Andrew Goldstraw · Prevention of Future Deaths report

On 27th January 2020 I commenced an investigation into the death of Andrew Goldstraw, aged 43. The investigation concluded at the end of the inquest on 6th February 2020. The jury concluded that Mr Goldstraw intended to take his own life and that he did so as a result of:

1 a Ligature suspension

The jury returned a narrative conclusion in which they set out their concerns regarding the care that Mr Goldstraw received whilst at HM Prison, Winchester between 23rd October and 14th November 2018. In particular, the jury found that an ACCT should have been opened, the absence of which would more than minimally have helped to prevent his death. The ACCT would have resulted in awareness of his risk factors and would have created better cross service communication. The jury found that Mr Goldstraw was suffering from an adverse psychological state due to a combination of drugs (Spice) and psychoactive medication (Fluoxetine and Mirtazapine) taken prior to his death. If an ACCT has been opened he would not have had access to / in possession mediation.

Mr Goldstraw was found hanging in cell B3-03 at HM Prison, Winchester at 7.23am on 14th November 2018. There was no dispute that he had taken his own life using a ligature made from torn bed linen.

At the time of Mr Goldstraw’s death, healthcare was provided by Central and North West London NHS Foundation Trust.

A. The computer system used by CNWL is known as SystmOne. Healthcare staff working on Reception when Mr Goldstraw first arrived at the prison had access to his previous medical notes and history (around 240 pages in all) stored on SystmOne. The records contained numerous references to suicidal ideation and previous attempts at deliberate self-harm. Mr Goldstraw had attempted to take his own life on several previous occasions, the most recent of which was only three months prior to his arrival at the prison. However, despite a proliferation of entries making reference to his mental health history the mental health nurse who had access to SystmOne was seemingly unaware of the relevant entries. Had he been, he said he would have opened an ACCT.

B. SystmOne makes it difficult for a doctor or mental health nurse to ascertain the key information needed to undertake a risk assessment and to decide whether or not to open an ACCT. Too much reliance is placed on the individual prisoners presentation and how he answers a series of pre-set questions.

C. At best, SystmOne makes it difficult for a mental health nurse to ascertain the relevant information and at worst it actively misleads them. For example, a search can be made of the "journal" section but this would rely on the exact words being searched (such as "suicide" or "deliberate self-harm") and it would then be necessary to go through the various entries (in Mr Goldstraw's case spread over 111 pages) using the "Key Word Search" function. Further, the functions that would (on the face of it) serve to assist in this situation (such as the "Summary" page or "Active Problems" section) were not populated with the information relevant to an accurate assessment. It was conceded by the legal representatives acting on behalf of CNWL that the "Summary" section is "very limited in its contents" and is not routinely used by healthcare staff within the prison in order to gain an insight into a prisoner's past medical history.

D. The "Active Problems" section of SystmOne is subdivided in to a number of distinct areas and it appears to be wholly inadequate in terms of identifying key areas of concern such as the risk of suicide or deliberate self-harm. The only information contained in the "Active Problems" section of SystmOne in Mr Goldstraw's case was four years out of date. None of the relevant information was contained in "Active Problems" but a great deal of irrelevant information was there!

E. The "Communications" section of SystmOne contains a chronological record of correspondence with the hospital, GP surgery and psychiatric units. However, the "Key Word Search" facility does not function at all and short of going through all of the correspondence there is no way of identifying the key information needed to undertake an effective risk assessment. The "Communications" section in Mr Goldstraw's case amounted to 83 pages. Although the relevant information concerning Mr Goldstraw's mental health issues was contained within the "Communications" section of SystmOne there was no way of easily extracting it.

F. Accordingly, a busy, under pressure mental health nurse or doctor is very likely to struggle to find the relevant entries using SystmOne, which may explain why (in Mr Goldstraw's case) too much reliance was placed on how he presented during interview. A prisoner who chooses not to disclose his true state of mind or suicidal ideation is unlikely to come to the notice of the healthcare staff whose job it is to identify the risk that he may pose to himself because SystmOne does not facilitate this.

G. There also appeared to be a lack of training in relation to the effective use of SystmOne. In particular, it was not clear whether any steps had been taken to ensure that the staff who were working at the prison at the time of Mr Goldstraw's death had been retrained or had their competencies assessed in light of the failures identified. There is a real concern that some staff are still failing adequately to carry out assessments of a prisoners risk of suicide / deliberate self-harm.

H. The Head of Healthcare at HM Prison, Winchester has indicated that she intends to provide (in conjunction with the Prison Governor) a joint learning bulletin to all staff, stressing the importance of sharing information, most notably in reception and during the early days in custody. However, this does not address the technical shortcomings of SystmOne which present a matter of considerable concern, even if healthcare staff undertake all reasonable steps to ascertain a prisoner's previous mental health history as part of the prison induction process.