Ms Laura Galligan:

I thank the Chairman and members for the opportunity to address the committee and discuss current issues relating to the recording and classification of Garda homicide figures. I am a senior crime and policing analyst in An Garda Síochána. I work within the Garda Síochána analysis service, GSAS, and have been attached to special crime operations since June 2016. From 2012 to 2016, I worked in the position of the senior scientist in the Office of the State Pathologist and had sole responsibility for the management and running of the laboratory, where I performed all histopathological tests on samples from State post mortems performed by the State pathologists. I also accompanied the forensic pathologists to crime scenes and assisted them in post mortem examinations. My educational background includes a five-year honours degree in biomedical science and a postgraduate qualification in forensic medicine. I am currently undertaking a professional certificate in governance.

In July 2016, my line manager, Ms West, assigned me responsibility for a ten-year domestic homicide review from 2007 to 2016 as requested by the Garda national protective services bureau. Ms West had from the outset expressed general concerns about the data quality on the PULSE system. To prepare for the assessment of how the ten-year review could be undertaken, I carried out an initial overview analysis of information contained on PULSE in regard to the homicide categories. This involved assessing how easy it would be to recognise a homicide with a domestic element.

In September 2016, a meeting was held with the chief superintendent of the Garda national protective services bureau to discuss the action plan of the task. With prior permission from Ms West, I suggested working in collaboration with the Office of the State Pathologist as, from my previous experience, I was aware the records of that office may contain more information on whether there was a domestic element to the death, which would greatly assist me in undertaking the ten-year review. The suggestion was very much welcomed by the chief superintendent and other members of the Garda national protective services bureau.

I had met the State pathologist, Professor Marie Cassidy, to ascertain her interest in the work and wrote to her on 8 September 2016 requesting an agreement to collaborate data and information to carry out the review between the Office of the State Pathologist and the Garda Síochána analysis service. On 14 September 2016, Professor Cassidy wrote to me to confirm her agreement to the collaboration. I would like to take this opportunity to thank Professor Marie Cassidy, Dr. Linda Mulligan, Dr. Michael Curtis, Dr. Margot Bolster and all staff in the Office of the State Pathologist for accommodating the review as well as the professional support and advice they have continually offered me.

I compiled a domestic homicide review proposal which was sent to Ms West and the chief superintendent in the Garda national protective services bureau. I set out the methodology I would utilise to conduct as comprehensive an analysis as possible and designed a template qualitative data sheet.

GSAS produces an internal homicide spreadsheet every month. I was the analyst responsible for the spreadsheet in September to December, inclusive, in 2016. In the four months I produced the spreadsheet, I noted the absence of a number of homicides from 2016 which were recorded as non-crime incidents of sudden death.

I commenced my analysis at the beginning of October 2016, utilising records from the Office of the State Pathologist, information on PULSE and the homicide spreadsheets produced monthly by GSAS for internal use. When I reviewed the post mortem report and the circumstances surrounding the death, with consultation with the relevant forensic pathologist if required, I scrutinised PULSE for the recording of the death and other data quality parameters. For no particular reason, I began with the year 2014 but quickly realised that there were inconsistencies evident with homicide and death classification as well as many data quality issues. I immediately raised my concerns with my line manager, Ms West.

Following a period of three weeks in October 2016 when I was abroad on training for An Garda Síochána, I resumed the homicide review in November 2016. Ms West asked me specifically to consider a period of three years to ascertain any inconsistencies and the varying issues I had identified. We decided to conclude with a sample of three years as it was pertinent that the issue be raised with Garda management as soon as possible. On 24 November 2016, I emailed to GSAS management an 87 page report entitled Comparative Analysis of the Recording and Reporting of Homicide Incidents in the PULSE Database and the Office of the State Pathologist. It regarded 43 cases I identified within the time period of 2013 to 2015, not all of which were potential homicides. There were many areas of concern in regard to misclassification, under-reporting, over-reporting, delay in recording, and data quality issues. I reviewed 524 deaths in total for that period.

There were four main areas of concern: deaths that were pathologically considered to be a homicide but recorded in the incorrect crime category and-or a non-crime category; deaths that were recorded at a later date as a homicide and-or recorded as occurring in the incorrect year; deaths deemed suspicious by virtue of cause of death and in which criminal involvement would be determined by Garda investigation but which were recorded in a non-crime category or no category; and deaths that had varying data quality issues such as the deceased not being marked as such on PULSE, an incorrect "weapon used" or date of birth being recorded for the deceased, or the incorrect Garda district or division being recorded for the death.

In mid-January 2017, a review group comprising myself, Ms West and officers from policy development, implementation and monitoring, PDIM, was created. PDIM stated that two of the 43 deaths with which I had raised issue in the report from 2013 to 2015 would not be considered in the review for reasons that cannot be publicly disclosed due to legislative frameworks, and 41 cases were, therefore, the subject of review between PDIM and GSAS. As mentioned by Ms West, nine meetings were held between January and March 2017 attended by individuals from PDIM, Ms West and me. Robust discussions were had at the meetings and Ms West and I voiced our concerns at what we deemed a very serious issue of the recording and classification of deaths within An Garda Síochána. However, we were not given the professional respect we deserved and, at times, I felt we were belittled and treated very poorly. At no stage during the nine meetings was the methodology I utilised questioned by members of PDIM or referred to as weak.

During those months, I began interrogating the PULSE database to try to determine a little more background on how deaths can be and are recorded in An Garda Síochána. I identified that there are 16 incident categories on PULSE under which a death could potentially be recorded. I considered this a significant number of ways to record a death and the categories were extremely varied in their descriptions according to the PULSE creation manual. I continued to analyse the PULSE database and discovered that there are no identifiable parameters on PULSE for fatal or non-fatal hit-and-run incidents. On inspection of some fatal hit-and-run incidents, I discovered they were all recorded as traffic collision fatal incidents, which is a non-crime category, but a criminal investigation is initiated from the time of such incident occurring. I raised these issues among many others during the nine meetings with PDIM but we were quickly told they were not up for discussion and only the 41 cases were to be reviewed. We felt that the various issues I was identifying were critical to highlighting the facts of misclassification in the recording of all types of deaths.

In January 2017, having continued the analysis of deaths in 2012, I wrote an 81 page report on various cases with serious anomalies and sent it to Ms West. Areas of concern were consistent with the previous report and others identified were: deaths recorded in non-death categories such as missing persons, public order offences and attention and complaints; deaths being counted twice; and deaths not recorded for significant periods of time after they had occurred. Although this report was completed, PDIM gave strict instructions not to be sent the report as it was not reviewing any more than the 41 cases from the three-year period.

Within this timeframe, I identified other issues that I immediately raised to various individuals in management. I raised an incident of serious concern to the Garda national protective services bureau upon discovering that the partner of a deceased person was in a new domestic violent relationship. I have since highlighted several such cases to the Garda national protective services bureau and that led to the establishment of Operation Devise, in which I have been involved. I highlighted an incident that I had discussed with the chief superintendent in the Garda national protective services bureau which was recorded as a non-crime incident but was a domestic murder-suicide incident. Following inquiries with the relevant division by the chief superintendent, the incident was appropriately reclassified that evening.

In February, I had a routine meeting with my manager, Ms West, during which I expressed my concerns at GSAS continuing to produce the monthly homicide spreadsheet in view of our being unable to guarantee the accuracy of PULSE. It was agreed with Ms West that I would review January 2017, in collaboration with Professor Marie Cassidy and the Office of the State Pathologist, to compare the cases carried out and the corresponding PULSE incident recorded. In March 2017, I wrote a report entitled Data Quality Issues with Death Classification 2017 which was sent to management. There were many areas of concern in terms of misclassification, under-reporting, over-reporting, delay in recording, and data quality issues.

There were three unrecorded homicides as of the end of February 2017. One of the unrecorded homicides was raised at a senior leadership meeting the next day. It still remained recorded incorrectly for a further six months, despite being highlighted numerous times. There were cases with the following concerns: homicides misclassified; suicides misclassified; hit-and-run incidents recorded as non-crime traffic fatal incidents; suspicious deaths under sudden death classification; and natural deaths containing PULSE data quality issue.

At the end of April 2017, I was extremely surprised when it was announced in the media that the review into homicides was complete, and the conclusion was that there were only "minor misclassifications". It was at the public Policing Authority meeting on 27 April that we became aware that the review team had written a report. Neither Ms West nor I, as part of the review team, had any input nor were our views taken into consideration in the drafting of the report. On receipt of the report on 8 May, it became immediately clear why the report was not shared with us. On reading of the report, I felt very personally insulted by the commentary regarding my work and my methodology. As previously mentioned, at no stage during the nine meetings was any criticism voiced regarding the inappropriateness of my methodology. It had now been concluded that the methodology I had utilised in comparing the records of the OSP and the PULSE system was "inherently weak", "inappropriate", "confined" and "restricted". The days between 8 and 10 May were troublesome, in that PDIM brought significant pressures to sign off on the report that we had only just received. Like Ms West, I professionally and ethically could not agree to a report I knew was completely inaccurate and misleading. I remained assured that there were homicides misclassified and incorrectly recorded.

In May 2017, I wrote a report entitled Homicide and Suspicious Death Incidents 2017: PULSE Categorisation & Office of the State Pathologist Categorisation, which was sent to management. There were many areas of concern continuing in misclassification, under-reporting, over-reporting, delay in recording, and data quality issues, such as 33 homicides recorded as of May 2017, three of which did not occur in 2017 but would be counted by the CSO; six unrecorded homicides recorded as follows: non-crime incidents of person misadventure - sudden death; non-fatal incidents of assault causing harm; and non-fatal incidents of discharging a firearm. There were a number of fatal hit-and-run traffic incidents in 2017, all of which are unidentifiable in the recording on PULSE. These incidents are routinely recorded as a non-crime incident of traffic - traffic collision FATAL when a criminal investigation is ongoing. In the email sending this report, I stated:

There are many victims going unnoticed by this organisation the longer this goes on. When these cases are being recorded as a non-crime category of Sudden Death or even Discharging a Firearm, is there any follow up with the victim’s family or liaison with a FLO. Is there even a FLO being assigned? The deceased victims have not been given the respect and duty of investigation into their deaths that they deserve. Sadly, the majority of 2017 homicides are of a domestic nature, again going unreported.

This is 2017 in four months 2013 to 2015 we are now satisfied that PDIM have reviewed, but see nothing but ‘minor inaccuracies’, which is completely incorrect. I am not going to lose my work ethic or determination over this, and one didn’t build Rome in a day. However, I cannot continue to identify incidents of homicide as we continue in 2017 that are not recorded correctly or not knowing whether there is an adequate investigation being carried out into that person’s death.

In June 2017, I wrote a report entitled Homicide and Suspicious Death Incidents 2017 – Month Ending May 2017, which was sent directly to senior Garda management. There were the same areas of concern continuing in misclassification, under-reporting, over-reporting, delay in recording, and data quality issues, as well as new cases emerging that had issues, such as 18 homicides recorded as of May 2017, three of which did not occur in 2017 but would be counted by the CSO; six unrecorded homicides, still unchanged from previous report; ten suspicious deaths being recorded under a non-crime classification, although a criminal investigation was initiated; and a number of fatal hit-and-run traffic incidents in 2017, all of which were recorded as a non-crime incident of traffic - traffic collision FATAL.

In July 2017, I wrote a report entitled Homicide and Suspicious Death Incidents 2017 – Month Ending June 2017, which was sent directly to senior Garda management again. There were the same areas of concern continuing in misclassification, under-reporting, over-reporting, delay in recording, and data quality issues, as well as new cases emerging that had issues, such as 20 homicides recorded as of May 2017, three of which did not occur in 2017 but would be counted by the CSO; nine unrecorded homicides recorded in non-crime or a non-fatal crime incident type; ten suspicious deaths being recorded under a non-crime classification, although a criminal investigation was initiated; and a number of fatal hit-and-run traffic incidents in 2017 recorded as a non-crime incident of traffic collision FATAL, when a criminal investigation was ongoing.

In August 2017, I wrote a report entitled Traffic: Dangerous Driving Causing Death Incidents: PULSE Categorisation analysis 2003-2017, which was sent to management. There were many areas of concern in the recording of the dangerous driving death incident type over the years 2003 to 2017.

I also wrote a report in August 2017 entitled Review of 2006 Homicides – Murder and Manslaughter: PULSE Categorisation Analysis, and sent this to GSAS management for onward transmission. This was an overview document regarding the adherence to crime counting rule 3.1 in 2006. As per headquarters directive 139/03, section 3.1 states, "A reclassification within homicide occurs where a murder is reclassified to manslaughter when a charge of manslaughter commences or when a murder charge results in a conviction for manslaughter". This review analysed conviction records and detection status for all homicide incidents as well.

In October 2017, I wrote two reports entitled Homicide and Suspicious Death Incidents 2017 – Month Ending September 2017, and An Analysis of Death Incidents – September 2017, which were sent directly to senior management. There were the same areas of concern continuing in misclassification, under-reporting, over-reporting, delay in recording, and data quality issues, as well as new cases emerging that had issues such as 37 homicides recorded as of May 2017, seven of which did not occur in 2017 but would be counted by the CSO; six unrecorded homicides recorded in non-crime or a non-fatal crime incident type - a number of the previous highlighted misclassified homicides had been reclassified to a homicide so there were new homicides; 11 suspicious deaths being recorded under a non-crime classification although a criminal investigation was initiated; and six fatal hit-and-run traffic incidents recorded in 2017 as a non-crime incident of traffic collision FATAL. I also had queries as to why in some deaths with concerning details the State Pathologist was not requested when a crime scene was held at the death. There were deaths that should be classified as suicide or that had a suicide incident recorded which resulted in double counting of the death, according to the detail in the narrative. There were varying data quality issues.

In November 2017, I wrote a report entitled Homicide and Suspicious Death Incidents 2017 – Month Ending October 2017, which was sent to management. There were the same areas of concern continuing in misclassification, under-reporting, over-reporting, delay in recording, and data quality issues, as well as new cases emerging that had issues, such as 40 homicides recorded as of October 2017, seven of which did not occur in 2017 but would be counted by the CSO; five unrecorded homicides, recorded in non-crime or a non-fatal crime incident type; 13 suspicious deaths being recorded under a non-crime incident, although a criminal investigation was initiated; a number of fatal hit-and-run traffic incidents in 2017 recorded as the non-crime incident of traffic collision FATAL; 18 dangerous driving causing death incidents recorded for 2017, four of which did not occur in 2017, one had been invalidated, and one had three fatalities but only one crime incident when there should have been one for each victim; and 95 deaths in October with varying types of data quality issues.

I have continued to highlight to the current day cases that I believe are of concern in areas of homicide, traffic fatal incidents, suicide, sudden death and, indeed, homicide attempts. The classification of death incidents on PULSE requires an urgent review and needs to be simplified. The review of the classification comparing PULSE and the OSP is based on the pathological classification, but the ultimate arbitrator of the classification of death is the coroner. Unfortunately, coroners do not publish statistics. I have been directed to complete a review of all files in the OSP between 2003 and 2017, totalling more than 3,000 deaths, for classification purposes and investigation review using the methodology from the initial review.I will endeavour to complete this as quickly as is possible and also to complete the requested domestic homicide review that was initially requested by the Garda national protective services bureau. I will continually strive as an analyst in An Garda Síochána to ensure every death is recorded correctly, not just for statistical purposes but mainly for the victims and their families.

I thank the Chairman for the invitation to appear before committee to provide clarity and context to the work that we do.