Vol. 8: Opioid deaths and hospital admissions

We need inclusion health teams working across 'islands' of care.

Hello and welcome back to Antidotum.

This paper is a sorry indictment of how people who use drugs are served by the current system. We should feel angry. Then we should take that anger and turn it into determined, purposeful action to create some change. 

There is much hand-wringing about the appalling drug-related deaths in the UK and this is the type of paper that offer plenty of insight into the multiple interwoven factors that need to be addressed. Certainly, when it comes to providing healthcare — especially related to physical health problems that result in acute admissions — we need a wee bit more than we’re currently offering. 

Yes, we’ve moved on in some ways. In others, not so much. Photo by Austrian National Library on Unsplash

Opioid-related deaths during hospital admissions or shortly after discharge in the United Kingdom: A thematic framework analysis of coroner reports. Dan Lewer, Thomas D Brothers, Magdalena Harris, Kirsten Rock, Caroline S Copeland.  PLOS ONE. 2023 Apr 19;18(4):e0283549–e0283549 https://doi.org/10.1371/journal.pone.0283549

This study investigated deaths between 2010 and 2021 where an opioid was detected in the toxicology testing in the National Programme on Substance Abuse Deaths database which holds coroner reports for deaths following psychoactive drug use. (‘National’ in this case meaning England, Wales, and Northern Ireland.) The team threw in another factor: the deaths had to be during a hospital admission or within 14 days of discharge from a hospital. They included medical (78.5%) and psychiatric admissions (18.2%).  

Overall, they found reports on 121 deaths. Out for these, 42 people died during the admission and 79 died in the 2-week window after discharge. Other key numbers were that the median age at death was 40 years; 73% were male; and other sedatives were detected in 73% of cases. These set the scene for the thematic analysis of the reports and the authors report on potential causes of fatal overdoses falling into three areas. The first was hospital policies and action. There were four themes within this area: concealing drug use; discharge against medical advice (DAMA); discharge into unsafe areas such as the street or hostels; and expectations of low quality care. The second was high-risk use of sedatives where there was evidence of increased use of sedatives during acute illness and loss of tolerative to opioids. And the third was declining health where this reduced access to treatment for substance use (for instance due to mobility problems); and also acute exacerbations of long-term conditions. 

Reflections: 

Firstly, let’s deal with the limitations. All studies have them, in some shape or form, and we have to give them due consideration. As the team drily note “NPSAD is an incomplete database” and this happens on at least a couple of levels.¹ Some coroners don’t submit reports and those that do are clearly limited in the detail provided. This means that an awful lot of deaths and details are missing — given a quantitative study of post-hospital opioid-related deaths gave a number of 1088 deaths then this database may represent a sample of just 7%.² 

I’m going to concentrate on a couple of linked points related to hospitals, discharges, and the interaction with the community. 

Ensuring access to adequate OAT.

The authors recommend in the practice and policy section that attention needs to be paid to “providing sufficient OAT [opioid agonist therapy] during periods of acute illness”.

This is critical. (There’s an argument, but I won’t go there, that we need to do more to ensure we are proving sufficient OAT for everyone but that’s another story.) It has been identified as a factor in discharge against medical advice and, as the authors highlight, it’s going to be part of the picture in covert use while people are in hospital. There can be a loss of tolerance during a long admission. Many community drug teams up and down the country will take calls from their local hospital to give advice — there is certainly is a gap in knowledge but we have to be clear that education on its own is not enough. Of course, it seems obvious that it is important all healthcare staff have a basic grasp of the principle and practice of OAT. 

However, low level awareness doesn’t come within a hundred miles of adequate when it comes to tailoring clinical care in someone who is unwell enough to be admitted acutely. Clinical care in people dependent on opioids is complex and I’d argue that it needs some specialist clinical input, not just a foundation year doctor who can, on a good day, spell buprenorphine. We need to find a way to build in that clinical expertise where needed and I think it needs to be available directly and immediately to people admitted to hospital. Given the difficulties across the sector in recruiting and retaining clinicians this is no small matter. 

Improving discharge planning

After discharge, patients may have reduced mobility and be unable to attend community pharmacies or OAT clinics, and therefore use illicit drug dealers who are willing to visit clients’ homes. Some cases in our study were unable to access OAT due to illness and died after using illicit drugs. In these cases, hospital staff may need to work with community drug services to arrange home visits and delivery of medication. 

As a GP, I am certainly prepared to say, with some confidence, that discharge planning is regularly appalling. It’s not just people who use drugs. Discharges from hospital are often “sudden and unsupported”, in the authors’ phrase, and it’s a function of an NHS system that is being run hot, more or less at breaking point. A crush of patients arrive on a medical assessment unit and there is a sudden need to free up beds. This routinely happens on Fridays to anticipate the admissions received at weekends. In most cases, family and social contacts rally around, and the GP practice may get involved to sort scripts and tidy up medical concerns. That’s not necessarily an option for PWUD who have weaker social ties and are sleeping on the streets…

The quote above from the paper does, to my reading, slightly infer that the responsibility lies with the hospital. But what about the community drug services here? What are the policies of the community service around hospital admissions and discharges? Do they have any provision to help people for Friday evening or weekend discharges? What, if any, in-reach is happening? 

Discharges shouldn’t be perceived as ferrying a person from one island of care to another island. 

People who use drugs face multiple hits when it comes to discharges. They don’t attend their GP in any case and it is not routine for drug treatment services to be notified of discharges in the same way as the GP will be (though I suspect this is hugely variable across the country). 

Unfortunately, in a large number of cases there is no discharge planning as people ‘discharge against medical advice’. (There’s an argument the language needs to change around this phrase as it puts the blame squarely on the patient — something like ‘failed admission’ might be a better term.) A qualitative study in the US found several main themes for why someone might DAMA. These were: inadequate management of withdrawal, inadequate pain management, boredom/social isolation, and stigma. 

I am prepared to wager that providing extra education for staff at hospitals won’t be enough — its real benefit could be in catalysing some change. We also need to address some of the systemic difficulties and I do have a suggestion. It’s not a new idea but I find it hard to see how we will genuinely improve care without it — we need an integrated approach that does the education but also provides the clinical expertise and social support at the point of need.  

The inclusion health team

I’d like to see the development of inclusion health teams. These will need inclusion health clinicians and they should work across hospital units and communities — some places have hospital in-reach but we need a system that is much more coherent. This will present a challenge to commissioners as they will need to be strong links between the NHS and the non-NHS providers but we can’t continue with the ‘islands of care’ mentality. 

If you’ve not worked in a hospital it is important to convey just how fragmented care can be within the hospital, never mind the challenges of working with community agencies. Hospitals are highly siloed with specialised teams working in their own niches — there are few generalists now and even fewer staff who work across the boundaries. There is an argument that an inclusion health clinician is another specialist but I think their work can be framed in a generalist way that brings together the clinical and social needs of people with very complex needs. 

An inclusion health team could work with people who use drugs, people who use alcohol (many hospitals already have alcohol teams), the homeless, and several other vulnerable groups. As part of the team there should be people with lived experience and part of the strategy should be to address the social isolation and stigma that results in inadequate care.

One might think that this is exactly what you’d expect an Integrated Care System to be considering… it’s early days for them yet but it’s this kind of wicked problem, across multiple sectors, that needs strategic commissioning to address.

A final thought: a key metric of success locally would be the discharge against medical advice rate. Yes, it’s a proxy marker and there should certainly be consideration given to harder endpoint metrics such as mortality, length of stays, and re-admission rates. But the DAMA rate in people who use drugs is utterly damning and a complete scandal. 

1. Is there anything uglier, in data science at least, than an “incomplete database”? That’s the first action point from this study — we need better data. 

2. This paper was in PLOS Medicine: https://doi.org/10.1371/journal.pmed.1003759