Prostate cancer testing in Sweden : the interplay between cost and effectiveness

Abstract

As a major cause of male deaths in Sweden, prostate cancer constitutes an essential public health issue to the society. Early detection through an organised testing program with the prostate-specific antigen (PSA) test and systematic biopsy (SBx) has not been adopted in most countries due to potential harms from over-diagnosis and over-treatment of low risk cancers. The magnetic resonance imaging (MRI) and a novel serum-based reflex test Stockholm3 are possible two approaches to tackle this problem. This doctoral thesis aimed to characterise the societal economic burden due to prostate cancer in Sweden and assess the cost-effectiveness of prostate cancer testing using MRI with or without the reflex Stockholm3 test.

Study I characterised and illustrated the resource utilisation in the diagnostic and care pathways of prostate cancer in Sweden during the calendar year 2016. A prevalence-based cost-of-illness approach was applied to quantify the resource utilisation and related costs by care type in Stockholm using register-based data. Direct healthcare resources used in the primary, outpatient, inpatient, palliative care and the pharmaceuticals were valued by their unit costs. Informal care and productivity losses were valued by the human capital method. The societal costs in Stockholm were estimated to be €64 million, of which the direct healthcare, informal care and productivity losses accounted for 62%, 28% and 10%, respectively. The extrapolated costs to Sweden were estimated to be €281 million. An average costs of €1,510, €828 and €271 per prevalent case were calculated for the direct healthcare, informal care and productivity losses, respectively. The results were sensitive to the exclusion of primary care visits for those without a diagnosis of prostate cancer and the proxy good method for valuing informal care.

Study II assessed the cost-effectiveness by microsimulation for: (i) no screening and quadrennial PSA screening of prostate cancer for men aged 55-69 years from a lifetime societal perspective using; (ii) SBx alone; (iii) MRI and targeted biopsy (TBx) for men with a positive MRI result; (iv) MRI and the combined targeted and systematic biopsies (TBx/SBx) for those who had a positive MRI result; and (v) SBx for men with a negative MRI result and the combined TBx/SBx for those who are MRI positive. Based on the test performance estimated from the data included in a recent Cochrane review, the screening strategies could reduce prostate cancer related mortality by 8-10% compared with no screening, but resulted in incremental cost-effectiveness ratios (ICERs) that were classified as high costs per quality-adjusted life year (QALY) gained in Sweden. MRI-based screening with either TBx or the combined TBx/SBx had a lifetime reduction in the biopsy episodes by approximately 40%, compared with screening using SBx alone. These two MRI-based strategies were associated with lifetime reductions in detecting International Society of Urological Pathology Grade group 1 (GG=1) cancers by 17% and 11%, respectively, and both strategies yielded strong dominance over alternative screening strategies. MRI-based screening with TBx was found to have the lowest ICER relative to no screening. This ICER would lead to a 25% reduction when substituting the background health state values reported by the World Health Organisation (WHO) with a value set measured from the Swedish general population.

Study III evaluated the cost-effectiveness comparing: (i) no screening and three quadrennial MRI-based screenings with the combined TBx/SBx on men with a positive MRI result given (ii) positive PSA test value; (iii) positive Stockholm3 test at a reflex threshold of PSA≥1.5ng/mL; and (iv) positive Stockholm3 test at a reflex threshold of PSA≥2ng/mL. Based on the data from the STHLM3-MRI invitation-to-screening trial, the adjustment for the test performance using data from the Cochrane review, and employing a lifetime societal perspective, all screening strategies were associated with a prostate cancer mortality reduction by 7-9%. The ICERs of MRI-based screening strategies in relation to no screening were classified as a moderate cost per QALY gained in Sweden. In comparison with screening without Stockholm3 test, MRI-based screening with Stockholm3 at a reflex test threshold of PSA≥2ng/mL predicted a lifetime reduction of MRI examinations and biopsy episodes by 60% and 9%, respectively, and was considered as the optimal choice for prostate cancer screening. The results were robust in the one-way and probabilistic sensitivity analyses.

Study IV further assessed the cost-effectiveness of prostate cancer screening using a microsimulation approach for: (i) no screening; (ii) traditional screening pathway using PSA and SBx; and (iii) MRI-based screening using the combined TBx/SBx on men with a positive MRI result. Test performance was estimated by the evidence from the STHLM3-MRI trial with model-based imputations. Applying a lifetime healthcare perspective, the quadrennial screening strategies reduced prostate cancer related deaths by 6-9%. Compared with the traditional PSA screening pathway, the MRI-based screening with the combined TBx/SBx halved the MRI examinations and reduced cancer over-diagnosis by approximately 50%. The use of MRI and subsequent combined TBx/SBx for screening resulted in an ICER that was classified as moderate cost per QALY gained in Sweden and has high likelihood to be more cost-effective than the traditional PSA screening pathway. Expanding the screening ages to 50-74 years would increase the ICER by approximately 34%.

In conclusion, substantial economic burden was estimated for prostate cancer in Sweden, with the main costs from the direct healthcare and informal care provided to the patients. This doctoral thesis contributes to the characterisation and illustration of the resource utilisation and costs alongside the diagnostic and care pathways and provides point references for future economic evaluations in prostate cancer testing and treatment. In the context of screening for men aged 55-69 years and compared with no screening, the incorporation of MRI in the screening program with or without a reflex Stockholm3 test yielded reductions in prostate cancer mortality and over-diagnosis over a lifetime period. Assessing cost-effectiveness from a healthcare perspective and using the background health state values from the Swedish general population, the MRI-based screening resulted in higher QALYs and ICERs that are classified as a moderate cost per QALY gained in Sweden. This doctoral thesis suggests that MRI is considered to be more effective and cost-effective in the population-based screening leveraging the evidence from the screening-by-invitation trial than using estimates from diagnostic patient cohorts and MRI-based screening demonstrates higher probability to be cost-effective than the traditional PSA screening pathway. Screening with MRI can be considered as a cost-effective choice for early detection of prostate cancer in Sweden.