Do chronic illnesses also increase administrative costs?

The premium of a health insurance plan must cover not only medical costs but also the insurer's administrative costs. Over the past decades, with improved data availability, health economics literature has thoroughly demonstrated that medical costs vary significantly at an individual level. However, a corresponding analysis of individual administrative costs is lacking. In theoretical analyses of health insurance, administrative costs are often assumed to be either constant per capita or not considered at all.

There are two main reasons why administrative costs have received little attention in research so far: First, it is inherently challenging to attribute specific components of administrative costs to individual policyholders. Many tasks performed by an insurer, such as calculating premiums or negotiating with healthcare providers, are not directly tied to individual insured persons. Nonetheless, it seems plausible that individuals who are frequently ill generate higher administrative costs than those who are generally healthy. This is because many tasks, such as billing, claims processing, and collections, are more likely to be performed for sick individuals. However, even in these cases, it can be difficult to accurately assign these costs to specific individuals. Second, there is often a lack of data. In many cases, only aggregate costs, such as those for infrastructure or personnel, are available, which do not provide insights into costs at the individual level.

We focused on one component of administrative costs that is not affected by the aforementioned challenges: customer interactions, which can be easily attributed to individual policyholders. Even from the descriptive analysis, interesting insights emerge. Gender plays a role only at younger ages. Women of childbearing age contact their insurance provider significantly more often than young men. However, at older ages, the frequency of contact is higher regardless of gender. We did not further investigate the reasons behind these differences in our study.

Our main question was whether individuals with chronic illnesses—who often incur high medical costs—also have more frequent interactions with their health insurance provider. To make a causal inference, we need a control group to show what would have happened to customer interactions if the person had not developed a chronic illness. We track the trajectory of customer interactions over two years for customers who do not develop a chronic illness, represented by the red line in the chart above. Additionally, we use two other control groups. The pink line shows the interaction trajectory of customers who incur no healthcare costs at all, capturing customer interactions unrelated to illness (e.g., address change requests). The dark blue line represents customers classified as chronically ill one year earlier than the main group of interest (light blue). Even these three curves alone reveal that customers with a chronic illness consistently have more interactions with their insurer. Compared to the dark blue curve, which includes customers classified as chronically ill at time 0, the assumption that this relationship is causal appears highly credible: Before developing a chronic illness, these customers have nearly the same number of interactions as those without a chronic illness. Interactions increase even before classification, peak at the time of classification, and remain at a high level thereafter.

To ensure the reliability of our results, we examined specific categories of customer interactions that are clearly linked to illnesses—such as inquiries from doctors or hospitals. For these types of interactions, the increase among chronically ill individuals was even more pronounced. In contrast, for requests unrelated to health conditions, such as address changes, there were no differences between groups. However, customer interactions account for only a portion of administrative costs. What about the remaining administrative costs? To investigate this, we analyzed aggregate administrative costs at the insurer level. The Federal Office of Public Health publishes annual administrative costs per insurer. As a measure of the health status of the insured population, we used risk equalization payments. Our analysis confirms the positive relationship observed earlier, even at the insurer level. In a related, already published paper, we also demonstrated that this relationship holds true in four other countries.

As the population continues to age, it’s likely that administrative expenses will rise. Yet despite this demographic trend, the share of administrative costs within total health insurance spending has actually declined in recent years. Our findings indicate that interactions between insurers and their customers are increasing—even when no illness is involved. For health insurers, this evolving landscape offers a promising opportunity to guide their customers more effectively and efficiently through the healthcare system, ultimately enhancing both care quality and patient experience.