2

I'm a highschool student conducting a small research on the contributions of computational modelling to medical advancement. I've found examples relating to its uses in tracking diseases, drug testing and implant design development for the introduction part. For the discussions, I'm searching for a suitable case study to conduct to support my claim. Any good examples that you would reccomend? I would also like to learn about some limitations of computational modelling. Are there cases where it could have been applied/cannot be applied, where it could potentially help? I would really appreciate it if you could cite some sources while answering.

3
  • 2
    This isn't the best place for your question. In fact, stackexchange (SE) in general isn't the best. Questions that ask for a list of ideas are better suited for discussion forums such as Quora and Reddit. SE is a Q&A site, not a discussion forum; it's better suited to questions with definitive answers. You can try Computational Sciences (scicomp.stackexchange.com), but their help section suggests your question won't be a good fit there either.
    – Carey Gregory
    May 18 at 3:58
  • 1
    However, instead of closing your question I'll give it a chance and convert it to a community wiki, which means the rules are a bit relaxed on what can be asked and answered. But I think you'll be better off asking in the other places I suggested. And surely there are computational sciences discussion forums somewhere? You should search for those in particular. I just don't foresee your question generating a lot of useful answers here because it's a very niche subject.
    – Carey Gregory
    May 18 at 4:00
  • I understand. Thanks a lot! May 18 at 5:01

1 Answer 1

4

I recommend looking into protein folding.

Since the 1960s, computer modelling has been used to (attempt to) predict the structure of proteins and the steps they go through to get there. Low-resolution simulations have tracked the progress of simple proteins, but because of the very large number of degrees of freedom of movement within a polypeptide chain, large molecules have barely been studied even with the world's most powerful purpose-built computers.

This research is has a direct link or will lead to serious advantages in many fields: fighting cancer, immunology, fighting amyloid conditions causing degeneracy of brain function, the production of artificially engineered organisms for the production of medications and chemicals etc..



fMRI processing.

This is still a relatively new science despite having been around de-facto since the 1930s when differential magnetic properties for blood and tissues were observed, the first map of blood-flow being observed in the 1990s. There are still immense challenges in refining the techniques, the machinery - and the software attributes that produce the best results in a given circumstance. Requires a degree in physics/medical-physics/mathematics and talent with code to work in the field.



CT-scans.

Winning a Nobel Prize for its pioneers in 1979, this is well established and the software is refined and capable. There are in addition, many different modes of data-visualisation which are increasing in sophistication.



Patient information processing and retrieval.

This may seem trivial on the surface, but the efficiency and interconnectivity of such systems between general practitioners, hospitals and pharmacies has been a great factor in the increase of consistency in certain areas of treatment. Though many hospitals use the chart system that was in use before computers arrived, many of the corporate hospitals use the computer system for prescription of medicines and payment of the bills as well as electronic filing of the patient details.

The electronic record system helps in the case of insurance claims and also when the physicians need to refer to the previous history of the patient. The electronic records are safer and last for a longer time than the regular records maintained on paper. They can also be subject to statistical analysis, and regular monitoring for signs of possible widespread contaminants/disease-spread patterns.

And perhaps an increasingly critical part and ever needing to be reassessed and adjusted to new conditions and threats: computer security and threat analysis.



Computational-Epidemiology.

Varies in sophistication, monitors patterns of keywords on social media and the like for alert-signs of likely disease spread. Projections of spread from this and hospital/physician reports are processed to produce predictions of spread and course of the likely societal-level system effects.

Machine-learning and AI are just starting to be used in the marketplace.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.