QJM Advance Access published online on October 8, 2008
QJM, doi:10.1093/qjmed/hcn130
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The potential effects on fracture outcomes of improvements in persistence and compliance with bisphosphonates
From the 1General Practice Research Database, Medicines and Healthcare Products Regulatory Agency, London, UK and 2Novartis Pharma AG, Basel, Switzerland
Address correspondence to Dr T.P. van Staa, General Practice Research Database, Medicines and Healthcare Products Regulatory Agency, 1 Nine Elms Lane, London SW8 5NQ, UK. email: Tjeerd.vanstaa{at}GPRD.com
Received 14 February 2008 and in revised form 14 September 2008
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionFundingReferences |
|---|
Background: Osteoporotic fractures are a substantial public health burden worldwide. Evidence from studies show that treatments, such as bisphosphonates, which reduce the risk of fractures when taken regularly and long term, are being used irregularly and suboptimally. The objective of this study was to quantify the additional number of hip fractures prevented by improving persistence and compliance with bisphosphonates.
Methods: The study population included patients prescribed alendronate or risedronate in the UK General Practice Research Database. Individualized probabilities of fracture and death during bisphosphonate therapy and of treatment persistence and compliance were estimated by age, gender, dosage, calendar year and clinical risk factors using Cox regression. Persistence was calculated by measuring repeat prescribing and compliance by estimating the medication possession ratio. A unique patient-based decision model was then developed using these probabilities. By varying the persistence and compliance probabilities in the simulation, the fracture outcomes with different scenarios were then evaluated. The outcomes were simulated over a 4-year period (maximum of three years of bisphosphonate use followed by 1 year of offset). It was assumed that the bisphosphonate users had experienced similar fracture reductions as observed in clinical trials.
Results: The study population included 44 531 patients. Modelling showed that improvement of the 3-year persistence by 10% (over current persistence) would prevent an additional 14.4 hip fractures per 10 000 patients with weekly treatment. If weekly was substituted with yearly treatment (refill once a year), an additional 68.4 hip fractures (per 10 000 patients) would be prevented. If 3-year persistence improved by 10% with yearly treatment, an additional 78.5 hip fractures would be prevented compared to monthly bisphosphonates. The effects of this substitution were largest in elderly patients and in women.
Conclusion: Improvements in treatment persistence and compliance may improve the impact of bisphosphonates in reducing the risk of fractures. Yearly administration may also improve the impact on fracture risk reduction, unless long-term persistence is substantially reduced.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionFundingReferences |
|---|
Osteoporotic fractures represent a huge public health burden, with millions of people affected around the world.1 In recent years, various treatments have been developed to reduce the risk of fracture in patients with osteoporosis.2 To be effective in protecting against fracture, treatments need to be taken long term. There is now evidence from several studies that persistence with osteoporosis treatment (i.e. continuation of treatment over time) is suboptimal with a substantial proportion of patients discontinuing bisphosphonate treatment within 1 year of starting.3,4 Patients may also take bisphosphonates at irregular intervals rather than the recommended weekly or daily regimens.4 Compliance can be defined as the proportion of the total treatment time that is covered by the dispensed medication. It has been reported that patients with lower compliance have increased rates of fracture compared to those using it continuously.5,6
Bisphosphonates are the most commonly used treatments for osteoporosis and large clinical studies have shown that they reduce the risk of hip fracture.2 Hip fractures are associated with considerable morbidity and mortality and reduced quality of life. Poor therapeutic adherence with daily bisphosphonates has been improved by the introduction of weekly regimens, although adherence levels remain suboptimal. Bis-phosphonate regimens with dosing intervals beyond a week have been developed to address this issue.7 Ibandronate given at monthly intervals was found to reduce the risk of vertebral fractures although no effect was observed on the risk of non-vertebral fractures.8 Recently, a large clinical study found that patients treated with zoledronate, a once-yearly bisphosphonate infusion, had a statistically significant lower risk of hip fracture as well as vertebral and non-vertebral fracture compared with patients using placebo.9 Given the increasing availability of bisphosphonate with different dosing regimens, it is of interest to evaluate the effects of changing the dosing regimen of bisphosphonates and improving treatment persistence and compliance on the reduction of osteoporotic fractures. The objective of this study was to quantify the potential effects of improving persistence and compliance with bisphosphonates. This was done by developing a decision model that was based on actual observations in a large cohort of bisphosphonate users in actual clinical practice.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionFundingReferences |
|---|
Study population
The data in this study were obtained from the General Practice Research Database (GPRD). GPRD comprises the computerized medical records maintained by general practitioners (GPs). GPs play a key role in the UK health care system, as they are responsible for primary health care and specialist referrals. Patients are semi-permanently affiliated to a practice, which centralizes the medical information from the GPs, specialist referrals and hospitalizations. The data recorded in the GPRD since 1987 include demographic information, prescription details, clinical events, preventive care provided, specialist referrals, hospital admissions and their major outcomes.10
The study population consisted of women or men aged 18 years or older who started treatment with the bisphosphonates alendronate or risedronate. Start of treatment was defined as the first bisphosphonate prescription at least 12 months after start of GPRD data collection. Daily alendronate was first available for routine use in the UK in 1995, weekly alendronate in 2000, daily risedronate in 2000 and weekly in 2003. Patients with a history of Paget's disease were excluded.
Overall design of the model
The decision model estimated the effects of improving persistence and compliance with bisphosphonates. The probabilities in the study population for long-term persistence (i.e. continuation of bisphosphonate treatment over time), compliance [i.e. medication possession ratio (MPR)], fracture and mortality were estimated. In order to estimate the potential impact of improving persistence and compliance, their probabilities were then varied in a simulation model. For example, the observed 3-year persistence of 35% with bisphosphonates was increased by 20–55%, resulting in a longer exposure to bisphosphonates. In order to estimate the additional number of fractures prevented due to longer exposure, we assumed that the fracture effects of bisphosphonates [i.e. the relative rates (RRs)] were similar to those observed in clinical trials. Using the clinical trial RRs, we then estimated the probabilities of fracture if the patients would have been unexposed to bisphosphonates. If the effect of a drug, i.e. the RR, is known, the underlying (unexposed) event probability can be estimated by dividing the event probability as observed in the exposed GPRD patients through the RR. As an example, if the observed event probability during exposure was 3% and the RR of drug effect was 0.6 (i.e. 40% reduction of event probability due to drug exposure), the probability during non-exposure would have been 5%. Improvements in persistence and compliance would prevent additional fractures as the time period of non-exposure would be smaller.
The probabilities for persistence, compliance, fracture and mortality were individualized, i.e. estimated by age, gender, dosage, calendar year and clinical risk factors including fracture history (at any time before the start of bisphosphonates as recorded in the records), recent use of oral glucocorticoids and number of non-bisphosphonate prescriptions in the preceding 3 months.
Probabilities of fracture or death
The model included the following seven outcomes: hip fracture, clinically symptomatic vertebral fracture, radius/ulna fracture, humerus fracture, rib/sternum/tibia/fibula fracture and other (non-osteoporotic) fracture and death. The individual probabilities for each type of fracture or death were based on the probabilities observed in this cohort during bisphosphonate treatment estimated using Cox proportional hazards models.
Treatment persistence
The persistence of weekly/daily bisphosphonate treatment was calculated in this population by measuring repeat prescribing within the 3 months following the end of a prescription. If no prescription was given during this time-window of 3 months, the patient was considered to have stopped treatment. Using Cox proportional hazards models, these probabilities were calculated for each month of follow-up. For each set of patient characteristics, the Cox proportional hazards model allows the calculation of an individual's probability of an event (i.e. survivor function).
We also considered the effects of treatment with yearly bisphosphonates. It was assumed that treatment persisted for the 1 year following an infusion. After this, the percentage of patients that took a new infusion was similar to the persistence level at that time with weekly bisphosphonates.
Treatment compliance
In this study, we determined persistence by evaluating repeat prescribing within a 3-month period. It was found that the use of a shorter time-period substantially reduced the estimated long-term persistence, as patients frequently had gaps in their prescribing. In order to address these gaps, we also measured compliance (MPR). Compliance with bisphosphonate treatment was measured by dividing the expected duration of bisphosphonate use with all previous prescriptions (based on the number of tablets prescribed) by the time elapsed since start of bisphosphonates. Patients with ‘very good’ compliance (MPR >0.90) were considered the reference group. Patients with a MPR below 0.90 were categorized into tertiles. Time-dependent Cox proportional hazards models were used to estimate the RRs of fracture with each of the tertiles of lower compliance during bisphosphonate treatment. This was done for each of the six fracture types included in the model.
The analysis evaluated the effects of changing the percentage of patients in the ‘very good’ compliance group (with corresponding reductions in the percentages of patients with lower compliance) during bisphosphonate treatment. The fracture risks in each compliance group were multiplied by the RRs estimated in Cox proportional hazards model. In the analysis of yearly bisphosphonates, patients were considered to be in the ‘very good’ compliance group for 1 year after each infusion.
Efficacy of bisphosphonate treatment
It was assumed that the bisphosphonate users had experienced similar fracture reductions as those observed in clinical trials. Thus, the fracture incidence without bisphosphonate use would have been the rate observed in our study population divided by the clinical trial efficacy (RR). We used the results of a meta-analysis on the efficacy of the bisphosphonates on the risk of fractures. Data for RR in severe osteoporosis for risedronate and alendronate were used, as data were available on different fracture types.11 The fracture efficacy estimates for these two drugs were combined, each RR estimate was weighted by the inverse of its variance. The following estimates for bisphosphonate efficacy were used: 0.58 for clinically symptomatic vertebral fractures, 0.53 for hip fractures, 0.62 for radius/ulna fractures and 0.75 for humerus fracture and rib/sternum/tibia/fibula fractures and 1.00 for other (non-osteoporotic) fractures.11 Following discontinuation of treatment, it was assumed that the efficacy of treatment decreased towards null in a linear manner over time. The maximum period for efficacy after treatment discontinuation was taken as the minimum of 1 year or the duration of prior treatment (e.g. the efficacy reduced towards null over 6 months after stopping treatment if a patient was previously treated for 6 months; if the patient was treated for 2 years, the offset period was 1 year). In the analysis of yearly bisphosphonates, it was assumed that their efficacy in reducing the risk of fractures was similar to that of daily/weekly bisphosphonates.
Modelling the outcomes
The study population included 44 531 patients who were prescribed either alendronate or risedronate. It was found that 58.3% of the patients continued bisphosphonate treatment for more than 1 year. Patients using weekly bisphosphonates were more likely to continue bisphosphonate treatment compared to patients using daily bisphosphonates. It was found that the majority of patients had very good compliance. Elderly patients were found to be more compliant than younger patients (72% of the patients aged 80–89 years had a MPR >0.90 and 62% of the patients aged <60 years). The RR of any fracture was 1.01 with good compliance, 1.06 with medium and 1.23 with bad compliance compared with patients with very good compliance. For hip fractures, these RRs were 1.01, 1.11 and 1.38, respectively.
Using the individual mortality, fracture, persistence and compliance probabilities, the outcomes were simulated over a 4-year period (maximum 3 years bisphosphonate use followed by maximum of 1 year of offset of the effects of bisphosphonates). In each analysis, we estimated the number of fractures prevented by bisphosphonates. The outcomes were simulated at each 28-day period, based on the standard duration of treatment following a prescription of daily or weekly bisphosphonates. The individual probabilities of fracture and mortality were adjusted in the model for increasing age and fracture occurrence. The simulations were conducted for different scenarios of treatment persistence and compliance.
From the study population, 10 000 patients were randomly sampled and their age and risk factors were used to estimate their individual probabilities for persistence, compliance, fracture and mortality. In another simulation, the levels of persistence and compliance were then modified. Each scenario of persistence and compliance was repeated 20 times using data from a different random samples of 10 000 patients from the study population. Each simulation estimated the number of fractures prevented due to bisphosphonate treatment. Using the results from all the simulations, linear regression was used to estimate the association between level of persistence or compliance and number of fractures prevented.
In order to establish an estimate of the total impact of improvements of persistence and compliance, each of the different fracture types was weighted according to the 1-year post-fracture mortality observed in the study population. Life expectancy has previously been used to standardize outcomes of different medical interventions in other studies.12 Patients with a hip fracture had the highest 1-year mortality and, therefore, hip fracture was considered the reference fracture. For another fracture type with half the mortality as hip fracture (i.e. RR of 0.50), two events of this fracture type would be considered ‘equivalent’ to one hip fracture. The following weights for hip-fracture equivalents were used: clinically symptomatic vertebral fracture, 0.63; radius/ulna fracture, 0.28; humerus fracture, 0.59; rib/sternum/tibia/fibula fracture, 0.59; and other (non-osteoporotic) fracture, 0.54. In a sensitivity analysis, alternative weights for hip-fracture equivalents were used based on reported cost-equivalence (weights of 0.17, 0.14, 0.21, 0.26 and 0.04, respectively).13
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionFundingReferences |
|---|
Table 1 shows the baseline characteristics of the study population. The mean age was 71 years and 81.6% were women. About 82% of the study population started on a weekly bisphosphonate, although this percentage varied over calendar time (due to the later market introduction of weekly bisphosphonates).
|
The 1-year persistence for weekly bisphosphonate in the model was 56.7% and the 3-year persistence was 35.3%. Table 1 shows the additional number of hip fractures prevented (or equivalent) with various levels of improvements in persistence and compliance with weekly bisphosphonates. An improvement of 3-year persistence by 20% (from 35.3% to 55.3%) would prevent an additional 28.7 hip fractures (or equivalent). A similar improvement in compliance resulted in fewer additional hip fractures prevented. The largest effects of persistence and compliance improvement in weekly bisphosphonate users were observed in elderly women (Table 2).
|
As shown in Table 3, the efficacy of bisphosphonates was an important predictor of the magnitude of fracture risk reduction. A bisphosphonate that only reduced the risk of vertebral fractures and not other types of fractures would prevent fewer fractures than weekly bisphosphonates, even when it substantially improved the persistence.
|
Table 4 shows the additional number of hip fractures prevented (or equivalent) with improvements in persistence with yearly bisphosphonate treatment (assuming equal efficacy). An additional 68.4 hip fractures (per 10 000 patients) would be prevented with yearly bisphosphonate treatments (compared to weekly bisphosphonates) if long-term persistence remained unchanged. If long-term persistence were to improve by 20%, it would prevent the occurrence of an additional 88.6 hip fractures. These effects of yearly bisphosphonates were related to both the full compliance and persistence in the 1-year following the administration.
|
Table 5 shows the results of the sensitivity analyses. A longer offset of the effects of bisphosphonates after treatment discontinuation increased the effects of improvements in persistence.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionFundingReferences |
|---|
This study quantified the potential effects of improving persistence and compliance with the use of bisphosphonates of varying dosing frequencies. These effects were most noticeable in elderly women. The improvement of treatment persistence and compliance was strongly related to the efficacy of bisphosphonates.
Long-term persistence to bisphosphonate therapy has been found to be suboptimal and a substantive proportion of patients take daily or weekly bisphosphonates irregularly.3,4 Given this lower persistence and compliance in actual clinical practice, the efficacy of bisphosphonates may be smaller in actual clinical practice compared to that observed in clinical trials. Analyses that are based on clinical trial efficacy may, therefore, have overestimated the cost-effectiveness of bisphosphonates. Alternative dosing regimens have been evaluated, including ibandronate given orally once a month and zoledronate given intravenously once a year.8,9 Although these bisphosphonates may improve long-term persistence and compliance, our study found that the efficacy in the reduction of fracture risk is important, particularly that of the risk of hip fractures. A less efficacious drug may have reduced benefit on fracture risk reduction even with improved persistence and compliance.
In our study population, the long-term persistence improved with the weekly dosing regimen compared to the daily regimen and this improvement has been also reported in other studies.3 Adherence with medical therapy for chronic disease is a complex and multifactorial problem. Many attempts have been made with interventions to improve medication adherence. A review concluded that the most effective interventions for improving adherence were often complex and also only had modest effects.14 It would be of interest to establish whether yearly infusion with a bisphosphonate would further improve the long-term persistence of bisphosphonates. One advantage of yearly administration would be that the complete dose would be administered and that compliance would be 100% for a year. It has been reported that the risk of fracture was increased in patients who had gaps in their bisphosphonate prescribing (i.e. had lower compliance).5,6,15 But further data are needed to establish whether patients will return every year for repeat infusions. Bisphosphonates need to be taken long-term in order to be effective in protecting against fracture.
The main strength of this study is that the data used by the decision model (on persistence, mortality and risk of fracture) were based on actual data observed in a large population in actual clinical practice. The data were obtained from the GPRD that has been widely used in fracture studies.16–18 Another strength of this study is that it used individual estimates for persistence, mortality and risk of fracture, taking into account the heterogeneity in risks in actual clinical practice.
There are several limitations of this study. An important limitation of this study is that it was based on prescribing by GPs. We did not have information on the actual drug intake by patients and it was assumed that the patient had actually taken the drug following the prescription. Also, it was assumed that this efficacy of the studied bisphosphonates in actual clinical practice was similar to that observed in clinical trials. The effects of bisphosphonates may have been lower in actual clinical practice if patients did not take their medication after receiving their prescription. The consequence is that we would have over-estimated the effects on fracture of improving persistence with weekly bisphosphonates and under-estimated the differences between weekly and yearly bisphosphonate regimen. There is also limited clinical evidence for the duration of effects of bisphosphonates after treatment discontinuation. An offset period of 1 year was used, but the sensitivity analysis found that a longer offset period would increase the effects of persistence improvements. Another limitation of this study is that it was assumed that patients who have good compliance to therapy are essentially the same as those who have poor compliance (save for the level of compliance). But there may be numerous social and medical reasons for poor adherence, which may make the patient groups dissimilar. Our decision model assumed that an improvement in compliance would reduce the fracture risks in patients with low compliance, but this may be an over-estimate. Our decision model also did not consider the switching to other non-bisphosphonates treatments following treatment discontinuation. Fracture outcomes could be better if all patients would switch to another effective treatment after bisphosphonate discontinuation.
In conclusion, improvements in treatment persistence and compliance may improve the impact of bisphosphonates in reducing the risk of fractures, unless the efficacy is reduced. Yearly administration may also improve the impact on fracture risk reduction, unless long-term persistence is substantially reduced.
|
Funding |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionFundingReferences |
|---|
Novartis Pharmaceuticals.
Conflict of interest: The views expressed in this paper are those of the authors and do not reflect the official policy or position of the Medicines and Healthcare products Regulatory Agency, UK.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionFundingReferences |
|---|
1. Cooper C, Campion G, Melton LJ. Hip fracture in the elderly: a worldwide projection. Osteoporos Int (1992) 2:285–9.[CrossRef][Web of Science][Medline]
2. Cranney A, Guyatt G, Griffith L, Wells G, Tugwell P, Rosen C, Osteoporosis Methodology Group and The Osteoporosis Research Advisory Group. Meta-analyses of therapies for postmenopausal osteoporosis. IX: summary of meta-analyses of therapies for postmenopausal osteoporosis. Endocr Rev (2002) 23:570–8.
3. Cramer JA, Gold DT, Silverman SL, Lewiecki EM. A systematic review of persistence and compliance with bisphosphonates for osteoporosis. Osteoporos Int (2007) 18:1023–31.[CrossRef][Web of Science][Medline]
4. Seeman E, Compston J, Adachi J, Brandi ML, Cooper C, Dawson-Hughes B, et al. Non-compliance: the Achilles’ heel of anti-fracture efficacy. Osteoporos Int (2007) 18:711–9.[CrossRef][Web of Science][Medline]
5. Caro JJ, Ishak KJ, Huybrechts KF, Raggio G, Naujoks C. The impact of compliance with osteoporosis therapy on fracture rates in actual practice. Osteoporos Int (2004) 15:1003–8.[CrossRef][Web of Science][Medline]
6. van den Boogaard CH, Breekveldt-Postma NS, Borggreve SE, Goettsch WG, Herings RM. Persistent bisphosphonate use and the risk of osteoporotic fractures in clinical practice: a database analysis study. Curr Med Res Opin (2006) 22:1757–64.[CrossRef][Web of Science][Medline]
7. Cooper C. Beyond daily dosing: clinical experience. Bone (2006) 38(4 Suppl. 1):S13–17.
8. Chesnut CH III, Skag A, Christiansen C, Recker R, Stakkestad JA, Hoiseth A, et al. Oral ibandronate osteoporosis vertebral fracture trial in North America and Europe (BONE). Effects of oral ibandronate administered daily or intermittently on fracture risk in postmenopausal osteoporosis. J Bone Miner Res (2004) 19:1241–9.[CrossRef][Web of Science][Medline]
9. Black DM, Delmas PD, Eastell R, Reid IR, Boonen S, Cauley JA, et al. HORIZON pivotal fracture trial. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med (2007) 356:1809–22.
10. Walley T, Mantgani A. The UK general practice research database. Lancet (1997) 350:1097–9.[CrossRef][Web of Science][Medline]
11. Stevenson M, Lloyd Jones M, de Nigris E, Brewer N, Oakley J. The clinical effectiveness and cost-effectiveness of prevention and treatment of osteoporosis. (2003) December. Accessed 10 May 2004. National Institute for Clinical Excellence. [http://www.nice.org/page.aspx?o=115573].
12. Wright JC, Weinstein MC. Gains in life expectancy from medical interventions – standardizing data on outcomes. N Engl J Med (1998) 339:380–6.
13. Melton LJ 3rd, Gabriel SE, Crowson CS, Tosteson AN, Johnell O, Kanis JA. Melton weighing fractures. Cost-equivalence of different osteoporotic fractures. Osteoporos Int (2003) 14:383–8.[CrossRef][Web of Science][Medline]
14. McDonald HP, Garg AX, Haynes RB. Interventions to enhance patient adherence to medication prescriptions: scientific review. JAMA (2002) 288:2868–79.
15. Siris ES, Harris ST, Rosen CJ, Barr CE, Arvesen JN, Abbott TA, et al. Adherence to bisphosphonate therapy and fracture rates in osteoporotic women: relationship to vertebral and nonvertebral fractures from 2 US claims databases. Mayo Clin Proc (2006) 81:1013–22.
16. van Staa TP, Wegman S, de Vries F, Leufkens HGM, Cooper C. Use of statins and risk of fractures. JAMA (2001) 285:1850–5.
17. van Staa TP, Leufkens HGM, Abenhaim L, Zhang B, Cooper C. Use of oral corticosteroids and risk of fractures. J Bone Miner Res (2000) 15:993–1000.[CrossRef][Web of Science][Medline]
18. van Staa TP, Dennison EM, Leufkens HGM, Cooper C. The epidemiology of fractures in England and Wales. Bone (2001) 29:517–22.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||