Chapter 9 The implementation challenge

The approach of relying on passive diffusion of information to inform health professionals about safer practices, is doomed to failure in a global environment in which well over two million articles on clinical issues are published annually

This was a quote from my doctoral thesis in 2008 and still applies today.

I am biased but to me implementation is the most important thing we need to get right; more important than improvement and more important than innovation.  We can create all sorts of new ideas, products, and policies but if they are not implemented they are all a waste of time, efforts and resources.

In healthcare we are drowning in ways in which we could improve; numerous interventions and solutions, lots of research and guidance.  We create standards, alerts and ‘must do’ notices and targets.  There have even been repeated alerts published and disseminated in relation to the same topics in patient safety.  This should tell us that the method of ‘telling people just to do it’ isn’t working.

In the early days of patient safety many of us fell into the trap of disseminating guidance and expecting change to happen simply as a result, but after a while we have been forced to admit that things didn’t turn out as we had originally intended and planned.

Implementation is a complex process not a one off event.

Vincristine error

If we study vincristine errors it provides us with a window into the complexity and difficulties faced in relation to implementation. The following detail is mainly taken from the excellent summary of the quest to eliminate intrathecal vincristine errors, a 40 year journey by Noble and Donaldson (2010).

In 2004, in the UK, a review was conducted to assess the level of implementation of guidance for the safe administration of vincristine.  This was three years after the revised guidance had been issued and disseminated.  Despite two very high profile deaths related to vincristine error in the UK, the review revealed only 50% compliance.  Despite these tragic events sustained change had not occurred across the system.

The specific risks of inadvertent administration of vincristine were clearly recognised from the early experience in the 1960s. However, to date around 60 reported cases of intrathecal vincristine errors are known to have occurred with several others assumed as unreported (Noble and Donaldson 2010).  Since they started articles have been written, alerts sent out, guidance disseminated but further deaths occur.  This tells us about the effectiveness of the current approach to change.

The ‘nirvana’ of patient safety is to identify a design solution – one that makes it impossible for mistakes to happen.  Implementation of the design change for vincristine has been an enormous uphill task and we are still not quite there yet in terms of total compliance (Noble and Donaldson 2010).

Rather than wait for the design solution, alternative solutions have been sought.  In recent years, the use of a minibag to deliver vincristine has become increasingly discussed and advocated.  This is not a full physical design solution or forcing function but offers an interim measure.  It involves not using syringes to deliver vincristine at all and instead diluting the drug in a minibag, working on the premise that it would be virtually impossible to deliver this to a patient through a spinal needle.  An alert issued by the National Patient Safety Agency in 2008 stated that no incidents had been reported where a minibag had been used (NPSA 2008).  The World Health Organisation published an alert immediately following the death of a patient in Hong Kong in July 2007 and recommended using the minibag to prevent errors in the absence of the more formal design solution (WHO 2007).

In China, in 2007, two drugs normally given intrathecally (methotrexate and cytarabin hydrochloride) were contaminated with vincristine at the factory level (Noble and Donaldson 2010).  At first, a few children in Shanghai and Guangxi Zhuang Autonomous Region suffered neurological symptoms. Paralysis was common, and these incidents increased the world total from 58 to 251 overnight.  Eventually the outbreak was believed to have led to 193 patients across China having intrathecal vincristine unintentionally. The Chinese government recalled the drugs and closed the plant.

As Noble and Donaldson report, unusual events like this do not get reported via incident reporting systems and even the best design solution may not have prevented such an unpredictable and ‘upstream’ form of error. Yet, this story keeps us attentive to the on-going dangers in every aspect of healthcare, from factory to frontline and adds another vulnerability to this long-running story of unsafe care and the complexities of implementation of good practice.

Noble DJ, Donaldson LJ (2010) The quest to eliminate intrathecal vincristine errors: a 40 year journey BMJ Qual Saf 19:323-326

The checklist

The use of checklists are another interesting example of the difficulties of implementation. It remains astonishing to me today that the final WHO surgical checklist designed had the approval of over fifty different countries from across the globe. To all come together and agree the core components of what should and should not be included in a simple one page checklist is astounding.  It is important to remember this achievement in the light of the criticisms that have followed since.

Checklists in themselves do not prevent or reduce harm; it is the way in which the checklists are used that can do this.  Properly used, the checklist ensures that critical tasks are carried out and that the whole team is adequately prepared for a surgical operation.  However what we find is that rather than full implementation there is mere compliance and there is huge variability in use and implementation.

People think that if you implement the main bits patients will still receive the total benefit.  The checklist to work requires all of the component parts to be implemented but in the main around half of it is used.   In some organisations it has been found that components of the checklist were missed out or incomplete, for example the sign out and debrief post checklist were particularly poorly carried out.  The risk with this is that if you only implement some of the checklist then you may end up by receiving none of the benefits.

If you are in any doubt as to whether checklists can improve patient safety all you need to do is read The checklist manifesto Published by Profile Books in 2010, written by Atul Gawande.

Chapter 10 will talk about what we can do differently for implementation to increase the success of implementing and embedding patient safety solutions and interventions.