Pharmacoepidemiology and health economics of adherence to pharmaceutical fracture prevention

Sammanfattning: Background: Osteoporosis is a disease characterized by weak bone, affecting hundreds of millions of people worldwide, predominantly postmenopausal women. The main clinical consequence of the disease is bone fractures and the lifetime risk of any fracture has been estimated at ~55% in Norwegian women. Hip and vertebral fractures are the two most serious fracture types, associated with substantial pain, disability, and even death. Even though there is consensus that patients at high risk of fracture should be treated, there is still a troubling treatment gap that shows few signs of closing. Only 6.6% of untreated patients receive treatment after their first fracture and there are ~225,000 untreated individuals with a bone mineral density indicative of osteoporosis in Sweden. An equally noteworthy aspect of undertreatment is poor adherence (compliance and persistence) to treatment, i.e. how patients and physicians adhere to dosing instructions and treatment regimens. Many patients stop filling prescriptions at pharmacies prematurely (refill non-persistence) and this is a cause for concern with respect to effective fracture prevention. There are also reports that dispensings at pharmacies are too few and far between to provide adequate drug exposure (measured as refill compliance). Oral alendronate, a bisphosphonate, constitutes ~80% of all osteoporosis treatments and is generally recommended for 3-5 years. Treating osteoporosis have in most industrialized countries been estimated to be cost-effective (compared with no treatment) but this depends on several factors, such as the risk of the patient population, drug costs, treatment effectiveness, and the treatment alternatives being compared. Treatment adherence is often not factored into such cost-effectiveness analyses. Objectives: This thesis aims at addressing pharmacoepidemiologic and health economic aspects of poor compliance and persistence to osteoporosis treatment by both establishing the extent of the problem and consequences for fracture risk in a Swedish setting, as well as investigating how it can be incorporated into the health economic framework to inform reimbursement decisions and regional priorities for recommended prescription standards. The topics of health-economic value or treatment persistence are by no means specific to the Swedish setting. Therefore, even though the included publications are based on Swedish data, the background and findings are also often put in an international context, or entirely without reference to geography. Methods & papers: Three of the articles used Swedish register data on pharmacy dispensings, diagnosis codes, and mortality. Repeat dispensings at pharmacies by 57,000 individuals were used to estimate refill persistence and refill compliance as an approximation of true drug exposure. Paper I investigated the proportion of patients starting an osteoporosis treatment that stopped their treatment prematurely at different time points, as well as the implications on the risk of fracture in groups stratified by refill persistence. Paper II addressed how automatic generic substitution (for off-patent medication) influence persistence to treatment of oral bisphosphonates. A natural experiment was devised for the years 2006-2009 where an off-patent medication was compared to an on-patent medication to isolate the effect of generic substitution. The effect on persistence for patients getting their first medication refill substituted at the pharmacy was also investigated. Paper III, amended with a new analysis in a larger dataset, investigated the residual effect after treatment with bisphosphonates on fracture risk and explored whether a healthy adherer effect (i.e. that patients with an inherently lower fracture risk stay longer on treatment) confounds the association between refill persistence and residual anti-fracture effect. Paper IV proposes a health economic simulation model framework for incorporating adherence and studying the important drivers of cost-effectiveness in this context. Main conclusions: • Refill persistence to typical oral osteoporosis medication estimated from pharmacy dispensing in Sweden is poor, with ~50% stopping treatment within 12 months. Prescription refill gaps among persistent patients appears to be a margnial problem, with 96% of patients having access to >80% of intended doses. • Poor refill persistence to osteoporosis treatments is associated with an increased fracture risk in an exposure-dependant manner. • Automatic generic substitution of alendronate tablets at pharmacies was likely causing reduced treatment persistence to treatment during 2006- 2009. Patients who had their alendronate product substituted at the first prescription refill had 25% higher risk of stopping their treatment. This topic should be revisited in more recent data and for other therapeutic areas. • It is likely that treatments shorter than 6 months with oral bishposphonates has little effect on fracture risk. • Oral bisphophonates taken for at least 12 months may confer a residual effect of 20-35% on the risk of any fracture for up to 5 years after stopping treatment. It is not clear if and how such a residual effect wanes with time after stopping treatment. The health economic implications of residual effect can be considerable, depending on the context. • There is a statistically significant inverse relationship between time on bisphosphonate treatment and post-treatment fracture risk. This finding supports an assumption that the magnitude of a residual effect depends on the preceding time on treatment with bisphophonates in health-economic evaluations. • Incorporating treatment adherence into a health economic evaluation in osteoporosis can have a substantial impact, but is context specific. The choice of accounting for or disregarding adherence to treatment may have an impact on both treatment recommendations, priorities, reimbursement, and prices of treatments for osteoporosis. Poor persistence to osteoporosis treatments causes increased morbidity and mortality. Improving persistence to osteoporosis treatments would confer substantial health benefit for both patients and society. The clinical and health-economic consequences of persistence to osteoporosis treatments should not be disregarded when setting priorities and drug prices.