Ethical by Design: Shaping the Future of Healthcare Tech
‘’Your voice carries! Patient confidentiality is vital. Discuss cases privately. Thank you for ensuring our patients’ trust and privacy.’’ More than 20 years ago, I noticed that sign posted in a hospital elevator. I vividly remember it because two healthcare personnel (HCP) were in the elevator at the time, openly discussing one of their patients. Clearly, patients’ privacy was not an ethical concern for them.
Today, there is a greater awareness about handling patients’ data, and patient trust is not so easily earned. Ethics ratings of healthcare groups have significantly declined in the US over the last few years. For instance, in 2019, 65% of US adults rated the honesty and ethics of medical doctors “very high” or “high,” but in 2023, the rating fell to 56%. Nurses, pharmacists, dentists, and psychiatrists followed the same downward trajectory.
AI and Ethical Dilemmas
If patients and their families have concerns about their privacy, they might be interested to know that HCPs find themselves amid their ethical dilemmas. One study showed that 90.8% of physicians and 67.7% of nurses were often involved in moral dilemmas.1
The issue of ethics is not only in the human realm. There is a great deal of discussion on ethics regarding AI, especially AI in healthcare. Primary concerns, or ‘ethical challenges,’ in Chapter 12 by Gerke et al. of the book ‘Artificial Intelligence in Healthcare’ are ‘informed consent to use, safety and transparency, algorithmic fairness and biases and data privacy.’2
An End to Algorithmic Discrimination: The New OCR Rule
This year, the Office for Civil Rights, U.S. Department of Health and Human Services (HHS), published its final rule, ‘applying Affordable Care Act to algorithms used by covered entities, including health care providers and payers.’ This latest regulation mandates that healthcare organizations must not use patient decision support tools in a discriminatory manner. It explicitly mentions discrimination regarding race, color, national origin, sex, age, and disability.
It is important to note that this rule includes manual and automated decision tools, including algorithms, advanced analytics, and SaMD used in medical devices. Furthermore, any entity using a tool manufactured or developed by a third party must thoroughly vet that tool’s compliance measures to avoid HHS prosecution.
Plus, monitoring should be ongoing. Clinics purchasing a new tool must consider whether they have the proper resources and training to identify and mitigate risks and actively prevent discriminatory outcomes.
How Does the New Regulation Impact Patient Care?
The regulation applies to all organizations covered under existing healthcare laws—from smaller, single-practice medical offices to the largest hospital. Patients are now protected from discrimination resulting from misusing information gathered or reconstructed by an algorithm or software in all medical settings.
While this is excellent news for patients, algorithms are based on the information they are fed. The most accurate algorithms are based on the most extensive amounts of patient information (variables). This process becomes tricky following the input of sensitive variables. Take age as a variable, for example. If an algorithm is trained mainly on data from younger women, it may inaccurately predict that an older woman has low success rates. Predictions should depend on many other health factors.
For this reason, the HCPs using decision support tools must learn to identify and assess any potential biases that may lead to discriminatory outcomes. Then, they must take an active role in mitigating their effect. HCPs must be educated in using and assessing AI and ML technologies and look closely at their human prejudice as reflected in the algorithm.
Healthcare organizations can implement staff education through workshops, ongoing training, and policy development to manage data derived from algorithms. As with all forms of progress, education is critical. The more the HCPs become aware of the information they feed into AI tools and how data is processed, the more benefits they can reap from decision support tools.
Ensuring Ethical Practices in Healthcare Tech
At their core, health tech companies focus on improving healthcare and patient outcomes. The OCR rule has given health tech companies the mandate to find a means to enhance their offerings further by making them accountable for the information their algorithms produce.
The data itself, which is heterogeneous and collected from diverse populations in various settings, will be inclusive. The more inclusive the data, the less biased or discriminative the algorithm will likely be.
An essential feature of any tech model is ongoing retraining and calibration of models. Not only is there a need for regular adjustments and improvements to maintain accuracy and effectiveness, but such a model should remain current, accurate, and unbiased.
It is now incumbent upon all healthcare tech partners to be forthright in the data they compile. The most obvious way to accomplish this is by adopting transparency as a practice and working with third parties that use transparent algorithms. Once data and data sources are transparent, identifying and determining biases that can lead to discriminatory outcomes will be easier.
Ethical by Design: The commitment
In light of the new OCR rule, companies serious about compliance and ethical practices will view this rule as an opportunity. For AIVF, the OCR rule highlights our existing GDPR and HIPPA-compliant practices.
AIVF follows all the practices above and is dedicated to continuously developing its approach to meet current challenges. This continuous improvement process includes updating models, using calibration techniques, and leveraging the latest research.
We understand our responsibilities and are committed to ethical and safe practices in the fertility care field. We believe these practices help patients on a smoother, quicker, and more accessible path to parenthood.
The new OCR rule is a terrific step forward for both IVF clinics and patients. It ensures that the algorithms used to help make life-generating decisions are designed by specialists dedicated to the highest standards—as they should be. Visit our website to discover how AIVF uses AI to improve fertility care worldwide.
- Grosek, Štefan, et al. “The first nationwide study on facing and solving ethical dilemmas among healthcare professionals in Slovenia.” Plos one15.7 (2020): e0235509.
- Gerke, Sara, Timo Minssen, and Glenn Cohen. “Ethical and legal challenges of artificial intelligence-driven healthcare.” Artificial intelligence in healthcare. Academic Press, 2020. 295-336.