Written evidence from Anthony W. Fox,  BSc MBBS MSc MD(Lond) FRCP DipPharmMed MCSFS DipFHI Cbiol, Birkbeck, University of London and

EBD London Ltd


Are Regulation 28 Reports and their legal framework performing well ?



I am a physician with interests in the areas of clinical pharmacology, clinical toxicology and forensic medicine; this submission is based upon a piece of research that will comprise a dissertation for a Master’s degree in Law at Birkbeck, University of London.  My reason for submitting this evidence is because it addresses point no.3 of the Call for Evidence, viz., ‘Ways to strengthen the Coroners’ role in the prevention of avoidable deaths’.  The only active vehicle for this is the Reg.28 report, which this investigation shows to be quite haphazard.  An abstract follows which summarises the study, and makes some recommendations that follow from it.



H.M.Coroners in England and Wales have a duty to write a Regulation 28 (Reg.28) report when they identify causes for concern that, if addressed, could prevent future deaths.  These reports are provided for by Statute, Regulation, and Guidance no.5 from the Chief Coroner, and are made on a standard template.  The report is addressed to whomever the Coroner believes can take the necessary action to prevent future deaths, and must be copied to the Chief Coroner.  There is a presumption of publication by the Chief Coroner of both of the reports themselves and the responses received from addressees. 


Since 2013 publication has taken the form of web pages with reports being sorted into various categories of fatality.  The 50 most recent reports from the three categories ‘Child deaths’, ‘Alcohol, Drugs and Medications (AD&M)’, and ‘Railways’ were selected for study. A database was constructed comprising demographic data of the deceased, the report source (i.e., Coroner’s Area) medical cause of death, circumstances of death, Coroner’s concerns, addressee, response from addressee, and the types of remedial action (if any). 


The findings are that: i) Coroners generate Reg.28 reports at very varied rates, ii) there is a large deficit of published addressees’ responses to published reports written, iii) addressees from large organizations (especially NHS entities and Network Rail) are more likely to respond to a Reg.28 report, iv) substantive remedial actions are found in only a subset of the published addressee’s responses, and v) there is a gender imbalance with males predominating in all three categories; this imbalance extends even to neonates.  The legal framework is often vague in distinguishing the Coroner’s duty from discretionary powers, leading to subjective decision-making.


Thus, one conclusion that the Reg.28 report system is haphazard in many respects.  This does not serve the future public health and safety well, and risks undermining the altruistic motive of bereaved families, and the confidence of the wider public, to support the process of inquest. 


Recommendations for process improvement are:


1.  There should be a national, consistent rule about when a Reg.28 report should be written.  The Regulations and Chief Coroner’s Guidance should be amended accordingly.


2.  There should be a mechanism of enforcement of responding to reg.28 reports.  This might include powers for Coroners to follow-up with those ignoring a Reg.28 report, and sanctions for those in contempt.


3.  While recognising the need to balance parties’ rights against the proper exercise of public law, some protections for those responding to Reg.28 reports may be needed if civil liability is deterring responses.


4.  A national archive of inquest proceedings (without or without a Reg.28 report) would be a useful tool for creating law reports, audit, promoting national consistency in Coroners’ practices, and enhancing the performance of Reg.28 reports in preventing further deaths.


There are two further research recommendations: i) the methods of this study could be expanded to the all published Reg.28 reports (rather than a sample), and ii) the reasons for gender imbalance (even among neonates) should be explored.



AD&M:- Alcohol, drugs and medications

CJA2009:- Coroners and Justice Act 2009

Reg. 28:- Coroners (Investigations) Regulations 2013, Regulation 28.

RPFD(s):- Report(s) on Prevention of Future Death(s)


Table of Contents


1.  Introduction …………………………………………………………………..              4

2.  Research question ……………………………………………………………              7

3.  Methods ………………………………………………………………………              7

4.  Findings ………………………………………………………………………              9

5.  Analysis ………………………………………………………………………              15

6.  Conclusions ………………………………………………………………….              22

7.  Recommendations ….......................................................................................              23

1.  Introduction


The Regulation 28 (Reg.28) Report


1.1 When Coroners, as part of their investigations, discover information indicating a risk of future deaths then CJA2009 s.32, Sch. 5, Para 7(1) requires that a report should be sent to those believed to be able to mitigate that risk.  A written response is required, and copies of both are sent to the Chief Coroner, whose archive shows that, in practice, almost all Reg.28 reports are issued after inquests (see below).  These reports have a standard template (for an example, see Annex 1).[1] 


1.2 Prima facie, the secondary regulation[2] reflects an elevation of Reg. 28 reports into a Coroner’s duty, whereas, formerly, they were only required by a rule.[3]  The Chief Coroner’s Guidance no.5 (para 5) provides that Reg.28 reports:

‘… should be intended to improve public health, welfare and safety.  They should be clear, brief, focused, meaningful, and, wherever possible, designed to have practical effect.’

The Guidance further elaborates (para 10) that the Coroner’s duty arises when something revealed during her/his investigation causes a concern that ‘circumstances creating a risk of further deaths will occur, or will continue to exist, in the future.[4]  Moreover, the concern need not be directly related to the death of the deceased per se, but could involve tangential issues, e.g., the coincidental discovery of unlicenced firearms in the home of a person committing suicide by other means.[5]  However, the Guidance fails to advise when to originate a Reg.28 report beyond that there should be a ‘relatively low threshold’ which, itself, is without definition.  Thus, it is for the Coroner her/himself to decide whether the duty has arisen, essentially making the ‘duty’ discretionary, rather than mandatory.


1.3 Reg. 28 reports are addressed to those whom the Coroner ‘believes’ can take action to prevent further deaths.  The addressees may have become obvious during the inquest. 


1.4 The Coroner must copy a Reg.28 report to the Chief Coroner, who anticipates receiving and publishing more than 600 per year.[6]  The Regulation gives the Chief Coroner a discretion whether to publish, and if so, then a power to edit and redact.[7] The same applies to responses received.


The Reg.28 Report Template


1.5 A template for Reg.28 reports is provided by the Chief Coroner.  Box 6 of the template (see Annex 1) reads:

              “In my opinion action should be taken to prevent future deaths and I believe you [AND/OR your organisation] have the power to take such action.”

Furthermore, a response is demanded, again in stereotypical wording in Box 7:

                You are under a duty to respond to this report within 56 days of the date of this report, namely by [insert date]. I, the coroner, may extend the period.

              Your response must contain details of action taken or proposed to be taken, setting out the timetable for action. Otherwise you must explain why no action is proposed.

Coroners are not encouraged by the Guidance to specify precisely the remedial action that should be taken.[8] 


Publication and its purpose


1.6 The primary purpose of the Reg. 28 report is therefore to notify those likely to be able to create change to do so, to document their response, and to publish that correspondence.  This is not intended as a ‘name and shame’ process[9], which would threaten the principle that Coroners do not assign culpability, but are finders of fact.


1.7 Importantly, Reg.28 reports are the only way that cases from Coroners’ courts are officially reported: there are no equivalents of law reports, no published judgments, and no national repository of transcripts from modern-day inquests.[10]  Some appeals concerning inquests are reported as judicial reviews in the High Court, but these are usually about disputed verdicts or complaints about procedure, and are only tangential (at best) to the public health aspirations of the Reg.28 report.[11]  Lastly, there is nothing in the Guidance, the Regulation, or the Act[12] suggesting that Coroners have any responsibility for following up on addressees’ responses or actions.


2.  Research question


2.1 Is the Regulation 28 report system efficiently contributing to the objective of improving public health and safety, and, if not, then how might this process be improved ?


3.  Methods


Database source


3.1 Regulation 28 reports are published,[13] and The Chief Coroner’s Office has stated that its aim is to upload all reports made since 25 July 2013.  These reports are categorised, and a single report may be filed under more than one category (Table 1). Redactions by the Chief Coroner’s office were minor, and usually the personal names (but not always the appointed post) of those who were not interested persons,[14] had not given evidence at the inquest or, occasionally, those signing responses from addressee organizations. 




3.2 Cursory inspection suggested that unique cases might be allocated to more than one category by the Chief Coroner’s Office; for example, all 38 cases in a new type of category ‘Wales’ (see Table 1) also appear in one of the others.  Therefore, three categories were chosen with the aim that they should overlap as little as possible: Child death, Alcohol, drugs and medications (AD&M), and Railways.  The earliest of the consecutive 50 most recent cases in each category were 11 Feb 2019 (Child deaths), 17 Sep 2019 (AD&M), and 21 Oct 2013 (Railways), and the most recent cases are shown in Table 1.


3.3  For Child deaths, and following an unexpected observation among the subset of neonates in the initial sample (see below), a further ad hoc sample was taken.  This comprised 29 neonates found among the remaining 116 available Child Death reports.   These were assigned case numbers C51 – C79. Only one report (R1) had an unexpired date for the addressees’ responses, i.e., after the sample cut-off date (10 June 2020).   The database was locked July 2, 2020.



Table 1: Published and categorised Regulation 28 reports on 10 June 2020 (n=3861).  Note that unique cases may appear in more than one category

Category of Report

Earliest report

Number in  category

Most recent

Work / Health and Safety

5 Aug 2013


17 Apr 2020

Alcohol, drug and medication

27 Nov 2013


24 Apr 2020

Care Home Health-related

13 Aug 2013


12 Feb 2020

Child death

6 Jan 2015


5 Feb 2020

Community Health care and Emergency Services

30 Jul 2013


29 May 2020

Emergency services

9 Jan 2019


27 Jan 2020

Hospital death 

30 Jul 2013


28 May 2020

Mental Health related deaths

20 Aug 2013


27 Jan 2020

Other related deaths

27 Aug 2013


5 Feb 2020

Police related deaths

23 Sep 2013


6 Feb 2020

Product related deaths

21 Oct 2013


17 Jan 2020


21 Oct 2013


29 May 2020

Road / Highway Safety

1 Aug 2013


17 Dec 2019

State custody deaths

9 Aug 2013


24 Apr 2020


12 Feb 2015


24 Apr 2020


26 Feb 2019


24 Apr 2020






4.  Findings


Regulation 28 Report origins.


4.1  Given their discrete areas of responsibility, Coroners’ practices can be compared on a geographical basis.  The geographical distribution of Coroners’ offices and the origins of Reg. 28 reports (both for all sampled reports and each of the three chosen categories) are shown on the maps in Annex 2.  Overall, the distribution of Reg.28 reports exhibits geographical voids compared with the locations of Coroners’ offices.   However, within the Railways category, clusters of reports in London, Bedford and Hertford correspond to the highest density of railways in the country, and the West and East coast mainlines, respectively.


4.2  Twenty-seven of 80 Coroners’ Areas had no published reports in any of the three selected categories. Figure 1 shows the distribution of standardised report rates among the 53 Coroners’ areas in England and Wales where at least one report originated.  This is a bimodal distribution: the standardised report rates are not random variation around consensus values.[15]  These suggest that Coroners differ in their practice with regard to generating Reg.28 reports (see below).


Demographics of the deceased


4.3  The median age among the three Reg.28 report categories was (years): Child deaths (5.5), AD&M (41), Railways (26.5).  All three categories exhibited a female minority (n=50 each): Child deaths 19 (38%), AD&M 20 (40%), and Railways 11 (22%).  


4.4  For reasons discussed below, the subset of neonates among Child deaths was also examined for gender imbalance.  In the initial sample (n = 50) there were 14 neonates (28%), of whom only five (36%) were female.  A further ad hoc search among all remaining available Child Death Reg. 28 reports (regardless of date) identified a further 29 neonates, of whom only nine (31%) were female.  In aggregate (n = 166 Child death reports), 46 concerned neonates

Figure 1. The variation in Reg.28 report rates for the categories Child death, AD&M, and Railways among 53 Coroner’s Areas in England and Wales.  The report rate is standardised by using the number of cases reported to each Coroner in 2017 as the denominator.  In addition, there were 27 Coroners’ Areas with no published Reg.28 reports. ___________________________________________________________________________


(26%) of whom only 14 (32.6%) were female.   The female : male neonate ratios were, therefore, 1:1.8, 1:3.2, and 1:3.3 for the initial, ad hoc, and aggregate subsets, respectively.


Actions requested.


4.5  Most Reg. 28 reports carried only the template language in Box 6 with regard to addressees’ actions that would be appropriate (section 1.3 above and Annex 1).   The exceptions were (number of reports):  Child Deaths (7), AD&M (3), and Railways (5) or 10% of all 150 reports.


4.6  Instead, Coroners seemed to use Box 5 of the template ‘Coroner’s concerns’ to suggest or indicate what sorts of specific actions might be needed.  For example, Case C30 had a medical cause of death of ‘1a. ? Sudden infant death syndrome’ and the Coroner’s concerns were that, during pregnancy:

              - the deceased infant’s mother never saw the same midwife twice,

              - there were communications problems among midwives,

              - no advice against co-sleeping had been given, and

- in the context of alcohol intoxication there would have been the possibility of a criminal offence. 

This was all addressed to the relevant NHS Trust, who (unusually, see below) responded with some concrete organizational actions that they believed would address the Coroner’s concerns.


4.7  In other cases, however, the suggested actions are less specific.  For example, the Coroner’s concerns in case D46 included that particular drugs interact, with additive or supra-additive adverse effects including sedation and respiratory depression.  The report was addressed to (the Westminster) Minister for Health and Social Care. The response was that the risks were well-known and well-publicised (product labelling, professional guidances, academic syllabi, textbooks, the British National Formulary, and was included in one of their regular circulars).


4.8  However, sometimes, the stated Box 5 concern(s) indicate(s) remedial action fairly obviously. For example, among Railways cases, there are numerous examples of Coroners’ concerns about particular locations, such as:

              -  A particular hole in a particular fence,

              -  Inadequate signage on a specified railway station platform,

              -  The poor condition of a particular railway crossing with trip hazards,

              -  Failed audible warnings at a particular level crossing, or

              -  Failure of the metal blinds, leaving a particular underground railway station                                           unlocked overnight. 

All these clearly indicate specific actions, mostly relating to physical equipment, even if occasionally indicating systemic issues along the whole railway line (e.g, inadequate measures for staff to respond to CCTV monitoring).



Types of addressee


4.9 All Reg. 28 reports had one or more addressees.  These were classified into government (Westminster or local), NHS entities (Trusts, hospitals health boards in Wales, clinical commissioning groups), professional bodies (including Royal Colleges, General Medical / Nursing Councils), independent safety institutes,[16] law enforcement (police, H.M. Prison and Probation Service, National Offender Management Service), Railway companies / authorities[17], and other[18] (see Table 2).


4.10 Examples of incorrect addressees were rare.  In one case (Chief Coroner no. 2019-0296)[19] a response from a Government department pointed out that trees are the responsibility of local authorities, and, in another (2018-0258), West Midlands Railway responded that issues relating to platform markings at Hemel Hempstead station should be raised with its owner, i.e., Network Rail.  However, overall, there was little evidence for misaddressing Reg.28 reports. 



Responses to Regulation 28 reports.


4.11 Table 3 shows the number of published responses from each class of addressee for each Reg. 28 report category, and the overall published response rates for each type of respondent.  For the three selected categories of Reg. 28 reports, the overall published response rate (number of published responses / number of addressees) was 37.4%.  Reg. 28 reports without any published response were (% of unique cases): Child deaths (48%), AD&M (36%), and Railways (36%). 



Child deaths









NHS entities





Professional bodies





Independent safety institutes





Railway companies etc.





Law enforcement
















Table 2.  Addressees for Reg. 28 reports in each of the three selected categories. The greater number of addressees than reports (N=150) is because many reports had multiple addressees.  




Child deaths



Total addressees*

Overall Response rate (%)







NHS entities






Professional bodies






Independent safety institutes






Railway companies etc.






Law enforcement



















Table 3.  Responses published from Reg. 28 report addressees in the three selected categories.  *The numbers of addressees are drawn from Table 2.


NHS entities and railway companies / authorities led the response rates.  Of all individual addressees, Network Rail (n = addressed 26 times) had the highest reported response rate (73%).  The only report with an unexpired due date was R1; had the data been adjusted for this, then an increment of 0.47% would have been added to the overall published response rate from Railway authorities (46.5% versus 46.0% shown in Table 3).  




Respondents’ actions.


4.12  Those responses proposing no change to processes or physical equipment / facility were easiest to identify.   Responses, when published, often contained multiple elements and sorting and categorizing these relied upon subjective criteria (discussed further below).  In particular, distinguishing between a specific remedial action relating to the Reg.28 report fatality (e.g., mending a particular level crossing), and a wider system improvement (e.g., providing CCTV observation of level crossings along the whole railway) was often not clear. 


The proportions of reports that represented something more than mere acknowledgement, and indicated changes to systems or physical equipment or facilities were compared with the total numbers of responses (‘Substantive response’), as well as with the total numbers of addressees (‘Effective response’), as shown in Table 4.


Child deaths



Actual or identified change to physical equipment or facility




Actual or identified change to process specifically related to the unnatural death




Actual or identified change to a system process that contributed to the unnatural death.




No identified change in any physical equipment or process, but acknowledging report.




Substantive report rate (%)




Effective response rate (%)





Table 4.  Actions contained in 88 addressees’ responses to 150 published Reg.28 reports in the three selected categories.   Some responses contained multiple elements.  The substantive response rate is the number of addressees’ responses in the first three rows, divided by the total number of published responses for that report category.  The effective response rate is the number of substantive responses divided by the number of addressees. ________________________________________________________________________


5.  Analysis


General findings


5.1  This study sampled the most recent 50 published Reg. 28 reports in three of the Chief Coroner’s categories: Child Death, AD&M, and Railways.  The geographical origins of Reg.28 report, as proxy for individual Coroners’ practices, were compared as were the publication rates standardised for the size of each Coroner’s Area.  Few reports carried specific requested actions.  Among a wide variety of addressees, published response rates were partial, and among published responses, those reflecting substantive actions were a small proportion of the total number of Reg. 28 reports published.


5.2  Coroners have jurisdictions that are confined to specific geographical areas.  Therefore, variation of Coroners’ practices must underlie a difference between the geographical distribution of Reg.28 report origins and the geographical distribution of Coroners in England and Wales.  This was unlikely to be due to random chance for three reasons.  Firstly, the frequency distribution of report origins was bimodal (unlike a random distribution; Figure 1).  Secondly, there were no reports from 27 Coroners’ Areas; if 150 reports distributed randomly into only the observed subset of 53 out of 80 Coroners’ Areas, then the probability is obviously infinitesimal.[20]  Thirdly, for the Railways reports, the cluster of Reg. 28 reports at London probably reflects the greatest density of railways in the country, and is rational, not random. 


5.3  The fact that almost all published Reg. 28 reports contain only the stereotypical wording with regard to recommended actions (Box 6) is consistent both with the Guidance,[21] and obiter in Clegg [1996].[22]  Coroners cannot be expected to be experts in, say, civil engineering nor to be able to select an optimal technical solution from among many alternatives that may exist, in order to mitigate a particular hazard.  Placing the onus for design of precise actions onto the addressee would seem to have the reasonable rationale of delegation to the party that is best qualified.


5.4  The responsiveness of addressees varied widely.  The overall response rate of 37% (Table 3) suggests that either responses are simply not being published, or that Box 7 of the Reg. 28 report template is widely ignored.  Publication bias is returned to below, but the policy at the Chief Coroner’s Office for publishing responses is the same as that for the original Reg. 28 reports.[23]  The yet further small proportion of the published responses indicating some substantive action or response (Table 4), pending no huge publication bias, suggests that the Reg. 28 report system has a very sub-optimal yield.


Reflection of the legal framework by these data


5.5  This study has found 27 Coroners’ Areas without a single published report in these three Reg.28 report categories.   Differences between Coroners in Reg.28 reporting rates would doubtless have a much bigger influence on what actually gets published than any (unlikely) bias at the Chief Coroner’s office (see below).  It is a Coroner’s statutory duty to write a Reg.28 report when s/he has concerns (if addressed) which may prevent further deaths.[24] However, the Chief Coroner’s Guidance dilutes this duty, seeing it essentially as discretionary: ‘It is for each Coroner to decide on a case by case basis whether he or she has a statutory duty to make a report.[25]  


5.6  There is external evidence suggesting that Coroners indeed have idiosyncrasies about writing Reg. 28 reports.  This includes that many Coroners are indifferent towards future public health and safety as a purpose of the inquest.  In 2010-2011, McGowan interviewed 32 Coroners in England and Wales (about a third of the then total).  When asked about the Coroner’s purposes, only half of the sample made any reference to the public health, only six saw improving public health as a primary purpose, and only six made any reference to the prevention of further deaths. Of these few latter, only one Coroner mentioned the (then) Rule 43 report as an active role in public health, although a few others saw a passive role, e.g., by providing data to the Office of National Statistics.[26] 


5.7  The de facto discretionary nature of Reg.28 report writing by Coroners is perhaps emblematic of the generally discretionary implementation of their powers, with potential for inconsistent decision-making.[27]  Roberts et al[28] demonstrated that, among 64 Coroners, one inconsistent decision was between natural and unnatural causes in a test set of sixteen scenarios, most of which arise commonly.  That fundamental decision governs whether to open an investigation, whether to close it with or without an autopsy, and whether to proceed to inquest.[29]  While the Reg.28 report may be viewed as being at the ‘end of process’, whether it is actually written could depend not only on the Coroner’s discretion after the inquest, but also depends upon other subjective decisions (e.g., whether to close the inquiry after autopsy).  


5.8  The Coroner’s local domain is subject to very little audit and their guidance at the national level is vague (including the ‘duty’ to write a Reg.28 report; see above).  Moreover, as local authority appointees, Coroners are not subject to discipline by the Chief Coroner and can only be removed ‘for incapacity or misbehaviour’ by agreement between the Lord Chancellor with the Lord Chief Justice.[30]  It is hard to imagine how individual decision-making about the exercise (or not) of discretionary powers could become a matter of ‘incapacity or misbehaviour’.


The legal framework and addressees


5.9  Addressees are subject to the mandatory language in Box 7 of the template, the Regulation,[31] and the Act:[32] a response to the Reg.28 report is demanded.  When received, Coroners have a duty to forward addressees’ responses to the Chief Coroner, where there is the same publication policy as for the Reg.28 reports themselves.[33],[34]  The deficit found in this study between published responses compared with published reports could be due to a variety of reasons.   Firstly, it might be administrative: at some place in the process, for currently unknown reasons, processing responses is deprioritised;  collating a response with a Reg.28 report that is already published may be an administrative task that takes longer than simply publishing the report.  Another possibility is that addressees simply ignore the instruction (potential reasons for this, relating to tort liability, are discussed below).  Meanwhile, the legal framework does nothing to encourage addressees’ responses.  There is nothing in the Regulations requiring Coroners to follow up with unresponsive addressees, there is no sanction for failing to forward responses received to the Chief Coroner, and, above all, there are no sanctions for addressees who simply ignore Reg.28 reports.  These data cannot distinguish between these contributory causes for the deficit.  However, yield of published responses to Reg.28 reports is small, and this needs further investigation at the operational level.


5.10  When considering the subset of Reg. 28 reports that actually garner responses, Table 4 illustrates, as far as what is published, yet further inefficiency of the Reg. 28 process because responses do not always indicate substantive actions.  On a few occasions this was because the addressee is inappropriate, but in others, sorting responses into the substantive and non-substantive must be acknowledged to be a subjective process.  A potential solution for this is proposed below.  


Study limitations.


5.11  Chronological change and small sample bias are inextricably linked in this study, and both can cause artefacts to be mistaken for findings.  The sample was the most recent 50 published Reg. 28 reports in each category, as an effort to observe current practice.  Even so, the most recent 50 reports for Child Deaths, AD&M, and Railways have accrued in 16.5, 9.5, and 79 months, respectively.  Reducing these chronological differences could only be improved by either reducing sample size (which would increase the hazard of artefact) or lengthening the observation period for the first two categories (which might not represent contemporary practice).  However, in the case of gender imbalance among Child deaths (see below), the initial sample accurately signalled the gender imbalance in the enlarged sample; this suggests, if only for this subset of Reg.28 reports, that there was no chronological change and that initial sample size (n = 50) was adequate (see Annex 3).


5.12  Publication bias for the Reg.28 reports seems unlikely.  The database has three features that would be inconsistent with an hypothesis that the Chief Coroner’s office has withheld many from publication.  First, there are reasonably even intervals (of different sizes) for publication dates within all three Reg. 28 report category periods.  Secondly, these Reg.28 reports have a gender imbalance that matches that of inquests (see below).  Thirdly, there would seem to be little motive for selective publication because the Chief Coroner’s Guidance specifically indicates a ‘presumption of publication’, unless there is some special reason for exemption.[35] 


Comparison with other research in this area


5.13  Gender imbalance in this sample of Reg. 28 reports confirms, and provides an extension to, previous findings.  Maclean found a male predominance among cases reported to Coroners,[36] and the proportion of those cases then being advanced to inquest.[37]  The present data extends those findings slightly, in that the compound decision-making leading to a published Reg. 28 report does not seem to amplify the pre-existing gender imbalance at the inquest stage. The imbalance in this sample is therefore consistent with Maclean because a Reg.28 report (with rare exceptions) follows an inquest.


5.14 Maclean hypothesizes a variety of reasons for these gender imbalances.[38] These include that women are more likely to be known to their general practitioner, and therefore have a higher probability of death certification suitable for registration (and, thus, a lower probability of being reported to the Coroner).  However, among cases that are reported to the Coroner, female fatalities are more likely to be found to be due to natural causes than in males; hence, the imbalance becomes greater for proportions proceeding to inquest (at a rate of about 2:1 male : female).[39] There is some medical explanation for this: in England and Wales, the male suicide rate is about three-fold that for women, and hanging and other violent methods are far more popular amongst men than women.[40]  Evidence for this in the present database includes, for example, among the Railways category of Reg. 28 reports, that there were only three women among the 12 suicide verdicts. Thus, more inquests for men than women might be expected, and consequently more Reg. 28 reports, too.  Other analogous hypotheses could be offered for accidental deaths (more railway trespassers are male than female, men are more likely than women to engage in driving recklessly and under the influence of drugs, etc.).


5.13 However, these medical explanations are largely restricted to post-pubertal behaviours, and do not explain the male : female imbalance both in the initial sample of Child Death reports (n = 50), and, in particular, amongst the neonates.  This was not small sample artefact because the additional ad hoc sample confirmed the male predominance.  This even extended to neonates, where the Reg.28 reports describe either an occasional case of neonatal sepsis, or (the large majority) obstetrical disasters (Annex 3); neither of those have a biological gender imbalance.  While the aim here is not to engage with the public health issues themselves (e.g., the clinical problem of suicide or (say) prescribing safety), the persistence of this gender imbalance in the very young, suggests some unknown bias in Coroners’ practices. 


Public health and safety is an issue that concerns bereaved family members


5.14  For bereaved families, their interest in preventing further deaths, may be part of their closure following a tragedy, and is often altruistic:

              I don’t want something like that to happen to anybody else’s Mum.’,[41] and

              He was just like you said.  They need to learn, so that nobody gets treated like my brother.[42]

While expressed from a personal viewpoint, these are expressions by the bereaved of an interest in future public health and safety.  This has been echoed by the Chief Coroner, recognising that one common benefit of an inquest and Reg. 28 report is:

‘A bereaved family wants to be able to say: His death was tragic and terrible, but at least it shouldn’t happen to somebody else.’[43]

The obvious question arises that whether these needs of the bereaved (and the wider public) are currently being met by the watering down of the Coroner’s regulatory duty into an essentially discretionary power by the Guidance.[44]   Even if a Coroner felt that a narrative verdict was res ipsa loquitor, remedial action is less likely when there is neither addressee nor publication of a Reg.28 report.  The inadequacies of the Reg. 28 report process, were they known more widely, could undermine the confidence of bereaved families and the wider public in the process of inquest.


5.15 McIntosh points out that a common interpretation of Article 2, ‘Right to life’ of the European Convention of Human Rights includes mandatory investigations (in the form of an inquest in England and Wales) when ‘a state actor is alleged’ to have been a causative agent in an unnatural fatality.[45]  Smith LJ, in obiter, has extended that concept with ‘the duty to protect life extends to organs of the state, such as hospital authorities, to make appropriate provision and to adopt systems of work to protect the lives of patients in their care.[46],[47]  If it is written and published, a Reg. 28 report, and its addressee’s response (if any), have the potential to contribute exactly to these goals.  However, neither McIntosh nor Smith LJ gave specific consideration to the Reg.28 report as a vehicle for this purpose.


5.16 The use of a template for Reg.28 reports is the result of a Chief Coroner’s Guidance[48], and may be seen as part of the Government’s general effort to improve consistency of practice among coroners across England and Wales.  This motive contrasts with the vagaries of the legal framework discussed above.  Importantly, there remains no standard guidance as to when the template should be used.  This is an important lacuna in the Chief Coroner’s Guidance (see above).


5.17 An issue that has attracted little academic consideration is tort liability for addressees.  Responding to a Reg. 28 report has a potential for generating evidence that might need to be disclosed in litigation.[49]  Network Rail and NHS entities were found to be the most likely to respond to Reg.28 reports; these are large national actors with resources supporting permanent legal teams who deal with Reg.28 reports and litigation on a routine basis.  Other types of addressee lack such resources.  Thus, say, when a GP surgery or a sports club finds itself in the unfamiliar position of being an addressee of a Reg. 28 report, legal advice might be ad hoc.  In that insecure position, the addressee might be advised not to make any response, so as to avoid the risk of creating further perjorative evidence in future civil law proceedings.  Moreover, such advice is easy when there are no sanctions for failing to respond to a Reg.28 report, and when Coroners have no duty to follow up when Box 7 of their Reg.28 report is ignored (see above).  A counter-argument is that a substantive response in the public domain, describing concrete remedial actions taken, could help counter a claim of neglect of duty of care in some future fatality, even if not in connexion with the deceased discussed in the Reg. 28 report received.  This subject is returned to below in Recommendations. 


6.  Conclusion


6.1  Reg. 28 reports are not operating in the manner described by the Act[50], Regulations[51], or the Chief Coroner’s Guidance[52] and are haphazard in many respects.  Decision-making to originate a Reg. 28 report has geographical bias, probably attributable to non-uniform attitudes of Coroners.  It is probable that addressees often ignore their duty to respond to Reg. 28 reports.  The published corpus of Reg. 28 reports exhibits a gender imbalance, and this is least explicable for neonates.  However, while publication bias of Reg. 28 reports seems unlikely, this database cannot exclude such bias for addressees’ responses (when they exist at all).  The current practice is not compatible with the motivations of many bereaved families (and the broader public) to contribute to the inquest in the interests of future public health and safety. While it would appear that the system is working better for the Railways category than for Child deaths or AD&M, even then an effective published response rate of less than 50% suggests ample scope for process improvement.




7. Recommendations


7.1  First of all, this study used a sample of three selected Reg. 28 report categories.  Extending this study to all report categories could reduce the probability of sampling bias and confirm that the inefficiencies in the Reg. 28 report process are not new.  If those inefficiencies do not extend to other Reg.28 report categories, then this only narrows (and does not eliminate) the question of why it exists for these three.   An improvement in data management for that wider study could also be the use of an independent adjudication committee for some of the more subjective parts of the analysis, e.g., the classification of respondents’ actions (see Table 4). 


7.2  The gender imbalances identified in these samples should be confirmed in a larger set of published Reg. 28 reports.  While the ‘adult’ categories of AD&M and Railways appear to conform to known national gender imbalances in suicide, methods of suicide, and other male behaviours (see above), there are no known gender differences among medical causes of death sufficiently large to create the imbalance among neonates.[53]   The latter may also be true for other Reg. 28 report categories.  Standardisation against patterns of case referrals to Coroners would be a useful starting point for investigating the stage at which the bias develops.  An important aspect in this might also be to consider whether natural and unnatural death are insufficiently well-defined, for which there exists a small literature.[54]


7.3 A national, consistent rule about when to file a Reg. 28 report needs to be developed and instituted by amending the Regulations and the Chief Coroner’s Guidance.  This should reinforce that Reg.28 reporting is a Coroners’ statutory duty, and not at her/his discretion.[55] The drafting of that regulation would have to be careful, but it might include that a report should be written for any situation where any type of remedial action could be imagined by any reasonable person to mitigate a risk of future deaths under the same circumstances.


7.4  Next, enforcement of responses to Reg. 28 reports deserves consideration with sanctions for those who ignore Box 7. Haphazard, sporadic Reg.28 reporting risks missing opportunities to improve future public health and safety, does not encourage bereaved families to take part in inquests, and has a poor yield.  The Regulation anticipates that all reports should garner a response but there are no real mechanisms to enforce that duty.  Presuming neither that Coroners are neglecting their duty to forward addressees’ responses to the Chief Coroner, nor that the Chief Coroner has adopted a policy of less than a presumption of publication (which would be unlikely neglect of duty by both), and if addressees will not comply because of the current, weak legal framework, then sanctions should be invented for addressees, and Coroners should be required to follow-up on those in contempt. 


7.5 If it is litigation that deters addressees from responding to Reg.28 reports, then that still does not counter the value of Reg.28 reports to the future public health and safety.  As in other areas of tort law (e.g., defamation), there must be recognition of the need to balance protection of addressees’ rights against the proper conduct of public law.  Addressees can already apply for redactions of their responses before publication (and this study cannot assess the effectiveness, in practice, of that safeguard).  The Reg. 28 report itself is unattractive as evidence in civil proceedings because it will contain evidence heard at inquest, and thus hearsay which is incapable of cross-examination at some remote future time and place.[56]  However, this will not always be true for addressees’ responses to Reg.28 reports, when they themselves are parties in future litigation.  The Civil Evidence Act [1995] s.12 and Civil Procedure Rules [1999] provide for new Practice Directions in civil proceedings.  One potential compromise might be that a Reg. 28 report and its response would be exempt from disclosure in litigation concerning the subject deceased, but a response (mandated as above) could be discoverable in the case of some future fatality concerning the same addressee. 


7.6  Lastly, a national, public, searchable, electronic archive of inquest proceedings could be created.  Currently, transcripts of inquests are made but are not available to the public, and could support law reporting.  This would provide a further avenue for research into the Reg. 28 report system in general, as well as into the particular public health hazards that may have been uncovered.  An archive would also enable audit.  Audit would provide a comparative base to encourage consistency in decision-making in creating Reg.28 reports, provide a tool to measure the yield at each step of this system, and better pursue future public health and safety.  



Annex 1:  An example Report to Prevent Future Deaths, in its template.


N.B.: This example is unusual in that Box 6 carries the standard wording as its first sentence, but then continues with some specific information.  The three sentences of Box 7 appear in all Regulation 28 reports. Redactions are used here because of the intent to illustrate the template, and not to consider the clinical aspects of the published report.



Information Classification: CONTROLLED


Re: [redacted], deceased




1. The College of Policing 2. The Home Secretary




I am an assistant coroner for the coroner area of [redacted].




I make this report under paragraph 7, Schedule 5, of the Coroners and Justice Act 2009 and regulations 28 and 29 of the Coroners (Investigations) Regulations 2013.




The investigation was commenced on the [redacted] and concluded by an inquest before a jury on the [redacted].


The cause of death recorded by the jury was:- 


1(a) Use of cocaine, episode of altered behaviour including self-harm, exertion, excitement, the use of x26 Taser Device and restraint.




On the [redacted], who been acting in a paranoid and psychotic manner, ingested as substantive amount of cocaine before jumping from a first floor window of a friend’s home. He was in possession of a large knife with which he stabbed a woman in her garden before walking in the roadway and was seen to be slashing with the knife at his own throat and neck. The police arrived and, following a confrontation with [redacted], Tasered him on three occasions. He suffered a cardiac arrest at the scene and was rushed by ambulance to a local hospital where he was pronounced dead by medical staff. The jury, in dealing with the cause of death found specifically “excessive of cocaine taken resulting in paranoid and erratic behaviour with the use of the Taser having more than a trivial impact on [redacted] cardiac arrest”.




During the course of the inquest the evidence revealed matters giving rise to concern. In my opinion there is a risk that future deaths will occur unless action is taken. In the circumstances it is my statutory duty to report to you.


The MATTERS OF CONCERN are as follows.  – 


It was said by a number of witnesses that the Taser is not a device without risk but that there is limited data as to its effects upon individuals (both generally but more particularly in those classified as vulnerable). In evidence it was clear that there is no understanding about the potential for incremental risk with multiple Taser activations and no training provided as to the maximum number of activations nor of their duration which is appropriate or safe. The evidence was that the training given to police officers in this aspect is as set down by the College of Policing and that it is silent as to the potential incremental risk of multiple and or sustained activations (the so called ‘detention under power’). It was clear from the evidence of (an intensivist consultant) that a Taser does carry a risk – despite, he said, the claims of the manufacturers - but the extent of that risk is far from clear. Two forensic pathologists gave evidence and confirmed their joint opinion that the Taser caused (together with other things) [redacted] death in that it played a more than minimal, trivial or negligible part. Although I found as a fact that the training given to the police officers was appropriate I did so ONLY upon the basis that it was given based upon the limited knowledge presently available. I am concerned, based upon the evidence that was led before the jury, that there is insufficient independent data as to the lethality of Taser use and that, therefore the advice and training provided to police officers may be deficient or incomplete.




In my opinion action should be taken to prevent future deaths and I believe you [AND/OR your organisation] have the power to take such action. 


Perhaps by a wholesale review of the effects of multiple Taser activations and the effects of sustained activations (whether in isolation or in combination) so that fuller and more comprehensive advice, guidance and training can be given to those officers who are authorised to carry Tasers.




You are under a duty to respond to this report within 56 days of the date of this report, namely by 2nd April 2020 I, the coroner, may extend the period.


Your response must contain details of action taken or proposed to be taken, setting out the timetable for action. Otherwise you must explain why no action is proposed.




I have sent a copy of my report to the Chief Coroner and to the following Interested Persons:- The Family of the Deceased, [redacted] , [redacted]. I am also under a duty to send the Chief Coroner a copy of your response. 


The Chief Coroner may publish either or both in a complete or redacted or summary form. He may send a copy of this report to any person who he believes may find it useful or of interest. You may make representations to me, the coroner, at the time of your response, about the release or the publication of your response by the Chief Coroner.





    [Signature block redacted]






























ANNEX 2: Geographical Distributions of Coroners and Reg. 28 Reports in England and Wales (most recent 50 reports in categories of Child death, Alcohol, Drugs and Medications, and Railways).  See Table 1 for earliest and latest reports in each category.

Figure A1: Distribution of Coroner’s offices in England and Wales.



















ANNEX 2 (contd.)





















ANNEX 2 (contd.)






















ANNEX 2 (contd.)























Annex 3.  Absence of effect of increasing sample size of published Reg. 28 reports on the observed gender imbalance among neonates.



C1 – C50

C51 – C79

C1 – C79





Neonates (%)

14 (28.0)

29 (25.0)

43 (25.9)









Female : male





Annex 3 Table. The female : male ratios for neonates in the initial sample of Child deaths (C1 - C 50), the ad hoc further sample (C51-C79), and for all neonates regardless of date of published Reg. 28 report (C1 – C79).  There is no evidence for a chronological artefact, and there is evidence that the initial sample size (n = 50) was adequate to detect this gender imbalance.





[1] Chief Coroner.  Guidance no.5, Annex A, per Regulation 28 (ibid). [2013, revised 2016].

[2] Coroners (Investigations) Regulations 2013 (SI 2013 No.1629), Regulations 28-29.

[3] Coroners Rules 1984, Rule 43.

[4] Thornton P. Guidance no.5 Reports to Prevent Future Deaths’, para 10(2)-(3). London: Chief Coroner’s Office, as revised 2016.

[5] Ibid, para 19.

[6] Ibid, para 48.

[7] Ibid, para 52, and Coroners (Investigations) Regulations 2013 (SI 2013 No.1629), Regulations 28(5)(a) and 29(7)(a).

[8] Chief Coroner.  Guidance no.5, [2013, revised 2016], para 31.

[9] Ibid., paras 25-26.

[10] From about the 14th - mid-18th century, Inquest Rolls were forwarded to the Justiciar or (latterly) the King’s Bench, and have now found their way into the National Archive at Kew. Some were printed by the Harleian and Selsden Societies in the 19th century.

[11] Judicial review of Coroner’s verdicts and procedures is beyond the scope here.  See: Le Sueur A, Sunkin M, and Murkens JEK Public Law, Text, Cases and Materials Oxford: OUP 3rd edn 2016; pp.699-740.

[12] Chief Coroner.  Guidance no.5, [2013, revised 2016], Coroners (Investigations) Regulations 2013 (SI 2013 No.1629), and Coroners and Justice Act 2009 (CJA2009).

[13] https://www.judiciary.uk/related-offices-and-bodies/office-chief-coroner/https-www-judiciary-uk-subject-community-health-care-and-emergency-services-related-deaths/, accessed June – August 2020, passim.

[14] Coroners and Justice Act 2009, s.47(2)-(5), (7).

[15] The probability of 150 reports randomly distributing into only these 53 Coroners’ Areas is p = [(53/80)150] ≈ 1.5 x 10-27.  For any subset of 53 Coroners’ Areas drawn from the total of 80, the result is still infinitesimal, with p ≈ 10-6 for the approximately [80! / (53! x (80 – 53)!)] ≈ 1021 combinations.

[16] Health Safety Investigation Branch, National Institute for Health and Clinical Excellence, Clinical Care Quality Commission, and British Standards Institute.

[17] Network Rail, Transport for London, London Underground, train operating companies, H.M. Inspector of Railways, and the Rail Safety and Standards Board.

[18] The last comprised a wide variety of low frequency reports e.g., GP surgeries, a martial arts school, the Scout Association, private corporations, Safeguarding Children Boards, CAfCASS, and a family of the deceased who were simply interested parties.

[19] See below for case numbering.

[20] See also fn.44, page 17.

[21] Chief Coroner. Guidance no.5: Reports to Prevent Future Deaths (revised 2016), see: paras 24, 30–34.

[22] R v HM Coroner for Wiltshire ex parte Clegg [1996] 12 WLUK 7, 161 JP 521, Phillips LJ, penultimate para.

[23] Chief Coroner. Guidance no.5: Reports to Prevent Future Deaths (revised 2016), see: paras 52-54.

[24] Coroners and Justice Act 2009, s.7, Schedule 5.

[25] Chief Coroner.  Guidance no.5: reports to prevent future deaths. London 2016 (revised), para 8.

[26] McGowan CR.  Frustration of Purpose: Public Health and the future of death investigation in England and Wales.  Thesis: University of London, 2012,  p.135.

[27] See also:  Freckleton I. Death investigation, the coroner and therapeutic jurisprudence.  J Law Med (2007); 15: 1-13.

[28] Roberts ISD, Gorodkin LM, Benbow EW.  What is natural cause of death ? A survey of how coroners in England and Wales approach borderline cases.  J Clin Pathol 2000; 53: 367-373.

[29] The Coroners and Justice Act 2009, ss.4, 14.

[30] The Coroners and Justice Act 2009, Schedule 3, Part 4, paras 13-14.

[31] The Coroners (Investigations) Regulations 2013, Part 7, s.29(1).

[32] Coroners and Justice Act 2009,  Schedule 5, para 7 (2).

[33] The Coroners (Investigations) Regulations 2013, Part 7, s.29(6)(a).

[34] Coroners and Justice Act 2009, Schedule 5, para 7 (2).

[35] Chief Coroner.  Guidance no.5: Reports to prevent future deaths. London, revised 2016, paras 48, 55.

[36] Mclean, M. (2015) The Coroner in England and Wales; Coronial Decision Making and Local Variation in Case Outcomes, Doctoral thesis, University of Huddersfield, p.101.

[37] Ibid., p.122 - 124.

[38] Ibid., p.124.

[39] Ibid., p.125.

[40] Office for National Statistics.  Suicides in the UK: 2018 registrations. London (2019). See: https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/suicidesintheunitedkingdom/2018registrations, accessed 23 July 2020.

[41] Fox AW, personal observation.  St.Pancras Coroner’s Court 2011. 

[42] Fox AW, personal observation. West Brompton Coroner’s Court (London West Coroner’s Area) 2015.

[43] Chief Coroner. Guidance no.5 Reports to Prevent Future Deaths. London, revised 2016, para 2.

[44] Ibid, para 2.

[45] Mcintosh, S. (2012) ‘Fulfilling Their Purpose? Inquests, Article 2 and Next of Kin’, Public Law, 3: 407–415, and Human Rights Act [1998].

[46] Obiter in reference to Humberstone, 21, referring to Savage v South Essex Trust, [2009] 1 AC 681 see fn.64.

[47] R (on the application of Humberstone) v Legal Services Commission (The Lord Chancellor intervening) [2010] EWCA Civ 1479

[48] Chief Coroner.  Guidance no.5, Annex A [2013, revised 2016].  See: https://www.judiciary.uk/wp-content/uploads/2013/09/guidance-no-5-reports-to-prevent-future-deaths.pdf, accessed 19 Jun 2020.

[49] For a general discussion on disclosure within the Civil Procedure Rules see: Allbon E, Kaur Dua S.  Elliot & Quinn’s English Legal System. London, Pearson, 19th edition, 2018: pp.611 et seq.

[50] Coroners and Justice Act 2009, Schedule 5, para 7.

[51] The Coroners (Investigations) Regulations 2013, Part 7, s.28-29. 

[52] Chief Coroner. Guidance no.5 Reports to Prevent Future Deaths. London, revised 2016, paras 44 -49, 52-54.

[53] Office for National Statistics. Infant mortality (birth cohort) tables in England and Wales. London, 2019.  See:https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/datasets/infantmortalitybirthcohorttablesinenglandandwales, accessed 23 July 2020.

[54] E.g., Harris A, Walker A. Interpretation of 'Unnatural Death' in Coronial Law: A Review of the English Legal Process of Decision Making, Statutory Interpretation, and Case Law. Med Law Rev; 2019: 27(1):1-31.

[55] Chief Coroner.  Guidance no.5 Reports to Prevent Future Deaths. London, revised 2016, para 8.

[56] Civil Evidence Act [1995] ss.1-4.