Preventable maternal deaths in England and Wales, 2013-2023: a systematic case series of coroners’ reports

20

Coroners in England and Wales have a duty to write Prevention of Future 21 Deaths (PFDs) reports when they believe that action should be taken to prevent similar 22 deaths. We aimed to characterise learnings from reports involving maternal deaths. 23 Design: Systematic case series 24 Setting: England and Wales 25 Population or Sample: Database of all coroners’ PFDs published between July 2013 26 and 1 August 2023. There were 4435 reports at the time of data collection. 27

A reproducible computer code developed from the Preventable Deaths Tracker 28 (https://preventabledeathstracker.net/) was used to download all published PFDs from 29 the Judiciary website. Reports were searched for keywords related to maternal deaths . 30 Case information was extracted into pre-specified domains and compared to other data 31 on maternal deaths. 32 Main Outcome Measures: Case demographics, causes of deaths, risk factors, coroner 33 concerns and organisational responses. 34

Twenty nine reports involved a maternal death. The median age at death was 35 33.5 years (IQR 29-36 years) and 76% of deaths occurred in hospitals. The most common 36 cause of death was haemorrhage. Coroners frequently voiced concerns around failure to 37 provide appropriate treatment (57%), and failure of timely escalation (38%). Only 38% of 38 PFDs had published responses. When organisations did respond to the coroner , 80% 39 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 3 reported that they implemented changes , including publishing new local policies, 40 increasing training, or committing to increased staffing. 41

PFDs highlight gaps in obstetric care which, if appropriately addressed, 42 and regularly and routinely monitored, could prevent similar deaths. 43 Funding: None 44

maternal deaths, obstetrics, maternity, mortality, preventable 45 46 Funding 47 No funding has been obtained to undertake this study. 48 49 50 51 52 53 54 55 56 57 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 4

58 Global maternal deaths remain high and have stalled or worsened in 150 countries since 59 20001. In 2020, the global maternal mortality rate was 223 deaths per 100,000 live births, 60 which corresponds approximately to one maternal death every 2 minutes 1. In the United 61 Kingdom (UK), between 2019 and 2021, the Mothers and Babies Reducing Risk through 62 Audits and Confidential Enquiries (MBRRACE) programme reported 241 women died 63 during or within 42 days of their pregnancy, a rate of 11.7 deaths per 100,000 maternities2. 64 However, maternal deaths have not reduced in the past decade in the UK, and there is 65 inequality in mortality rates for women from areas of economic deprivation and ethnic 66 minorities2. As a result, the UK government commitment to ending maternity-related 67 preventable deaths of both mothers and their children by 20303. 68 69 Coroners in England and Wales have a duty to report when they believe that action should 70 be taken to prevent future deaths4. These reports are called Prevention of Future Deaths 71 reports or PFDs4,5, which are sent to addressees who must respond to the coroner within 72 56 days. PFDs are listed by the National Health Service (NHS) Patient Safety Strategy 73 as an official source of data 6, but it is unclear how they are used, if they are monitored, 74 and what action is taken to prevent similar fatalities in the NHS. A detailed analysis of 75 PFDs implicating maternal deaths could highlight factors contributing to maternal mortality 76 and continuing inequalities in care. This would help inform clinical staff managing 77 pregnanies and public health policy by highlighting points for improvement in care. 78 79 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 5 The aim of our study was to systematically determine the number of maternal deaths 80 reported by coroners in PFDs and to characterise these deaths in terms of demographics, 81 risk factors and causes of death s. In addition, we aimed to explore concerns raised by 82 coroners in these reports and understand what actions were reported by organisations in 83 their responses to the coroner. 84 85 86 87 88 89 90 91 92 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 6

93 The systematic case series was designed and a study protocol was developed and 94 preregistered on an open repository7. 95 Data collection, screening and eligibility 96 We used a reproducible, openly available code written by FD to download all portable 97 document format (pdf) documents from the judiciary website published from inception 98 (July 2013) to 1 August 2023. The code was reproduced from the Preventable Deaths 99 Tracker ( https://preventabledeathstracker.net/) and is available here: 100 https://github.com/francescodernie/coroner_PFDs. 101 102 We used the OCCRP Aleph tool8 to create a repository of the PFDs to process the reports, 103 allowing them to be machine-readable. We then conducted keyword searches using the 104 following terms: ‘pregnancy’, ‘pregnant’, ‘maternal’, ‘post -partum’, ‘partum’, ‘natal’’, 105 ‘perinatal’, ‘antenatal’, ‘obstetrics’, ‘gestation’, ‘parturition, ‘birth’, and the positive control 106 word ‘coroner’ (to identify documents that could not be read automatically). 107 108 Cases were included when a maternal death occurred, using the World Health 109 Organisation (WHO) definition 9: deaths of a female due to any causes related to or 110 aggravated by pregnancy or its management, both during pregnancy, childbirth or within 111 42 days of its resolution. We also included late maternal deaths defined as “the death of 112 a woman from direct or indirect obstetric causes, more than 42 days but less than one 113 year after termination of pregnancy”. 9 Deaths were excluded if the incidental causes of 114 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 7 death were unrelated to the pregnancy (e.g. a road traffic accident) as per the WHO 115 definition.9 116 117 Data extraction 118 Data was extracted from the included PFDs in line with established methodology that has 119 been used in other case series of PFDs10–12. Demographics, causes of deaths, pregnancy 120 outcomes and risk factors, coroner concerns and organisational responses were 121 extracted manually by two authors (JJ, DL) and reviewed by FD. Risk factors extracted 122 were based on existing literature13, including gestational diabetes, hypertensive disease, 123 obesity, cardiac disease, haematological disorders, psychiatric illness, epilepsy, aged 124 35 years, recreational drug use, complications from previous pregnancy, 125 multiple pregnancies, hepatic disease, and congenital anomalies of the child. 126 The numbers of maternal deaths (up to 1 year from birth) from MBRRACE-UK2 were also 127 extracted from each MBRRACE -UK report between 2014 -2021, to compare to the 128 number of PFDs in the same time period. Rolling averages over 3 year periods were 129 taken to match the presentation of the MBRRACE-UK data. 130 131 Data analysis 132 The number of maternal deaths reported in coroners’ PFDs and their rates as a proportion 133 of the maternal deaths reported in MB RRACE-UK were calculated over time. Medians 134 and interquartile ranges (IQRs) were calculated for continuous variables (e.g age) and 135 frequencies were reported for categorical variables (e.g sex, location of death, and 136 coroner jurisdiction area). 137 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 8 138 We calculated the years of life lost (YLL) 14 for each case (where age was reported) by 139 extracting their remaining life expectancy from the ONS cohort life tables15. The cause of 140 death determined by the coroner in each case was assigned to categories stipulated in 141 the 2019 -21 MBRRACE -UK report2. Two investigators (JJ, DL) also assigned the 142 International Statistical Classification of Diseases and Related Health Problems 11th 143 Revision (ICD-11) numeric codes for the causes of death to each PFD16. 144 145 Directed content analysis17 was used to collate and evaluate concerns raised by coroner 146 and classify them. Concerns were classified by one author (JJ) and ambiguities clarified 147 with other authors (DL, FD). 148 149 To calculate response rates to PFDs, we used the 56 -day legal requirement to classify 150 responses as “early or on time” (on or before the due date), “late” (after due date), or 151 “overdue” (response was not available on the Judiciary website as of the time of 152 extraction). We calculated the average response rate and frequency for recipients. The 153 content of responses were classified by the type of change reported if applicable. 154 155 Missing data 156 Coroners have a duty to write PFDs4,5, but this is not mandatory or enforced. Thus, data 157 is constrained by the working practices of coroners who may vary in their thresholds for 158 writing a PFD report. Furthermore, PFDs must be sent to the Chief Coroner’s Office who 159 will assess the report before publication. In some instances, reports may go unpublished 160 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 9 as they have not been sent to the Chief Coroner’s Office, get lost in the email inboxes or, 161 in very rare cases the Chief Coroner will choose not publish the report at the request of 162 family members or if the report poses a risk to the public18. We can therefore only analyse 163 the publicly-available PFDs. 164 165 Patient involvement 166 No patients were directly involved in this study, however our research team constantly 167 engages with bereaved families and friends who have lost a loved one through our 168 platform the Preventable Deaths Tracker ( https://preventabledeathstracker.net/). This 169 engagement has guided the development of our platform and helped identify new 170 research projects on specific areas of death prevention. 171 Software and data sharing 172 We used R (version 4.1.1) to create the openly available code to download and screen 173 the pdf documents19. The OCCRP Aleph tool, an openly available platform was used for 174 document storage and investigation management, as well as conducting the keyword 175 searches8. Microsoft Excel was used for data extraction and analysis. Figures were 176 created using Datawrapper20. 177 178 179 180 181 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 10

182 There were 29 coroners’ PFDs involving maternal deaths between July 2013 and August 183 2023 in England and Wales (0.7% of all available PFDs, n=4435). When compared to 184 data from MBRRACE-UK, PFDs represented only a small fraction of total maternal deaths 185 (Table 1). Direct comparison was difficult due to MBRRACE data including all UK deaths, 186 whereas PFDs only cover England and Wales. 187 Table 1. Maternal deaths reported in coroners’ PFDs published in England and 188 Wales between July 2013 and August 2023 compared with deaths reported by 189 MBRRACE in the UK between 2014 and 2021. 190 Average over 3 years* Deaths reported by MBRRACE in the UK Deaths reported in coroners’ PFDs in England and Wales Percentage of MBRRACE-UK deaths written into a PFD (%) 2013-15 528 9 1.7 2014-16 511 9 1.8 2015-17 522 13 2.5 2016-18 522 8 1.5 2017-19 475 10 2.1 2018-20 518 5 1.0 2019-21 552 6 1.1 2020-22 NA 6 NA 2021-23 NA 8 NA . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 11 Total 2013-21 2774 29 1.1% *A three year time period is used to align with MBRRACE reporting (Table 2.1 and 2.5, MBRRACE -191 UK Confidential Enquiries into Maternal Deaths 2013-21 reports)2 192 Demographics 193 The median age of death was 33.5 years (IQR 28.5-36.0 years, n=20). The median years 194 of life lost was 54 years per death (n=20, total=1051 years). One death occurred in a 195 person aged 35 years. 196 Deaths most often occurred in hospitals (75. 9%) followed by other community settings 197 outside of the home (10.3%), and one in the home (3.4%). The location of death was not 198 reported in three cases (10.3%). 199 Deaths occurred in 23 coroners’ areas out of a total of 82 areas in England in Wales, 200 most frequently occurring in London Inner North (17.2%) (Supplementary Table S1). 201 The Chief Coroner’s Office (CCO) categorise PFDs into one or more types of deaths of 202 which the most (78%) common was ‘Hospital death’ (Supplementary Table S2). 203 204 Causes of deaths 205 206 Figure 1. Causes of death s reported by coroners in PFDs involving maternity 207 deaths in England and Wales between July 2013 and August 2023. Causes of death 208 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 12 from the 2019-2021 MBRRACE report were used to categorise deaths. Note: one case 209 had >1 cause mentioned. 210 211 The most frequent causes of death (Figure 1) reported by the coroner were haemorrhage 212 (27.5%, n=8), followed by early pregnancy deaths (20.6%, n=6, which included 213 complications of ectopic pregnancies and terminations), and suicide (20.6%; n=6). 214 Further details of causes of death and ICD -11 coding can be found in Supplementary 215 Table S2. 216 44.8% (n=13) cases reported none of the 15 factors associated with high risk 217 pregnancies. 10.3% (n=3) of women had previous psychiatric history, and this was related 218 to their cause of death. Other risk factors, including previous clotting disease and multiple 219 pregnancies only occurred in a small number of individual (n=1) cases. 220 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 13 The majority (55.2%, n=16) of deaths occurred postpartum, including psychiatric causes 221 and those occurring after abortion, or surgery for ectopic pregnancy. Antenatal deaths 222 occurred in 24.1% (n=7) of women and intrapartum deaths (occurring during or within 24 223 hours of labour) occurred in 20.6% (n=6) of cases. 224 In 9 PFDs (31.0%) it was explicitly stated that pregnancy went on to lead to a live birth. 225 Four (13.8%) of deaths involved ectopic pregnancies, three (10.3%) involved terminations 226 of pregnancy, and one (3.4%) involved a miscarriage (<24 weeks). Two (6.9%) cases 227 involved antenatal deaths of the mother in which the child was presumed to have died. In 228 10 cases, the outcome for the child was not mentioned. 229 Coroner’s concerns 230 121 c oncerns were raised by coroners . The most common concern was regarding 231 providing appropriate treatment (48.2%, n=14), the failure to escalate (37.9%, n=11), 232 recognition of risk factors (31.0%, n=9), and lack of training (31.0%, n=9) (Supplementary 233 Table S3). Specific lessons are discussed below in the next section. 234 Responses to PFDs 235 PFDs were sent to 53 organisations. These organisations included NHS trusts (n=19) and 236 professional bodies such as the General Medical Council or Medical Royal Colleges 237 (n=13). Only 3 7.7% of PFDs received a response from the organisation to which they 238 were sent . When organisations did respond, 8 .00% (16 out of 20) of rep orted new 239 changes including publishing new local policies, increasing multidisciplinary training in 240 obstetric scenarios, or committing to increasing staffing levels (Supplementary Table S4). 241 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 14 The relation of types of concerns to whether change was initiated can be found in Figure 242 2. Specific actions taken in response to concerns can be found in Supplementary Table 243 S1 and “Specific Lessons” above. 244 245 Figure 2. Concerns raised by coroners (left) in Prevention of Future Deaths reports 246 involving maternity deaths in England and Wales published between July 2013 and 247 August 2023 and the actions reported by organisations in their responses (right). 248 249 250 251 252 SPECIFIC LESSONS 253 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 15 Gaps in national guidance 254 PFDs highlighted gaps in national maternal care guidance. One case (2017 -0005) 255 concerned poor follow-up for discoloration found on amniocentesis, for which there was 256 no existing guideline. The patient later presented with chorioamnionitis and died. The 257 coroner’s suggestion to send such samples for immediate microbiological analysis was 258 acknowledged by the Royal College of Obstetricians and Gynaecologists’ ( RCOG) 259 response with updated guidelines21. 260 261 Another patient (2021-0371) died due to sepsis from a retained fetus following feticide. 262 The coroner highlighted the lack of guidance surrounding treatment of such infections, in 263 particular whether antibiotics suffice, or if obstetric intervention is required. The hospital 264 trust identified a similar death in the region and changed local guidelines. However, in lieu 265 of national guidelines no specific treatment recommendations could be made. 266 267 A 2016 PFD detailed the death of a patient from bowel obstruction following previous 268 bariatric surgery (2016 -0213), where surgical causes of her symptoms were not 269 adequately considered during pregnancy nor was a specific obstetric plan made . It was 270 noted that there are no national guidelines for obstetric planning for pregnant women with 271 previous bariatric surgery, a patient subset that is increasing in number. No published 272 response from t he RCOG, General Medical Council or Care Quality Commission was 273 available so there were no responses to the concerns raised in this PFD. 274 275 National protocols not followed 276 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 16 In a 2020 case (2020 -0162), misoprostol administration to induce labour following 277 intrauterine fetal death led to uterine rupture and maternal death. The hospital in question 278 updated its misoprostol dose to match national guidelines and mandated medical reviews 279 before misoprostol administration to multiparous mothers. 280 281 Four cases highlighted deficiencies in major obstetric haemorrhage protocols. In two 282 cases (2015-0288; 2023-0095), local guidelines did not reflect national recommendations 283 and were updated accordingly. A specific problem was blood products not being available 284 near the maternity unit. In the other two cases (2022 -0228; 2017-0020), local guidelines 285 were not followed, leading to delays in diagnosis and treatment. 286 287 One death (2021-0418) resulted from a ruptured ectopic pregnancy where thrombolysis 288 was administered for suspected pulmonary embolism (PE). While MBRRACE guidelines 289 recommend a FAST scan in all women of child -bearing age suspected of a PE, this had 290 not been integrated into local trust policy. 291 292 Another PFD (2015-0414) detailed failure to prescribe adequate doses of clexane for a 293 mechanical valve contributing to development of fatal thrombosis. The report expressed 294 concern from doctors that this was a recurring issue. There was no published response 295 from the hospital trust or National Institute for Health and Care Excellence ( NICE), so it 296 is unclear if action was taken. 297 298 Communication issues 299 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 17 In case (2015 -0413), the patient attended the emergency department seven days 300 postpartum, but obstetricians were not involved in her care, and information sharing 301 between general practitioners, out of hours care and the obstetric department was poor. 302 In another case (2019-0281), the ambulance crew did not communicate that the patient 303 was pregnant to ED staff. In a third case (2019-0027), the obstetric department was not 304 forewarned about a pregnant patient's arrival to ED, leading to delays in care. Two actions 305 were reported by trusts in their responses , including the addition of a prompt to record 306 pregnancy status on call sheets and the development of a set of criteria to determine if 307 an obstetric call needs to be initiated prior to ambulance arrival. 308 309 Poor communication between teams was highlighted in the psychiatric care of pregnant 310 women. Two PFDs reported failures to coordinate multidisciplinary care for pregnant 311 patients between their medical and psychiatric teams (2015-0418, 2017-0055). Delays in 312 accessing psychiatric care were found to directly contribute to the death of another patient 313 (2022-0303). 314 315 Lack of resources or staff-cover 316 Concerns have also been raised regarding staffing and resource allocation. In one case, 317 there was no formal method of getting assistance when the first consultant called could 318 not attend an emergency (2019-0453). Another challenge appears to be regions with poor 319 availability of perinatal health clinics. This was the case in one PFD (2014-0239), in which 320 the mother was not able to access community care that met both her and her son’s needs. 321 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 18 The report claimed that in the region mentioned, 50% referrals to the nearest Mother and 322 Baby Perinatal Mental health in-patient Unit are declined due to distance. 323 324 325 326 327 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 19

328 Main Findings 329 We identified 29 maternal deaths reported by coroners in PFDs. There was a broad range 330 of causes of death, occurring mostly in the postpartum period in hospitals . Coroners 331 raised significant concerns spanning all stages of pregnancy, but only 38% of PFDs 332 received a response. PFDs frequently highlighted gaps in national guidance, lack of 333 consistency in local guidelines, and problems with communication. 334 Strengths and Limitations 335 This study uses reproducible methods from previous research of PFDs involving other 336 types of deaths12. However, to the best of our knowledge, it is the first published study of 337 PFDs involving maternal deaths reported by coroners’ in PFDs. 338 We compared the data from coroners with the MBRRACE-UK initiative – an established 339 national data source on maternal deaths in the UK 2. Our findings illustrate the ability of 340 coroners’ reports to provide unique case-level insights into issues in care, systems , and 341 processes, which complements larger scale epidemiology research such as MBRRACE-342 UK. 343

of the PFD data are well -established22, including inter -coroner and inter -344 regional variability in the publication of reports and the information reported in them . In 345 the maternal setting, specifically, the type of deaths identified by coroners may be biased 346 by those sent for autopsy. This has resulted in a small sample size of maternal deaths 347 which may not be representative of maternal deaths as a whole in the UK. 348 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 20 PFDs also do not consistently report established factors contributing to maternal health 349 inequality, including ethnicity, socio-economic status, and previous parity. It is difficult to 350 conclude that risk factors were not present in a PFD case, as there is no requirement for 351 coroners to consistently report them. 352 Interpretation 353 Maternal deaths continue to be major global health issue 23. As emphasised in a recent 354 call to action by the International Network of Obstetric Survey Systems (INOSS), one of 355 the key steps to address stagnating maternal mortality rates is learning from case-based 356 analyses of maternal deaths24. National analyses in the USA 25 and China 26 suggest that 357 over 80% of maternal deaths may be preventable. 358 359 Coroners in the UK are able to report maternal deaths to MBRRACE-UK 2, which collates 360 these deaths for the confidential enquiry into maternal death and morbidity. Separately, 361 coroners have a duty to write PFDs, but we found that only around 1% of maternal deaths 362 in the UK reported by MBRRACE were written into a PFD. This is an underestimate of 363 the true number of maternal deaths where action ought to be taken by organisations. 364 365 Coroner reports can provide unique insights into the systems and processes that can go 366 wrong and lead to preventable deaths. An analysis of coroner reports of maternal deaths 367 in Ontario, Canada 27 found that (when physical injury was excluded), the two most 368 common causes of death were haemorrhage and suicide, both of which were in the top 369 three causes of death seen in our study27. 370 371 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 21 Many of the concerns raised in PFDs were reflected in the most recent MBRRACE -UK2 372 report, including failures in providing appropriate treatment, recognising risk factors, and 373 communication issues, which is an important aspect of maternal care highlighted in the 374 wider literature28. Similar concerns found in our analysis and MBRRACE -UK included a 375 need to have accessible electronic records, better shared management care across 376 multidisciplinary teams, and identification of care coordinators when multiple teams are 377 involved. A thematic analysis of maternal deaths in London, the region with the highest 378 number of maternity -related PFDs, conducted by the London Maternity Network’s 379 Maternal Morbidity and Mortality Working Group, had similar messages including 380 improved adherence to protocols and improved access to care for women with complex 381 medical needs29. 382

383 PFDs are an under-recognised source of data for improving maternal care and reducing 384 preventable deaths. Organisations that receive PFDs from coroners may be failing to take 385 action as there is no mechanism to follow up on missing responses or ensure that 386 reported actions are implemented. Using PFDs, we identified issues in the provision of 387 care and gaps in national guidance and policy. We have created a reproducible method 388 for collecting and analysing reports (https://preventabledeathstracker.net/), so that PFDs 389 can be more widely used as a learning tool to prevent future deaths. To improve access 390 to such reports, the Chief Coroner’s Office should consider updating their categorisation 391 of deaths and include ‘Maternal deaths’ as an official classification for PFDs. 392 393 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 22

394 None. 395 Disclosure of interests 396 JJ and DL declare no interests. FD works as a doctor in the National Health Service 397 (NHS). GCR has a fixed-term contract of employment at the University of Oxford to teach 398 evidence-based medicine and supervise research. GCR is the Director of a limited 399 company that has provided consultancy for the private sector. GCR travel expenses have 400 been reimbursed for speaking at conferences and events, and she has received a 401 speaker's fee for providing training and speaking at coronial law events. GCR receives 402 fees from subscriptions to a personal Substack publication. 403 404 Contribution to authorship 405 GCR established the methodology used to perform case series ’ of PFDs 406 (https://preventabledeathstracker.net/). JJ and DL conceived of the study and wrote the 407 protocol. FD wrote the code used to download the pdf documents and performed the 408

screening using the OCCRP Aleph too l. JJ and DL performed data extraction 409 and analysis. All authors interpreted the study findings and contributed to writing and 410 reviewing the manuscript. All authors accept responsibility for the paper as published. 411 412 Details of Ethics Approval 413 This study uses publicly available information, for which ethics committee approval is not 414 required. Both the General Data Protection Regulation (GDPR) and the Data Protection 415 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 23 Act (2018) no longer apply to identifiable data that relate to a person once they have 416 died(30). 417 418 Funding 419 No funding has been obtained to undertake this study. 420 421 Data availability 422 Protocols and study materials used for data synthesis are openly available on the Open 423 Science Framework (https://osf.io/7h4j6/). Demographic information from all PFD cases 424 is openly available on the Preventable Deaths Tracker 425 (https://preventabledeathstracker.net). The code used to download all the PFD pdf 426 documents is openly available (https://github.com/francescodernie/coroner_PFDs). 427 428 429 430 431 432 433 434 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 24

435 1. Trends in maternal mortality 2000 to 2020: estimates by WHO, UNICEF, UNFPA, World 436 Bank Group and UNDESA/Population Division [Internet]. [cited 2024 Mar 25]. Available 437 from: https://www.who.int/publications/i/item/9789240068759 438 2. Knight M, Bunch K, Felker A, Patel R, Kotnis R, Kenyon S, Kurinczuk JJ (Eds. ) on behalf 439 of MBRRACE-UK. Saving Lives, Improving Mothers’ Care Core Report - Lessons learned 440 to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal 441 Deaths and Morbidity 2019-21. Oxford National Perinatal Epidemiology Unit, University of 442 Oxford; 2023. 443 3. Foreign, Commonwealth & Development Office. Policy Paper: Ending preventable deaths 444 of mothers, babies and children by 2030 [Internet]. 2021. Available from: 445 https://www.gov.uk/government/publications/ending-preventable-deaths-of-mothers-babies-446 and-children-by-2030 447 4. The Coroners Rules 1984. Rule 43. 448 1984.http://www.legislation.gov.uk/uksi/1984/552/article/43/made (accessed 16 Jun 2019). 449 5. Coroners and Justice Act 2009. 450 2009.https://www.legislation.gov.uk/ukpga/2009/25/pdfs/ukpga_20090025_en.pdf 451 (accessed 16 May 2019). 452 6. The NHS Patient Safety Strategy: Safer culture, safer systems, safer patients (July 2019) 453 https://www.england.nhs.uk/wp-454 content/uploads/2020/08/190708_patient_safety_strategy_for_website_v4.pdf. 455 7. Jindal J, Launer D, Richards GC, Dernie F. Preventable deaths involving Pregnancy and 456 Maternity [Internet]. OSF. [cited 2024 5]. Available from: https://osf.io/7h4j6/ 457 8. OCCRP Aleph. https://aleph.occrp.org. 458 9. World Health Organisation Global Health Observatory: Maternal Deaths [Internet]. [cited 459 2023 Nov 5]. Available from: https://www.who.int/data/gho/indicator-metadata-registry/imr-460 details/4622. 461 10. Dernie F, France HS, Thomas ET, Bilip M, DeVito NJ, Ferner RE, et al. Preventable deaths 462 involving opioids in England and Wales, 2013–2022: a systematic case series of coroners’ 463 reports. Journal of Public Health. 2023;45(4):e656–63. 464 11. France HS, Aronson JK, Heneghan C, Ferner RE, Cox AR, Richards GC. Preventable 465 Deaths Involving Medicines: A Systematic Case Series of Coroners’ Reports 2013–22. 466 Drug Safety. 2023 Apr 1;46(4):335–42. 467 12. Bremner BT, Heneghan C, Aronson JK, Richards GC. A systematic narrative review of 468 coroners’ Prevention of Future Deaths reports (PFDs): A tool for patient safety in hospitals. 469 Journal of Patient Safety and Risk Management. 2023;28(5):227–36. 470 13. Nair M, Knight M, Kurinczuk JJ. Risk factors and newborn outcomes associated with 471 maternal deaths in the UK from 2009 to 2013: a national case-control study. Bjog. 472 2016;123(10):1654–62. 473 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 25 14. Martinez R, Soliz P, Caixeta R, Ordunez P. Reflection on modern methods: years of life lost 474 due to premature mortality—a versatile and comprehensive measure for monitoring non-475 communicable disease mortality. International Journal of Epidemiology. 2019;48(4):1367–476 76. 477 15. Office for National Statistics (ONS). Past and projected period and cohort life tables: 2020-478 based, UK, 1981 to 2070. 479 https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/lifeexpe480 ctancies/bulletins/pastandprojecteddatafromtheperiodandcohortlifetables/2020baseduk198481 1to2070. 482 16. International Classification of Diseases, Eleventh Revision (ICD-11), World Health 483 Organization (WHO) 2019/2021 https://icd.who.int/browse11. Licensed under Creative 484 Commons Attribution-NoDerivatives 3.0 IGO licence (CC BY-ND 3.0 IGO). 485 17. Kleinheksel AJ, Rockich-Winston N, Tawfik H, Wyatt TR. Demystifying Content Analysis. 486 Am J Pharm Educ. 2020;84(1):7113. 487 18. Chief Coroner. Prevention of Future Deaths Reports Publication Policy. 488 https://www.judiciary.uk/wp-content/uploads/2021/11/pfd-publication-policy-9-11-21.pdf. 489 19. Dernie F. Coroner PFD github. https://github.com/francescodernie/coroner_pfds. 490 20. Data Wrapper. https://www.datawrapper.de. 491 21. Navaratnam K, Alfirevic Z, Royal College of Obstetricians and Gynaecologists. 492 Amniocentesis and chorionic villus sampling: Green-top Guideline No. 8 July 2021: Green-493 top Guideline No. 8. BJOG. 2022 Jan;129(1):e1–15. 494 22. Zhang Q, Richards GC. Lessons from web scraping coroners’ Prevention of Future Deaths 495 reports. Medico-Legal Journal. 2023;0(0):00258172221141284. 496 23. Khalil A, Samara A, O’Brien P, Coutinho CM, Quintana SM, Ladhani SN. A call to action: 497 the global failure to effectively tackle maternal mortality rates. The Lancet Global Health. 498 2023;11(8):e1165–7. 499 24. van den Akker T, D’Souza R, Tura AK, Nair M, Engjom H, Knight M, et al. Prioritising 500 actions to address stagnating maternal mortality rates globally. The Lancet. 2024 Feb 501 3;403(10425):417–9. 502 25. Trost SL, Beauregard J, Njie F, et al. Pregnancy-Related Deaths: Data from Maternal 503 Mortality Review Committees in 36 US States, 2017-2019. Atlanta, GA: Centers for 504 Disease Control and Prevention, US Department of Health and Human Services; 2022. 505 26. Liang J, Dai L, Zhu J, Li X, Zeng W, Wang H, et al. Preventable maternal mortality: 506 Geographic/rural-urban differences and associated factors from the population-based 507 maternal mortality surveillance system in China. BMC Public Health. 2011 Apr 508 19;11(1):243. 509 27. Aflaki K, Vigod SN, Sprague AE, Cook J, Berger H, Aoyama K, et al. Maternal Deaths 510 Using Coroner’s Data: A Latent Class Analysis. Journal of Obstetrics and Gynaecology 511 Canada [Internet]. Available from: https://doi.org/10.1016/j.jogc.2024.102349 512 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint Preventable maternity deaths: a case series 26 28. Brennan RA, Keohane CA. How Communication Among Members of the Health Care 513 Team Affects Maternal Morbidity and Mortality. Journal of Obstetric, Gynecologic & 514 Neonatal Nursing. 2016;45(6):878–84. 515 29. London Maternity Clinical Network, Maternal Morbidity and Mortality Working Group. 516 London Maternal Mortality Thematic Review for 2017. NHS England and NHS Improvement 517 [Internet]. 2019. Available from: https://www.england.nhs.uk/london/wp-518 content/uploads/sites/8/2019/11/London-maternal-mortality-thematic-review-2017_18th-519 July-2019-.pdf 520 30. Health Research Authority (HRA): Consent and Participant Information Guidance [Internet]. 521 [cited 2024 Mar 28]. Available from: https://www.hra-522 decisiontools.org.uk/consent/principles-523 deceased.html#:~:text=In%20legal%20terms%2C%20the%20General,person%20once%20524 they%20have%20died. 525 526 527 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 10, 2024. ; https://doi.org/10.1101/2024.07.09.24310137doi: medRxiv preprint