Data Security Protocol

Why is data security important? Compliance with Institutional Review Board (IRB) guidelines An IRB is a group designated by an institution to approve, monitor, and review research involving human subjects to assure appropriate steps are taken to protect the rights and welfare of those subjects. It is a federally registered body. Non-compliance can jeopardize: Funding Research progress Organization’s reputation This protocol aims to follow Harvard’s guidelines for security of personally identifiable data in research http://www.security.harvard.edu/research-data-security-policy Protection of human subjects Field projects often collect personally identifiable information (PII) from respondents PII + other sensitive information (e.g., financial or medical data) = RISK PII

Overall principles for data security Use Cold-room computers, passwords and encryption: PII should only be viewed on cold-room computers that are password-protected and are equipped with TrueCrypt Pick strong passwords for files and computers. Rule of thumb: more than 10 characters, alpha, numeric, caps and non- caps, and symbols should be included (all). No dictionary words. Share verbally and keep record of passwords in a secure location. Ensure physical security: Keep data in a physically secure location Store, transmit, and use PII separately as much as possible: Separate personally identifiable information from the dataset as soon as possible (while maintaining respondent id link). Store and transmit PII separately from rest of data and use only de-identified data for analysis as much as possible. Obtain confidentiality agreements: Confidentiality agreements should be signed and kept on record for anyone who handles PII (surveyors, data entry operations, project staff) Data entry operations Many want to use a mnemonic device that will made passwords easier to remember.

Data security for new projects: Stage 0 Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage and transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public All Research Assistants/Associates and anyone else who will have access to data with PII should: Take the course (Citi or NIH) on human subjects research and send the certificate of completion to your IRB coordinator Read JPAL/IPA human subjects manual and Data security checklist Read the IRB requirements for the project Protect data on computers: Use cold room computer with Password protection and TrueCrypt Use secure file transfer and encryption for sending PII

Data security for new projects: Stage 1a Before data collection Stage 1: Data protection in the field Stage 2: Secure data transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public Rest of survey Unique ID PII and Consent Unique ID PII and Consent Unique ID Structure the physical survey packet into the “PII-Consent section” and the “Questionnaire section”, so they can be separated Ensure that you have a field for the Unique ID Code on every page of the survey packet. It is CRITICAL that each page of the survey has the CORRECT unique ID code so that you can match up the questionnaire to PII if it is necessary later Ensure you have a secure location to keep hard copies of surveys, with the identifying information separate from the rest of the survey Consider pre-printing all the surveys with the Unique ID Code on each page to avoid risking mistakes by surveyors Examples of insecure locations: cardboard boxes on the floor of the office (vulnerable to pests, spills, theft) Examples of secure locations: Locked metal file cabinet that only the research assistants and project manager have access to

Data security for new projects: Stage 1b Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage and transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public PII Survey Paper surveys received from surveyors should be physically separated into PII-Consent section and the rest of the questionnaire. These two sections should be stored and transported separately Ensure that data entry operators have signed a Confidentiality Agreement Once data has been double-entered, receive datasets on disc (NOT email). PII and rest of data should be stored in separate discs. Confirm that data entry operators have removed the data from their computers

Data security for new projects: Stage 2 Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage and transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public Transfer data from data entry to disc to password protected cold room computer and encrypt immediately Make 3-5 encrypted copies of the original data and store on at least 2 secured servers or computers Send encrypted data through a secure file transfer protocol (SFTP) such as Accellion (HKS) or WinSCP (NBER) Sending data containing PII over email or Dropbox needs to be avoided

Data security for new projects: Stage 3 Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage and transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public Data analysis does NOT require PII (e.g. no need for names, addresses, etc in analysis) Data analysis does NOT require PII (e.g. no need for names, addresses, etc in analysis) Data analysis does NOT require PII (e.g. no need for names, addresses, etc in analysis) Maintain two separate datasets: first which contains PII and the unique id code and a second which contains the unique id code and the rest of the data (make sure both contain the respondent id code) Keep the dataset containing personally identifiable information encrypted Decrypt and download only the second dataset (the one without personally identifiable information) for cleaning and analysis onto your computer If you need to view the PII, then you should use a cold room computer.

Data security for new projects: Stage 3 Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage and transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public Data analysis DOES require PII Download the encrypted file onto a password-protected USB key or other storage device. Transfer the file in encrypted form to a password-protected cold room computer As long as the data you are working with directly uses PII, you will need to work on a cold-room computer that is password-protected. You may not transfer the data containing PII to other computers. There may be ways to de-identify the data and retain the elements needed for analysis, giving you more flexibility on where you clean and analyze data.

Data security for new projects: Stage 4 Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage and transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public Once data analysis is finished, hardcopies of surveys need to be destroyed in a secure manner (e.g., shredded) within 5 years of completion of the study Once all data is received for cleaning and analysis and secure back-up of the files has been confirmed, completely delete the file from any field computers (make sure all data has been transmitted from the field before deleting files) You may consider ‘wiping’ your hard drive of these files using a program such as Eraser (http://eraser.heidi.ie/)

Data security for new projects: Stage 5 Before data collection Stage 1: Data protection in the field Stage 2: Secure data storage transmission Stage 3: Environment for analysis Stage 4: Field wrap-up Stage 5: Making data public Multiple team members need to review the dataset before it is released publicly, preferably ones who are familiar with the survey instruments and data collection The potential negative repercussions of making on mistake and releasing PII on a public database can be huge (imagine leaving a social security number in a public medical procedures database) Always get PI approval before making data public

Data security for existing projects People: Ensure requirements are met for all team members who have access to PII: Read IRB requirements for the project Certification of completion for the IRB training course is on file Protect data on computers with passwords Sign Confidentiality agreements Digital data: Take inventory of all digital data in the project. For the files that contain PII: Separate PII from non-PII data Encrypt datasets with PII Assess if PII is needed for analysis and if so, use cold room computer Hardcopies Ensure that hardcopies are stored in an appropriate and secure place. Once analysis is finished, check with PI to get permission to destroy hardcopies (within 5 years) Using a commercial shredding machine or giving the hardcopies to a reputable office services company Scans Scans of hardcopy surveys should follow the same protocol as Digital Data Scan first page separately from the rest of the survey makes running do-files with hardcoded file paths harder to run but for most projects this should not be an issue since PII is not typically used in analysis

Sample Confidentiality Agreement   As a member of the research team for the Center for Microfinance (CMF),I understand that I may have access to confidential information about individuals participating in surveys conducted by CMF or partner banks, NGOs and institutions. By signing this statement, I am indicating my understanding of my responsibilities to maintain confidentiality and agree to the following: I understand that all information about study participants obtained or accessed by me in the course of my work is confidential. I agree not to divulge, publish, or otherwise make known to unauthorized persons or to the public any information obtained in the course of data collection or data processing that could identify the persons who participated in the study, unless specifically authorized to do so by office protocol or by a supervisor acting in response to applicable law or court order, or public health or clinical need.

Sample Confidentiality Agreement   I understand that I am not to read information or records concerning study participants, or any other confidential documents, nor ask questions of study participants for my own personal information but only to the extent and for the purpose of performing my assigned duties as a staff member, volunteer or employee of CMF. I agree to notify my supervisor immediately should I become aware of an actual breach of confidentiality or a situation which could potentially result in a breach, whether this be on my part or on the part of another person. I agree to return all data in my possession to my supervisor upon terminating work with CMF or upon being requested by a supervisor to do so and I understand that failure to do so may result in legal action. I understand that a breach of confidentiality may be grounds for disciplinary action, and may include termination of employment. Name: ________________________ Signature: ________________________ Date of Signature: ________________________

True Crypt walk-through True Crypt = Box created on your computer used to hide (encrypt) files You can: Send these “boxes” like a normal file Disguise them to look like something else You have to go through True Crypt to both put things inside the box (encrypt) and take things out (de-encrypt)

Encryption and un-encryption in ideal world Cold room computer Encryption and un-encryption in ideal world Networked computer Password- Protected USB Encrypted PII Un-encrypted SFTP Does not need PII in analysis PII stays encrypted Rest of data unencrypted Rest of data Unencrypt PII PII SFTP Needs PII in analysis Rest of data Unencrypt Rest of data

Data Security Checklist   All project staff have take IRB course and sent certifications Survey structured with PII-Consent detachable from Main Questionnaire Field staff sign a confidentiality agreement before working with data/surveys Using IRB approved consent form Unique ID code written on every page PII-Consent separated from Main Questionnaire prior to data entry Hard copies stored in a secure location Only using cold room computer for management and analysis of PII data Make 3-5 backup copies (encrypted) of the original data Transfer encrypted files using file transfer system Store backup copies on a secured server Confirm data entry operators have removed data from their computers Destroy hard copies and PII within 5 years of end of project