Stockholm3 test to estimate the risk of prostate cancer: a single technology assessment
Health technology assessment
|Published
The objective of this STA was to appraise the evidence provided in the submission file from A3P Biomedical (submitter) concerning the use of the Stockholm3 test for estimating the risk of prostate cancer compared with using only prostate-specific antigen (PSA).
Summary
In the PDF file, prices are redacted for reasons of confidentiality for the manufacturer. In the text below, the slagging appears as X.
Introduction
Prostate cancer is the most common form of cancer in men in Norway. In recent years, over 5,000 new cases of prostate cancer have been diagnosed each year. The Division of Health Services at the Norwegian Institute of Public Health was commissioned in May 2021 to conduct a single technology assessment (STA) of the Stockholm3 test to estimate the risk of prostate cancer. Stockholm3 is a multi-parametric blood test utilizing protein analyses, genetic analyses, clinical data, and an algorithm, to estimate the risk of having a clinically significant prostate cancer (defined as Gleason Score ≥ 7) in biopsy.
Objective
The objective of this STA was to appraise the evidence provided in the submission file from A3P Biomedical (submitter) concerning the use of the Stockholm3 test for estimating the risk of prostate cancer compared with using only prostate-specific antigen (PSA). We also appraised the cost-minimisation analysis and budget impact analysis provided by the submitter.
Method
The appraisal included assessment of the quality of the literature search, the submitter’s summary of the prognostic accuracy of the Stockholm3 test, and the cost-minimization and budget impact analyses provided by the submitter. This implies that NIPH has not performed additional literature searches, any compilations of effects and safety, nor conducted any supplementary health economic analyses.
Results
Clinical effectiveness
The submitter included eight studies in their literature review. However, these eight studies do not represent eight independent samples. All studies compared Stockholm3 with PSA and were conducted in Sweden and/or Norway. None of the studies has been conducted by an independent research group.
All studies excluded patients with previous prostate cancer diagnosis. One of the studies was a randomized controlled trial. The others were prospective paired diagnostic studies or observational studies. The largest study (Grönberg 2015) which was a prospective, population-based, paired, screen-positive, diagnostic study with more than 58,000 participants reported that use of the Stockholm3 model could reduce the number of biopsies by 32% (95% CI 24–39), could avoid 44% (35–54) of benign biopsies, and reduce unnecessary biopsies by 37%. It was not feasible to pool the results from the studies presented in the submission, due to variations in designs and research questions.
Health economics
Testing with Stockholm3 entails an additional direct costs of XXXX NOK per test. However, the submitted analysis indicates that it could be a cost-saving strategy compared to PSA testing alone. The analysis shows cost-savings of 1,400 NOK per man. The submitted budget impact analysis also indicates cost savings each year after implementation of Stockholm3.
Discussion
The documentation of the diagnostic accuracy provided by the use of Stockholm3 is considered to be uncertain. This is partly due to variability in the design and outcomes of the relevant studies which implied that it was not feasible to pool studies for overall assessments. Furthermore, none of the studies were conducted solely by independent researchers.
There are several substantial uncertainties in the health economic analyses. The major driver of the results are differences in the proportions of positive test results with the Stockholm3 test compared to those with PSA tests alone. Only non-published data are used in the analysis for estimating these differences. Assumptions in the analysis on the further diagnostic work-up, surveillance and treatment of patients after referral to urologists are also uncertain and mainly based on input from the clinical experts consulted by the submitter. Furthermore, the current practice in terms of diagnosis and treatment of prostate cancer varies and is under continuous development. This makes it difficult to assess whether the analysis provides a relevant basis for clinical practice in the next coming years.
Based on the one-way deterministic sensitivity analyses that the provider has chosen to perform, the analyses indicate that Stockholm3 will still be a cost-saving alternative compared to the PSA strategy. However, if two or more of the parameters associated with substantial uncertainty in the model are changed simultaneously, Stockholm3 is no longer cost saving compared to PSA testing. The budget impact analysis is also associated with substantial uncertainty.
Conclusion
The clinical documentation submitted does not allow clear conclusions on to which extent a work-up which includes the Stockholm3 test will provide improved detection of clinically significant prostate cancer and avoiding unnecessary biopsies and cancer treatment compared to current clinical practice with PSA as the initial test.
The submitted cost-minimisation analysis indicates that Stockholm3 could be a cost-saving strategy compared to standard practice with PSA-testing. However, there are substantial uncertainties in the structure and assumptions applied in the model, and the results of the analysis should therefore be interpreted in light of these caveats.
