Module 5 Group Case Therac 25
Relevant facts about Therac-25
Therac-25 was a machine designed for cancer treatment.
Therac-25 was mostly relied on machine's software for control and safety.
There were six serious accidents involved.
There were 3 patients actually died directly because of Therac-25
Therac-25 marketed by Atomic Energy of Canada Limited (AECL).
The Therac-25 case has been described and analyzed extensively by Nancy Leveson
The number 25 is designed referred to its power, which was greater than its predecessors, the Therac-6 and the Therac-20
Stakeholders
AECL
CGR of France
Programmers who designed the software
Assurance Officers.
Alternatives
Patients undergoing treatment with Therac-25
A bug in the software could cause treatment settings to not be saved, thereby giving the wrong treatment to some patients
Alternative 1: More rigorous software testing and debugging
Pros
Who is harmed?
The AECL will have to spend more money on testing and debugging
Cancer patients awaiting treatment will have to wait longer to receive it
More work is involved for the programmer
Due to the cost of additional work on testing, each Therac unit will cost more for hospitals
Alternative 2: Release Therac-25 as is, with no further testing
Cons
There is still room for errors
Programmers may not want to be a part of it
Bad to already have it known the program has hurt people
May eventually still not work
Less chance for patients to be killed or injured due to software bugs
Reduces the amount of complaints and legal trouble
Reduces errors in the software
Allows a good base for the next programmer to improve upon
Who benefits?
Doctors benefit from operating on patients safely without violating the Hippocratic Oath to "do no harm"
The AECL would benefit from not being legally responsible for 6 cases of radiation overdoses
Patients using the device benefit from a device that does not harm them
Pros
Cons
Who benefits?
Who is harmed?
The AECL will have it's image and reputation harmed by the machine killing it's patients
Hospitals and doctors will fail in their duty to do no harm to patients
If the machine malfunctions patients could be given lethal doses of radiation
The programmer who designed it because they knew it could hurt people and still went through with the idea so they may not have a future job
The AECL will save money by not having to hire additional programmers
The science community as a whole because they can improve on the software to make it better
Patients receive treatment immediately without waiting on a testing phase for the Therac machine (assuming the machine is operating normally)
The programmer could be seen as a valuable member and could be asked to improve on the software with a team since one person was able to go so far, it would be seen as better with a team
Less people for the company to pay to improve on it
Treatments can be immediately
Still has a chance to help with the treatments
If the software works properly, people will be treated
Can cause a lot more legal trouble
Little to no business due to the knowledge that it can hurt people and potentially kill them
People will continue to be hurt
If the software works incorrectly then people could be hurt or killed
Alternative 1 or 2?
We chose Alternative 1 because it is the ethical choice when dealing in software that could harm people. Programmers following a code of ethics for software that deals with medical devices is almost as important as doctors following their own ethical oaths. Viewing the scenario with ethical theories, we would argue that the time, money, and effort spent on ensuring the Therac-25 software would not outweigh the cost of having the machine kill people. Seeing this through an act utilitarian lens, the less harm your software causes, the more happiness there will be.
Policy Changes that would prevent this dilemma
Let the public know they could be taking a risk using the software on themselves.
Have multiple programmers collaborate with one another to solve all of the issues with Therac-25.
Do not allow public to use software unless it is known to be full proof.
What if everyone took Alternative 1?
What if everyone took Alternative 2?
The world would be a much safer place with less unintentional software errors. Software may cost more to produce if everyone always chose option 1, but bugs would be minimized to great effect. This would be a net benefit for society.
People can get treatments immediately but they will risk their own safety because there is a flaw in the software that can give a giant dose of radiation that could harm or even kill them.