跳到正文
原文
Nature 健康护理主题·· 9 小时前精选AI 评分75

德国门诊低价值甲状腺功能检测分布与集中度

Physician-level variation and concentration of low-value thyroid function testing in german ambulatory care

AI 导读

研究发现,低价值甲状腺功能检测在德国医生中高度集中,约30%的医生贡献了80%的低价值检测。GPs贡献了超过80%的低价值检测。该研究为减少不必要的检测提供了数据支持。

推荐理由

研究揭示了德国低价值甲状腺功能检测的分布和集中度,为减少不必要的检测提供了数据支持。

正文 · 原文

Abstract

According to recent guidelines, the measurement of free triiodothyronine (fT3) and free thyroxine (fT4) in the routine monitoring of patients with known hypothyroidism does not offer added value beyond thyrotropin (TSH) testing and is considered low-value care. Understanding variation in low-value testing is essential for informing targeted de-implementation strategies. This study examines physician-level variation and concentration in low-value thyroid function testing in Germany using nationwide ambulatory statutory health insurance data from 2020, comprising 48.2 million cases of common blood tests in 22.2 million patients. We analysed ordering patterns of thyroid function tests (TSH, fT3, fT4) across specialties and individual physicians and assessed the distribution of testing volumes to identify concentration among high-utilising physicians. In total, 12.8 million patients received thyroid function tests (19 million cases), most commonly single TSH measurements (74%) and complete thyroid panels (17%). Among patients with hypothyroidism, 1.3 million TFTs were classified as low-value (0.9 million patients, 84% female), with 80% ordered by general practitioners. Low-value testing was highly concentrated: 27% of physicians accounted for 80% of all low-value thyroid function tests. Higher physician-level volumes of low-value testing were associated with an increased use of complete thyroid panels. These findings indicate substantial unwarranted variation in care and suggest that interventions targeting high-utilising physicians may be particularly effective in reducing unnecessary testing and improving the efficiency of healthcare delivery. The analytic approach used in this study can be applied to administrative data in other healthcare systems.

Introduction

Thyroid function tests (TFTs) rank among the most frequently performed laboratory investigations in many European countries and the United States1,2,3. By quantifying thyroid hormone concentrations, these tests assist in detecting functional abnormalities such as hypothyroidism or hyperthyroidism. Despite their clinical utility, evidence suggests that TFTs are often performed without a clinical indication according to guidelines. As a result they have become a prominent example of overused diagnostic services4,5,6,7. Overuse of diagnostic testing may cause patient harm through avoidable blood sampling, contribute to diagnostic cascades following incidental findings, and impose a preventable financial burden on health systems8,9,10.

Choosing Wisely (CW) initiatives and clinical guidelines across several countries recommend restricting common thyroid testing in patients with hypothyroidism to the measurement of the thyroid-stimulating hormone (TSH) and advise against routine ordering of triiodothyronine (fT3) or free thyroxine (fT4) tests11,12,13,14. Nevertheless, studies from various healthcare settings report that between 10% and 6% of TTs are low-value care, defined as testing with limited expected clinical benefit that does not outweigh costs and potential harms4,5,6.

Thyroid function testing in patients with hypothyroidism often involves complete panels, with fT3 and/or fT4 frequently being ordered alongside TSH4,5,15. In addition, there is substantial variation in the use of low-value TFTs and in the test profiles ordered across physicians and specialties. Higher testing volumes were associated with a greater proportion of results within the reference range, indicating potential overuse16. While existing studies quantify the extent of TFTs and the proportion of low-value tests, comparatively little is known about how such testing is distributed across individual physicians and specialties17,18. In particular, large-scale analyses that investigate whether low-value TFTs are concentrated among a subset of physicians remain scarce. Understanding this variation is relevant for the design of tailored de-implementation strategies, as interventions adapted to specific target groups are likely to be more effective.

Consistent with prior research we hypothesise that medical specialties and individual physicians vary in their frequency of ordering low-value TFTs and that a subset of physicians accounts for a disproportionate share of these tests. Identifying such variation may inform targeted practice-level interventions that aim to reduce low-value thyroid testing. We therefore address the following two research questions:

  1. 1.

    How do low-value thyroid function test ordering patterns in patients with hypothyroidism differ across medical specialties in Germany?

  2. 2.

    How are low-value thyroid function tests distributed among physicians, and to what extent are they concentrated within a group of high-utilising physicians?

Methods

Description of the dataset

In 2020, approximately 73 million persons were insured under the statutory health insurance in Germany (88% of the German population19). We utilised statutory health insurance claims data for this population provided by the Central Research Institute of Ambulatory Health Care in Germany (Zentralinstitut für die kassenärztliche Versorgung, Zi). The Zi database pools the nationwide ambulatory billing information of the 17 regional Associations of Statutory Health Insurance Physicians (Kassenärztliche Vereinigungen, KV). This database includes all persons within statutory health insurance who had at least one contact with the ambulatory health system during the observation period. For our analysis, data from 2020 were used, and information from 2019 was available to assess eligibility. Data was available for 15 of the 17 KVs, covering approximately 84% o ambulatory statutory health insurance patients (Bavaria and Bremen excluded19).

Ambulatory claims data in German statutory health insurance are documented at the case level. Each case includes all services billed for an insured person within a calendar quarter, together with the corresponding diagnoses and provider and patient information. Patients in our dataset may appear in multiple cases across different calendar quarters or with different physicians. Ambulatory services are coded according to the Physicians’ fee schedule (EBM20) with individual services identified by their specific fee schedule item (GOP). Diagnoses are coded using the German Modification International Classification of Diseases (ICD-10-GM21) and are available at a quarterly level. Physicians are identified by their lifelong physician identifier (LANR), indicating specialty. Patient information includes age, sex, and county of residence.

We included cases involving billing for fT3, fT4, TSH tests (GOP 32321, 32320, 32101), or one of the 30 most common blood tests, which accounted for approximately 75% of all blood tests (supplementary file, Sect. 1). To ensure computational feasibility given the size of the dataset, analyses of concomitant testing were restricted to these common blood tests. Data on total T3 and T4 measurements were not available as their reimbursement eligibility was discontinued in 200722.

The initial dataset comprised 75.4 million cases of TFTs and blood tests in 28.2 million patients, which we restricted to 69.3 million cases in 26.2 million adults (aged 20 years or older). Within these, 65.6 million cases in 25.2 million patients had complete information on age, sex and county of residence. We then retained only cases in which the ordering physician could be uniquely matched by the LANR. Records with pooled or proxy identifiers, unknown specialty or without a unique match were excluded. After applying these physician criteria, 48.2 million cases in 22.2 million patients were included in the analysis, corresponding to 85% of eligible adults and 70% of cases. Thus, the final set included data on TFTs and common blood tests for approximately 70% of statutorily insured adults in 2020 in Germany.

Operationalisation of low-value thyroid function tests

Low-value thyroid function testing was defined as fT3 and/or fT4 testing in adults with known hypothyroidism. Hypothyroidism in a patient was considered known if in addition to the quarter under review for low-value TFTs, hypothyroidism was diagnosed in one of the preceding four quarters (ICD-10-GM E03 or E89.0, confirmed diagnoses).

This definition follows international recommendations on low-value thyroid diagnostics issued by the CW initiatives in Australia, Canada and USA, which are based on evidence-based guidelines developed by Garber et al. (commissioned by the American Association of Clinical Endocrinologists in association with American Thyroid Association11,12,14,23). Building on these recommendations and related work by Schwartz et al.24, we developed and validated the low-value care indicator for fT3/fT4 testing in German statutory health insurance claims data, using an expert delphi panel25.

Data analysis

For the first research question, we described the distribution and frequency of TFTs by quantifying absolute counts and testing profiles defined as one of the seven possible combinations of the three TFTs within a single patient case. We then examined testing patterns in patients with hypothyroidism, including low-value fT3/fT4 testing by age and gender, and analysed these patterns by the medical specialty ordering the TFT profiles. We limited comparison by specialty to the six most relevant specialties with at least 10,000 patients with hypothyroidism who received low-value TFTs in 2020, covering 95% of such patients.

For the second research question, we examined variation in low-value TFT ordering at physician level based on the LANR associated with each case, allowing patients to be attributed to multiple physicians where applicable. For every physician, we calculated the proportion of patients with hypothyroidism who received low-value TFTs, using as the denominator all patients with hypothyroidism attributed to the respective physician in 2020. The cumulative distribution of low-value testing by physicians treating patients with hypothyroidism was assessed using pareto-type analyses, estimating the proportion of physicians responsible for 50% and 80% of all patients with hypothyroidism who received low-value TFTs26,27. Physicians were primarily ranked by their absolute number of patients receiving low-value TFTs to reflect the potential impact of targeted de-implementation. As a complementary analysis, we additionally ranked physicians by their individual proportion of patients with hypothyroidism receiving low-value TFTs, which is more sensitive to small denominators among physicians treating only few eligible patients. Additional analyses compared TFT profile distributions between physicians with high and low levels of low-value test utilisation to determine whether their ordering patterns differed with respect to the types of TFT panels selected, rather than solely in the number of patients tested.

Data handling and ethics

All analyses were performed using SQL (Dremio Version 3.1.128 and R (Version 4.4.229). We adhered to the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) reporting guidelines and followed the framework for Descriptive Epidemiology (RECORD Checklist available in supplementary file, Section 230,31. All methods were carried out in accordance with relevant guidelines and regulations. The study was based on de-identified statutory health insurance claims data collected for administrative purposes in Germany. Use of the data for the present study was based on permissions (application § 75 SGB X) of the governing legal supervisory authority of each participating KV. As no direct patient contact occurred and the data did not allow identification of individual patients by the researchers, ethical approval and informed consent were not required in accordance with applicable regulations governing the use of secondary data.

Results

In 2020, a total of 12.8 million patients (58% of the 22.2 million patients included) received at least one thyroid function test, accounting for 19.1 million TFT cases. Of these cases, 15.1 million (79%) were performed alongside other common blood tests. Among all ambulatory laboratory services, TSH was the fifth most frequently billed test overall, while fT4 and fT3 ranked 24th and 27th, respectively. 96% of all TFT cases comprised either a single TSH test (74%), a complete thyroid panel (fT3 + fT4 + TSH, 17%), or a combination of the free hormones (fT3 + fT4, 5%).

Within this overall testing volume, a substantial proportion of TFTs was performed in patients with known hypothyroidism. A total of 4.3 million patients in our dataset met the criterion for known hypothyroidism. Among these patients, 2.2 million received 3.7 million TFTs. Of these tests, 1.3 million TFTs in 900,000 patients included measurements of fT3 or fT4 and were classified as low-value care according to our indicator definition. TFTs containing fT3 or fT4 measurements accounted for 26% of cases in the general population and 34% among patients with hypothyroidism (Fig. 1). Low-value testing was predominantly observed in women, who accounted for 80% of all identified patients with hypothyroidism and 84% of the low-value tests, consistent with the female predominance of hypothyroidism. The proportion of patients with hypothyroidism receiving low-value tests declined with increasing age (supplementary file, Section 3).

Fig. 1

Thyroid profiles. Distribution of thyroid profiles in the general population and among patients with known hypothyroidism. Abbreviations: fT3 free triiodothyronine, fT4 free thyroxine, TSH thyrotropin.

Most TFTs were ordered by general practitioners (GPs), who accounted for 81.3% of all included TFTs. The remaining orders were distributed across other specialties, including gynaecology 3.4%, haematology/oncology 2.2%, internal medicine 1.8%, urology 1.7%, nuclear medicine 1.5%, and nephrology 1.5%. The residual TFT cases were attributable to specialties contributing less than 1% each. A similar distribution was observed for low-value TFTs in patients with hypothyroidism, which were predominantly ordered by GPs (82.1%), followed by nuclear medicine (5.1%), gynaecology (2.9%), and endocrinology (2%, supplementary file, Section​ 3).

Marked differences in low-value testing behaviour were observed across medical specialties. Among specialties treating at least 10,000 patients with hypothyroidism, specialists from endocrinology and nuclear medicine exhibited the highest shares of patients receiving low-value TFTs (87.5% and 81.7%, supplementary file, Section 3). In these two specialties, complete thyroid panels (fT3 + fT4 + TSH) were ordered most frequently, whereas the other included specialties predominantly requested TSH-only measurements (Fig. 2). Across all specialties, complete thyroid panels represented the most common type of combined TFTs.

Fig. 2

Thyroid profiles by specialty. Distribution of thyroid function test profiles ordered in patients with known hypothyroidism, stratified by medical specialty of the ordering physician. Included are specialties who ordered low-value thyroid function testing for more than 10,000 patients with hypothyroidism in 2020, sorted by the share of single TSH tests. Median number of TFT cases per physician: 61. Abbreviations: fT3 free triiodothyronine, fT4 free thyroxine, TSH thyrotropin.

Variation in low-value testing at the physician level differed substantially between specialties (Fig. 3). Specialists from endocrinology and nuclear medicine displayed the most homogeneous ordering behaviour, with low coefficients of variation (0.34 and 0.36), reflecting their overall high rates of low-value testing. The remaining specialties exhibited coefficients of variation above 0.8, indicating substantial heterogeneity in ordering practices.

Fig. 3

Boxplot. Physician-level variation in low-value thyroid function testing among patients with hypothyroidism, stratified by medical specialty of the ordering physician (LANR). Included are specialties ordering low-value TFTs for more than 10,000 patients with hypothyroidism and physicians treating more than 10 such patients in 2020 (97% of all low-value cases within the six specialties). Patient counts represent the sum of patients receiving low-value testing per physician. Specialties are sorted by the number of patients receiving low-value thyroid testing.

This heterogeneity was further reflected in the concentration of low-value testing among a limited number of physicians. Among physicians treating at least one patient with hypothyroidism in 2020 and ranked by the absolute number of patients receiving low-value TFTs, 27% ordered low-value TFTs for 80% of all patients receiving such tests. 50% of all patients receiving low-value TFTs had these tests ordered by only 9% of physicians. Within general practice, 31% of GPs ordered low-value TFTs for 80% of affected patients, whereas 11% did so for 50% (Fig. ​4). When physicians were instead ranked by their individual share of patients receiving low-value TFTs, 25% accounted for 50% and 53% for 80% of affected patients.

Fig. 4

Concentration curve. Concentration curve of low-value thyroid function testing among patients with hypothyroidism, showing the cumulative share of patients plotted against the cumulative share of physicians, stratified by medical specialty of the ordering physician (LANR). Patients may be attributed to multiple physicians if they received low-value testing from more than one physician. Shares based on all physicians who documented at least one patient with hypothyroidism in 2020.

Figure 5 visualises GP-level ordering behaviour by presenting the distribution of TFT profiles ordered by GPs in patients with hypothyroidism, stratified by the number of their patients with documented hypothyroidism receiving low-value care TFTs in 2020. GPs in the top 10% of this ranking, defined as the 10% of GPs ordering low-value TFTs for the highest absolute number of patients, ordered thyroid panels containing fT3 or fT4 4.4 times as often as the 70% of GPs with the lowest number of such patients. Distributions for other specialties are available in the supplementary file, Section​ 3.

Fig. 5

Test distribution. Distribution of thyroid function test profiles ordered by general practitioners, stratified by the number of their patients with hypothyroidism receiving low-value care in 2020. Included are all practitioners who documented at least one patient with hypothyroidism in 2020. Practitioners were ranked according to the number of their patients with hypothyroidism receiving low-value testing and grouped into three categories: top 10%, 10–30%, and bottom 70% of practitioners. As multiple practitioners could share the same cut-off value, the exact number of practitioners included in each group varies slightly. Abbreviations: fT3 free triiodothyronine, fT4 free thyroxine, TSH thyrotropin.

Discussion

Using nationwide ambulatory claims data, we found that low-value thyroid function testing in patients with hypothyroidism is common and unevenly distributed across physicians. Approximately one third of TFTs in this population were classified as low-value care. This low-value testing was highly concentrated: just below 30% of physicians accounted for 80% of patients receiving low-value TFTs. The extent of this skewness is unlikely to be explained by patient-level variation alone, suggesting that low-value thyroid testing is driven by a relatively small subset of physicians that represent a potential target for focused quality improvement and de-implementation efforts.

The majority of laboratory orders in this study provided value to patients. However, 34% of TFTs in patients with hypothyroidism were classified as low-value care according to our definition. This prevalence is consistent with our previous work in a subset of German statutory health insurance6​. One of the central findings of this study is the pronounced concentration of the identified low-value thyroid testing among a small subset of physicians. This concentration was less pronounced when physicians were ranked by their individual low-value testing share, highlighting the complementary distinction between relative testing propensity and absolute potential for reducing low-value testing. Importantly, the combined perspective shows that physicians with high absolute volumes of low-value testing also tended to exhibit less selective ordering patterns. Among GPs, the 10% f practitioners with the highest number of patients receiving low-value TFTs accounted for more than half of all low-value TFTs ordered by GPs and ordered thyroid panels containing fT3 or fT4 4.4 times as often as the 70% f GPs with the lowest number of such patients. These GPs also had approximately 4.3 times the overall TFT volume per physician. Combined, the higher testing volume and greater use of low-value profiles resulted in an approximately 19-fold higher number of low-value cases per physician. This combination of high testing volume and a high propensity for low-value testing is particularly relevant for de-implementation, as even modest changes in ordering behaviour among these physicians could affect a substantial number of low-value tests. GPs are particularly relevant for interventions aimed at reducing low-value TFTs because they account for over 80% f all low-value TFTs. Other specialties showed comparable patterns. These findings align with prior work demonstrating between-practice variation in thyroid testing. Vaidya et al. reported a wide range in TSH testing rates and stated that over 99% f the variation was attributable to differences between practices rather than sampling variability. Even after adjusting for hypothyroidism prevalence and deprivation, only 23.8% f this variation was explained16. Several mechanisms may plausibly contribute to the comparatively high absolute volume of potentially low-value TFT ordering observed in primary care. Thyroid testing has historically been perceived as a low-threshold and relatively low-risk investigation and may therefore be embedded in routine diagnostic templates or follow-up habits. Physician-level variation has also been documented in other areas of ambulatory care, including imaging for low back pain and antibiotic prescribing for respiratory tract infections32,33.

Another central finding of this study is that 62% of all low-value tests were ordered within complete thyroid panels (TSH + fT3 + fT4). This is in line with other research that reports the majority of fT3/fT4 testing being performed alongside TSH4,5,15. In routine clinical practice, requesting a complete thyroid panel is often perceived as the fastest way to obtain a broad biochemical overview. However, since additional parameters are often unnecessary and primarily increase the likelihood of incidental findings, requesting a panel can lead to repeat tests, additional consultations, referrals, treatment adjustments, and increased costs for the healthcare system. The dominance of complete profiles among low-value tests in our data may be particularly relevant, as measuring additional thyroid parameters without indication increases the opportunity for incidental abnormal findings and subsequent diagnostic cascades.

Our study offers a strong empirical basis for future research and contributes to the discussion of potential drivers of low-value thyroid testing in several ways. The variation in TFT ordering at the physician level suggests that physician behavior, practice organisation, and local norms play a major role in shaping testing patterns. The physician-level variation in TFT ordering may be explained by differing awareness and adherence to national policies, variability in computer system use, staff authorised to advise TFTs, practice protocols, risk aversion and training differences34,35,36,37. A large German survey by Wangler et al. showed that while most GPs consider evidence-based guidelines as beneficial, 28% repoted using them rarely and 10% neve38. Such gaps between guideline endorsement and routine use may contribute to the use of more extensive laboratory testing in situations where it is not supported by guidelines. Notably, physicians with high absolute counts of patients receiving low-value TFTs tend to rely disproportionately on thyroid test profiles that include low-value tests, suggesting that differences in test selection rather than patient volume drive this pattern. Perceptions of overuse may further reinforce such patterns: in a Canadian survey of inpatient general internal medicine clinicians, only 15% considerd themselves high utilisers of diagnostic tests, whereas 73% believedthat their peers ordered too many tests39.

The high share of low-value tests ordered as part of complete thyroid panels suggests an important role for system-level factors. In particular, differences in preconfigured electronic ordering systems may contribute to variation in the prevalence of low-value TFTs. Future research could explore their role in shaping testing behavior and assess whether modifications might reduce unnecessary thyroid function testing.

The comparatively high and homogeneous ordering rates observed in endocrinology and nuclear medicine may also reflect a combination of specific issues in referral cases, standardised diagnostic pathways, or standard laboratory protocols. Specialist care often involves treating patients with more complex or atypical thyroid disease progression patterns, which may justify more extensive examinations in selected cases. However, these mechanisms cannot be clarified on the basis of claims data.

Beyond system-level factors, societal influences and patient demands may also contribute35. An exemplary driver of patient demands is the availability of commercial market at-home TFTs directly to consumers, encouraging patients to request these tests from their physicians free of charge40,41 and social media posts promoting medical tests42. Improving health literacy and shared decision-making may therefore represent complementary strategies to reduce low-value testing. Clear communication about the limited incremental benefit of additional thyroid hormone measurements in stable hypothyroidism, as well as potential downstream consequences of broad testing, could help align expectations with evidence-based care.

Addressing low-value thyroid function testing requires that evidence is translated into action, including the unlearning of unsupported testing practices43. Interventions such as education, guidelines, or clinical decision support systems have demonstrated effectiveness in reducing inappropriate TFTs by more than 20% i several studies44,45,46. The results of this study may support interventions for reducing low-value care in several ways. A recent systematic review reported that most interventions target broad physician populations rather than those physicians who contribute most to inappropriate testing47. The identified pattern of disproportionate test ordering offers a clear basis for prioritising de-implementation efforts. A hypothetical reduction of testing rates in this group to the level observed among the low-ordering physicians would lead to substantial reductions in low-value testing overall. Given the high absolute number of low-value TFTs observed, even moderate reductions could translate into cost savings for the healthcare system, in addition to potential reductions in downstream diagnostic cascades. Interventions may follow a minimum principle, focusing on the few high-utilising physicians that account for the majority of low-value tests, or a maximum principle, aiming to reach the largest possible number of such physicians within economic limits. In both cases, targeting physicians with extensive low-value test volumes, most notably GPs, can substantially improve the cost–benefit ratio of intervention programs. This approach may also be applicable to other types of low-value care interventions.

Beyond such targeted interventions, financial incentives may provide another chance for reducing low-value testing. In German ambulatory care, physician services are generally remunerated on a fee-for-service basis within specialty-specific budget constraints, with reduced payments once defined thresholds are exceeded. Laboratory services are reimbursed to the laboratory physicians performing them, while ordering physicians are incentivised to limit laboratory expenditure through a laboratory efficiency bonus based on overall laboratory costs per case. However, this mechanism is not indication-specific.

Measures of low-value testing could potentially complement this approach by adding an indication-based quality dimension to predominantly volume- and cost-oriented reimbursement mechanisms. For example, future studies could evaluate, whether pay-for-performance approaches could make laboratory efficiency bonuses conditional on both overall laboratory expenditure and validated measures of test appropriateness, rather than volume alone. This could allow higher-volume but predominantly appropriate testing to remain eligible for incentives, while lower-volume testing with a high share of low-value use could still be penalized. Such use would require careful validation before being linked to reimbursement to avoid unintended incentives in laboratory ordering and coding behaviour.

Strengths and limitations

This study draws on a large, nationwide dataset of ambulatory statutory health insurance data, covering approximately two thirds of all statutorily insured adults in Germany and 19 million TFTs. Its extensive scope allows for detailed analyses of physician-level variation in test ordering across specialties and the approach may be utilised to inform low-value care de-implementation (phase two of the Choosing Wisely De-Implementation Framework48).

However, several limitations should be considered when interpreting the findings. First, the dataset provided by the Zi is limited to ambulatory claims data from statutory health insurance. Consequently, the diagnoses used in our operationalisation do not cover diagnoses in the hospital or other healthcare settings. Notably, previous research that included hospital diagnoses reported comparable prevalence estimates6. Claims data also do not include laboratory values themselves, which limits the assessment of whether additional hormone measurements were clinically triggered by abnormal TSH results.

Second, our operational definition of low-value tests is based on guideline recommendations and administrative data rather than detailed clinical information. In certain clinical situations (e.g. symptoms, suspected central hypothyroidism) measuring fT4 and/or fT3 may be clinically appropriate. These scenarios are not captured in claims data and may lead to a slight overestimation of low-value tests. Thus, our population-based classification of low-value testing should be interpreted as indicating potentially low-value testing rather than a definitive assessment of clinical appropriateness at the individual patient level. Conversely, missing codes for the diagnosis of hypothyroidism in primary care records may lead to an underestimation of low-value TFTs. Münch et al. analysed the quality of diagnostic documentation in German general practice records and reported partially incorrect recording, particularly underreporting, of thyroid disorders25,49.

Further limitations of the data relate to the completeness of all variables required for the descriptive analysis and results of laboratory requests. Based on physician linkage and completeness across variables, 15% of patients and 30% of potentially relevant cases were excluded due to missing information. Missing physician attribution may differ between specialties, for example due to different billing routes, which cannot be assessed with the available data, and could therefore affect absolute specialty-specific case volumes and the observed specialty distribution. By contrast, distortion of relative physician-level utilisation patterns would require systematic differential missingness of low-value and other TFTs within individual physicians, for which we have no indication. Although the dataset comprises a large and nationally representative sample of statutory health insurance data, the results may not be generalisable to the privately insured population or to health systems with different reimbursement structures and testing practices. Finally, the COVID-19 pandemic affected healthcare utilization in 2020 and has influenced absolute testing volumes, although there is no indication that it systematically affected TFT ordering patterns or the appropriateness of testing among patients with hypothyroidism.

Conclusions

This study identifies substantial physician-level variation in the ordering of potentially low-value TFTs. A small proportion of physicians accounts for a disproportionately large share of low-value test orders. This pattern is observable across specialties. GPs are particularly relevant for interventions aimed at reducing low-value TFTs due to their large patient volume and their contribution to the overall testing burden. Against the backdrop of limited resources available for improvement initiatives, our results provide a basis for prioritising physicians with the highest volumes of low-value testing. Targeting this subset may offer the greatest potential impact for interventions. The analyses of physician-level variation may also inform the design of strategies for reducing other forms of low-value care.

Data availability

The data that support the findings of this study are available from the Central Research Institute of Ambulatory Health Care in Germany (Zi), but access is subject to restrictions related to the use of claims data. Data access may be requested from the Zi via email at zi@zi.de.

Abbreviations

CW:

Choosing Wisely

fT3:

free triiodothyronine

fT4:

free thyroxine

GP:

General practitioner

KV:

Association of statutory health insurance physicians (Kassenärztliche Vereinigung)

LANR:

Lifelong physician identifier (Lebenslange Arztnummer)

TFT:

Thyroid function test

TSH:

Thyroid-stimulating hormone (Thyrotropin)

Zi:

Central research institute of ambulatory health care in Germany (Zentralinstitut für die kassenärztliche Versorgung)

References

  1. NICE, National Institute for health and care excellence. Thyroid disease: Assessment and management. NICE guideline. Reference number: NG145. (2023). Available at https://www.nice.org.uk/guidance/ng145/chapter/Context

  2. Markwart, R., Storch, J., Bleidorn, J. & Freytag, A. Utilisation and trends of laboratory testing and point-of-care testing in primary care in Germany: An analysis of claims data. BMJ open.14, e093780 https://doi.org/10.1136/bmjopen-2024-093780 (2024).

    Article  PubMed  PubMed Central  Google Scholar 

  3. Medicare Fee for Service for Parts A & B | CMS. (2025). Available at https://www.cms.gov/data-research/statistics-trends-and-reports/medicare-fee-for-service-parts-a-b

  4. Kluesner, J. K. et al. Analysis of current thyroid function test ordering practices. J. Eval Clin. Pract.24, 347–352 https://doi.org/10.1111/jep.12846 (2018).

    Article  PubMed  Google Scholar 

  5. Gill, J. et al. Evaluation of thyroid test utilization through analysis of population-level data. Clin. Chem. Lab. Med.55, 1898–1906 https://doi.org/10.1515/cclm-2016-1049 (2017).

    Article  CAS  PubMed  Google Scholar 

  6. Hildebrandt, M. et al. Quantifying low-value care in Germany: An observational study using statutory health insurance data from 2018 to 2021. Value in Healthhttps://doi.org/10.1016/j.jval.2024.10.3852 (2024).

    Article  PubMed  Google Scholar 

  7. Müskens, J. L. J. M., Kool, R. B., van Dulmen, S. A. & Westert, G. P. Overuse of diagnostic testing in healthcare: A systematic review. BMJ Qual. Saf.31, 54–63 https://doi.org/10.1136/bmjqs-2020-012576 (2022).

    Article  PubMed  Google Scholar 

  8. Piessens, V., Delvaux, N., Heytens, S. & Aertgeerts, B. an de Sutter. Downstream activities after laboratory testing in primary care: an exploratory outcome of the ELMO cluster randomised trial (Electronic Laboratory Medicine Ordering with evidence-based order sets in primary care). BMJ Open.12, e059261 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  9. Ganguli, I. et al. Prevalence and cost of care cascades after low-value preoperative electrocardiogram for cataract surgery in fee-for-service medicare beneficiaries. JAMA Intern. Med.179, 1211–1219 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  10. Hueber, S. et al. Consequences of early thyroid ultrasound on subsequent tests, morbidity and costs: An explorative analysis of routine health data from German ambulatory care. BMJ Open.13, e059016 (2023).

    Article  PubMed  PubMed Central  Google Scholar 

  11. Choosing Wisely Australia. The Endocrine Society of Australia. Recommendations. Available at https://www.choosingwisely.org.au/recommendations/esa5

  12. Choosing Wisely America - AAFP. Don’t order a total or free triiodothyronine (T3) level when assessing levothyroxine (T4) dose in hypothyroid patients. (2025). Available at https://www.aafp.org/pubs/afp/collections/choosing-wisely/142.html

  13. Schübel, J., Voigt, K. & Uebel, T. Elevated TSH values in primary care. DEGAM-Guideline. AWMF-Register-Nr. 053 – 046. (2023). Available at https://register.awmf.org/assets/guidelines/053_D_Ges_fuer_Allgemeinmedizin_und_Familienmedizin/053-046k-eng_S2k_Erhoehter-TSH-Wert-in-der-Hausarztpraxis_2023-07.pdf

  14. Choosing Wisely Canada. Endocrinology and Metabolism. Five Tests and Treatments to Question. (2023). Available at https://choosingwiselycanada.org/recommendation/endocrinology-and-metabolism/

  15. Gupta, S. et al. Are we using Thyroid Function Tests Appropriately? Indian J. Clin. biochemistry: IJCB. 26, 178–181 (2011).

    Article  CAS  Google Scholar 

  16. Vaidya, B. et al. Variability in thyroid function test requests across general practices in south-west England. Qual. Prim. Care. 21, 143–148 (2013).

    PubMed  Google Scholar 

  17. Kiel, S., Ittermann, T., Völzke, H., Chenot, J. F. & Angelow, A. Frequency of thyroid function tests and examinations in participants of a population-based study. BMC Health Serv. Res.20, 70 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  18. Thiyagarajan, A., Koenen, N., Ittermann, T., Völzke, H. & Haug, U. Trends in the use of thyroid diagnostics and treatments between 2008 and 2019 in Germany. Sci. Rep.14, 26710 (2024).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Bundesministerium für Gesundheit (BMG). Mitglieder und Versicherte der gesetzlichen Krankenversicherung (GKV). Statistik über Versicherte, gegliedert nach Status, Alter, Wohnort und Kassenart (Stichtag: 1. Juli des jeweiligen Jahres). Available at https://www.bundesgesundheitsministerium.de/themen/krankenversicherung/zahlen-und-fakten-zur-krankenversicherung/mitglieder-und-versicherte.html (2004–2024).

  20. Bundesministerium für Gesundheit (BMG). Einheitlicher Bewertungsmaßstab - EBM. Available at https://www.bundesgesundheitsministerium.de/service/begriffe-von-a-z/e/einheitlicher-bewertungsmassstab-ebm (2024).

  21. Bundesinstitut für Arzneimittel und Medizinprodukte (BfArm). ICD-10-GM. Internationale statistische Klassifikation der Krankheiten und verwandter Gesundheitsprobleme, German Modification. Available at https://www.bfarm.de/DE/Kodiersysteme/Klassifikationen/ICD/ICD-10-GM/_node.html (2024).

  22. Kassenärztliche Bundesvereinigung (KBV). Richtlinien zur Abrechnung: Regelungen und Hinweise zur Beauftragung und Abrechnung von Laboruntersuchungen. Laborkompendium. Available at https://www.kbv.de/html/82.php (2014).

  23. Garber, J. R. et al. Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Endocr. practice: official J. Am. Coll. Endocrinol. Am. Association Clin. Endocrinologists. 18, 988–1028 (2012).

    Article  Google Scholar 

  24. Schwartz, A. L., Landon, B. E., Elshaug, A. G., Chernew, M. E. & McWilliams, J. M. Measuring low-value care in Medicare. JAMA Intern. Med.174, 1067–1076 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  25. Pioch, C. et al. Selecting indicators for the measurement of low-value care using German claims data: A three-round modified Delphi panel. PLOS ONE. 20, e0314864 https://doi.org/10.1371/journal.pone.0314864 (2025).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Schwartz, A. L., Jena, A. B., Zaslavsky, A. M. & McWilliams, J. M. Analysis of Physician variation in provision of low-value services. JAMA Intern. Med.179, 16–25 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  27. Miyawaki, A. et al. Primary care physician characteristics and low-value care provision in Japan. JAMA health forum. 6, e251430 (2025).

    Article  PubMed  PubMed Central  Google Scholar 

  28. Dremio Corporation & Dremio. Available at https://www.dremio.com (2025).

  29. R Core Team. R: A language and environment for statistical computing [Computer software]. R Foundation for Statistical Computing, (Vienna, Austria, 2021). Available at https://www.R-project.org/

  30. Benchimol, E. I. et al. The REporting of studies conducted using observational routinely-collected health data (RECORD) statement. PLoS Med.12, e1001885 (2015).

    Article  PubMed  PubMed Central  Google Scholar 

  31. Lesko, C. R., Fox, M. P. & Edwards, J. K. A Framework for descriptive epidemiology. Am. J. Epidemiol.191, 2063–2070 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  32. Tan, A., Zhou, J., Kuo, Y. F. & Goodwin, J. S. Variation among primary care physicians in the use of imaging for older patients with acute low back pain. J. Gen. Intern. Med.31, 156–163 (2016).

    Article  PubMed  Google Scholar 

  33. Hueber, S., Kuehlein, T., Gerlach, R., Tauscher, M. & Schedlbauer, A. What they see is what you get: Prescribing antibiotics for respiratory tract infections in primary care: Do high prescribers diagnose differently? An analysis of German routine data. PLOS ONE. 12, e0188521 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  34. Hardwick, R. et al. Exploring reasons for variation in ordering thyroid function tests in primary care: a qualitative study. Qual. Prim. Care. 22, 256–261 (2014).

    PubMed  Google Scholar 

  35. Chapman, A. J. et al. Barriers and facilitators to reducing overuse of thyroid function testing: a mixed-methods study. BMJ Qual. Saf.https://doi.org/10.1136/bmjqs-2025-019042 (2025).

    Article  PubMed  Google Scholar 

  36. Lam, J. H., Pickles, K., Stanaway, F. F. & Bell, K. J. L. Why clinicians overtest: development of a thematic framework. BMC Health Serv. Res.20, 1011 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  37. de Sutter, A. I., de Meyere, M. J., de Maeseneer, J. M. & Peersman, W. P. Antibiotic prescribing in acute infections of the nose or sinuses: a matter of personal habit? Fam. Pract.18, 209–213 (2001).

    Article  PubMed  Google Scholar 

  38. Wangler, J., Rusert, R. & Jansky, M. Leitlinienorientierung in der hausärztlichen Versorgung – Befunde einer quantitativen und qualitativen Studienreihe zur Leitliniennutzung von niedergelassenen Allgemeinmediziner*innen. Präv Gesundheitsf. 20, 194–201 (2025).

    Article  Google Scholar 

  39. Bodley, T., Kwan, J. L., Matelski, J., Darragh, P. J. & Cram, P. Self-reported test ordering practices among Canadian internal medicine physicians and trainees: a multicenter cross-sectional survey. BMC Health Serv. Res.19, 820 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  40. FEMNA Health. Schilddrüsentest für Zuhause. Misst TSH, fT3, fT4. Available at https://femna.de/products/schilddruesentest (2025).

  41. Careshop360. Schilddrüsen Test (TSH, FT4, FT3). Available at https://www.careshop360.de/Schilddruesen-Test-TSH-FT4-FT3 (2025).

  42. Nickel, B. et al. Social media posts about medical tests with potential for overdiagnosis. JAMA Netw. open.8, e2461940 (2025).

    Article  PubMed  PubMed Central  Google Scholar 

  43. Flynn, R., Cassidy, C. & Scott, S. D. Learning to forget: deimplementation and the science of sustainability in healthcare. BMJ Qual. Saf.https://doi.org/10.1136/bmjqs-2025-019666 (2026).

    Article  PubMed  PubMed Central  Google Scholar 

  44. Bradshaw, A. B., Bonnecaze, A. K., Burns, C. A. & Beardsley, J. R. Impact of an Interprofessional collaborative quality improvement initiative to decrease inappropriate thyroid function testing. Hosp. Pharm.56, 481–485 (2021).

    Article  PubMed  Google Scholar 

  45. Gilmour, J. A. et al. Promoting resource stewardship: Reducing inappropriate free thyroid hormone testing. J. Eval Clin. Pract.23, 670–675 (2017).

    Article  PubMed  Google Scholar 

  46. Krouss, M. et al. Free the T3: Implementation of best practice advisory to reduce unnecessary orders. Am. J. Med.135, 1437–1442 (2022).

    Article  PubMed  Google Scholar 

  47. Pioch, C., Hildebrandt, M., Goetz, G. & Vogt, V. Effectiveness and mechanisms of interventions to reduce low-value thyroid function tests: a systematic review. Syst. Rev.https://doi.org/10.1186/s13643-026-03119-8 (2026).

    Article  PubMed  PubMed Central  Google Scholar 

  48. Grimshaw, J. M. et al. De-implementing wisely: Developing the evidence base to reduce low-value care. BMJ Qual. Saf.29, 409–417 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  49. Münch, C. et al. Qualität der hausärztlichen Diagnosedokumentation in Patientenakten – Eine Analyse am Beispiel von Schilddrüsenerkrankungen. ZEFQ115–116, 56–62 (2016).

    Google Scholar 

Download references

Funding

Open Access funding enabled and organized by Projekt DEAL. for this study was provided by the Federal Joint Committee (G-BA) grant 01VSF19038. The publication of study results was not contingent on the sponsor’s approval or censorship of the manuscript. The authors declare that they have no conflicts of interests.

Author information

Authors and Affiliations

  1. Institute of General Practice and Family Medicine, Jena University Hospital, Friedrich Schiller University Jena, Bachstraße 18, 07743, Jena, Germany

    Meik Hildebrandt & Verena Vogt

  2. Center for Health Services Research, Jena University Hospital, Friedrich Schiller University Jena, Bachstraße 18, 07743, Jena, Germany

    Meik Hildebrandt & Verena Vogt

  3. Institute of General Practice, Friedrich-Alexander-University Erlangen-Nürnberg (FAU), Erlangen, Germany

    Susann Hueber

  4. Department of Primary Care - General Practice, Medical University Lausitz – Carl Thiem, Cottbus, Germany

    Jeannine Schübel & Karen Voigt

  5. Department of General Practice, Faculty of Medicine, University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany

    Jeannine Schübel & Karen Voigt

Authors

  1. Meik Hildebrandt
  2. Susann Hueber
  3. Jeannine Schübel
  4. Karen Voigt
  5. Verena Vogt

Contributions

M.H. and V.V. conceived of the presented idea and design. M.H. performed the analytic calculations and wrote the initial draft with support from VV. M.H., S.H., J.S., K.V. and V.V. contributed to the interpretation of the data and the final version of the manuscript. V.V. supervised the project. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Meik Hildebrandt.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Hildebrandt, M., Hueber, S., Schübel, J. et al. Physician-level variation and concentration of low-value thyroid function testing in german ambulatory care. Sci Rep 16, 38903 (2026). https://doi.org/10.1038/s41598-026-81364-6

Download citation

  • Received: 14 April 2026

  • Accepted: 16 November 2026

  • Published: 20 November 2026

  • Version of record: 20 November 2026

  • DOI: https://doi.org/10.1038/s41598-026-81364-6

Keywords

来源:Nature 健康护理主题 · nature.com