[SystemSafety] Computer Systems and the Law

[Prof. Dr. Peter Bernard Ladkin]

Five years ago, Bev, Martyn, Harold Thimbleby and I published a study of the "common law 
presumption" that computers are reliable unless there is evidence to the contrary. 
https://journals.sas.ac.uk/deeslr/article/view/5143/

Harold just sent me a photo taken in Mike Parson's talk to the SSS'25 yesterday, in which he 
referred to the work.

Let me call the "common law presumption" the Presumption. The Presumption  replaced PACE 1984 S. 69, 
Evidence from computer records, https://www.legislation.gov.uk/ukpga/1984/60/section/69/enacted when 
this was repealed in 1999.

Barrister Paul Marshall was asked by then-Minister Alex Chalk back in 2020 for a suggestion for how 
computer evidence should be handled in court. Paul consulted us and others. We sent Minister Chalk 
our suggestions, and in 2021 published them https://journals.sas.ac.uk/deeslr/article/view/5240 as 
well as a shorter Briefing Note in 2022 https://journals.sas.ac.uk/deeslr/article/view/5476 However, 
when the then-government was asked in Parliament about re-legislating, they replied that there were 
no plans to do so.

The current government appears to be thinking differently. As Derek recently noted, the UK MoJ have 
just issued a Call for Evidence on it, indicating they are seriously considering legislating 
differently.

The reason for all of this is the Horizon scandal. Post Office prosecutors made heavy use of the 
Presumption, as well as not disclosing documentation (such as logs and records of known errors) to 
defendants. For some years now, it appears there are divergent views on whether the Presumption was 
the main reason for the scandal (as the Court of Appeal ruled, prosecutions bringing the legal 
system into disrepute) or the lack of disclosure. I believe that handling the Presumption is a 
different issue from dealing with lack of disclosure, and argued so at 
https://abnormaldistribution.org/index.php/2025/01/22/addressing-the-common-law-presumption-of-computer-reliability-is-a-separate-issue-from-addressing-failure-of-disclosure/ 


I've recently started up my blog again after a makeover. There are three recent posts I'd like to 
mention.

Retired barrister Stephen Mason, one of the co-editors of the practitioner's handbook on electronic 
evidence, has compiled an essential reading list and has kindly allowed (encouraged?) me to publish 
it. You can read it at 
https://abnormaldistribution.org/index.php/2025/01/23/horizon-articles-in-the-deeslr-journal-and-relevant-handbook-material/ 


Finally, many people here may not be aware that "reliable system" is since 2023 a technical term in 
the law of England and Wales. With a meaning which is not identical to the technical meaning of 
"reliable system". Uh, oh. 
https://abnormaldistribution.org/index.php/2025/02/04/reliable-system-as-a-legal-term-and-as-a-technical-term/ 


This business, of presuming computers and their associated actions are operating reliably unless 
there is evidence that they are not, is not confined to English law. Here is an example from German 
legal proceedings, in Bielefeld.

I have been advising a lawyer and his client in Bielefeld, an assistant doctor accused by the state 
prosecutor of "negligence leading to death" ("fahrlässige Tötung") of a patient in 2020. The doctor, 
let me call her Doctor B, allegedly entered wrong information on dosage about medication into an 
electronic patient medication record (EPR). Whatever she entered was reentered manually (by another 
assistant doctor, also accused, call him Doctor I) and this record transferred to another hospital 
to which the patient had been transferred for an orthopaedic operation. Six different doctors (at 
least) "saw" that record before the medication was started (four days after admission, and two or 
three days after transfer) and no one noticed it was wrong. The patient developed sepsis, and died a 
few days later. The overdose of this medication for rheumatism, MTX, contributed. The doctor was 
charged in 2022, I believe. The initial court date was in late 2024. I read about it in the local 
newspaper, contacted the law firm representing the three doctors charged, and Doctor B's lawyer got 
back to me. He asked me for a list of points about the computer involvement which needed to be 
clarified. Doctor B tried to clarify them with her local SysAdmin (she is still working at the same 
hospital) and wasn't able to get very far.

I was informed that, at the second court date in January 2025, an IT specialist gave a witness 
statement in which he said there was nothing wrong with the computers.

On the contrary, I thought, there is a lot about the computer software architecture and connections 
and accuracy of data transfer as well as the usability of the software which need to be clarified 
before anyone can say that Doctor B entered the wrong information negligently.

I felt I actually needed to be present and hear what was said. The third court date was this last 
Monday (when I was supposed to be on the train to SSS'25). I attended court.

The charges against one of the three accused had already been dismissed at the second session in 
January against a "Geldauflage". There is a mechanism in German law whereby charges can be dismissed 
against a payment, Geldauflage, which covers minimal court costs, and the majority of which goes to 
other needy institutions specified by the judge. It is not a fine; the charges are dismissed. It is 
more like a compulsory charitable contribution. It is not trivial. Typical, I understand, is 
three-months' salary.

 From what I was able to ascertain, the prosecutor had no information by means of which he could 
prove (to my mind, obviously not to his), in any shape or form, that Doctor B entered the wrong 
information.

The head of a rheumatism clinic gave expert evidence. He said "everyone knows" that MTX should only 
be given once weekly, but there is no way you can directly input this information to the EPR 
software being used. You have to use a workaround (to put it politely) which involves writing such 
information in a Note. Doctor I was asked if he transcribed any information in such a Note and he 
said no. So whether or not Doctor B input the right information or not, assuming what Doctor I said 
is correct, the wrong information followed Doctor I's transcription to the second hospital, where 
the MTX was given at the wrong dosage and the patient died, partly as a consequence. As far as I 
could see, the only way you could find out what Doctor B actually input is by going into the system 
archives at the time she input the data and before Doctor I had transcribed them.

If people had been able to access system archives, they'd surely have done so. There wouldn't be a 
prosecutor saying "you put in the wrong information" and a defendant saying " no I didn't". In the 
absence of those archives, with Doctor I's admission that he didn't transcribe any Note, I don't see 
any way at all of resolving prosecutor's accusation and defendant's denial.

This wasn't even discussed.

The proposal was made by the defence to dismiss the charges against a Geldauflage. Judge and 
prosecutor agreed. That is what happened. Charges were dismissed. Thereby was the second witness 
(also a medical doctor) discharged without giving evidence.

There is more to this, and I am writing it up. What I have written above is consonant with what 
appeared in the local newspaper yesterday, with my additional comments about the apparent lack of 
any interest in the computer-evidential aspects. When I've written it up, I'll ask Doctor B if she 
is OK with it. I found her an exceptionally sympathetic person.

Mike Parsons made the point last year in April, while discussing the Horizon scandal, that 
computers/computerisation can hurt people even if the systems are not viewed as "safety related". 
When I wrote up my Monday's experience yesterday and sent it around to a small group of colleagues, 
one replied that his daughter, who is an anaesthetist, had observed to a professional meeting that 
now that records are electronic nobody gives them a quick look-over any more. Whereas with paper 
there used to be a summary sheet on top that would at least be glanced at. Similarly, in the 
Bielefeld case there was an electronic record that nominally went through six doctors' "oversight". 
I doubt this will be the last time this happens.

PBL

Prof. Dr. Peter Bernard Ladkin
Causalis Limited/Causalis IngenieurGmbH, Bielefeld, Germany
Tel: +49 (0)521 3 29 31 00