Tony Duncan

Chief Nurse’s Office 
Trust Headquarters 
1st Floor Administration Building 
Maudsley Hospital 
Denmark Hill 
London 
SE5 8AZ 

07/01/2026 

PRIVATE AND CONFIDENTIAL 

Dear HM Senior Coroner Alison Hewitt, 

Re: Tony Montana Duncan 
Date of birth: 20/08/1989 
Date of death: 04/07/2024 

Thank you for your Regulation 28 Report dated 15th October 2025, setting out your 
concerns  to  be  addressed.  I  would  like  to  begin  by  expressing  our  deepest 
condolences to the family and friends of Mr Duncan on their loss.  

The concerns set out in your PFD report were as follows: 

1.  The  Deceased  presented  to  the  South  London  and  Maudsley  NHS  Foundation 
Trust’s  psychiatric  liaison  team  which  was  operating  within  the  Accident  and 
Emergency  Department  of  King’s  College  Hospital,  with  a  referral  letter  from  his 
General  Practitioner  which  sought  possible  admission  and  medication  review.  The 
Deceased  was  known  to  the  Trust,  and  he had  been  the  subject  of  a  safeguarding 
referral  and  a  self-referral  shortly  before  his  attendance  at  the  hospital.  From  the 
information available to the psychiatric liaison team, it was apparent that: 
(i)  The  Deceased  had  a  chronic  and  persisting  mental  health  condition  which  was 
usually  controlled  by  medication  but  which,  when  not  controlled,  could  give  rise  to 
suicidal ideation; he had previously been helped by periods of detention / voluntary 
admission to hospital, 
(ii) By May 2024, there was evidence that he was suffering an acute deterioration in 
his mental health which he subsequently reported was because he had not been 
properly compliant with his prescribed medication for a number of weeks, and 

Page | 1 

 
 
 
 
 
 
 
 
 
 
 (iii)  The  Deceased  recognised  the  deterioration  in  his  mental  health,  that  he  was 
suffering specific suicidal ideation relating to jumping from London Bridge, and that he 
needed help from mental health services, including by voluntary admission to hospital; 
he sought help by making a  self-referral to the Trust via the Single Point of Access 
service and by attending his GP and the hospital. 

2.  When  the  Deceased  attended  the  hospital,  the  Accident  and  Emergency  team’s 
triage notes included express reference to his specific suicide plan and attached the 
GP’s letter of referral. The Deceased was then assessed by a psychiatric liaison nurse 
who concluded that his presentation was as a result of psycho-social stressors rather 
than mental illness; she was not concerned about the risk of suicide  because he had 
no  plan  or  intent;  and  she  referred  the  Deceased  to  the  homelessness  team  and 
discharged him back to the care of his GP. The nurse did not take any steps to review 
the  Deceased’s  medication  or  consider  admission,  or  escalate  these  matters  to  a 
doctor,  nor  did  she  involve  the  Crisis  or  Home  Treatment  teams  for  follow  up  / 
immediate safeguarding. Despite there being a recognised risk to self and to others, 
both of which the Deceased himself said he could not control, there is no evidence of 
any risk assessment documentation being completed. 

3. The Deceased was subsequently seen in the Accident and Emergency Department 
by a Social Worker from the homelessness team. The Deceased insisted that he was 
not homeless and that he had attended the hospital for help with his mental health, 
without  which  he  would  jump  from  London  Bridge.  The  Social  Worker  immediately 
passed  this  information  to  members  of  the  psychiatric  liaison  team  who  he  found, 
together,  in  their  office.  Subsequently,  whilst  still  in  the  department,  the  Deceased 
became  agitated  and  abusive,  which  behaviour  was  a  recognised  aspect  of  his 
behaviour  when  he  was  unwell.  It  seems  he  later  left  the  department  and/or  was 
escorted out as he was being abusive; the records show that at least one member of 
the  psychiatric  liaison  team  was  aware  of  this  development  but  took  no  action  to 
prevent the Deceased from leaving or to encourage him to stay in order to re-assess 
him, nor to alert the Crisis and/or Home Treatment teams, the GP, or the Deceased’s 
family as to the situation. 

4. Following the report of the Deceased’s death, South London and Maudsley NHS 
Foundation Trust’s own review highlighted various concerns about the operation of its 
Single Point of Access service but neither that review, nor the evidence provided to 
the inquest from the Consultant Psychiatrist who was responsible for the psychiatric 
liaison team in King’s College Hospital, identified any concerns about the management 
of the Deceased by the psychiatric liaison team on the 4th July 2024. This may suggest 
that there were systemic as well as operational factors which led to the Deceased not 
receiving the help and support, he needed on the 4th July 2024. 

Before we address the concerns you have raised, I would like to apologise that there 
was  no  review  available  at  inquest  into  the  care  and  treatment  provided  by  King’s 

Page | 2 

 
 
 
 
 
 
 College Hospital Liaison Psychiatry. In accordance with Trust mortality processes, a 
mortality review had been carried out, and the incident had been referred for an After-
Action  Review  (AAR),  which  is the  Trust’s  main  investigatory  response  to  a  patient 
safety incident. However, the AAR spanned two boroughs and was taken forward by 
Lambeth  Governance  Team  with  no  representation  from  Southwark.  This  was  an 
oversight  and  meant  that  the  AAR  did  not  focus  on  learning  from  the  parts  of  Mr 
Duncan’s care which had been provided by the King’s College Hospital (KCH) Liaison 
Psychiatry  team,  which  is  a  service  in  the  Southwark  Directorate.  Governance 
arrangements for AARs have been discussed at the Trust Patient Safety Committee 
on 13th November 2025 and are being reviewed.  

Therefore, an AAR into the care and treatment provided by KCH Liaison Psychiatry 
was carried out retrospectively, on 19 November 2025, and is attached. The review 
identified several areas of learning and improvement, and it has also established facts 
that  the  Trust  did  not  put  forward  at  the  inquest,  which  is regrettable.  Although  the 
Trust appreciates that the coroner has made findings of fact based on the evidence 
heard at the inquest, the Trust still considers it important to set out its position. The 
Consultant  Psychiatrist  covering  KCH  Liaison  Psychiatry  on  the  day  of  the  incident 
may have been able to put forward some of these points during his verbal evidence at 
the  inquest  but  unfortunately became  seriously ill days  before  the  inquest  and  was 
unable to attend. 

In response to the concerns raised: 

1. Mr Duncan had a diagnosis of personality disorder (Type B, emotionally unstable 
traits, also referred to as borderline personality disorder or EUPD). This is a chronic 
and  persistent  mental health  condition  and  had  been  diagnosed following  a  two-
week admission to hospital in 2016. He had been stable since then and managed 
by his GP in primary care, except for a brief review by a SLAM community team in 
2022. Since 2016, his GP had prescribed his medication, namely the antipsychotic 
olanzapine. Olanzapine can be helpful for sleep and agitation as well as psychotic 
treatment  of  emotionally 
symptoms  and  whilst  sometimes  used 
unstable/borderline  personality  disorder,  it  is  not  a  strongly  evidence-based 
treatment; NICE guidelines recommend that medication is used in the treatment of 
personality disorder on a short-term basis only, or for the treatment of co-morbidity. 
It  is  not  clear  discontinuation  of  medication  was  the  main  causative  factor  in  Mr 
Duncan’s relapse as one might suspect in a psychotic illness. During the Psychiatric 
Liaison  Nurse’s  (PLN)  assessment  in  KCH  Emergency  Department  (ED),  Mr 
Duncan did not present with signs or symptoms of psychotic or mood disorder, or 
with  agitation  or  sleep  disturbance  which  might  indicate  the  need  for medication. 
We acknowledge the GP requested a medication  review in the referral letter and 
that the PLN did not address this. Medication review is often more usefully carried 
out with a full treatment history, and this is more suitable for the Community Mental 

the 

in 

Page | 3 

 
 
 
 
 
 Health  Team  (CMHT)  setting;  this  was  another  important  reason  to  ensure 
community follow up.  

We acknowledge the safeguarding referral and self-referral to Lambeth Single Point 
of Access prior to presentation at the ED provide further evidence that Mr Duncan 
was  experiencing  significant  and  persistent  need/distress.  The  safeguarding 
referral  contains  information  suggesting  possible  paranoia;  if  the  assessing  PLN 
had  access  to  this,  an  exploration  of  the  reasons  behind  the  presentation  to  the 
police could have elicited important information about his mental state. The referral 
was  uploaded  to  an  unusual  part  of  the  electronic  notes  which  is  not  always 
routinely  checked,  nor  was  it  referenced  in  the  main  body  of  the  notes; 
recommendations for managing such information have been made by the AAR as 
it is crucial staff have easy access to it.  

Mr Duncan had last been admitted to hospital in 2016; this was initially a voluntary 
admission, subsequently an emergency Section 5(2) was used to detain him until a 
full  Mental  Health  Act  assessment  could  be  organised.  He  was  found  not  to  be 
detainable  and  self-discharged  against  medical  advice.  Shortly  after  this  he 
requested to be readmitted and when this could not be accommodated, he caused 
damage  to  Trust  property  by  smashing  the  windows  of  the  ward.  When  in  the 
community, he was less agitated than on the ward and was subsequently managed 
by a CMHT and then primary care without further intervention from acute services. 
It is therefore not clear admission had been helpful, and this is not uncommon in 
patients  with  personality  disorder  and  one  of  the  reasons  why  the  benefits  of 
admission  should  be  weighed  against  potential  harms  of  admission  to  hospital. 
NICE  guidelines  state  alternatives  to  admission  must  be  considered  and  likely 
harms resulting from admission ought to be discussed with the patient. Mr Duncan 
initially seemed to be able to engage with community treatment and therefore this 
would have been an appropriate plan. These principles likely shaped some of the 
decision making by the PLN. 

2. The AAR found that on review of the SLAM electronic record, the PLN considered 
admission  but  concluded  this  was  not  indicated.  NICE  guidelines  mandate  the 
consideration of alternatives to admission in the treatment of borderline/emotionally 
unstable  personality disorder.  She  discussed  her  opinion  that admission  was  not 
indicated  with  Mr  Duncan  who,  according  to  the  record,  seemed  initially  in 
agreement. However, the PLN did not document any exploration of the discrepancy 
between the presentation to the GP and the presentation to her. Whilst this change 
in  presentation  is  not  unusual  in  itself  -  as  suicidal  ideation  is  dynamic  and  can 
fluctuate rapidly, particularly in response to emotional containment - exploration of 
this could have been helpful in formulating a better understanding of the triggers 
and  mitigating  factors  for  suicidal  ideation.  Training  needs  around  assessing 
suicidal ideation are discussed below.  

Page | 4 

 
 
 
 
 
 
 Addressing social stressors, as the PLN did through referral to the homeless team, 
is an important aspect of holistic mental health care, and Mr Duncan was also in 
agreement  with  this  plan  initially.  The  PLN’s  plan  was  to  refer  to  a  CMHT  after 

review by the homeless team.  

Referral by  the  PLN to  a  doctor  was  not  clearly indicated,  as  this pathway  is for 
those who may require admission under the Mental Health Act, or changes to their 
medication,  or  for  other  complexities  as  deemed  by  the  assessing  clinician.  The 
AAR  explored  onward  referral  to  doctors  by  PLNs  and  this  was  thought  to  be 
working well, with senior psychiatric doctors available 24 hours a day. Mr Duncan 
initially presented as calm and without signs or symptoms of affective disorder or 
psychosis. Later, when Mr Duncan became agitated, referral to a doctor to consider 
next steps (including potential referral to a crisis team) may have been indicated, 
and his self-discharge without further review or discussion was a lost opportunity to 
review  the  assessment  and  offer  further  support.  Had  Mr  Duncan  been  found  to 
have capacity to make decisions around treatment and care, as he was in the initial 
assessment, there would have been no grounds to detain him and stop him if he 
insisted on leaving. However, it would have been useful to review his mental state 
again, given that his presentation appeared to change while he was in the ED. It 
does not appear that the Liaison Psychiatry team were informed by the ED team 
that Mr Duncan was trying to leave the ED and self-discharge, until such time as he 
was  being  escorted  out  by  security.  The  AAR  recommends  that  potential  self-
discharges  must  be  flagged  to  the  Liaison  Psychiatry  Team  by  Emergency 
Department  colleagues  early  and  there  must  be  consideration  whether  further 
assessment is warranted to ensure self-discharge is safe. 

Ideally,  the  PLN  would  have  sought  Mr  Duncan’s  consent  to  contact  a  named 

person/carer, ideally his mother with whom he lived, but did not do this; the AAR 
has made a recommendation to address this omission by embedding ‘carer contact’ 

in  the  Liaison  Psychiatry  departmental  handover  board;  this  must  be  done  and 
documented before patients can be discharged. The Trust  is accredited under the 
Triangle of Care initiative led by the Carers Trust and endorsed by NHS England, 
which seeks to implement six key standards required to achieve better collaboration 
and  partnership  with  carers,  including  identification  of  carers  at  first  contact;  the 
implementation of this in the ED can be  difficult for reasons outlined in the AAR, 
and this extra flag is intended to provide further operational support for future patient 
cases.  The  AAR  did  note  that  adult  patients  with  mental  capacity  to  make  the 
relevant decision may well decline or refuse a request to contact their family, but in 
this case there is no documentation that this discussion took place. Family members 
are  often  able  to  provide  useful  collateral  information  which  can  assist  in  care 
planning, even if the patient does not permit the clinician to share information about 
them. However, if a patient refuses to allow contact, it may not be possible to make 
this contact. This should be noted in the electronic record.  

Page | 5 

 
 
 
 
 
 The AAR reviewed the electronic record which showed the PLN did complete the 
risk assessment tool and documented a brief risk assessment in the ‘Events’ section 

of  the  SLAM  electronic  record.  However,  she  did  not  explore  the  discrepancy  in 
suicidal  intent  and  planning,  especially  such  that  was  evident  in  the  difference 
between her and the GP’s assessment, and there is not sufficient evidence in the 
notes that she carried out safety planning with Mr Duncan. It is possible that she 
intended to do this following review by the homeless team, as what they can offer 
often  might  affect  what  can  be  discussed  around  a  safety  and  crisis  care  plan. 
Although  it  is impossible  to  accurately predict  suicide  in  an  individual,  modifiable 
risk  factors  for  suicide  should  be  identified  and  addressed  and  safety  planning 
should  be  carried  out.  Furthermore,  risk  assessment  should  involve  a  carer,  if 
possible. Training on personalised risk assessment and management was released 
by NHSE in September 2025, and the AAR recommends that such training should 
be  mandatory  for  clinicians.  The  Trust  is  one  of  ten  mental  health  organisations 
taking  part  in  a  national  pilot  through  the  NHS  England  and  Royal  College  of 
Psychiatrists  Culture  of  Care  Programme  –  Personalised  Approach  to  Risk.  The 
pilot aims to enhance how we approach,  assess, and manage the risk of suicide. 
This  work  aligns  with  the  NICE  guidance  for  Self-harm,  which  states  that  risk 
assessment tools should not be used to predict suicide. Further information can be 
found  here:  Culture  of  Care  Programme and  here:  NCISH  |  Implementing  a 
personalised approach to risk.  

3. The AAR found that Mr Duncan was ambivalent about whether he wanted help with 
his housing. At some points during his treatment episode at KCH, he said he wanted 
to be seen by the homeless team, at others he did not. His mental distress increased 
after he was seen by the homeless team and he then stated he would jump into the 
Thames if he were not admitted to a psychiatric ward. A recommendation made by 
the  AAR  is  that  the  Liaison  Psychiatry  team  and  KCH  homeless  team  should 
consider  seeing  patients  together,  especially  patients  with  complex  emotional 
needs/borderline/emotionally  unstable  personality  disorder  where  differences  of 
opinion between different teams can be marked due to the phenomenon of ‘splitting’ 

of affective states.  

When  Mr  Duncan  became  abusive and  was escorted  out  of  the department,  the 
PLN team  seemed  to  be  of  the  opinion,  perhaps  on  the  basis of  the  information 
provided  to  them  by  the  ED team,  that  this was  also  a  self-discharge  made  with 
mental  capacity.  The  original  plans  to  refer  on  to  a  CMHT  and  finalise  the 
management plans around discharge were then lost. This was a missed opportunity 
to offer further support. The AAR recommends that any Liaison Psychiatry patient 
wishing  to  self-discharge  from  the  ED must  be  flagged  to  the  Liaison  Psychiatry 
team  at  the  earliest opportunity  and  the  Liaison  team must  then  consider further 
assessment  and  whether  the  decision  to  self-discharge  is  appropriate. The  Trust 
endorses  the  Triangle  of  Care,  as  discussed  earlier,  and  the  PLN  should  have 

Page | 6 

 
 
 
 
 requested  consent  to  contact  Mr  Duncan’s  named  person/carer  as  discussed 

above.  

4. The Trust would like to reiterate that we are sorry there was no review of the care 
and treatment provided by the KCH Liaison Psychiatry team available at inquest. A 
Mortality Review had been completed, and the plan had been to carry out an AAR. 
Governance  arrangements  for  AARs  that  involve  multiple  directorates  are  under 
review  by  the  Trust’s  Patient  Safety  Committee.  The  Consultant  Psychiatrist  for 
Liaison Psychiatry provided a witness statement for the inquest, but this was not an 
investigatory review of this incident.  

Since Mr Duncan’s death the Trust has introduced several new systems to address 

the  challenges raised  by patients  presenting to  ED.  The  Liaison  Psychiatry  team 
carries out multiple clinical safety huddles in ED each day. These are brief, daily, 
multi-disciplinary  team  meetings  to  quickly  review  patient  safety,  share  urgent 
information, identify risks (like high-risk behaviour or medication issues), plan care, 
improve  teamwork,  and  resolve problems. This follows the  team  handover  which 
facilitates rapid risk review and shared decision making between ED and Liaison 
Psychiatry teams.  

The AAR recognises the difficulties of providing optimum mental health care in the 
ED  environment  and  makes  a  recommendation  around  the  development  and 
opening  of  a  dedicated  mental health  urgent  &  emergency  care  (UEC)  centre  at 
The Maudsley Hospital. This had already been planned for June 2026. The model 
is  designed  around  time,  space,  privacy,  and  dignity,  permitting  clinicians  to 
formulate  care  and  treatment  plans  collaboratively  with  patients  and  carers.  A 
growing evidence base supports the model. The  available evidence shows these 
deliver improved patient and staff experience; reduce 12-hour breaches in ED; and 
alleviate wider ED activity. SLAM has therefore developed this service with support 
from South East London Integrated Care Board (SEL ICB). 

KCH  has  also  launched  a  new  ED  Low  Intensity  Area  (LIA)  in  partnership  with 
SLAM.  The  LIA  space  offers  a  calm  and  supportive  environment  for  suitable 
patients  who  would  otherwise  wait  in  the  busy  environment  of  the  main  ED. 
Operating 24/7, it currently has capacity for six patients. The LIA is a continuum of 
the ED, but patients are kept in a less stimulating environment. Patients who are 
moved  into  LIA  have  already  been  assessed  and  have  a  plan  in  place,  but  they 
need to wait to have it enacted. These plans may include referral for a psychiatric 
admission, referral to an associated team such as the homeless team or addictions 
care  team,  with  ongoing  care  planning  following  the  assessment,  or  referral  to 
Recovery House in Lewisham, where they can be offered  a maximum of 7 nights 
stay,  as  an  alternative  to  admission  for  people  who  feel  unable  to  return  home. 
However, the Recovery House is not suitable for people who are homeless so would 
not have been an option for Mr Duncan.  

Page | 7 

 
 
 
 
 
 
 5. It is our sincere hope that this response letter and the accompanying AAR provide 
sufficient reassurance that the Trust has taken appropriate action to address your 
concerns in order to prevent future deaths. The Trust is committed to improving its 
systems and processes in the interests of patient safety wherever possible and we 
are grateful for this opportunity to reflect on the service we deliver and how it can 
be improved for patients like Mr Duncan to prevent such a tragic event occurring 
again.   

Yours sincerely 

Chief Nurse   

Chief Medical Officer 

Page | 8