You are here

Article Text

Article menu

Download PDFPDF

Original article
Variations in death notification of nursing home residents to Australian Coroners
  1. Joseph Elias Ibrahim1,
  2. Chebiwot Kipsaina1,
  3. Cathy Martin2,
  4. David Leo Ranson3,
  5. Lyndal Bugeja1
  1. 1 Department of Forensic Medicine, Health Law and Ageing Research Unit, Monash University, Melbourne, Victoria, Australia
  2. 2 Department of Epidemiology and Preventive Medicine, The Alfred Centre, Monash University, Melbourne, Victoria, Australia
  3. 3 Victorian Institute of Forensic Medicine, Forensic Services, Southbank, Victoria, Australia
  1. Correspondence to Professor Joseph Elias Ibrahim, Health Law and Ageing Research Unit, Department of Forensic Medicine, Monash University, Melbourne, VIC 3006, Australia; joseph.ibrahim{at}monash.edu

Abstract

Objectives To examine the impact of changes to the reporting requirements in coronial legislation on the nature and frequency of nursing home resident deaths reported to Coroners.

Design National retrospective study of a population cohort of nursing home resident deaths.

Setting Accredited Australian nursing homes between July 2000 and June 2013.

Participants Residents who died in nursing homes accredited by the Aged Care Standards and Accreditation Agency reported to Coroners.

Main outcome measures We explored three death-reporting models in the nursing home setting: comprehensive model, selective ‘mechanism of death’ model and selective ‘age of death’ model. These models were examined by manner of death subgroups: natural, falls-related and other external causes using the outcome measure of deaths notified to the Coroner per 1000 residents. We used an interrupted time series analysis using generalised linear regression with a negative binomial probability distribution and a log link function.

Results The comprehensive model showed the proportion of reportable deaths due to natural causes far exceeded those from falls and other external cause. In contrast, the selective notification models reduced the total number of reportable deaths. Similarly, the selective ‘age of death’ model showed a decline in the reportable external cause deaths.

Conclusions Variation in the causes, locations and ages of persons whose deaths are legally required to be notified to Coroners impacts the frequency and nature of deaths of nursing home residents investigated by Coroners. This demonstrates that legislation needs to be carefully framed and applied to ensure that the prevention mandate of Coroners in Australia is to be achieved.

  • older people
  • residential institution
  • mortality

https://doi.org/10.1136/injuryprev-2017-042689

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Coroners’ legislation in many Australian jurisdictions sets out their public health and safety responsibility in formulating recommendations for interventions to prevent deaths. Coroners rely on medical practitioners appropriately identifying and notifying reportable deaths to the Coroner for independent investigation.1 2 In common with the rest of the community, unexpected and unnatural deaths resulting from an injury or complications of clinical care occur among nursing home residents.3–5 In Australia, legislation mandates such deaths are reportable to Coroners, whose investigation holds some of the most detailed information about these events.6 7

    However, formal medicolegal death investigations may cause intrusion into family bereavement particularly when they occur in a resource-constrained environment. Over-reporting of deaths to Coroners may exhaust resource capacity leading to limited investigations.2 8 In contrast, overly selective reporting may lead to under-reporting of deaths with a consequent loss of disease and injury prevention information.8 9

    In Australia over the past decade, three changes in the reporting requirements have been seen (table 1). One jurisdiction transitioned from a comprehensive model10 where all deaths in aged care were reported to the Coroner and replaced by a conventional approach where all unknown, unnatural (ie, injury) deaths are reported to the Coroner, which was the standard for most other jurisdictions at that time. A second jurisdiction became selective with legislative amendments that excluded the deaths of persons 72 years and older who died of an injury attributable to old age.11 A third jurisdiction focused on improving the reporting and the depth of investigation of fall-related deaths12 and the remaining five jurisdictions maintained the prior conventional death notification approach.13–17

    View this table:
    Table 1

    Definitions of model and year of legislation changes to death notification

    Very little is known about jurisdictional differences in nursing home mortality rates, and globally very few studies have explored the causes of premature deaths in nursing homes3–5 let alone the possible temporal or geographical variation.

    Determining jurisdictional differences in mortality rates of injury-related (external) deaths of residents from falls-related injury, choking, suicide or homicide is a challenging task since any numerical differences may be due to a lack of risk adjusting the data for differences in age and sex, deficits in quality of care delivery or reporting practices. Examining the nature of variation in healthcare outcomes that have been adjusted for demographic differences provides insight into the quality of care service delivery.18–21

    The present study examined the impact of legislation changes on the reporting of all deaths and death injury rates of nursing home residents in different Coroners’ jurisdictions. Understanding these differences will both assist in legal and healthcare regulation policy development regarding resident safety and welfare in nursing homes and inform optimal practice of death investigation in the health, law and aged care sectors.

    Method

    Study design

    We treated the legislative changes in reporting of nursing home deaths to the Coroner as a ‘natural experiment’ to determine whether different or restricted models of reporting could alter the capacity of Coroners to contribute to death and injury prevention policy in nursing homes through in-depth independent investigation processes and inclusion of recommendations in Coroners’ findings.

    Definitions

    Although death notification criteria vary slightly between jurisdictions, all six states and the two territories in Australia require unexpected, unnatural or violent deaths, those resulting from an accident or injury or resulting from a medical procedure to be notified to the Coroner. On the basis of existing Coroners’ systems in Australia,10–17 we explored three observed legislative and procedural changes in the reporting of deaths as applied to the setting of nursing homes (table 1).

    Data sources

    The National Coronial Information System provided data on the deaths of residents of nursing homes that occurred between 1 July 2000 and 30 June 2013, where Coroners’ investigation was completed at 31 December 2014 and the death occurred while the deceased was a resident of an Aged Care Standards and Accreditation Agency accredited nursing home.22 Population data by age and sex of nursing home residents were drawn from the Australian Institute of Health and Welfare (AIHW) as at the end each financial year.

    Data were collected on the incident including the year of death, mechanism and medical cause(s) according to the International Classification of Diseases 10th Revision as assigned by Australian Bureau of Statistics and the temporal relationship of the incidents leading to death. We disaggregated the data into subgroups by manner of death: natural, falls-related and other—external causes (ie, all other injury deaths except falls, such as choking, drowning and transport-related deaths).23

    Statistical analysis

    Interrupted time series analysis using generalised linear regression with a negative binomial probability distribution and a log link function was used to model these deaths over time. The analysis was undertaken at the resident level, which allowed for adjustment of age and sex, considered ‘a priori’ as confounding factors of interest. The number of deaths notified to the Coroner was offset by the population per age and sex group for each of the models using the data from AIHW. The population data provided information on the number of residents in permanent nursing home care, by sex, age group and state on the 30th of June each year. The calculated linear estimates were exponentiated back to the original units and the rate was multiplied by 1000. This is the main outcome measure of the study and the unit is ‘deaths notified to the Coroner per 1000 residents’.

    To assess the effect of the legislation changes of notifying deaths to the Coroner on the rates, we compared the midpoints of the estimated pre-legislative and post-legislative change linear slopes. The midpoint was chosen because the beginning and end of the periods were thought to be biased by the delayed implementation of legislative changes to the notification of deaths to the Coroner.

    Analysis was undertaken using Stata (StataCorp 2015. Stata statistical software: release 14. College Station, TX: StataCorp LP).

    Results

    Population characteristics

    Detailed population characteristics have been presented in a previous study.3 Briefly, during the study period, 21 672 deaths of nursing home residents were notified to Australian Coroners. The mean age (SD) at death was 84.8 (±16.6) years, with majority (47.9%, 10 ­383) in the age group 85–94 years. The majority of deaths were determined as natural causes (84.6%, 18 390), followed by deaths from falls (12.4%, 2679), with the remaining determined as due to ‘other—external’ causes (2.8%, 610) (table 2).

    View this table:
    Table 2

    Population characteristics of resident deaths in nursing home notified to coroner 2000–2013

    Relation between notification models and nursing home deaths reporting rates

    Notification of all-cause deaths

    As expected with the comprehensive model, there was an elevated reporting rate for all cause deaths, adjusted for age and sex during the pre-legislation change period, which dropped post-2004 (figure 1A). Comparison of the midpoint ratios suggested reporting rate post-change was approximately one-fifth of the reporting rate during the pre-legislation period (midpoint ratio 0.23, 95% CI 0.19 to 0.29; p<0.001) (table 3). As shown in figure 1B, there was a similar result in the analysis of the reporting system of the age-focused model (midpoint ratio 0.47, 95% CI 0.31 to 0.71; p<0.001). Although there is an apparent downward trend in reporting rates using the falls–focus model, the midpoint ratio comparison was not statistically different (figure 1C). Between 2002 and 2012, the conventional model showed that reporting rates for all-cause deaths generally increased over time (figure 1D).

    Figure 1
    Figure 1

    All-cause deaths by reporting model.

    View this table:
    Table 3

    Midpoint comparisons of pre-policy and post-policy changes for deaths notified to the coroner per 1000 residents (midpoint ratio) adjusted for age and sex

    Notification of natural cause deaths

    Adjusted for age and sex, the reporting rates for natural cause deaths by the comprehensive model increased pre-change and then decline post-change to the legislation. The midpoint comparison showed that, overall, reporting rates following the changes was almost one-fifth of the previous rates (midpoint ratio 0.22, 95% CI 0.15 to 0.32; p<0.001) (figure 2A, table 3). This declining trend was similar with the age–focus model figure 2B, which had reporting rates after the legislative changes (2010) of almost half the previous rates (midpoint ratio 0.46, 95% CI 0.24 to 0.88; p=0.02). A declining reporting trend was also noted in the falls–focus model, but the midpoint comparison was not statistically different (figure 2C). The analysis showed that, similar to all causes of death, reporting rates for natural causes using the conventional model increased over time (figure 2D).

    Figure 2
    Figure 2

    Adjusted natural cause deaths by reporting model.

    Notification of deaths related to falls

    Reporting rates of falls-related deaths by the comprehensive and falls–focus models increased prior to the legislative changes and continued afterwards (figure 3A, C). In both these models, midpoint analysis was statistically significant (p<0.001) (table 3).

    Figure 3
    Figure 3

    Adjusted falls-related deaths by reporting model.

    In contrast, the age–focus model (figure 3B) showed a declining trend following the legislative changes. The midpoint analysis showed that this declining reporting rate showed a decrease in midpoint ratio by approximately 14% (midpoint ratio 0.14, 95% CI 0.05 to 0.39; p<0.001). Comparable with the comprehensive model, reporting falls-related deaths using conventional model showed an upward trend over time (figure 3D).

    Notification of other—external cause deaths

    Between 2000 and 2004, there was an increasing rate in reporting of other—external cause deaths using the comprehensive model, while post-2004, the analysis suggested a constant level of reporting over time (table 3). There was not a significant difference in the midpoint ratio comparisons (figure 4A).

    Figure 4
    Figure 4

    Adjusted other external cause deaths by reporting model.

    The opposite was observed in the age–focus model. Here, reporting rates for other—external cause deaths before the introduction of the legislative changes were relatively constant; however, following the changes, the reporting rates declined over time. Nevertheless, the midpoint ratios were not statistically different (figure 4B). When reporting rates were examined using the falls–focus model, the analysis showed a slight upward trend before the changes in 2008. After the changes, there was an apparent downward trend in the overall reporting rates, although the midpoint analysis comparing the two periods is not statistically different (figure 4C). Similar to natural and falls-related deaths, the conventional model showed an increasing trend in reporting of other—external cause deaths over time, adjusted for age and sex (figure 4D).

    Discussion

    Statement of key findings

    The current study is the first in Australia to examine the impacts of the variation in the notification requirements to Coroners of deaths of nursing home residents. The comprehensive model where all deaths of nursing home residents were notified to Coroners showed the proportion of deaths due to natural causes far exceeds those due to falls and other—external cause. It could be argued that a comprehensive death notification system increases the probability that mortality data regarding nursing home deaths will be complete and accurate. The comprehensive model may also limit under-reporting since the need to make the sometimes complex decision as to whether a death is reportable to the Coroner is removed. However, the comprehensive model significantly increases the workload of the Coroners’ Court and their medical investigators, which brings into play fiscal matters including resourcing and cost–benefit issues. Despite the small numbers, when death notification requirements changed from this comprehensive to conventional model, there was an increase or constant notification of external cause deaths. In addition, the proportion of deaths within the external cause categories did not change.

    In contrast to the comprehensive model, the selective notification models reduced the total number of deaths notified. The jurisdictions that altered their legislation towards an age-focused reporting system showed a decline in the external cause deaths reported. Focusing on a selective model has merit in that it tailors the notification process and may result in resource saving to the police, forensic pathology and the coroner.

    Strengths and limitations

    A key strength of our study was the use of national data over a 12-year period, which enabled the determination of death notification trends over time. One limitation of the study was that it included only deaths of nursing home residents notified to Coroners. It is possible that other notifiable deaths may have occurred but have not been notified perhaps because they were misclassified as non-reportable natural causes. This is particularly likely for deaths from other—external cause deaths such as complications of clinical care including therapeutic drug toxicity or choking given the small numbers. Other limitations include the possible influence of environmental changes and medical therapeutic developments that occurred over the 12-year analysis period of the study. Implementation of new nursing home accommodation design, changes and developments in nursing equipment and other resident-assist devices together with the introduction of new drugs and therapies could also be an explanation for some aspects of these trends.

    The available population data in the analysis were based on point and time (as at 30 June of the year), thus the population variation across the year could not be determined. Finally, the findings should be interpreted with caution as the time trends were calculated based on a short period of timeframe.

    In terms of generalisability, the principles of the different reporting models would apply throughout the world. However, the impact will vary according to differences in the health, aged care and legal systems.

    Implications

    Variation in death notification to Coroners creates gaps in the scope and quality of mortality data, particularly in terms of its national representativeness, comparability and completeness. This limits the ability to identify trends in premature and preventable deaths and intervention strategies. Three possibilities exist that may individually or collectively explain differences in jurisdictional variation.

    First, differences in population characteristics are recognised as one explanation for variation in outcomes. To account for this, we adjusted for age, sex and nursing home population in our analysis. Second, another potential explanation for the differences is a possible lack of consistency in the application of the legislative requirements for reporting death. Each Coroner is an independent judicial officer and therefore their determination that a reported death is in fact reportable may be influenced by the application of permissible judicial discretion and inevitably this may vary due to individual subjective interpretation.1 2 Third, it has been shown that medical practitioners may be unaware of some of their statutory obligations to notify particular deaths to the Coroner or be confused by the legislative language that defines such deaths.1 2 24–26

    Consequently, the type of notification model raises a number of policy considerations for a State’s Justice and Health systems including resource availability and the modern focus on injury and death prevention within the Coroners’ jurisdiction. In the comprehensive model, a large number of deaths, mostly due to natural causes, were identified and only a small number of external cause deaths. This dilution could affect the focus of investigations into external cause deaths and raises concerns about the cost-effectiveness and efficacy of the Coroners’ enquiries. In the absence of other factors, for example, deaths in custody, the law does not require natural cause deaths to be notified to Coroners. However, in the case of selective models, where the focus is on falls, the investigation may produce little new information, given the existing substantial research evidence base about interventions that can prevent or limit the injury associated with falls.27 28 Finally, the discrepancies in other—external cause deaths demonstrated across the four models highlighted the gaps and variations that still exist in the death notification/reporting processes across the Coroners’ jurisdictions in Australia.

    While we are unable to draw definitive conclusions about the variation in reporting between jurisdictions, it is worthy of more detailed examination. It is too simplistic to attribute the differences seen to medical death certification practices. Improving accuracy of death notification to Coroners requires reducing under-reporting of external cause deaths by the health sector and, reducing over-reporting of natural cause deaths.6–9 25

    A benefit of the comprehensive model of death notification is it reduces the potential for intentional under-reporting. Factors contributing to under-reporting may include a culture of institutional pressure and loyalty to colleagues,7 29 concealment of cases of medical negligence or rarely criminal activities.30 It may also be that some clinicians euphemistically under-report unavoidable deaths in an attempt to spare the deceased’s family. These clinicians believe this protects families from the increased distress arising from engagement with the coroners’ jurisdiction including a possible autopsy as well as the potential delay in the release of the body for burial or cremation.7

    In contrast, although a selective reporting model may capture certain categories of external cause deaths, such as falls, it may limit identifying deaths from choking or complications of clinical care by diverting attention towards physical injury-related deaths.

    Conclusion

    There is variation in the reporting of the deaths of nursing home residents to Australian Coroners. These variations may reflect differences in population health, clinical practice, healthcare delivery in nursing homes or issues with the death certification process. Improving the lives of frail, vulnerable older people living in nursing homes requires a better understanding of this variation.

    What is already known on this subject

    • Unexpected and unnatural deaths resulting from injury or complications of clinical care occur among nursing home residents and are required by law to be reported to the Coroner.

    • Over the past decade, legislative and operational changes to the reporting requirements have been implemented in three coronial jurisdictions.

    What this study adds

    • This study demonstrates that variation to coronial legislation regarding death reporting requirements has a broad impact on the frequency and nature of deaths of nursing home residents reported to Coroners.

    • Legislative amendments to death reporting requirement can lead to a change in the depth of mortality data available regarding nursing home residents. This in turn can restrict the public health and safety mandate of Coroners and adversely influence the development of policy and operational procedures designed to prevent avoidable death and injury in nursing homes.

    References

    1. ↵
      2. Start RD ,
      3. Delargy-Aziz Y ,
      4. Dorries CP , et al
      . Clinicians and the coronial system: ability of clinicians to recognise reportable deaths. BMJ 1993;306:1038–41.doi:10.1136/bmj.306.6884.1038
      OpenUrlAbstract/FREE Full Text
    2. ↵
      2. Start RD ,
      3. Usherwood TP ,
      4. Carter N , et al
      . General practitioner’s knowledge of when to refer deaths to a coroner. Br J Gen Pract 1995;45:191–3.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      2. Ibrahim JE ,
      3. Bugeja L ,
      4. Willoughby M , et al
      . Premature deaths of nursing home residents: an epidemiological analysis. Med J Aust 2017;206:442–7.doi:10.5694/mja16.00873
      OpenUrl
    4. ↵
      2. Kennedy B ,
      3. Ibrahim JE ,
      4. Bugeja L , et al
      . Causes of death determined in medicolegal investigations in residents of nursing homes: a systematic review. J Am Geriatr Soc 2014;62:1513–26.doi:10.1111/jgs.12929
      OpenUrl
    5. ↵
      2. Ibrahim JE ,
      3. Murphy BJ ,
      4. Bugeja L , et al
      . Nature and extent of external-cause deaths of nursing home residents in Victoria, Australia. J Am Geriatr Soc 2015;63:954–62.doi:10.1111/jgs.13377
      OpenUrlCrossRefPubMed
    6. ↵
      2. Ibrahim JE ,
      3. Bugeja L ,
      4. Ranson D
      . Using existing information from medico-legal death investigations to improve care of older people in residential aged care services. J Law Med 2013;21:255–61.
      OpenUrl
    7. ↵
      2. Barnes M ,
      3. Kirkegaard A ,
      4. Carpenter B
      . Intake rigour: ensuring only "reportable deaths" become coroners' cases. J Law Med 2014;21:572–83.
      OpenUrl
    8. ↵
      2. Freckelton I ,
      3. Ranson D
      . Death investigation and the coroner’s inquest. Melbourne: Oxford University Press, 2006.
    9. ↵
      2. Charles A ,
      3. Ranson D ,
      4. Bohensky M , et al
      . Under-reporting of deaths to the coroner by doctors: a retrospective review of deaths in two hospitals in Melbourne, Australia. Int J Qual Health Care 2007;19:232–6.doi:10.1093/intqhc/mzm013
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      Australasian Legal Information Institute. Coroners Act South Australia. 2003http://classic.austlii.edu.au/au/legis/sa/consol_act/ca2003120/ (accessed Feb 2017).
    11. ↵
      Australasian Legal Information Institute. Coroners Act New South Wales. 2009http://www5.austlii.edu.au/au/legis/nsw/consol_act/ca2009120/ (accessed Feb 2017).
    12. ↵
      Australasian Legal Information Institute. Coroners Act Victoria. 2008http://www.austlii.edu.au/cgi-bin/viewdb/au/legis/vic/consol_act/ca2008120/ (accessed Feb 2017).
    13. ↵
      Australasian Legal Information Institute. Coroners Act Queensland. 2003http://www5.austlii.edu.au/au/legis/qld/consol_act/ca2003120/ (accessed Feb 2017).
    14. ↵
      Australasian Legal Information Institute. Coroners Act Australian Capital Territory. 1997http://www.austlii.edu.au/cgi-bin/viewdb/au/legis/act/consol_act/ca1997120/ (accessed Feb 2017).
    15. ↵
      Australasian Legal Information Institute. Coroners Act Northern Territory. 2016http://classic.austlii.edu.au/au/legis/nt/consol_act/ca120/ (accessed Feb 2017).
    16. ↵
      Australasian Legal Information Institute. Coroners Act Tasmania. 1995http://www.austlii.edu.au/cgi-bin/viewdb/au/legis/tas/consol_act/ca1995120/ (accessed Feb 2017).
    17. ↵
      Australasian Legal Information Institute. Coroners Act Western Australia. 1996http://www.austlii.edu.au/cgi-bin/viewdb/au/legis/wa/consol_act/ca1996120/ (accessed Feb 2017).
    18. ↵
      2. Cameron D ,
      3. Jones IG
      . John Snow, the broad street pump and modern epidemiology. Int J Epidemiol 1983;12:393–6.doi:10.1093/ije/12.4.393
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      2. Brody H ,
      3. Rip MR ,
      4. Vinten-Johansen P , et al
      . Map-making and myth-making in Broad Street: the London cholera epidemic, 1854. Lancet 2000;356:64–8.doi:10.1016/S0140-6736(00)02442-9
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      2. Wennberg JE
      . Time to tackle unwarranted variations in practice. BMJ 2011;342:d1513.doi:10.1136/bmj.d1513
      OpenUrlFREE Full Text
    21. ↵
      2. Wennberg JE
      . Understanding geographic variations in health care delivery. N Engl J Med 1999;340:52–3.doi:10.1056/NEJM199901073400111
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      Australian Government Department of Social Services. Aged care service list—Australia. 2014https://www.dss.gov.au/our-responsibilities/ageing-and-aged-care/tools-and-resources/aged-care-service-list (accessed Feb 2017).
    23. ↵
      Australian Bureau of Statistics. Feature article: Population by age and sex, Australia, states and territories (Report 3101.0). http://www.abs.gov.au/ausstats/abs@.nsf/0/1CD2B1952AFC5E7ACA257298000F2E76?OpenDocument (accessed Feb 2017).
    24. ↵
      2. Middleton S ,
      3. Buist MD
      . Medical-setting deaths and the coroner: laws, penalties and guidelines. Med J Aust 2014;201:679–81.
      OpenUrl
    25. ↵
      2. Neate SL ,
      3. Bugeja LC ,
      4. Jelinek GA , et al
      . Non-reporting of reportable deaths to the coroner: when in doubt, report. Med J Aust 2013;199:402–5.doi:10.5694/mja13.10246
      OpenUrl
    26. ↵
      2. Lakkireddy DR ,
      3. Gowda MS ,
      4. Murray CW , et al
      . Death certificate completion: how well are physicians trained and are cardiovascular causes overstated? Am J Med 2004;117:492–8.doi:10.1016/j.amjmed.2004.04.018
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      2. Cameron ID ,
      3. Murray GR ,
      4. Gillespie LD , et al
      . Interventions for preventing falls in older people in nursing care facilities and hospitals. Cochrane Database Syst Rev 2010;1:CD005465.
      OpenUrlPubMed
    28. ↵
      2. Day L ,
      3. Finch CF ,
      4. Hill KD , et al
      . A protocol for evidence-based targeting and evaluation of statewide strategies for preventing falls among community-dwelling older people in Victoria, Australia. Inj Prev 2011;17:e3.doi:10.1136/ip.2010.030775
      OpenUrlAbstract/FREE Full Text
    29. ↵
      2. Pritt BS ,
      3. Hardin NJ ,
      4. Richmond JA , et al
      . Death certification errors at an academic institution. Arch Pathol Lab Med 2005;129:1476–9. doi:10.1043/1543-2165(2005)129[1476:DCEAAA]2.0.CO;2
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      2. Smith J
      . The Shipman Inquiry: third report—death certification and the investigation of deaths by coroners. 2003http://webarchivenationalarchivesgovuk/20090808154959/http://wwwthe-shipman-inquiryorguk/tr_pageasp (accessed Feb 2017).

    Footnotes

    • Contributors JEI is the senior author and contributed to the conception and development of ideas, acquisition of data, drafting and critical revision of the article draft, and final approval of the version to be published. CK contributed to the acquisition and interpretation of data, drafting the article and final approval of the version to be published. CM contributed to the analysis and interpretation of data, drafting and critical revision of the article draft, and final approval of the version to be published. DLR contributed to the design, drafting and critical revision of the article draft, and final approval of the version to be published. LB contributed to the conception, development of ideas, acquisition and interpretation of data, drafting and critical revision of the article draft, and final approval of the version to be published. All authors contributed substantially to the intellectual substance of the manuscript and are in full agreement regarding its content.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Patient consent Not required.

    • Ethics approval Justice Human Research Ethics Committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement Unpublished data are not available as our ethics approval does not permit us to do so.