Introducing Lean in to an organisation takes work. Sometimes it flies. In other cases it is a long, hard slog, or is abandoned altogether. I’ve written about problems in using Lean in the past, including in a book review and a discussion of problems in Canada. In this post, I want to talk about the other side of this equation: what makes quality improvement work?
Healther Kaplan has spent a lot of time looking at QI success. Her findings give information about context and quality improvement, and relate well to the Lean literature. Combined with a review by Fulop and Robert, we know a great deal about what helps improvement to succeed. (Links to abstracts of the papers are at the end of this post. Please also see the method note. )
Leadership and Strategic Alignment
Fulop and Robert found leadership and strategic alignment to be one of the factors more consistently associated with success. The strength of the associations varies, but combining the various reviews gives the following factors:
- A supportive Board that takes an interest in progress on QI
- A Senior Leadership Team that sets clear priorities and aligns QI efforts with the priorities
- A culture that supports improvement, rather than blame
- Alignment of human resource policies and staff training to support QI work
Much of this is familar to Lean readers from Hoshin Kanri or World Class Management. Kaplan’s work suggests an interesting detail: her review found that senior management having a vision and strategy that was communicated to the rest of the organisation was important, but that top down planning and implementation was not a significant predictor of success. This ties in with David Mann’s view that senior managers need to be able to support the system, but don’t need to do everything themselves.
This also relates an often described idea that you will do better involving all of your staff in QI, than by keeping it in the hands of the fiew. The catchball process should help with this, by letting people influence strategy, and in particular how strategic aims are delivered.
Organisational Culture and QI Support
The value in aligning other mechanisms at the organisation’s disposal, such as training plans, personnel policies, recruitment and even financial systems, is in having each reinforce the other. This presumably also feeds in to organisational culture, one of the other variables that comes out of the reviews.
Generally, organisations where managers, doctors and other staff are more supportive of QI, do better. This is a chicken and egg situation: QI successes influence staff, but it can be difficult to produce improvements without staff buy-in. As most of the studies are cross-sectional (see Method Note below) they are silent on how you get from one to the other. There are glimpses of the process, however.
Organisations that have QI teams who report adequate resource tend to perform better. This may relate to both capacity to engage with clinical teams on improvement, their ability to offer training, and the general visibility of the work. Combined with supportive personnel policies and management alignment, it seems likely that this in itself creates a more positive organisational attitude to quality improvement.
Kaplan and colleagues report that physician leadership is associated with success, and that motivation and capability in teams are both important. Capability was more consistently associated with change than was motivation. This may mean that motivation on its own is often not enough – you need the skills, infrastructure and organisational support to make an impact.
Information resources and ability to obtain, display and analyse data also featured as positive factors in several of the studies.
Fulop and Robert reported that the organisational alignment factor remained important at team level, noting, ‘consistently positive findings in relation to ‘quality and coherence of policy’ and ‘supportive organisational culture’ , although they did note that this was based on a small number of studies. This suggests that, if teams know priorities and have the skills to tackle problems in a culture that offers them the safety to experiment, then they can and do make improvements.
These findings fit well with lived experience of Lean work. Organisational alignment; clarity of aims; promotion of learning over blame; a culture that supports improvement; good leadership at Board, Leadership Team and clinical team level; an adequate QI infrastructure; good use of information and maximising the influence of other methods to support change, such as human resource policies and training strategies, are all familiar to Lean practitioners. When introducing Lean approaches in to an organisation, or extending their use, paying attention to these factors will pay dividends.
Illustration courtesy of GDJ at openclipart.org
The main sources used for this blog post are:
Almost all of the available studies on context and QI are cross-sectional. The studies are snapshots of a point in time. This means they can only show association, not causation. What you see on a cross-sectional study does not always reflect history.
For example, if you looked at people with Chronic Obstructive Pulmonary Disease (COPD) you might find that smoking rates were not particularly high. We know, however, that smoking tobacco is a major cause of COPD. How can this be reconciled? COPD gets worse if people continue to smoke, and also because of repeated infections. People are routinely offered advice and support to stop smoking. A modest rate of smoking in people with COPD does not tell you anything about past smoking.
The same general issue arises in all ‘snapshot’ studies: it’s hard to know what happened previously. Even if you ask people, their recollections may be affected – biased – by what they know about the current state.
There are two other relevant issues – the effect of intervention, and confounding. Taking confounding first, most studies reviewed by Fulop and Robert did not include any analysis of how variable interacted. This can matter a lot. Some effects are heavily modified by the presence of other factors. You can even reach the wrong conclusions. For example, in the UK, a disproportionate number of cases of lung cancer occur in people wiht low incomes. When you adjust for smoking levels, however, the difference reduces enormously. Smoking levels are higher in poorer groups than in more affluent – it’s the smoking, not the poverty, that causes the cancer (although the reasons people smoke more may well be related to poverty).
Finally, showing that a is associated with b does not always mean that changing a then affects b. Medicine is full of examples of this. In Senile Dementia of the Alzheimer’s Type, for example, amyloid plaques are common. Removing the plaques should reduce symptoms, right? Well, maybe. Some studies have found ways of removing plaques without improving symptoms. This might be because damage has already happened, and ways of stopping the plaques in the first place might be better – but it does show that b does not always follow a. Studies where people systematically change aspects of an organisation, and then show increases in QI performance will be needed to convince some people.