Why did my life insurance application get flagged after a phone health scan?

Completing a health scan on a smartphone for a life insurance application is a fast, convenient process. But for some applicants, the result can be confusing: a flagged application, a request for more information, or a premium quote that is higher than expected. This outcome often leaves consumers wondering what just happened and whether the assessment was fair. The increasing use of new data sources from technology like a phone health scan in insurance underwriting requires a closer look at how these tools work, how they are regulated, and where the line is drawn between assessing risk and creating unfair outcomes.

According to a recent NAIC survey, 84% of health insurers are already using artificial intelligence (AI) and machine learning (ML), tools that power these new underwriting methods.

How phone scans create data for underwriting

When you use a smartphone camera for a health scan, it's typically using a technology called remote photoplethysmography (rPPG). This technique analyzes subtle changes in the color of the light reflected from your skin to estimate cardiovascular information, such as heart rate and respiratory rate. This digital data becomes a new input into a life insurer's underwriting models. The core of the issue is not the scan itself, but how its data is integrated into the complex world of phone health scan insurance underwriting.

Historically, underwriting has relied on a standard set of inputs: medical questionnaires, fluid-based tests (blood and urine), and paramedical exams. An algorithm using data from a phone scan is, in principle, doing the same thing: correlating specific data points with established mortality and morbidity tables to classify an applicant into a risk pool. The key difference is the nature of the data and the method of its collection and analysis, which is faster, more immediate, and algorithmically driven. This shift demands a new level of scrutiny to ensure the outputs are both accurate and fair.

Feature	Traditional Underwriting Data	Digital Phone Health Scan Data
Data Sources	Paramedical exams, lab tests, medical records	Smartphone camera (rPPG), user-reported data
Collection Method	In-person appointment with a medical professional	Remote, self-administered scan via an app
Immediacy	Slow; results can take days or weeks	Fast; data is generated in minutes
Data Type	Static, point-in-time measurements	Dynamic, real-time physiological estimates
Regulatory Framework	Well-established, based on decades of practice	Emerging; subject to new AI and data privacy rules

Unfair discrimination vs. risk-based pricing

The fundamental principle of insurance is risk-based pricing-the idea that the premium should reflect the level of risk. Individuals with characteristics that statistically correlate to longer, healthier lives pay lower premiums than those with characteristics suggesting higher risk. This is not, by itself, unfair discrimination.

The regulatory concern arises when new technologies, like algorithms analyzing phone scan data, inadvertently introduce bias. This can happen in several ways:

• Proxy Discrimination: The algorithm might use a data point that is not itself a protected characteristic (like race or gender) but is so closely correlated with one that it effectively discriminates. For example, if a model found a spurious correlation between a video background element more common in a specific demographic and a negative health outcome.
• Algorithmic Bias: The model may be trained on historical data that contains implicit biases, which the algorithm then learns and perpetuates at scale.
• Lack of Transparency: If the model is a "black box," it can be impossible for a carrier to explain to a regulator or a consumer exactly why a specific decision was made, undermining trust and accountability.

Regulators are focused on ensuring that any data used for underwriting, regardless of its source, has a clear, scientifically validated connection to risk-a concept known as "actuarial justification."

Industry applications and regulatory oversight

As insurers adopt these technologies, they are coming under increasing regulatory pressure to prove their models are fair, transparent, and accountable. The National Association of Insurance Commissioners (NAIC) is leading this effort.

NAIC Model Bulletin on AI

In 2023, the NAIC adopted its "Model Bulletin on the Use of Artificial Intelligence Systems by Insurers." This provides guidance for insurers to establish a formal program for using AI. The principles are often summarized by the acronym FACTS:

• Fair and Ethical: Systems should be designed and used in a way that respects consumer rights and avoids unfair discrimination.
• Accountable: Insurers are responsible for the outcomes produced by their AI systems, even if those systems are developed by a third-party vendor.
• Compliant: AI systems must comply with all existing insurance laws and regulations.
• Transparent: Insurers must be able to explain how their AI systems work to regulators and, in some contexts, to consumers.
• Secure and Robust: Systems must be protected from cyber threats and perform reliably.

The importance of auditable systems

For chief medical officers and compliance leaders, this means that adopting phone health scan technology is not just a technology decision-it is a regulatory and governance challenge. The ability to document and defend underwriting decisions made by an algorithm is now a core requirement. Carriers must be able to demonstrate to regulators that their models are based on sound science and are free from prohibited biases.

Current research and evidence

The scientific validity of phone-based health scans is an active area of research. A 2022 review published by the National Center for Biotechnology Information (NCBI) titled "Vital Signs Monitoring Using Smartphones" confirmed the potential of using smartphone cameras and other sensors to measure parameters like heart rate, respiratory rate, and even blood pressure. However, researchers (Islam, et al., 2022) note that accuracy can be influenced by factors like lighting conditions, skin tone, and user movement. For insurance underwriting, this means that the raw data from a scan must be processed through models that can account for these variables and have been validated against traditional medical measurements.

The future of phone health scans in insurance

The use of phone health scans in insurance underwriting is set to grow. As the technology's accuracy improves and regulatory frameworks become more established, these scans will offer a convenient and efficient alternative to traditional methods for many applicants. The focus will increasingly shift toward the governance and oversight of these systems. Carriers that can demonstrate a "compliance-first" approach, with robust data governance, model risk management, and transparent documentation, will be best positioned to innovate responsibly and build trust with both consumers and regulators.

Frequently asked questions

Q: Is a phone health scan as accurate as a visit from a nurse? A: Phone-based scans using rPPG are designed to estimate vital signs and are becoming increasingly accurate. However, their results can be affected by factors like lighting and user movement. They are generally used for risk classification and are not intended to replace a clinical diagnosis. Insurers must validate their specific tools against traditional measurements to ensure they are appropriate for underwriting.

Q: Can I refuse to do a phone health scan for my insurance application? A: Yes, applicants typically have a choice. Insurance companies are required to offer alternative paths for underwriting, such as a traditional paramedical exam or providing medical records. You should be able to choose the option you are most comfortable with.

Q: How can I be sure the algorithm used for my application was fair? A: Insurance regulators, guided by bodies like the NAIC, are requiring carriers to develop and maintain programs that test for and prevent unfair bias in their algorithms. Insurers must be able to demonstrate to regulators that their models are actuarially justified and do not result in illegal discrimination.

The challenge of deploying new underwriting technologies in a way that is both innovative and compliant is significant. For carriers and their compliance leaders, building a defensible, audit-ready program is not just a best practice but a business necessity. At Circadify, we are focused on providing the frameworks and tools to help the insurance industry navigate these complex regulations with confidence. To learn more about building compliant digital underwriting programs, explore our Compliance guides + regulatory insights.