Critical Issues
in Fraud Analytics
I S A C A - 2 0 1 5
Presenter:
Charles Faircloth, JD, CIG Faircloth Fraud Consulting
Critical Issues in Fraud Analytics
Introduction
1) Factors that drive fraud 2) Current fraud risks
Data breach fraud
Mobile & connected device fraud
3) Fraud Analytics Limits
Parameters of presentation:
Specifically covers:
Short horizon factors
Important for:
CIOs CISOs
System administrators IT managers
Focus of this review:
Fraud & fraud analytics in healthcare Health care primary target of fraud Critical areas of security risk & liability:
Fraud by data breach
Mobile & connected device fraud
Has wide application - all types businesses & agencies Provides references/review frameworks:
Normal IT operations Fraud incidents
Note: This review is not legal advice or counsel. - consult with your attorney.
Part 1:
Factors that
Drive Fraud
Introduction: Fraud in General
Fraud earliest reported urban crimes
Code of Hammurabi - 1,754 BC
Fraud - is a hidden crime
Uses deception to steal funds/data Data breach fraud (worldwide):
$3.5 trillion plus - lost revenue
Healthcare fraud (US):
10% total costs - $30 billion plus
3 Factors Drive Fraud:
Opportunity
Pressure
Rationalization
FRAUD
Rationalization
Factor 1 -
Opportunity
Opportunities for fraud 2 categories:
Security weaknesses Position advantage
Security weaknesses:
More mobile devices/apps - More problems. Healthcare behind in:
Anti-fraud & HIPPA security standards IEEE published 1stsecurity standards (2015)
Security weaknesses endemic to IT
Keeps CISOs employed & CEOs up at night!
Internal & external security weakness:
External fraud security weaknesses - more publicity Internal weakness - far more costly
Due to position advantage
Position advantage:
Increased opportunity for fraud - position of perpetrator Internal example:
CIO sells proprietary information to competitor.
External example:
Criminal hacker buys list of credit card numbers Commits multiple frauds at a point of sale.
Largest fraud losses:
Healthcare: Due to combination of:
Systemic security weakness (e.g., mutable audit trail)
+
IT administrator with position advantageExchange trading: Combination of:
Security weaknesses
+
Employee with position advantage between: Trading&
Back-office trade executionFraud Factor 2 -
Pressure
Usually financial pressure, brought on by:
Gambling - drug use - negative life events (e.g., divorce, bankruptcy)
Pressure - human resource issue Managers must be aware of employees:
Life events
Performance changes
Decreasing internal fraud pressure
Steps to counter pressures:
Conduct background checks at acceptable levels Before grant administrator/higher-system privileges
Require:
Administrators submit credit reports
At standard intervals & change of position
Employees report:
Civil suits criminal arrests & incidents
service of process debt collection at work place
Fraud Factor 3 -
Rationalization
“The human capacity for denial and rationalization
is always shocking, but never surprising.”
David Levy, PhD,
Humor in Psychotherapy Lectures
Part 2:
Current Fraud Risks
Data Breach Fraud
Data breach fraud - General
Methods of data breach changing
In healthcare (and most IT industries)
In 2014 - leading method data breach in healthcare: Criminal system attack
Surpassed:
Employee negligence - lost laptops for 1sttime Reference:
Fifth Annual Benchmark Study on the Privacy & Security of Healthcare Data Ponemon Institute (2015)
FBI Stats - Data breach fraud
According to FBI:
Criminals target healthcare databases
because they contain in one place:
PII - Personal identification information PCI - Personal credit information PHI - Protected health information
Ask yourself - 2 critical questions
Critical question 1:
Do you have PII, PCI or PHI on your system?
If answered is yes - know that:
PII / PCI / PHIs - primary targets of data fraud Across all enterprises
How secure is your data?
Criminals constantly work : to keep pace with data technology
Critical question - 2
Data Breach - Fraud pays
How much is your data worth to criminals? How is it monetized?
FBI statistics - fraud monetization (2015): Credit cards: $0.50 - $1.00(each) Healthcare data records: $60 - $70(each) Criminals obtain: Name, DOB, SS, Policy, etc.
Do the math:
How many healthcare records do you have?
1,000 - 10,000 - 100,000
$60,00 - $600,000 - $6 million
Total the number of records in your system
- Calculate the huge financial temptation to criminals Ka-ching!
Billions of dollars - Very tempting
Stolen data is quickly sold on the Internet:
To criminal organizations Using masking sites such as Tor
In a few minutes:
Your data - sold & resold All over the planet
Your stolen data - Used & Reused
Criminals use your stolen data:
To commit more frauds: Identity theft
Tax fraud
Medical device fraud Prescription fraud Other crimes
Part 2:
Current Fraud Risks
Mobile/Connected Device Fraud
Mobile/Connected Device Fraud
In healthcare & other large organizations that collect data
Mobile & connected device devices create: New data security & liability risks.
Connected devices rapidly in greater use
Mobile devices (e.g., cell phones) universally used
Healthcare - Connected Devices
Connected devices in healthcare:
Bluetooth-enabled insulin pump connected with a wireless internal glucose monitor Pump delivers insulin to a diabetic patient
using data from the monitor without human direction (except monitoring)
M/CD - Risks
Risk for mobile/connected devices - similar Connected devices have additional risk
They preform actions automatically without direct human intervention.
M/CD - Proactive security
To ensure a STRONG anti-fraud program:
1) Address fraud issues via contract provisions with all IT vendors 2) Develop & enforce a written mobile/connected device policy 3) Install mobile/connected device management software 4) Create, update, audit incident plans for all M/CD
5) Integrate personnel management, training, compliance audits
M/CD - 6 critical questions
Make sure you can answer YES to all these questions Do your M/CD devices:
1) Store & transmit data securely?
2) Accept software security updates to address new risks? 3) Avoid creating unauthorized access points of data? 4) Detect & avoid a new way or path to data theft?
5) Connect to institution's IT infrastructure so data are secure? 6) Have secure APIs - software and device connections?
M/CD - 4 Anti-fraud actions
Once you can answer YES to all 6 security questions:
Implement these
4
recommended antifraud actions -to mitigate fraud opportunities & weaknesses in M/CDRequire all IT vendor contracts include:
1) Data encryption
2) Authorized device-only networks 3) Physical security training and measures
M/CD - Data encryption
Healthcare providers (and other organizations) :
Should require IT vendors agreements ensure Data traffic of devices and applications be encrypted when communicating with:
1) The provider’s private network, 2) Any outsourced providers and 3) Any cloud systems.
M/CD - Provider audits
Vendor contracts must allow provider to audit to:
1) Verify data transmitted in appropriate level of encryption 2) Ensure encryption works on your network
3) Test & retest your data encryption
Authorized device-only networks
Require that IT contracts allow:
A single mobile or connected device to collect only data required for its intended operation
Only grant access to data generated by single devices: By authorized and authenticated individuals,
who need to handle the information
Physical security - training & measures
Device physical security - CRITICAL 5 effective steps - Ensure that:
1) Devices prevent data storage media accessed/removed 2) Devices difficult to take apart & display signs of tampering. 3) Data cannot be removed from device
(otherwise security for transmitted data useless) 4) Train personnel in physical security procedures 5) Test & audit employee physical security awareness.
Credential & password protection
2
steps to increase credential & password security:(after setup - before critical data used & transmitted) 1) Require vendors change default passwords/usernames
to meet organizational standards
2) All passwords/usernames - random & unassociated (Much more secure against hackers)
3 direct anti-fraud actions
1) Develop & train using written M/CD policy
2) Install mobile/connected device management software 3) Create, update & audit incident plans for M/CD
MC/D - Written policy
Your M/CD policy should at minimum provisions for:
Purpose Applicability Appropriate use
Management software (MDM)
For organizations with multiple employees:
MDM is a key component in device security
MDM will ensure:
Control of M/CD access to your infrastructure Securely manage data usage
Securely manage internal/external movement Control M/CD application management features
Incident plans
Important part - comprehensive risk management strategy
Your incident plan should designate:
Unit roles & authority with contact info attached
Flowcharting recommended (in addition to written plan)
Can’t quickly respond - if haven’t assessed weaknesses Document all incidents - no matter how minor
Audit devices singly
Final system considerations
M/CD networks must NOT be configured to allow:
Credentials & passwords be exposed in network traffic Audit at standard intervals
Vendor contracts should require:
Connected devices regularly updated: With improved security
Testing/verification of updates before being put into use (Usually standard contract language - be sure to check)
Part 3:
Fraud Analytics
Limits
Fraud Analytics Limits
Specific vs. Bulk data collection
In 2014, President Obama tasked Director of National Intelligence:
Determine feasibility of software targeting specific data transmissions
(Instead of using bulk collection of data transmissions)
ODNI referred to National Academy of Sciences Research Council.
The Research Council’s answer - in short - was NO:
(Targeted collection could not replace bulk collection. )
Fraud Analytics Limits - Targeted data
Many problems with targeted data: Too limited targeted view Wrong target source
Loss of possibly valuable data over time Lack of context
Lack of analytic resources
Research Council recommends:
Continuing use of bulk collection
Apply automated use controls - prevent privacy breaches
Fraud Analytics Limits - Targeted sourcing
If targeted sourcing of: Very large/mutable telecom databases is insufficient - Then targeted fraud analytics of large datasets
share the same problems.
Credit, healthcare, insurance, government & others Using fraud analytics on large datasets
should be aware of the limits of analytic tools.
Fraud analytics are like a searchlight
They brightly illuminate the focal point
Just beyond the beam - it is dark
References
Slide: 6 - "Fraud Statistics." Fraud Statistics. CAIF Coalition Against Insurance Fraud, Web. 01 June 2015.
<http://www.insurancefraud.org/statistics.htm#.VWyNbKxFDIV>.
Slide: 9 - Landwehr, Carl, and Tom Haigh. "Building Code for Medical Device Software Security." IEEE Cybersecurity Initiative. Institute
of Electrical and Electronic Engineers, Mar. 2015. Web. 04 June 2015. <http://cybersecurity.ieee.org/>.
Slide: 16 - "Criminal Attacks Are Now Leading Cause of Data Breach in Healthcare, According to Ponemon Study." Ponemon
Institute/ID Experts, Inc., 06 May 2015. Web. 01 June 2015.
<https://www2.idexpertscorp.com/press/single/criminal-attacks-are-now-leading-cause-of-data-breach-in-healthcare-ponemon>.
Slide: 17 - "FBI Health Care Fraud Release Collection." FBI. 17 June 2014. Web. 01 June 2015.
<http://www.fbi.gov/collections/health-care-fraud>.
Slide: 28 - Tannenbaum, William A. "Healthcare's 'Internet of Things' Should Be the 'Security of Things'" Healthcare IT News. 19 May
2015. Web. 01 June 2015. <http://www.healthcareitnews.com/blog/healthcares-internet-things-should-be-security-things>.
Slide: 35 - Rosciam, Michael, CPA. "Moving Violations: 3 Steps for Taming Mobile Threats." Thomas, Howell, Ferguson, PA., Web. 01
June 2015. <http://www.thf-cpa.com/blog/detail/moving-violations-3-steps-for-taming-mobile-threats#.VWyVyqxFDIV>.
Slide: 40 - "New Report Says No Technological Replacement Exists for Bulk Data Collection;." National-Academies.org. Office of the
Director of National Intelligence, 15 Jan. 2015. Web. 01 June 2015 <http://www8.nationalacademies.org/onpinews/newsitem.aspx? RecordID=19414>.