questions we receive about LfE.

I’m sometimes asked to demonstrate the “learning” from learning from excellence.  Often, LfE is regarded as a mere “pat on the back” for a job well done:  a harmless initiative, but not of any tangible utility.  What is the actionable intelligence?  Where is the “learning” which can be transferred from one situation to another?  This view of LfE is missing the effect of feedback on performance.

The core of LfE is a simple, formal positive feedback tool, which allows peers to show appreciation to each other.  Recipients of excellence reports are made aware of the positive effects of their actions.  This gives them the opportunity to reflect and think about why their actions were so well received.  This may prompt comments like “I was just doing my job”, yet staff members often go on to make changes in their future behaviours based on the new awareness they have of their positive actions.

So the principle type of “learning” in LfE is the same as the learning we experience from any type of feedback.  The main difference is that LfE is exclusively positive feedback – an extremely rare phenomenon in today’s NHS.

Adrian

On stone cutting in PICU

Standard

This guest blog post is from one of our PICU colleagues, Dr Heather Duncan. Heather reviewed some of our LfE reports, and wrote this blog post:

It’s not what we do but how we do it….

Reading the recent LfE reports reminded me of the story about the three stonecutters who are asked what they are doing. The first one says, “I am cutting a stone.” The second one says, “I am cutting this block of stone to make sure that it is square and it’s dimensions are uniform, so that it will fit exactly in it’s place in the wall.” The third stonecutter grins and replies, “I am building a cathedral.”

Each of us with our small tasks contributes to a much greater achievement. It is in particular not what we do, or even that we do, but how we do our work that we identify as excellent. Rather than stonecutting, in BCH the reports are about receiving and sharing patient feedback, being a team leader, a band 7, Admin colleague and consultant on both ordinary and difficult days in BCH. I have collected words from these IR2s about how we do our tasks that makes them notably excellent. Try reading them aloud. Calm, run smoothly, energy, positivity, helpful, resilient, considered, accepting, listened, sensitive, unruffled, go the extra mile, pleasure to work with, enthusiastic approach, empathy, compassion and respect. They are all descriptive words of how actions look and feel when we excel.

There are a few tasks in each of our roles that feel like cutting stone. Next time I come across one of those tasks I’m going to try to remember these positive words and see whether I can remember to do the routine with the energy and positivity; learning from excellence. Like the third stonecutter the tasks we perform every day are part of a whole and necessary for the grand picture of what we do, “building a cathedral.”

Heather Duncan