Impact of Norwegian pharmaceutical tender on expenditure and prescription rates of costly drugs and the limitations of various remission definitions in rheumatoid arthritis

Abstract

BACKGROUND AND OBJECTIVES Biologic and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs) are highly effective in treating rheumatoid arthritis (RA) but come with a high cost. Within the scope of this thesis, cost refers to the financial resources for the procurement of b/tsDMARDs, and no attention is placed on the associated expenses, e.g., storage, transportation, and administration of these drugs. A recommended cost-reducing strategy for these drugs is pharmaceutical tendering, which has been implemented in Norway for over a decade. Under the influence of a national tender system, this thesis explores the cost and proportion (i.e., the relative amount of given b/tsDMARD registrations compared to the total b/tsDMARDs registrations) of 13 different b/tsDMARDs used in treating RA from 2010 to 2019 in Norway. The goal of all anti-inflammatory drugs, including b/tsDMARDs, is to treat RA patients into stable remission or low disease activity. Many remission-assessing methods are available and recommended, albeit few are comparable. Therefore, this thesis also evaluates the different remissionassessing methods in-depth and questions the elements causing the discrepancy. METHODS The RA patients were monitored at ordinary outpatient clinics across ten Norwegian centers. Data concerning demographics, disease outcomes, and the prescribed treatment were collected from each center for each year. The b/tsDMARD users were subdivided into the categories: naïve users (those registered to receive a new b/tsDMARD without prior b/tsDMARD), non-naïve (those registered to receive a new b/tsDMARD with prior b/tsDMARD but are given a different b/tsDMARD), and current users (those registered on a b/tsDMARD). The proportion of 13 b/tsDMARD (with subgroups of subcutaneous vs. intravenous and biosimilars vs. non-biosimilars) was thoroughly assessed and compared with the tender rankings from each year. The total and average cost (using confidential tender offers) of b/tsDMARDs was calculated and subdivided according to the b/tsDMARD RA user groups. Complete data from a single center were used to calculate remission rates (i.e., the number of RA patients achieving remission divided by the total number of assessed RA patients) using different remission-assessing methods between 2015 and 2019. The variables from these methods were analyzed using linear and logistic regression. All studies were cross-sectional studies. RESULTS The overall b/tsDMARD treatment proportion ranged between 40% in 2010 to 45% in 2019. The average annual cost per b/tsDMARD per naïve RA user decreased from 13.0 thousand euros in 2010 to 3.2 thousand euros in 2019, which resulted in a 75% cost reduction. Non-naïve users had an estimated cost reduction of 64%, while current users had an estimated cost reduction of 47%. All cost reductions were even more prominent when variation in Norwegian currency was accounted for. The tender-winning b/tsDMARD was either the highest or second-highest in usage in nine out of ten years for b/tsDMARD naïve users, seven out of ten years for non-naïve users, and twice out of ten years for current users. The tender-winning drug was intravenous in eight out of ten years, but the average proportion of intravenous b/tsDMARDs for the different user groups was approximately 50% or lower. The tender-winning drug was a biosimilar in five out of six pharmaceutical tenders. On average, during the years with biosimilars, the biosimilars accounted for roughly 40%, 40%, and 20% of the total b/tsDMARD proportion for naïve, non-naïve, and current users, respectively. There was an approximate 40% difference in remission rate between the most stringent remission-assessing method (i.e., Boolean remission) and the Boolean remission without patients' subjective evaluation (i.e., without the variable patient global assessment (PGA)) when the same 502 RA patients from a single center in 2019 were evaluated. Among these patients, less than 30% achieved ≤10 PGA (0-100), while over 75% achieved the other stringent requirements of Boolean remission. Disease activity measuring methods incorporating PGA equally as the other variables in the algorithm had lower remission rates and were more associated with other patient-reported outcomes (e.g., pain) than those that did not include PGA or included it but with reduced impact in the algorithm. Regardless of the measurement method, pain (standardized coefficient β = 0.7, p < 0.001) was most strongly associated with PGA. CONCLUSION During the national pharmaceutical tendering in Norway in the period from 2010 to 2019, changes were observed. For the payer, the average annual cost per b/tsDMARD per patient decreased considerably and was most pronounced for patients starting naïve on b/tsDMARD. For the supplier, the pharmaceutical companies that provided the most inexpensive drug each year also had the highest or second-highest proportion among new b/tsDMARD prescriptions. Simultaneously, no findings of worsened disease activity, patient-reported outcome measures, or work capability were observed. With the saved expenditure (i.e., the total cost spent on acquiring b/tsDMARDs over a given time) and increasing b/tsDMARD treatment proportion, this thesis recommends being vigilant when starting patients on new b/tsDMARDs to avoid administering the medications for noninflammatory causes. As such, the thesis advises against using disease activity measuring methods that can interfere with that judgment. This thesis shows that the instruments used to determine remission substantially influence attaining remission. The patient's overall assessment of disease activity, which is strongly associated with pain and does not distinguish between inflammatory and noninflammatory pain, has a considerable impact on whether remission is achieved with certain instruments. This is important to consider when starting patients on b/tsDMARDs to prevent administering the medications for noninflammatory reasons. In that perspective, when treating patients with b/tsDMARDs, this thesis calls into question the use of methods for measuring disease activity that is heavily impacted by the patient's own evaluation of their disease activity.