SCORM data output formatting

I'm trying to work with SCORM 2004 data coming out of my company's LMS (which uses SCORM Cloud as a backend), and I'm struggling a bit with the formatting of the questions. Specifically, the Likert and Pick Many questions seem to have replaced spaces and other punctuation in the sub-questions and responses with underscores, like this:

Discuss_highlights_of_the_Massachusetts_law_related_to_Alzheimer_s_disease_and_related_dementias__including_the_continuing_education_requirement_for_clinicians_who_serve_adult_populations.___Strongly_Agree___Describe_guidelines_for_the_diagnosis_and_treatment_of_Alzheimer_s___related_dementia.___Strongly_Agree___Identify_care_and_management_best_practices_for_patients_with_Alzheimer_s___related_dementias_in_the_healthcare_settings.___Strongly_Agree

My real issue is that 3 consecutive underscores (___) are being used as a delimiter between the sub-question and the learner response. But due to punctuation used in some of the questions, strings like Alzheimer's - related dementia become Alzheimer_s___related_dementia, which is indistinguishable from the already convoluted delimited value pairs.

There are several other occurrences in this course, including Define & review and medical & nursing. And I have hundreds of other courses already published to the LMS that will likely have similar issues.

I've attached the SCORM debug file for further review.

I need to pull this data and put it into some kind of readable format for evaluation reporting, and these issues are really slowing me down. Aside from the obvious meticulous massaging of the data, any suggestions on how I can improve this situation? Is this something Storyline is doing or is it SCORM Cloud?

Thanks for any info you can provide.