Study guide
15.3 Data Ethics, Audit Trails & Secondary Data Use
AHIMA Certified Health Data Analyst CHDA
Track your progress as you read
Finished sections get a check mark.
1 Introduction & CHDA Exam Overview
0.1 About the AHIMA CHDA Exam 0.2 How to Use This Study Guide
2 Chapter 1: Healthcare Delivery Systems & Revenue Cycle
1.1 Healthcare Delivery Settings & Organizational Structures 1.2 Healthcare Revenue Cycle Management & Reimbursement Models 1.3 Regulatory Reporting & Accreditation Agencies
3 Chapter 2: Health Informatics, Datasets & AI
2.1 Electronic Health Records (EHR) & Health Informatics Foundations 2.2 Healthcare Datasets, Benchmarks & Population Health 2.3 Analytics Hierarchy & Artificial Intelligence in Healthcare
4 Chapter 3: Clinical Terminologies, Classifications & Standards
3.1 Clinical Classification Systems: ICD-10-CM/PCS & CPT/HCPCS 3.2 Clinical Terminologies & Ontologies: SNOMED CT, LOINC & RxNorm 3.3 Interoperability & Data Exchange Standards: HL7 v2, CDA & FHIR
5 Chapter 4: Business Needs Assessment & Scoping
4.1 Stakeholder Engagement & Analytics Needs Elicitation 4.2 Defining Metrics, KPIs & Healthcare Value 4.3 Analytics Project Management & Analysis Planning
6 Chapter 5: Data Sourcing, Lineage & Collection Methods
5.1 Healthcare Data Sourcing & Lineage 5.2 Healthcare Data Collection Methods & Architectures 5.3 Clinical & Financial Data Extraction
7 Chapter 6: Data Cleansing, Transformation & Validation
6.1 Healthcare Data Quality Assessment 6.2 Data Cleansing & Anomaly Remediation 6.3 Data Transformation, Mapping & Validation
8 Chapter 7: Exploratory Data Analysis & Descriptive Statistics
7.1 Healthcare Data Types & Scales of Measurement 7.2 Measures of Central Tendency & Dispersion 7.3 Exploratory Data Analysis in Clinical Populations
9 Chapter 8: Inferential Statistics & Hypothesis Testing
8.1 Sampling Techniques & Probability in Healthcare 8.2 Hypothesis Testing Fundamentals 8.3 Parametric vs. Non-Parametric Tests
10 Chapter 9: Advanced Statistical Modeling & Risk Adjustment
9.1 Correlation & Regression Analysis 9.2 Risk Adjustment & Healthcare Severity Modeling 9.3 Survival Analysis & Model Evaluation
11 Chapter 10: Healthcare Epidemiology & Study Methodologies
10.1 Epidemiological Measures of Disease Frequency & Association 10.2 Healthcare Study Designs & Methodologies 10.3 Confounding, Bias & Health Equity Analytics
12 Chapter 11: Healthcare Data Visualization & Dashboard Design
11.1 Principles of Effective Healthcare Data Visualization 11.2 Selecting Appropriate Healthcare Visualizations 11.3 Executive & Clinical Dashboard Architecture
13 Chapter 12: Healthcare Performance Measurement & CMS Quality Programs
12.1 CMS Quality Programs & Value-Based Purchasing 12.2 Clinical Quality Measures & Benchmarking 12.3 Financial Analytics & Case Mix Index Analysis
14 Chapter 13: Communicating Insights & Narrative Analytics
13.1 Translating Complex Analytics for Diverse Audiences 13.2 Contextualizing Limitations, Assumptions & Uncertainty 13.3 Narrative Reporting & Data-Driven Decision Support
15 Chapter 14: Healthcare Data Governance Frameworks & Stewardship
14.1 Enterprise Data Governance Frameworks 14.2 Data Stewardship & Master Patient Index (MPI) Management 14.3 Metadata Management & Business Glossaries
16 Chapter 15: Healthcare Privacy, Security & Data Ethics
15.1 HIPAA Privacy & Security Rules for Analysts 15.2 De-Identification Methodologies 15.3 Data Ethics, Audit Trails & Secondary Data Use
15.3 Data Ethics, Audit Trails & Secondary Data Use
Key Takeaways Explain This Topic AI
Secondary healthcare data use requires separate Common Rule and HIPAA analyses: a project designed to develop or contribute to generalizable knowledge may be research, while qualifying quality-improvement activities may fall within HIPAA health care operations; institutional review, exemption, consent, authorization, and waiver requirements depend on the activity and applicable rules.
Under 45 CFR § 164.512(i), an IRB or Privacy Board may grant a Waiver of HIPAA Authorization if the research poses minimal privacy risk, could not practicably be conducted without the waiver, and could not practicably be conducted without access to identifiable PHI.
Healthcare data ethics and algorithmic accountability operationalize the four bioethical pillars—Beneficence, Non-maleficence, Autonomy, and Justice—to proactively mitigate historical dataset bias (e.g., healthcare spending proxies) and satisfy HTI-1 algorithmic transparency rules.
The HIPAA Security Rule audit-controls standard (§ 164.312(b)) requires mechanisms that record and examine activity in systems containing or using ePHI; organizations may add risk-based automated surveillance to detect insider threats, inappropriate chart access, bulk extraction, and anomalous activity.
Under the HIPAA Breach Notification Rule (45 CFR §§ 164.400–414), an impermissible use or disclosure of unsecured PHI is presumed to be a breach unless an exception applies or a four-factor assessment demonstrates a low probability of compromise; individual, HHS, and—when 500 or more residents of a state or jurisdiction are affected—media notice follow distinct timing rules.
Last updated: August 2026
Data Ethics, Audit Trails & Secondary Data Use
As healthcare organizations increasingly leverage clinical data repositories for artificial intelligence development, real-world evidence generation, quality measurement, and clinical benchmarking, the ethical and regulatory boundaries governing secondary data use become paramount. Health data analysts must distinguish between operational quality improvement and human subjects research, apply robust ethical frameworks to mitigate algorithmic bias, maintain forensic audit trail integrity, and navigate federal breach notification mandates. Adhering to these standards ensures that data analytics protects vulnerable patient populations, preserves public trust, and withstands regulatory scrutiny.
1. Secondary Use of Health Data: Quality Improvement vs. Human Subjects Research
Secondary data use is defined as the utilization of clinical, financial, or operational data collected during primary direct patient care delivery for purposes other than direct healthcare provision. Secondary use spans clinical research, comparative effectiveness studies, institutional quality improvement, machine learning model training, and public health reporting. +---------------------------------------------------------------------------------------------------+
- | QUALITY IMPROVEMENT (QI) vs. HUMAN SUBJECTS RESEARCH |
- +-------------------------------------------------+-------------------------------------------------+
- | CLINICAL QUALITY IMPROVEMENT (QI) | HUMAN SUBJECTS RESEARCH |
- | ----------------------------------------------- | ----------------------------------------------- |
- | - Regulatory Domain: HIPAA Health Care Ops (HCO)| - Regulatory Domain: Common Rule (45 CFR 46) |
- | - Statutory Intent: Improve local clinical care | - Statutory Intent: Develop or contribute to |
- | delivery, patient safety, or facility workflow| generalizable scientific knowledge |
- | - Scope: Purpose/design determine classification| - Designed to contribute generalizable knowledge|
- | - Oversight: Institutional determination | - IRB/HRPP review or exemption as applicable |
- | - Common Rule and HIPAA analyzed separately | - Consent and HIPAA permission are distinct |
- | - HCO may apply when its conditions are met | - Authorization or waiver may address PHI |
- +-------------------------------------------------+-------------------------------------------------+
- The Common Rule & Delineation Criteria
The Federal Policy for the Protection of Human Subjects , codified as 45 CFR Part 46 (The Common Rule) , defines research as "a systematic investigation, including research development, testing, and evaluation, designed to develop or contribute to generalizable knowledge."
Quality improvement versus research: The label depends on the project’s purpose and design under the Common Rule definition, not merely whether it occurs at one site or may be published. Many activities designed solely to improve a local program are not research, but a systematic investigation designed to develop or contribute to generalizable knowledge may be research.
Institutional determination: Follow the organization’s Human Research Protection Program or IRB process for a determination of non-research, exemption, or required review. Do not let the analytics team self-exempt a borderline project.
HIPAA is a separate analysis: An activity can be health care operations under HIPAA yet still require a Common Rule determination, or vice versa. Document both bases.
Institutional Review Board (IRB) Review & HIPAA Authorization Waivers
When research uses PHI, investigators need an applicable HIPAA permission, such as individual authorization or a valid IRB/Privacy Board waiver. Common Rule consent and HIPAA authorization are related but distinct requirements, and exempt research can still involve HIPAA obligations. +---------------------------------------------------------------------------------------------------+
- | CRITERIA FOR IRB WAIVER OF HIPAA AUTHORIZATION |
- +---------------------------------------------------------------------------------------------------+
- [ CRITERION 1: Minimal Risk to Privacy ]
- The research involves no more than minimal risk to the privacy of individuals, evidenced by:
- (a) An adequate plan to protect patient identifiers from improper use and unauthorized disclosure;
- (b) An adequate plan to destroy identifiers at the earliest opportunity consistent with research; and
(c) Adequate written assurances that PHI will not be reused or redisclosed to third parties.
- │
- ▼
- [ CRITERION 2: Impracticability of Research Without Waiver ]
- The research could not practicably be conducted without the waiver or alteration (e.g., retrospective
cohort of 50,000 historical patients, many deceased, relocated, or lost to follow-up).
- │
- ▼
- [ CRITERION 3: Impracticability of Research Without PHI Access ]
- The research could not practicably be conducted without access to and use of the specific PHI.
Other Statutory Research Provisions under HIPAA
Preparatory to Research Exception (§ 164.512(i)(1)(ii)): Permits researchers to query and view identifiable PHI within the covered entity to develop research protocols, test hypotheses, or assess cohort feasibility, provided no PHI is removed from the covered entity's secure boundary.
Research on Decedents' Information (§ 164.512(i)(1)(iii)): Permits research exclusively using the health records of deceased individuals upon representation that the data is necessary for research and documentation of death.
2. Healthcare Data Ethics & Algorithmic Accountability
As healthcare analytics shifts from descriptive reporting to predictive and generative artificial intelligence, analysts bear a profound ethical duty to ensure algorithmic fairness, transparency, and accountability. +---------------------------------------------------------------------------------------------------+
- | THE FOUR PILLARS OF HEALTHCARE DATA ETHICS |
- +-----------------------------------+-----------------------------------+---------------------------+
- | 1. BENEFICENCE | 2. NON-MALEFICENCE | 3. AUTONOMY & JUSTICE |
- | (Maximizing Benefit) | (Primum Non Nocere - Do No Harm) | (Respect & Equity) |
- | - Design analytics to enhance | - Mitigate algorithmic bias | - Preserve patient data |
- | diagnostic accuracy & safety | - Prevent clinical misallocation | dignity & consent |
- | - Optimize resource allocation | - Eliminate privacy leakage | - Ensure equitable model |
- | - Reduce healthcare disparities | - Prevent automation bias | performance across races|
- +-----------------------------------+-----------------------------------+---------------------------+
- Mitigating Historical Bias in Machine Learning
Machine learning models trained on historical electronic health record data frequently encode, reproduce, and amplify historical societal disparities:
Proxy Variable Confounding (The Cost-Proxy Trap): In a landmark study published in Science (Obermeyer et al., 2019), a commercial risk-prediction algorithm widely used by insurers and health systems to allocate specialized care management programs utilized past healthcare expenditures (cost) as a proxy for illness severity and healthcare need. Because historical socioeconomic and racial disparities resulted in less money being spent on Black patients compared to White patients with identical illness burdens, the algorithm falsely generated lower risk scores for Black patients. Consequently, the proportion of Black patients enrolled in specialized programs was cut in half. Changing the target from cost to direct measures of health need substantially reduced the measured racial bias in the study.
Representation Bias: Underrepresenting racial minorities, pediatric populations, or rural cohorts in training datasets leads to severe performance degradation when models are deployed across diverse patient populations.
Fairness Metrics for Analysts: Analysts must perform subgroup validation using standardized fairness metrics, including Demographic Parity (equal positive prediction rates across groups) and Equalized Odds (equal true positive and false positive rates across demographic groups).
Algorithmic Transparency & The HTI-1 Final Rule
Promulgated by the Office of the National Coordinator for Health Information Technology (ONC, now the Assistant Secretary for Technology Policy / ASTP), the Health Data, Technology, and Interoperability (HTI-1) Final Rule establishes stringent transparency requirements for Decision Support Interventions (DSI) and predictive AI embedded in certified health IT:
Predictive DSI source attributes: Certified health IT developers subject to HTI-1 must make specified source attributes available for predictive decision support interventions supplied as part of certified health IT; this is not a blanket claim about every model in every EHR.
Required Metadata Elements: Training data demographics, intervention purpose, external validation metrics (AUROC, sensitivity, specificity), known performance limitations, and funding sources/conflicts of interest.
Honest Broker Systems
An Honest Broker is a designated, trusted independent individual or specialized technological platform that serves as an ethical and technical firewall between clinical systems and secondary researchers:
Architecture: The honest broker extracts identifiable clinical data, strips all direct identifiers, generates pseudonymous research IDs, and maintains the isolated master linkage crosswalk table in a secure, encrypted vault.
Function: Researchers receive an approved de-identified or limited dataset while the linkage key remains segregated and access-controlled, reducing re-identification opportunity while permitting authorized longitudinal refreshes.
3. Audit Trails & Access Monitoring
Under the HIPAA Security Rule (45 CFR § 164.312(b)), covered entities and business associates must implement hardware, software, and procedural mechanisms that record and examine activity in information systems containing or using ePHI. Retention and review schedules are established through applicable documentation rules, risk analysis, other law, and organizational policy. +---------------------------------------------------------------------------------------------------+
- | EHR AUDIT TRAIL LOGGING ARCHITECTURE |
- +---------------------------------------------------------------------------------------------------+
- [ 1. USER IDENTIFIER ] -> Unique User ID, Employee Role, and Assigned Department
- [ 2. EXACT TIMESTAMP ] -> Microsecond UTC timestamp synchronized via NTP
- [ 3. PATIENT IDENTIFIER ] -> Patient MRN, Account Number, and Enterprise Master Person Index ID
- [ 4. NETWORK & LOCATION ] -> Workstation Machine ID, IP Address, Physical Facility / Nursing Unit
- [ 5. ACTION / EVENT TYPE ] -> Chart View, Record Insert, Modification, Deletion, Print, SQL Export
- [ 6. CLINICAL CONTEXT ] -> Active Encounter ID, Clinical Relationship, Order Number
- │
- ▼
- ┌─────────────────────────────────────────────────────────────────────────────────────────────┐
- │ AUTOMATED AUDIT TRAIL SURVEILLANCE & AI ENGINES │
- │ - Continuous SIEM ingestion parsing millions of daily electronic access transactions │
- │ - Machine learning anomaly detection flagging suspicious behavioral patterns │
- └─────────────────────────────────────────────────────────────────────────────────────────────┘
- Automated Audit Log Surveillance Algorithms
Healthcare organizations generate millions of audit log rows daily. Modern health systems deploy automated surveillance algorithms to flag high-risk access violations:
VIP & High-Profile Patient Snooping: Automated triggers that generate immediate compliance alerts when workforce members access the medical charts of celebrities, elected officials, hospital executives, or patients admitted following widely publicized trauma events without an active clinical assignment.
Self, Family & Coworker Snooping: Algorithms cross-referencing employee HR files (matching employee surnames, home residential addresses, payroll tax IDs, and dependent lists) against patient charts accessed, detecting unauthorized curiosity viewing of coworkers, family members, or self-chart viewing.
Mass Data Scraping & Bulk Exfiltration: Detecting anomalous spikes in query volume, rapid sequential chart opens (e.g., accessing 200 charts within 15 minutes), bulk SQL table exports, or unauthorized copying to external USB media.
Temporal & Geographic Anomalies: Access occurring during unassigned shifts, while an employee is on approved vacation leave, or logins originating from anomalous geographic IP ranges.
Departmental Mismatch: Flagging access by workforce members whose department has zero clinical or administrative relationship with the patient (e.g., an orthopedic physical therapist viewing a labor and delivery psychiatric record).
Forensic Log Integrity & Retention
Integrity protection: Protect logs against unauthorized alteration or deletion using controls justified by risk. Access restrictions, separation of duties, tamper-evident storage, hashing, or WORM storage may be appropriate; HIPAA does not mandate one architecture.
Retention: HIPAA requires retention of specified compliance documentation for six years, but does not assign every audit event a universal six-year period. Set log retention from the risk analysis, investigation needs, applicable law, contracts, and approved policy.
- 4. The HIPAA Breach Notification Rule (45 CFR §§ 164.400–414)
- Definition of a Breach
Under 45 CFR § 164.402, a Breach is defined as the unauthorized acquisition, access, use, or disclosure of Protected Health Information in a manner not permitted under the Privacy Rule which compromises the security or privacy of the protected health information.
The Statutory Presumption: An impermissible use or disclosure of unsecured PHI is presumed to be a breach , unless the covered entity or business associate demonstrates that there is a low probability that the PHI has been compromised based on a formal four-factor risk assessment. +---------------------------------------------------------------------------------------------------+
- | THE FOUR-FACTOR BREACH RISK ASSESSMENT |
- +---------------------------------------------------------------------------------------------------+
- [ FACTOR 1: Nature & Extent of PHI ]
- Types of identifiers involved, clinical sensitivity (e.g., mental health, HIV status, substance
abuse records), and the likelihood of direct re-identification.
- │
- ▼
- [ FACTOR 2: Unauthorized Recipient ]
- The identity of the unauthorized person who received or used the PHI (e.g., another HIPAA covered
entity bound by statutory confidentiality vs. a malicious hacker or public internet forum).
- │
- ▼
- [ FACTOR 3: Actual Viewing or Acquisition ]
- Whether forensic examination confirms the PHI was actually accessed, viewed, or exfiltrated
(e.g., forensic analysis of recovered stolen laptop proving files were unaccessed vs. data opened).
- │
- ▼
- [ FACTOR 4: Extent of Mitigation ]
- The extent to which risk has been mitigated (e.g., immediate physical recovery, remote device wipe,
or obtaining a signed, legally binding certificate of destruction from the recipient).
Statutory Exceptions to the Breach Definition
A security or privacy incident is explicitly excluded from the legal definition of a breach if it meets one of three statutory exceptions (§ 164.402):
Unintentional, Good-Faith Acquisition: Unintentional access by a workforce member acting within scope of authority, with no further impermissible use or disclosure.
Inadvertent Disclosure Between Authorized Persons: Inadvertent disclosure between persons authorized to access PHI at the same covered entity, business associate, or organized health care arrangement (OHCA).
Inability to Retain Information: Good-faith belief that the unauthorized person to whom disclosure was made would not reasonably have been able to retain the information (e.g., a nurse hands a paper discharge summary to the wrong patient, immediately realizes the mistake, and recovers the document before the recipient reads it).
Breach Notification Timelines & Reporting Mandates +---------------------------------------------------------------------------------------------------+
- | MANDATORY BREACH NOTIFICATION TIMELINES |
- +-------------------------------------------------+-------------------------------------------------+
- | 1. INDIVIDUAL NOTICE | 2. MEDIA AND HHS NOTICE |
- | ----------------------------------------------- | ----------------------------------------------- |
- | - Written notice via first-class mail (or email)| - Media: 500+ residents of a state/jurisdiction |
- | - Sent without unreasonable delay and in no | - HHS: report breaches of 500+ within 60 days |
- | case later than 60 CALENDAR DAYS post-discovery| - Different HHS timing applies below 500 |
- | - Contact procedure required in the notice | - Publicly listed on HHS OCR Breach Portal |
- +-------------------------------------------------+-------------------------------------------------+
- │
- ▼
- [ BREACHES AFFECTING FEWER THAN 500 INDIVIDUALS ]
- Notify affected individuals within 60 calendar days of discovery
- Log electronically and submit to HHS OCR annually within
- 60 CALENDAR DAYS AFTER THE END OF THE CALENDAR YEAR
Business Associate Breach Notification: A Business Associate must notify the Covered Entity of any discovered breach without unreasonable delay and in no case later than 60 calendar days post-discovery (many BAAs contractually shorten this window to 24 to 72 hours to allow the CE time to investigate).
5. Master Ethics & Regulatory Compliance Summary Table
Governance Dimension Clinical Quality Improvement (QI) Human Subjects Research (Common Rule) HIPAA Breach Incident (500+ Individuals)
Primary Authority HIPAA Privacy Rule (Health Care Operations) Common Rule (45 CFR 46) & HIPAA Privacy HIPAA Breach Notification Rule (45 CFR 164.400)
Primary Purpose Improving institutional care delivery & safety Generating generalizable scientific knowledge Remediating unauthorized PHI compromise
Oversight Body Departmental Quality & Clinical Governance Institutional Review Board (IRB) HHS Office for Civil Rights (OCR) & Media
Authorization / Permission Confirm applicable health-care-operations permission Common Rule consent and HIPAA permission assessed separately Mandatory individual notice unless an exception applies
Review / Notice Timing Defined by operational governance Depends on review category, approval, and protocol Individual notice within 60 calendar days of discovery
Data Scope & Boundary Defined local improvement purpose and permissions Defined by the research protocol and data permissions Governed by the four-factor risk assessment and notice rules
Analytics Controls Validation, metric governance, and monitoring Protocol-specific validity, privacy, safety, and fairness review Investigation, risk assessment, mitigation, and documentation