Talk to a specialist
Home / Resources / HIPAA guide / HIPAA logging, audit controls and activity review

HIPAA · Practical business guide

HIPAA logging, audit controls and activity review

Build useful HIPAA audit controls by choosing meaningful events, validating collection, reviewing activity and preserving evidence for investigations.

In this guide
Logging becomes a control when someone can use the recordChoose events that explain access and changeValidate collection from source to reviewMake activity review a defined responsibilityProtect logs and preserve investigation integrityHypothetical example: a hospital investigates unusual record accessHypothetical example: a BPO links export activity across customersMeasure whether the process can answer real questionsCheck that an alert can become an investigationReference basisFrequently asked questionsDiscuss the next practical stepRelated HIPAA guides

Logging becomes a control when someone can use the record

HIPAA audit controls concern mechanisms that record and examine activity in systems containing or using electronic protected health information (ePHI). The Security Rule also requires procedures for regular review of information system activity, such as relevant logs and reports. Collecting data without coverage checks, review ownership or a response process leaves a significant gap between a technical feature and an operating safeguard.

Start with the questions the organisation may need to answer. Who accessed a record? Who changed a permission? Which account exported information? What happened before a system became unavailable? Can activity be linked across the application, identity service and supplier? These questions help select useful events and identify sources whose absence would limit an investigation.

The current rule does not prescribe one universal event list, review interval or retention period for every environment. HHS’s historical logging guidance explains the need to determine appropriate controls in context. Choose a justified approach based on the service, information, risks and other applicable obligations, and document it. Do not confuse the Security Rule’s documentation-retention requirement with a blanket mandate to retain every raw log for six years.

Choose events that explain access and change

Identify the significant actions within each ePHI workflow. Relevant events may include authentication, viewing or exporting records, changes to permissions, administrative actions, failed access and changes to logging itself. The precise design depends on system capability and risk. A clinical application may provide record-level events, while infrastructure logs provide a different view; neither automatically substitutes for the other.

For each event, determine whether it identifies the actor, action, object, time and outcome with sufficient clarity. A shared account may make an event impossible to attribute to an individual. A technical identifier may need a reliable mapping to the relevant application or user. Document these limitations and address them where reasonable and appropriate instead of assuming that a large volume of events is inherently useful.

Review exports and interfaces separately. An application might record a user opening a patient screen but not downloading a report containing many records. An API might use a service account that conceals the initiating user unless the application preserves that context. Understanding the workflow helps identify the information needed to reconstruct what happened.

Avoid collecting unnecessary clinical content in logs. Technical diagnostics can accidentally capture free text, identifiers or complete payloads. Define what is needed for the purpose and apply appropriate safeguards to any protected health information (PHI) in the logs. Logging should improve security without creating a widely accessible secondary copy of patient records that bypasses the application’s normal permissions.

Validate collection from source to review

A logging design has several links: event generation, transport, storage, indexing, access and review. Test each link using authorised synthetic activity. Confirm that the source produces the expected event, that it reaches the intended location and that an analyst can retrieve and interpret it. A green collector status does not prove that the application is sending the events the organisation needs.

Time matters. Check time synchronisation and understand time zones, especially in global support operations. A timeline that combines local times without context can create false conclusions about sequence or duration. Preserve the original timestamps and explain conversions used in analysis. Where clocks are unreliable, record the limitation instead of presenting a precise chronology unsupported by the evidence.

Monitor gaps and failures. A stopped agent, changed licence, full storage location or application update can silently remove coverage. Define how missing expected activity is detected and who owns the repair. A control that works only until the next software change is not maintainable. Include logging requirements in change review and supplier discussions.

Test access to older records within the justified retention period. Data may exist in an archive but be difficult to retrieve quickly or interpret without obsolete tools. Consider the time and expertise needed during an incident. Retention should include practical retrievability, appropriate protection and lawful disposal, not merely a statement that storage has been purchased.

Make activity review a defined responsibility

Define which team reviews which sources, what conditions warrant investigation and how the review is recorded. Routine automated detection, targeted human review and retrospective investigation serve different purposes. Use them in a proportionate combination rather than assuming one dashboard can answer every concern. The review process should reflect the organisation’s risk analysis and service model.

Provide context to reviewers. A large export may be normal for an authorised migration and unusual for a scheduling user. A late-night login may be expected for a global support shift and unexpected for another role. Useful detection depends on role, customer, change and incident information. Keep that context current without giving analysts unnecessary access to patient content.

Define escalation and follow-up. An alert should have an owner, an initial assessment and a route to the security or privacy incident process when warranted. Record why it was closed or escalated. Repeated false positives should lead to a reviewed adjustment, not an undocumented habit of ignoring the alert. Repeated benign explanations may still reveal a process that needs clearer authorisation.

Protect independence where practical. The person whose privileged activity is under review should not be the only person able to erase or interpret the evidence. Smaller organisations may use compensating oversight through another manager or provider. The chosen design should be explained in terms of risk and capability, with limitations visible to the accountable security official.

Place activity review within the wider safeguard programme using our HIPAA Security Rule guide and keep responsibilities explicit across technical and operational teams.

Protect logs and preserve investigation integrity

Logs can contain sensitive patient, workforce and security information. Limit access to people who need it, control export and protect the records against inappropriate alteration or deletion. Separate ordinary analysis from administration where appropriate. An attacker or insider who can modify both the source activity and the only record of it can undermine the organisation’s ability to investigate.

Define retention by source and purpose. Consider investigation needs, contractual terms, other applicable laws and operational constraints. Raw technical logs, review records and formal incident documentation may have different retention rules. Record the rationale and implement it consistently. Do not keep every item indefinitely simply because storage is inexpensive; unnecessary retention can create its own exposure.

When an incident occurs, preserve relevant evidence through a controlled process. Record who obtained it, when, from which source and what handling occurred. Use methods appropriate to the investigation and seek specialist help where necessary. Avoid editing the only copy of a log to make it easier to read; create an analysis copy while maintaining the original under appropriate controls.

Be candid about gaps. Absence of a recorded event is not always proof that an action did not happen. The source may not log it, collection may have failed or records may have expired. Explain what the evidence can and cannot support. That distinction is especially important when logs inform a breach assessment or a statement to a customer.

Hypothetical example: a hospital investigates unusual record access

A fictional hospital receives a concern that an employee viewed records unrelated to their duties. The application records access events, but the initial report contains technical account identifiers and no role information. Security works with the application owner and privacy team to establish the account mapping, relevant period and legitimate work context before drawing conclusions.

The review compares the activity with assignments and approved tasks. It checks whether access occurred through the named account or an emergency route and whether a report export was involved. The team restricts access to the investigation material and avoids distributing patient details to general management. Human resources and legal advisers are involved through the organisation’s established process where appropriate.

The evidence reveals a logging limitation: one reporting function records the report name but not the parameters needed to identify affected records reliably. The team documents that uncertainty rather than claiming a precise scope. It takes appropriate immediate action and uses the finding to improve the report’s audit design and approval process.

Afterward, the hospital adds a controlled test to confirm that relevant reporting events are captured. Reviewers receive better role context and a clearer escalation route. The example demonstrates that useful audit controls require both technical detail and careful interpretation; an event list alone does not establish whether access was authorised or which information was affected.

Hypothetical example: a BPO links export activity across customers

A fictional healthcare BPO runs quality reviews for several customers. A central monitoring tool records workstation events, but each customer application provides a different export log. During a review, analysts cannot determine whether a large local file came from an authorised quality task or an unrelated download. The technical sources exist, but they are not connected to a usable business context.

The BPO defines the information needed for a representative export investigation: worker identity, current customer assignment, approval where required, source application, export time and approved destination. It works with customer owners to obtain the records available under the service arrangement. Where a platform does not provide necessary detail, the limitation is recorded and alternative controls are considered.

A pilot uses synthetic data and an authorised test account. Analysts follow the event from application export to the controlled review workspace and confirm that the timeline uses consistent time references. A deliberately unapproved destination triggers the expected review. The exercise also checks whether analysts can see more patient information than they need for the investigation.

The resulting process includes customer-specific limitations and contacts. It does not claim that every application now has identical monitoring. Management can see which risks are covered by logs, which rely on prevention or supervision and which require further customer discussion. This is more credible than presenting a single monitoring dashboard as proof of comprehensive visibility.

Measure whether the process can answer real questions

Evaluate audit controls with short, repeatable scenarios. Can the team trace an authorised record access, a failed login, a permission change and a representative export? Can it identify a collection outage? Can it retrieve older evidence and explain its limits? These tests should be designed around actual risks and performed safely, without creating unnecessary access to live patient information.

Keep a coverage record naming the system, useful events, known gaps, owner, retention and review route. Link gaps to the risk-management process. A legacy limitation may require restricted use, additional supervision or a replacement plan. It should not disappear because another system produces abundant logs.

HHS’s audit protocol includes questions about procedures, review records, responsibilities and whether systems actually generate activity records. It is a useful source of evidence questions, although the historical protocol is not a replacement for the current rule. Use it to test the completeness of the operating process rather than to create a superficial checklist.

For a buyer of monitoring services, ask what the provider can observe, how quickly relevant findings reach the organisation, what evidence is retained and who performs privacy-sensitive interpretation. A supplier may provide detection and technical triage while the healthcare organisation supplies context and makes disclosure decisions. Clear responsibilities make the service more effective and prevent assumptions from becoming investigation gaps.

Include a handover test when analysts or providers change. The incoming team should locate the important sources, understand known limitations and follow an unresolved investigation without relying on undocumented knowledge held by the departing reviewer.

Use our HIPAA gap-assessment guide to structure the sample, findings and completion evidence when a logging process cannot answer a material question.

Check that an alert can become an investigation

Select a representative alert and follow it through assignment, contextual review, evidence retrieval and closure. Confirm that the analyst can contact the service owner and obtain the necessary facts without broad access to unrelated records. Then ask whether another reviewer can understand the decision from the retained record. If the answer depends on an undocumented conversation, improve the handoff. This test evaluates the operating value of the logging process rather than only the number of events collected or the visual appearance of a monitoring dashboard.

Reference basis

eCFR: 45 CFR Part 164, Subpart C; HHS: January 2017 Cybersecurity Newsletter; HHS: Summary of the HIPAA Security Rule; HHS: OCR Audit Protocol, updated July 2018.

Frequently asked questions

Does HIPAA require every log to be retained for six years?

Do not treat the documentation-retention rule as a universal raw-log requirement. Determine appropriate retention for each record category using the applicable rules, risks, contracts and operational needs.

Is installing a monitoring platform enough?

No. Validate source coverage, event quality, collection, retention, review and response. The platform supports the process but does not establish that the process operates effectively.

Can missing logs prove that no disclosure happened?

Not necessarily. Understand what was logged, whether collection worked and what records remain. Explain those limits when using logs to support an incident or breach assessment.

Discuss the next practical step

If investigations cannot answer who did what, start with the missing questions and relevant systems. Scope a logging review around event quality, collection, interpretation and response, with demonstrable tests and visible limitations.

Explore CyberVantage360 HIPAA assessment and implementation services, or contact CyberVantage360 to discuss your priorities.