Part 1: Lessons from Swedish treatment matrix implementation

The number of patients with obstructive sleep apnea (OSA) is rising and is expected to grow even more in the years ahead. As early as 2018, in Baveno, Italy, experts began discussing how to prioritize care for those with the most urgent need. Professor Ludger Grote was part of that working group. It was under his leadership when in 2021, Sweden became the first country to implement a national treatment prioritization matrix to guide patient care.¹

This interview shares insights from Sweden’s experience, which now informs efforts beyond its borders. The European Respiratory Society (ERS), under the leadership of Professor Winfried Randerath, has since proposed a similar matrix now being evaluated in clinical studies. Together, the Swedish experience and wider Baveno collaboration bring hope for a unified, simplified approach to OSA patient prioritization that ensures patients with the highest clinical need are treated first.

Q. In 2023, shortly after your paper on the Swedish treatment matrix was published in Diagnostics, we spoke with you about its early impact. Two years on, and four years after the initial implementation, how would you describe its reception among sleep physicians?

Professor Grote: We haven’t had hundreds of calls or emails. Interestingly, the reaction was more a lack of strong complaints. Over time, in talking with colleagues, I realized many physicians really appreciated having some kind of guidance on how to view sleep apnea. This moderate response may reflect that colleagues were already working in line with the matrix’s principles. So, for many, it was welcomed as confirmation of their clinical practice.

We also continue to see steady interest. Every month, about 100 people download our guideline paper and management plan from the website of our quality registry, SESAR, which adds up to around 1,200 downloads per year. We assume that it’s physicians and nurses entering the sleep field who are seeking it out.

Q. What aspects of the treatment matrix have worked well, and where have you seen challenges in implementation?

Professor Grote: Implementation did take time. The matrix structure has several levels, so people needed to familiarize themselves with it. Also, the visualization with green, yellow and red colors is not always immediately intuitive until you read the instructions.

Yet, my impression is that more and more colleagues, physicians and specialized nurses alike, now find it a helpful instrument. For example, rotating pulmonary physicians in our lab, who spend three to six months here, now keep the matrix permanently in their workspaces as standard reference material. For more experienced colleagues, the matrix provided reassurance. For younger colleagues, it helped them see the complexity of OSA without getting lost.

One key change for many using the matrix was realizing that patients with moderate to high AHI do not automatically need treatment. New insights on primary and secondary prevention effects of OSA treatment in the elderly reveal that we sometimes put pressure on people to use treatments that they cannot expect any major benefit from.

Q. What would you say is the main strength or benefit of the treatment matrix?

Professor Grote: It shows that diagnosis is relatively easy and helps in deciding on clinical relevance. The matrix isn’t rigid. It allows room for interpretation, requires talking to the patient to identify treatable traits and makes clear that not all OSA events must be treated.

Q. What limitations would you like to address in future versions?

Professor Grote: A major shortcoming is that we still use AHI as a central metric. Yes, AHI is a weak parameter for describing disease, but it still provides one important dimension: the frequency of events. So, I think it was a correct decision to keep AHI in it.

Secondly, in the matrix, symptoms are reduced to “yes” or “no,” but, in reality, are varied and require knowledge beyond what is in the matrix. We therefore added a separate table covering the broader OSA symptom spectrum. When it comes to risks and comorbidities, some aspects, such as traffic accident risk, are not specified, though they are clearly important. These have to be known by the treating physicians or healthcare personnel.

Psychiatric comorbidities such as depression and anxiety are also not explicitly mentioned, even though there is data showing sleep apnea treatment can help in these cases or even aid in prevention. The evidence here needs to be revisited so we can provide clearer guidance.

Q. The European Baveno group, which you are a part of, recently published a matrix similar to the Swedish one. How do the two approaches compare?

Professor Grote: Our Swedish matrix was influenced by the original 2018 Baveno work, which classified by symptoms and comorbidities. We added more dimensions and introduced a color scheme.

The original Baveno matrix did not include AHI, but the updated version then added different AHI thresholds, mainly the AHI ≥30 and a cardiovascular risk score. The Swedish matrix still uses the more traditional three AHI categories from 5 up to ≥30.

I think a main difference is that in the Swedish matrix, we would probably treat fewer patients compared to the modified Baveno classification. And the Swedish matrix also does not require all the detailed information needed for the SCORE risk matrix. A systematic comparison of these two matrices has not yet been performed, though.

Q. Could the Swedish matrix eventually be replaced by the Baveno version once it is finalized?

Professor Grote: We have not considered this yet. The Baveno matrix, for example, requires information on cardiovascular risk factors, including blood lipid assessments. This implies that the referring center must have already completed that work, but we do not believe that all our sleep centers would routinely take blood samples.

However, from the clinical perspective, if evidence shows Baveno is better and works well, we would, of course, replace ours with it.

Q. Oximetry-based measures like hypoxic burden, delta heart rate and pulse wave amplitude drops are gaining attention. In our last interview two years ago, you indicated it was premature to bring these measures into the picture. Do you see them being integrated into the matrix soon?

Professor Grote: I think we are getting closer, but we are not there yet. Many groups are working on normative values for these measures. The calculations, at least for hypoxic measures, are likely to be built into standard clinical equipment and software in the future, but that has not been fully realized as of today.

For example, we have had measures like time below 90% oxygen saturation (T-90) for quite a while. This “area under the curve” concept is not new, but T-90 has never replaced our thinking completely, probably because “below 90 percent” is too broad. Many patients with good lung function will not reach that threshold in a significant way yet still have substantial sleep apnea.

Pulse wave parameters are an area where I personally do a lot of research. Evidence for their predictive value has grown significantly in recent years, and I believe that incorporating more cardiovascular-oriented parameters during sleep will help us tailor treatment decisions more effectively in the future.

But right now, we still lack complete normative values and robust evidence. We are currently in the process of updating our guidelines and have decided it is not yet time to formally include these metrics in the matrix.

Q. Under your leadership, the guideline group has been impressively active over the last decade. It has published diagnostic guidelines (2018), treatment guidelines (2021) and management plan guidelines (2023). What are your next priorities?

Professor Grote: We are currently updating our diagnostic guidelines, and our main goal is to adapt them into relevant knowledge that helps primary care providers bring patients suspected of sleep apnea to the right diagnosis.

More specifically, we are working on condensing the 100-page guideline into a concise, digestible format that’s accessible for primary care. We also want to focus on telemedicine recommendations and on how to deal with co-morbid insomnia and sleep apnea (COMISA), treatment-emergent central sleep apnea (TE-CSA) and positional-dependent obstructive sleep apnea (POSA).