medical research that requires human analysis

[notned]

hi

i'm not a biochemist or molecular biologist, but i'm interested in something for a research project ...

 

could anyone give me any examples of areas of medical research that require analysis of the results by a human, rather than by a computer?

 

cheers

alex

[CaptainPanic]

I think that an expert eye is better than a computer at judging an X-ray picture to see if there is a bone-fracture or not.

 

But I am no expert.

[Arete]

The question is a bit vague, as any calculation a computer can do, a human can too. It's simply easier/more practical to do most statistics using a computer - I can do a basic linear regression with thousands of data points on a calculator, or I can do it with a single line of R code. On the other hand, any statistical analysis requires a human component to actually answer a hypothesis - you can do a whole bunch of complicated, but either assumptively incorrect or meaningless stats without a directive from a human.

 

A bigger concern for a research project involving medical research would be HIPAA (http://www.hhs.gov/ocr/privacy/) and internal ethics compliance. Any actual data collection, or bench-work relevant to human patients is going to require clearance. Unless you have an adviser who already has clearance, or data mining publicly available databases (ergo - needing to use a computer) you'll need ethics approval, which can take an impractical length of time for an undergraduate (or even some graduate) projects.

[notned]

hi

 

i suppose what i'm trying to say is: "is there any research results analysis that can ONLY be done by a human?"

 

cheers

alex

[Arete]

All of it is done by humans - a computer is simply a tool that humans use to analyze data. At least in my field (evolutionary biology/genetics/genomics) there's few practical ways left to analyze data that entirely eliminate computers. They're comparatively cheap (compared to say, a slab gel setup) and extremely effective.

 

Is there a motivation behind not wanting to use a computer at all? Presumably you have access to one - as you are using one now...

[notned]

haha, yes i have a computer ... i can't say what this is for unfortunately.

 

anyway, i understand that computers are tools etc, but i would have thought that in analysing results of research tests would initially be done by a computer (recognising patterns or counting cells etc) and then some of the samples validated by a human.

 

my question would be ... is there anything that can only be analysed by humans?

 

alternatively, could you give me five examples of analysis done by humans? i'd imagine these to be something like "count the number of stained cells in a slide" etc

 

cheers

alex

[Arete]

alternatively, could you give me five examples of analysis done by humans? i'd imagine these to be something like "count the number of stained cells in a slide" etc

 

 

Conducting counts or taking measurements would typically be considered to be data collection rather than analysis. I.e. making the counts from the slides would be data collection, determining that the count from treatment A was significantly different from the count from treatment B would be data analysis.

 

Plenty of measurements can only be taken by a human - i.e. taking/recording patient blood pressure or body temperature, checking bands on an electrophoresis gel, scoring genetic fragment analysis traces, meristic counts, colony counts, etc. but they would not typically be considered "analysis" as such.

[overtone]

Getting a computer to perform actual analysis is a pretty difficult matter - postulating mechanisms to explain newly discovered correlations, for one, is beyond the state of the software art in most medical fields.

 

For a ready supply of examples I would look at research into mental illness. The recent discovery of a single genetic factor underlying several apparently quite different mental disorders, for instance, was found and made visible with a combination of human organization and computerized statistical breakdown, but the computers have been no help in the subsequent analytical discussion.

 

Specifying research that can only be analyzed without a computer is a bit odd, anyway - like specifying construction projects that can only be accomplished without a hammer. They can be found, probably, but why the search?