How to deal with patient anonymity: Identified, quasi de-identified, & anonymous surveys

 

Participant Identification and Anonymity

Defining the level of participant identification and anonymity for surveys is one of the most critical parts of a survey project. It is also frequently overlooked AND improperly used! Poorly designed surveys are hard to fix once a project is started!

Interacting with Participants

What factors are critical to participant interaction?

Frequency of contact

  • Once

    • Identification (ID) may or may not be needed

  • Repeated

    • ID is needed

Anonymity

  • Anonymous: Participants AND their responses anonymous

  • Quasi Anonymous: Participants identified but their responses anonymous

  • ID but Confidential: Participants identified but responses are confidential

  • ID, Confidential, limited access: Limit who on research team can see

I1.PNG

Simple Public Survey

Publicly available: A link you publicize to a wide audience to get general answers

  • Example: How was the food at the hospital cafeteria on taco Tuesdays?

    • Characteristics-

      • Anonymous

      • Contact Once

    • Trade offs-

      • Flexible, click on a simple link to access

      • Not specific, can’t be used to track a single person over time

      • Limited study designs possible

I2.PNG

Simple REDCap Survey Project

  • To produce a publicly available survey only two steps are needed

    • Surveys must be enabled on the project setup page

    • The instrument of interest must also be enabled as a survey

I3.PNGI4.PNG
 

Once this is done the survey link can be accessed on the manage survey participants page. All responses will be anonymous in this case!

I5.PNG
 

Targeted Survey/Quasi Anonymous

You have an email list of a specified group that you are looking to gather anonymous feedback from

  • Example: Knowledge before and after a recent CPR training (Pre/Post)

    • Characteristics-

      • ID needed- Identifying list of emails for sending survey links

      • ID needed to link before and after responses

      • Participants RESPONSES need to be anonymous- Quasi anonymous (dependent on sample size and identifying questions asked of participants).

    • Trade offs-

      • Know who responded, but not what they responded

      • Can link responses over time, improved study design, delta vs average of sample.

i6.PNG
 

Quasi Anonymous Survey

i7.PNG

  • Participant List under Manage Survey Participants

  • Add participant emails

  • Email list can be copy and pasted, each entry separated by carriage return.

i8.PNG

  • Notice each email address has its own row with a link directly to that participants survey

    • Great for testing parts of your survey or opening a link to a single participant

  • Next notice you must choose the instrument and event

    • By default you are in the first survey, first event [initial survey]

i9.PNG

  • Now compose an invitation

  • Under info you see which survey you are in at what event

  • By default REDCap will select all emails in your participant list that have not yet received the first email

  • Feel free to use your own email to see how this works

i10.PNG

  • After your email has been composed and sent out you will be able to observe who has responded to the survey

  • Notice if you click on the responded green dot you will not be able to see an individual participants response! Quasi anonymous!

i11.PNG

  • Check out your data in “Data exports, reports, and stats”(under Applications)

  • Note there is no direct way to link the email to the actual participant responses. Data in the database exists without the email to identify response.

  • Email is only used in the background of REDCap (hidden) to track and link participant responses.

  • Keep in mind this principle does not hold up well with small sample sizes or studies where uniquely identifying information is collected.

    • i.e. if you only have a participant list of 10 and 3 people respond it will be much easier to identify respondents, compared to a sample of 100 where 50 people respond.

Full ID but Confidential

Randomized Control Trial: Patients enrolled in person and you need to follow up with surveys or future visits.

  • Characteristics
    • ID needed- Repeated patient contact
    • ID needed- Link before and after responses
    • Participants responses not anonymous but confidential
  • Tradeoffs 
    • Most flexibility with study design, tracking, buy in from participants
    • Careful planning required to manage a participants flow through data collection

I12.PNG

  • ID but confidential: Designate an email field for invitations

    • Email Field is created and associated with participants data

      • Example: Email is collected as part of contact information

    • Designate that email field to be used by REDCap to send emails and track patients’ responses.

      • Note the email field must be validated as an email field.

i13.PNGi14.PNG

  • Add a couple of fake records and enter email addresses (under “Add/Edit Records”

  • Now go to Manage Survey Participants

  • Now you can see the email addresses AND the study ID for the patient

  • Now when your participants respond, you can click on the green responded dot and view the participants responses as well

i15.PNG

  • Now when we go to “Data exports, reports, and stats” to view our data we will see the email field along with our participants’ responses and participant ID

i16.PNG

Surveys

  • Properly defining the level of anonymity for your participants and testing to ensure your project achieves that goal is the most critical part of using surveys!

  • Best ways to ensure this are deferring to your IRB for specifications on level of anonymity and testing well before moving to production.

  • Use your own email to test survey functionality thoroughly!