Ethics and Law in Health Informatics and Data Privacy

UPI Study › Courses › Healthcare Organization and Management › Free lessons › Lesson 3.4 Free with a free account

Ethics and Law in Health Informatics and Data Privacy

Healthcare Organization and Management · Chapter 3: Ethical Dimensions of Healthcare PracticeSection 3.4 · lesson 16 of 10510 min · free with a free account

← Ethics of Life, Death, and Healthcare FinancingNext lesson: Ethical Decision-Making Frameworks in Healthcare →

Earn 3 college credits for Healthcare Organization and ManagementRecommended for credit by NCCRS · Health 302 · self-paced · one payment, lifetime access
Enroll in Healthcare Organization and Management · $250 →Download the syllabus (PDF)

In this lesson

Data stewardship, de-identification, algorithmic bias, and accountable use of patient data.

  • Explain data stewardship and how consent weakens as data uses move away from care
  • Describe HIPAA's two de-identification methods and their limits
  • Identify how training data and proxy targets introduce bias into clinical algorithms
  • Explain why record browsing is a privacy wrong and how to enforce against it

Free study notes

Get this lesson's notes by email

The key points and 4 practice questions with answers, in your inbox. One email, free.

One email. Unsubscribe in one click.

Transcript

Everything said in the video, section by section (1,403 words). Read along, or search it.

Opening Case

A Free Tool Built From Patient Records

Blake Sutter, from an analytics company, has an offer for Larkfield. His firm will build a tool that flags patients at risk of getting worse after discharge. It would learn from six years of Larkfield's de-identified records. Larkfield gets the tool free. The firm keeps the right to sell what it builds to other clients. Kenji Watanabe, the chief information officer, likes the idea. It's free, it's legal, and it could help patients. So should Larkfield sign?

Data Stewardship

Holding Data on Patients' Behalf

Kenji takes the offer to Samuel Ndiaye, the compliance officer. Samuel starts with an idea called data stewardship. A hospital holds patient information on the patients' behalf, more like a trustee than an owner. Patients disclosed it because they had no real option if they wanted care. So 'may we use this?' is incomplete without a second question: would the person who gave it to us recognize this use? And compliance doesn't settle it. A use can be lawful and properly authorized, and still feel to patients like a breach of trust.

Data Uses

Primary, Operational, Research, Commercial, Development

Primary use is using data to care for the person it came from. Consent is implied: someone seeking treatment expects her information to be used to treat her. Secondary use is anything else, and it isn't one category. Operational use covers billing, scheduling, and quality reports. Most patients expect some of that. Research aims at general knowledge, and has its own review, with consent or a formal waiver. Commercial use shares data for value, and that's where patient expectations are furthest from the law. Development use, building tools like Blake's, is often labeled operational, though it looks more like research or commerce. The rule of thumb: the farther a use moves from the patient's own care, and outside the organization, the weaker the consent behind it.

De-Identification

De-Identified Is Not Anonymous

Blake's pitch rests on de-identification, removing what identifies a person, so the data can be used more freely. Under HIPAA, there are two methods: remove eighteen kinds of identifiers, the Safe Harbor method, or have an expert determine that the risk of identifying anyone is very small. But de-identified isn't anonymous. People are identified by combinations of traits, and a record can be matched against another dataset that has names attached. So the risk is reduced, not removed, and depends on what other data exists. Contracts must still ban attempts to re-identify and onward transfer. And calling such data anonymous overstates the protection.

Quick Check

Comparing Fall Rates Across Units

Quick check. Each month, Larkfield's quality team compares patient fall rates across Larkfield's units, to decide where to add safety rounds. What kind of data use is that? Pause, and decide. Operational use. It's quality measurement for running Larkfield, and most patients expect something like it. Research is tempting, because it involves analysis. But research aims at knowledge beyond the organization; this answers an internal management question.

Expectation Gap

Lawful Uses That Would Surprise Patients

There's a gap between what's lawful and what patients assume. Most expect their data to treat them, to support some administration, and not to be seen without a reason. Their expectations are usually narrower than the law, especially about data leaving the hospital, or any deal involving money. Where a use would surprise patients, Larkfield owes them clearer notice, a real choice, or a decision not to go ahead. Samuel's test is simple: could you describe the arrangement to the patients involved, in one plain sentence, without it hurting Larkfield? If not, it needs a better reason than being permitted.

Third-Party Sharing

Purpose, Binding Terms, and Who Benefits

Back to Blake's offer. Before data leaves Larkfield, Kenji asks three questions, ahead of the merits. First, what's the purpose, specifically? 'Improving care' is a category, not a purpose. What will be built, by whom, and for whose benefit? Second, what binds the recipient? Permitted uses, no re-identification, no onward transfer, security, audits, and destroying the data at the end. And who checks? Third, who benefits? Blake's firm gets a product it can sell, built from Larkfield's patients' records. So calling the tool free isn't accurate.

Algorithmic Bias

Training Data and Proxy Targets

Farah Siddiqui, the director of quality and patient safety, spots a deeper problem. Algorithmic decision support uses a model to score or rank patients for a decision. Bias gets in two ways. The first is training data: a model learns from past practice, so past inequality becomes a prediction. A model trained on one population may also misfire on another. The second is easy to miss: the outcome the model predicts. Often it's a proxy, because the real thing isn't recorded. In 2019, researchers led by Ziad Obermeyer studied a widely used tool. It predicted health costs, as a stand-in for health needs. Less had been spent on Black patients with the same needs. So at the same score they were sicker, and far fewer got flagged. A new target removed most of the bias. The model was accurate about costs. Its performance statistics didn't reveal the unfairness, because it did exactly what it was asked.

Evaluating Tools

Three Questions for Any Model

Blake's tool predicts readmission. That's a proxy too: readmission depends on how easily a patient can get back to the hospital. So it may miss risk in exactly the patients who struggle most to return. Farah writes down three questions for any tool. What does it predict, the real thing or a proxy? Who was it trained on, compared with our patients? And how does it perform for each group we track, not just overall?

Accountability

Transparency, Explainability, and Responsibility

Kenji adds two more terms. Transparency means disclosing that a tool is in use, for whom, and for what. Explainability means saying why it produced a particular output. They're separate. With complex models, being transparent without being able to explain is often the honest position. When a recommendation is wrong, who is accountable? Not the model; it can't be. And not only the nurse who followed it, if she had no way to judge it. The organization answers for deploying the tool, how it's used, and monitoring its effects. The user answers for her judgment, but only if she's given the information and authority to use it.

Record Browsing

Curiosity Access Without a Work Reason

Kenji's access audit raises a different issue: looking at records without a work reason. It's the most common privacy violation in hospitals, and it's rarely malicious. It's curiosity: a colleague, a neighbor, a relative, someone in the news, or a former patient. The patient had no real choice but to disclose, so looking without reason takes something from her without her knowledge. No harm is needed; the wrong is complete at the moment of access. Audits are necessary, but staff need to understand the duty, not just know access is logged. And enforcement must be consistent, because what's tolerated for senior people teaches everyone the real standard.

Quick Check

A Senior Surgeon Opens a Celebrity Chart

Quick check. An audit shows that a respected senior surgeon opened a celebrity patient's chart, out of concern. He had no role in her care. He changed nothing, and told no one. What should Larkfield do? Pause, and decide. Apply its usual response, the one anyone else would get. The access itself was the wrong, and his seniority can't change the standard. A quiet word is tempting, out of respect for him. But tolerating it for senior staff teaches everyone what the real standard is.

Digital Equity

Online Booking Leaves Some Patients Out

One more stewardship issue: equity in digital care. Larkfield plans to move appointment booking online. Every digital move excludes someone: people without a device, a reliable connection, a private space, confidence with technology, or the interface's language. Those groups overlap with the people already least well served. So Larkfield keeps a phone and front-desk route open, and measures uptake by group, so it can see who stopped coming.

Case Resolution

Sending the Offer to Data Governance

Back to Blake's offer. Kenji doesn't sign. He sends it to Larkfield's data governance committee: clinicians, legal, compliance, and a patient representative. He files it as a third-party arrangement, not operational use, with the commercial terms spelled out and an equity analysis set in advance. Review may take a quarter, and the firm may walk away. In exchange, Larkfield gets a decision it could describe to its patients in public. A governance committee that never says no isn't proof of good proposals. It's a sign it's being bypassed.

Key Takeaways

Stewardship, Proxies, and Equal Enforcement

To recap. Patient data is held for patients, so ask whether they would recognize the use. Consent weakens as a use moves farther from the patient's care, and outside the organization. De-identified isn't anonymous, so contracts still matter. For any model, check its target, its training population, and its results by group. And browsing records is wrong the moment it happens, for everyone, at every level.

About Healthcare Organization and Management at UPI Study

Healthcare Organization and Management (Health 302) is a self-paced online course of 105 video lessons in 18 chapters, from health systems and the law to ethics, leadership, teams, strategy, informatics, finance, quality, human resources, marketing and evidence-based practice. It carries a NCCRS credit recommendation for 3 semester credits.

Every lesson transcript is free on pages like this one, and the videos are free with a free UPI Study account. Enrolling adds a practice quiz after each lesson, the graded work and final exam that earn the credit, lifetime access, and an official transcript on completion. UPI Study has helped over 130,000 students with college credits. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Enroll now → Course details Download the syllabus All Health courses Every course with PRO →

Why take it here

Why students take Healthcare Organization and Management with UPI Study

How hospitals and clinics are runHealthcare Organization and Management covers how care organizations are led and run: health law and ethics, leadership and teams, strategic planning, health informatics, budgets, quality improvement, staffing and marketing. It is a core course in healthcare administration and health management degrees.
Credit at a fraction of the priceThe same 3 credits can cost $750 or more at a US college. Here it is one payment, self-paced, with lifetime access.
Finish your degree soonerTransfer credits reduce the courses you still owe, so you can graduate earlier or lighten a heavy semester.
Strengthen an applicationA completed college-level course on an official transcript shows admissions and scholarship committees you can do the work.
Study on your own scheduleNo start dates, no deadlines. Watch the lessons, practice, and take the final when you are ready.
130,000+students helped with college creditsACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Where the credit goes

Universities that work with NCCRS

Healthcare Organization and Management carries a NCCRS credit recommendation. These universities work with NCCRS; some are UPI Study's signed partners. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Thomas Edison State UniversityThomas Edison State UniversitySigned partnerUniversity of the PeopleUniversity of the PeopleSigned partnerCharter Oak State CollegeCharter Oak State CollegeSigned partnerStrayer UniversityStrayer UniversitySigned partnerColumbia Southern UniversityColumbia Southern UniversitySigned partnerIndiana TechIndiana TechSigned partnerPenn StatePenn StateWorks with NCCRSSouthern New Hampshire UniversitySouthern New Hampshire UniversityWorks with NCCRSSUNYSUNYWorks with NCCRS +many more cooperating universitiesFind my college →

Find my college → Not listed? Ask Athena in the corner of this page, or book a 15-minute call.

FAQ

Questions students ask us

About Ethics and Law in Health Informatics and Data Privacy

What are the two HIPAA de-identification methods?
HIPAA allows two methods. The Safe Harbor method removes eighteen specified kinds of identifiers. The Expert Determination method has a qualified expert conclude that the risk of identifying any individual is very small. De-identified data is not the same as anonymous data, because people can be re-identified by combining traits or matching records against other datasets, so contracts should still ban re-identification and onward transfer.
How can a healthcare algorithm become biased?
Bias enters in two main ways. A model trained on past practice learns past inequality, and it may misfire on a population unlike its training data. A model can also predict a proxy instead of the real target. A 2019 study led by Ziad Obermeyer found that a widely used tool, which predicted health costs as a stand-in for health needs, flagged far fewer Black patients who were equally sick.
What is the difference between primary and secondary use of health data?
Primary use means using data to care for the person it came from, and consent is implied. Secondary use covers everything else: operational uses like billing and quality reports, research, commercial sharing, and development of tools. As a use moves farther from the patient's own care, and outside the organization, the consent behind it weakens and patient expectations diverge more from the law.
Why is looking at a patient's record without a work reason a privacy violation?
Viewing a record without a work reason takes something from a patient who had no real choice but to disclose, and the wrong is complete at the moment of access, even if nothing is changed or shared. It is a very common hospital privacy violation, usually driven by curiosity. Enforcement should be consistent at every level, since tolerating it for senior staff teaches everyone the real standard.

About the Healthcare Organization and Management course

How many college credits is Healthcare Organization and Management?
Healthcare Organization and Management (Health 302) carries a NCCRS credit recommendation for 3 semester credits. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.
How much does Healthcare Organization and Management cost and what is included?
$250, paid once, with lifetime access: all 105 video lessons, a practice quiz after each, the graded work and final exam, and an official transcript when you finish. Students who want several courses choose UPI Study PRO instead.
Which universities accept UPI Study credits?
UPI Study has signed partner agreements with universities such as Thomas Edison State University, University of the People and Charter Oak State College, and works with universities such as Penn State, SNHU and SUNY. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities. Use "Find my college" on this page to check yours.
How long does Healthcare Organization and Management take?
It is self-paced with no deadlines. The 105 lessons total about 17 hours of video; most students add reading and practice time and finish in a few weeks to a few months, on their own schedule.
How are the exams done?
Quizzes and assignments are online in the course. The final exam is taken online with proctoring; a person on the UPI Study team checks assignments, usually within two days.
How do my credits reach my college?
When the course shows 100% complete, request the official transcript from your dashboard. It is sent to the college you name, and you can send it to more than one. Check your college first with "Find my college" on this page, or ask Athena.

Ready for the full course? Earn 3 credits in Healthcare Organization and Management.

105 video lessons, a quiz after each one, and the graded work that earns the credit. Recommended for credit by NCCRS.

Enroll in Healthcare Organization and Management →Download the syllabusEvery course with PRO →
Healthcare Organization and Management
3 credits · NCCRSEnroll →