What we learned in setting up a care bundle to support eating and drinking at risk

13 August 2020

Dr Peter Sommerville is a consultant in geriatric and stroke medicine at St. Thomas’ Hospital. He is a member of the multi-professional FORWARD team who, for the last five years, have been focussed on introducing and maintaining quality in the care of patients who are ‘eating and drinking at risk’. The project recently received the Royal College of Physicians’ Excellence in Care Quality Improvement Award.

Dysphagia is common in hospital and when there is an appreciable risk of choking, aspiration or pneumonia, people can be said to be ‘eating and drinking at risk’ (EDAR). This is usually considered when their swallow is unlikely to improve in a reasonable time frame, there is no safe and acceptable modified diet, and tube feeding has been ruled out. Decision making is complex: people’s preferences and capacity must be sought, as well as the views of others when best interests are discussed. The nature and prognosis of the dysphagia must be elucidated as must the practicality of different oral and tube feeding approaches. The risks are heterogeneous and do not pertain to oral feeding alone; many people with severe dysphagia may aspirate oral secretions and risk pneumonia even when tube fed. Moreover, people’s clinical condition and preferences may change over time.

If timely multi-professional assessment, good communication and the making of clear plans do not occur, patients in clinical settings may be kept nil by mouth (NBM) for an inappropriately long time, and people and their families may suffer undue distress.

All of the above recommends an approach to decision making and management of EDAR that is, collaborative, multidisciplinary, patient-centred and which is active rather than passive.  We identified cases in 2014 in our hospital where this did not occur and so we started a care bundle, initially based on work from another hospital, to try and support the care of people EDAR.  Today, the FORWARD bundle is used as standard in our hospital and has supported over 600 patients. Here are some of the lessons we learnt:

Start small, the first attempt doesn’t have to be perfect

Early versions of the bundle didn’t quite get it right first time for us and the care bundle now used is in its 18th iteration. We started with a pilot on just four wards. In retrospect we could have iterated faster if we had started smaller still –there isn’t anything wrong with starting a plan-do-study-act cycle with a single patient. 

Get ‘buy in’ early

We involved colleagues from across geriatrics, nursing, speech and language therapy (SLT), dietetics, palliative care and POPS early on. It’s good to get everyone together to ask ‘how do we solve this problem?’ but it’s better to ask ‘what are the problems?’ This helped with engagement, making it meaningful, and it makes sure that the right data gets collected.


Enhancing the ‘ownership’ of FORWARD within the SLT office ensured reliability. This engaged but concentrated group (should) get involved with almost all EDAR patients and so changes and developments in the bundle and its promotion can be quickly instituted.

Data drives improvements

We collected data about patients and usage right from the start. This helped us work out if we were improving anything (we used process measures such as completion of best interest meetings, capacity assessments, communication with families, swallow plans in discharge letter, and an outcome measure of time NBM). When things went in the wrong direction, we were able to do focused interventions. We found it harder to get qualitative data – try to get more than you think you need! Our results helped us make a case for expansion, and subsequently for a charitable grant. By keeping a balancing measure (rates of whole ward chest infections as FORWARD was instituted) we were able to provide risk assurance where it was required.

Checking boxes isn’t always the same as great clinical care

Care bundles ought to be straightforward (some say no more than 4 actions- ours is a little longer) and often appear to have a property of ‘completeishness’ when the boxes are ticked. However, when you actually measure the care that happened it may not be quite as complete. Bundles alone cannot make good care happen – they have to be supported by education, promotion and outreach.


Add new comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.