I've been talking to Mazi about some work. Namely changing the schema for lime_survey_sid, the table that stores saved answers, and is the only thing limiting questions to 1000.

In our situation, say we have a list of 1500 patients, with data, this go to may be 5 Doctors. next to the data is a set of questions:
does the patient smoke, on meds?, specific labwork value, you get the idea

representing this in limesurvey via a ruby tool to merge the csv list of patients and data with the survey is easy (and done)

however, limesurvey represents each thing as question data.. so let us say we have 5 questions per patient, and all of the data (text) is a 6th boiler plate question.

this leaves us with the ability to only gather information on 166 patients..

the reality is that we routinely work with lists of 1000 patients with perhaps 10 questions.

so table lime_survey_sid just doesnt work for us (though everything else does, with some ruby glue)

if the schema were changed to rows, with columns of:
id
submitdate
startlanguage
token
datestamp
startdate
saved <- new field indicating question sidXqidX....
saveddata <- the value of that question

then we have a table that is row based and essentially has the same functionality..

thoughts?

Hi Wrobym,

well.. you can change that but be alerted that this is a huge change in the engine. Essentially you will have to re-code half of LimeSurvey so I would not recommend that. At the moment everything is based the current way to store data. Also you have the problem that evaluating the data will be a pain.

I would suggest that you are using the token system and generate a token for each patient. Then have the doctor use these tokens to answer each questions for each patient.

For the change between each single survey you would have to customize LimeSurvey - especially the end url, so that after a survey is finished it automatically jumps to the next survey for that doctor.

You will need to change code in LimeSurvey but that is doable and affordable.

Or you can wait for LS2, depending how urgent it is.

As discussed before toucing the table structure isn't a very good approach because this affects the whole system.

Try to think of other solutions. You can easily copy a survey by using export/import. Why not setting up a survey for each doctor or something like this?

c_schmitz,
Thanks for the reply. Does LS2 database schema change so there are no more column based dependencies?

Survey to survey jump isn't really feasible as every patient needs to be seen on the same screen at the same time.

Hey Mazi,

unfortunately our constraints are:
the survey must contain every patient on screen at the same time (scrolling down is fine) so that the MD or nurse can go through everyone at one time then submit.

so again it is a list of patients with questions (on the same row) about that patient.. I have it working nicely in LS using importing and exporting.. Just hit the wall with the database schema..

I unfortunately don't see any other way around it except waiting for LS2..

This is the LS2 database schema:


For more specific questions about LS2 I recommend to ask in the LS2 subforum.

Did you try to use another database like Postgres?