Mathematical model can improve depression diagnosis
Mathematical models can be used to develop new methods to diagnose e.g. depression, and the models can provide important information to help develop better forms of treatment.
It is becoming increasing common to use mathematics in healthcare. One of the researchers using the mathematical approach to diseases is Professor Johnny T. Ottesen, who, among other things, is researching the hormonal stress system, also called the HPA axis or simply the stress axis. This hormonal system plays a role in connection with e.g. stress, depression and certain mental illnesses.
Johnny T. Ottesen's mathematical models can potentially provide new ways to diagnose depression, among other things, and perhaps help to develop improved forms of treatment.
Approximately 10% of people experience a severe depression during their lifetime. Many people are affected. It is not only the individual person who suffers. Their family is also affected. These diseases are therefore serious ailments that we hardly know how to handle” explains the professor.
Normally, mental illness and depression are diagnosed by a psychiatrist who asks a number of questions. The answers help the psychiatrist to conduct a psychiatric assessment.
"There can be a high degree of subjectivity in these assessments, and misdiagnosis sometimes occurs as a result. Many researcher, therefore, have tried to identify more objective bio-markers. Cortisol, which is often referred to as the stress hormone, is one substance that has received a great deal of attention. A wide range of studies have been conducted, but so far there has been no success using it as a bio-marker” says Johnny T. Ottesen.
Fluctuations in cortisol levels
The HPA axis is a complex system, where the three substances CRH, ACTH and cortisol affect each other and initiate processes that affect several reactions in the body.
When cortisol is released into the body and enters the cells, it increases the level of activity in the cells. This is a necessary mechanism for how we deal with stress. If we are under pressure, it is helpful that the energy level rises. However, it can be detrimental if the energy level does not subside again. If the energy level remains high, it can lead to what we generally refer to as stress.
One of the reasons that it has not been possible to use cortisol as a bio-marker, is that cortisol is not a constant. For example, cortisol production is influenced by circadian rhythm, and many rapid changes also occur during a normal day.
A model with two substances
The fluctuations in cortisol levels mean that it is difficult to obtain a clear picture of the situation if we only look at cortisol. The same is true if we only look at another of the substances in the HPA axis, ACTH, which stimulates the production of cortisol.
"Then I got the idea that we should not look only at cortisol, but also at ACTH. We should look at both substances simultaneously. The method cannot say with 100% certainty whether you are depressed, but it should be able to say it with 95% certainty. There is a very low level of uncertainty, so in that way it's a very good bio-marker” explains Johnny T. Ottesen.
His work led to a statistical model that provides a clearer picture than if only one of the substances is considered.
Potentially better diagnosis and treatment
The statistical model is not the only mathematical model that Johnny T. Ottesen has developed within this area. If the healthcare system adopts the models, they could have several beneficial effects.
"It provides a sophisticated diagnostic tool, but it also provides an opportunity for differentiated treatment because it would allow increased possibilities to treat people individually. It can also offer the pharmaceutical industry the opportunity to develop products that specifically target the mechanism that needs to be remedied. This is a slightly more long-term perspective” estimates Johnny T. Ottesen.
Although the models offer new methods to assess depression and provide clear indications of the relationships, there is still some way to go before they can be used in practice. That will require large-scale studies where even more people are tested.
In addition, the models are only based on so-called pure groups, which means that people with other conditions that could affect e.g. cortisol production, have been left out. These include people with diabetes, children and drug addicts.
But it's incredibly expensive to conduct the large and representative surveys that are required, and if the pharmaceutical industry then has to produce a targeted drug, it will be even more expensive.
"When the pharmaceutical industry develops a successful pill, it costs an average of DKK 10-15 billion, typically over a 10-year period. So, it is expensive, very expensive, and it’s not easy to get access to such vast amounts” notes Johnny T. Ottesen.
Mathematics plays a major role
Although there is a need for larger studies to support Johnny T. Ottesen's models, mathematics has found its way into the field of healthcare. In order to have certain medications approved, some of the clinical studies can be replaced with mathematical models, and the USA and EU are beginning to impose requirements for mathematical models.
"In this way, the models are not only increasing in popularity, they are also playing an increasingly vital role in the medical world. This does not mean that models have not been used previously, but they have usually been used in more limited areas. Now the models are beginning to play a major role - mainly abroad, but I hope that in future we will also make greater use of them in Denmark too. It is about maintaining a leading position internationally in healthcare and within the pharmaceutical industry” explains Professor Johnny T. Ottesen.