The findings of this study, which is based on the assessment of representative sample data of the German population, in the year 2019, showed that two per 100,000 adults fulfilled the criteria of diagnosis codes that could represent TIO, with 0.187 per 100,000 meeting more stringent operational criteria, including medications, advanced imaging, or tumor removal. Accurate identification of TIO cases in epidemiological studies is challenging because of the lack of a specific ICD-10 code. However, compared with the scarce data on TIO prevalence and incidence rates, our results fit within the expected ranges [20, 21]. In contrast to prior studies, we were able to focus on patients with a benign tumor diagnosis and we demonstrated good comparability with previously published epidemiological figures in the range of 0.1–0.47 per 100,000 and confirmed, as expected, a low number of cases. When we focused on group B + C, which according to our perspective, is probably the most realistic TIO population because it fulfilled at least one of the criteria of imaging, medication, or tumor removal, the 1 year prevalence and incidence rate (2019) were 0.187 per 100,000 persons and < 0.094 per 100,000 person years. If we had focused on patients with mesenchymal tumors, which can be, depending on the group, approx. 65% of the whole benign tumor population in the study, the range would have been even lower. However, the challenge in using PMT data is that a TIO diagnosis is often made many months before the causative tumor has been identified, removed, and classified as benign.

The patient characteristics are partly comparable with descriptions from available publications, insofar as this can be said for ultra-rare diseases with extremely low numbers of patients. The age distribution in the sixth decade of life reflects the expected ranges based on several case reports [10, 16, 21]. The age distribution of men in this study, [average age of 47.7 years (SD 25.5)] was consistent with other reports (Rendina et al.: 45.0; 39.0–54.0); however, in our study’s population, women were significantly older [average age of 57.2 years (SD 25.5)] than the women in the study by Rendina et al. [average age of 44.0 years (39.0–56.0)] [16]. In our analysis, there was a higher proportion of women to men (78:22), which was consistent with the Danish registry study with a women proportion of 60% [21]; however, the recent report by Rendina et al. states that, in contrast to the previous study, TIO occurs preferentially in adult men (55:45). Similar results were reported by Bosman et al., with 60% male population [10, 16]. These differences most likely result from the different ways in which the data were collected rather than a true difference in sex predilection between countries. The patients’ high burden of disease can be shown using TIO-associated comorbidities and clinical consequences. Patients suffered from bone and back pains, osteoporosis, and (pseudo-) fractures. The importance of pain has also been demonstrated in recent publications [16]. Up to 80% of the patients had at least one sick leave day, with an average of 40 to 300 sick leave days. In comparison, published sick leave days for the German population on average are 10.9 sick leave days per year, showing that TIO patients on sick leave took a longer time to recover [38]. More than 50% of patients had at least one prescription of analgesia, with an average daily dose of approx. 20 DDD per patient. Compared to the published analgesia doses of the German population (average of 2.6–8.8 DDD (N02/M01) per year, independent of age, sex, and morbidity) [39], our results showed that the DDD for TIO patients is clearly above this, and this factor should therefore be considered more closely in further analyses.

Strength and limitations of the SHI claims data analysis: Our findings must be interpreted in line with some limitations inherent to claims data. First, there was no specific code for TIO or the presence of a phosphaturic mesenchymal tumor (PMT). M83.8 is defined as other osteomalacia in adults and includes more than oncogenic osteomalacia or TIO. In patients aged  < 18 years, other codes must be used, which, in combination with the defined criteria, could indicate TIO. Accordingly, as with all rare diseases that lack specific coding, operational assumptions must be made. Second, the definition of tumor removal was a combination of predefined procedural codes (Supplemental material S1), which, according to expert opinion, could indicate the removal of a PMT, and an explorative analysis using rankings was made. It is possible that not all the tumor removal codes were identified. Moreover, information regarding successful tumor removal was not available in the claims data. The prescriptions of tumor-specific medications and laboratory examinations were documented only in the outpatient sector. If patients received low-price medication and laboratory examinations in an inpatient setting, the costs (lump sum) were included in the DRG. Furthermore, the results of the laboratory examinations could not be retrieved from the available administrative sources. Also, special tests such as FGF23 or 1,25-(OH)2D, though highly informative for TIO work-up, may not be accessible for every clinician seeing patients with hypophosphatemia. Within our observational period, advanced imaging was required, which in some cases needed approval from the insurance company or an individual funding request. Therefore, the number of patients who underwent these examinations might have been underestimated. Despite these limitations, this analysis had enormous strengths, and healthcare research studies with SHI claims data have a high level of acceptance in the healthcare system. The strength of the dataset results from the robustness of the data; hence, it is representative of the German population. This is supported by the results of patient characteristics and epidemiological key figures, which showed consistent results with the literature and expected ranges, despite a higher proportion of women. Approx. 70% of the patients (group A) received a specific laboratory examination and only 4% of the patients had at least one defined genetic test. These results also underpin robustness, because not each patient receives these specific examinations and test, therefore a further division into the defined subgroups was necessary.

The Danish study fortunately provided us with a template and shows that both claims and register data show a good comparability for both European countries. Adjusted for the German healthcare sector and the available dataset, we gain more insights into patient characteristics and were able to further improve subgroup identification through more available specific definable surrogates.

The challenges facing prescribers regarding the suboptimal coding system will probably end with the beginning of the obligatory coding system, alpha ID for orphans [41], in 2023 for the inpatient sector in Germany. The outpatient sector will follow at some point. In this study, epidemiological insights on German patients with TIO were obtained for the first time. We would very much welcome international efforts to facilitate the coding of TIO as a specific disease using other coding systems. It is recommended that this work be updated at regular intervals despite possible COVID-19 pandemic bias. In addition, it would be desirable to conduct more patients` care research studies that will provide deeper insights into the care reality of these patients. It is important to increase awareness of this rare disease to decrease its diagnostic delay and, hence, the loss of quality of life of patients with TIO.