Public inquiries sometimes feel like a version of Groundhog Day where it is repeatedly stated that ‘lessons will be learned’. One of the key reasons for an inquiry is to prevent similar events from reoccurring, but for that to happen, recommendations must be implementable and implemented.
Most of the hundred plus NHS inquiries that have been held since the Ely inquiry (the first NHS public inquiry) have highlighted the same areas for concern: inadequate leadership, system and process failures, poor communication, disempowerment of staff and patients. Either the NHS as an institution is unable or unwilling to implement the changes embodied in the recommendations; or the recommendations themselves are incapable of being implemented because of the way in which they are produced or expressed.
There is much evidence to show how other industries, including aviation and oil, have learned from past breaches in safety and adapted their practices successfully. So what prevents the NHS from implementing its own recommendations?
About the inquiries
I compare the ‘implementability’ of recommendations from the three NHS inquiries of Ely, Bristol and Mid Staffordshire. In terms of background, the Ely inquiry was set up in 1967 to look into allegations of various forms of misconduct on the part of members of the staff at the Ely Hospital, a psychiatric hospital in Cardiff [link Hilton]. It was chaired by Geoffrey Howe and reported in 1969.
The Bristol inquiry was set up in 1998 to examine the ‘excess deaths’ in paediatric cardiac surgery at the Bristol Royal Infirmary between 1984 and 1995, with Professor Sir Ian Kennedy as chair. It reported in 2001.
The (second) Mid Staffordshire inquiry was set up in 2010, into the serious failings at the Mid Staffordshire NHS Foundation Trust between 2005 and 2008 (see Smith and Chambers, in this issue). It was chaired by Robert Francis, who had chaired the first inquiry that reported in 2010. This second report focussed on the wider NHS system, reporting in 2013.
Everybody’s concern, but no one’s responsibility
My analysis of Ely, Bristol and Mid Staffordshire asks: is it clear who the recommendations are aimed at? And is the problem soluble in the sense of a clearly identified policy tool, or mechanism, or tool to implement, which suggest a clear course of action?
The first issue regards the clarity of the recommendations in the sense of: to whom they are aimed and who will take ownership of the problem? This may be seen in a very broad analysis of micro (individual patients and clinicians); meso (institution) or macro (system) level.
Some of the inquiries’ recommendations were clearly targeted. The most clear and explicit were those with an active voice: for example ‘the Care Quality Commission (CQC) should…’ rather than a passive voice of ‘consideration should be given to…’. The vaguest recommendations are seemingly aimed at everyone. For example, Francis’ first two recommendations that ‘require every single person serving patients to contribute to a safer, committed and compassionate and caring service’ may be seen as ‘everybody’s concern but no one’s responsibility’.
A rough estimate can be made of active and explicit recommendations, albeit with a fairly wide margin of error owing to difficulties of interpretation. My estimates suggest that just three of Ely’s forty‐four recommendations (7 per cent) have a clearly identified agent. For Bristol, it is thirty of 198 (15 per cent), of which some four are joint. Finally, for Mid Staffordshire, it is 118 of 290 (41 per cent), of which some twenty‐one are joint.
What? Or, the problem with sermons
The second issue concerns policy instruments or tools. Marie-Louise Bemelmans‐Videc
et al outline three broad types of policy instruments, which have been termed incentives, authority and persuasion but are usually called ‘carrots, sticks and sermons’.
The first type consists of incentive tools such as the conditional transfer of funds or charges and fines. The second type used coercion as their principal resource. Governments employed them through their hierarchical system, and their most common typologies include permissions, guidance, and compulsory actions. The third type of persuasion referred to a series of discursive strategies aiming to change behaviour through providing information or using moral‐based arguments, or sermons.
An example of a sermon is “The education and training of all healthcare professionals should be imbued with the idea of partnership between the healthcare professional and the patient (Bristol, para. 2).” Often, such sermons are not enforced by incentives or sanctions.
As is show on table 2 in the longer version of this article, sermons are the main policy tool accounting for some 89 per cent of Ely recommendations, compared to 66 per cent at Bristol and 63 per cent at Mid Staffordshire. Although Ely was a very different type of hospital in a different era, it seems that the Bristol sermons did not prevent the failings at Mid Staffordshire. This suggests either that sermons in general do not work, or that these sermons did not work, as they were either not implementable or implemented.
The data suggests that more could be done to arrive at practical recommendations of a series of practical doable steps that are clearly ‘owned’ by an identifiable agent. Pulling these issues together, given the large number of potentially responsible agencies, it is suggested that recommendations should be ‘active’ with a clearly identified agent and that a clear policy tool or mechanism should be identified rather than rely on a vague tendency to sermonise.
A longer version of this article originally appeared in the Political Quarterly journal.