Impact of antidepressants on the risk of suicide in patients with depression in real-life conditions: a decision analysis model.

BACKGROUND: The impact of antidepressant drug treatment (ADT) on the risk of suicide is uncertain. The aim of this study was to determine in a real-life setting whether ADT is associated with an increased or a reduced risk of suicide compared to absence of ADT (no-ADT) in patients with depression. METHOD: A decision analysis method was used to estimate the number of suicides prevented or induced by ADT in children and adolescents (10-19 years old), adults (20-64 years old) and the elderly (65 years) diagnosed with major depression. The impact of gender and parasuicide history on the findings was explored within each age group. Sensitivity analyses were used to assess the robustness of the models. RESULTS: Prescribing ADT to all patients diagnosed with depression would prevent more than one out of three suicide deaths compared to the no-ADT strategy, irrespective of age, gender or parasuicide history. Sensitivity analyses showed that persistence in taking ADT would be the main characteristic influencing the effectiveness of ADT on suicide risk. CONCLUSIONS: Public health decisions that contribute directly or indirectly to reducing the number of patients with depression who are effectively administered ADT may paradoxically induce a rise in the number of suicides.

Methotrexate in rheumatoid arthritis. An update.

Methotrexate has been approved for the treatment of refractory rheumatoid arthritis by several regulatory agencies, including the Food and Drug Administration. The tendency is now to prescribe it at earlier stages of the disease. Methotrexate is a well known antifolate. Its exact mechanism of action in rheumatoid arthritis remains uncertain. The polyglutamated derivatives of methotrexate are potent inhibitors of various enzymes, including dihydrofolate reductase and 5-aminoimidazole-4-carboxamide ribonucleotide transformylase. Inhibitory effects on cytokines, particularly interleukin-1, and on arachidonic acid metabolism, as well as effects on proteolytic enzymes, have been reported. Some of them may be linked to the antifolate properties of methotrexate. Overall, the drug appears to act in rheumatoid arthritis as an anti-inflammatory agent with subtle immunomodulating properties. Direct inhibitory effects on rapidly proliferating cells in the synovium have also been suggested. Methotrexate is usually given orally. Marked interindividual variation in its bioavailability has been found. Food intake has no significant effect on the pharmacokinetics of oral methotrexate. Methotrexate undergoes significant metabolism. The functionally important metabolites are the polyglutamated derivatives of methotrexate, which are selectively retained in the cells. Less than 10% of a dose of methotrexate is oxidised to 7-hydroxy-methotrexate, irrespective of the route of administration. This metabolite is extensively (91 to 93%) bound to plasma proteins, in contrast to the parent drug (35 to 50% bound). Methotrexate is mainly excreted by the kidneys. It undergoes tubular secretion and may thereby compete with various organic acid compounds. Early placebo-controlled trials demonstrated that weekly low dosage methotrexate produced early symptomatic improvement in most rheumatoid arthritis patients. Two meta-analyses showed that methotrexate is among the most efficacious of slow-acting antirheumatic agents, together with parenteral gold (sodium aurothiomalate), penicillamine and sulfasalazine. Furthermore, in the short term context of clinical trials, methotrexate has one of the best efficacy/toxicity ratios. There is little evidence that methotrexate, or any available slow-acting antirheumatic agent, is a true disease-modifying drug. However, the probability that a patient will continue methotrexate therapy over time appears quite favourable compared with any other slow-acting antirheumatic drug. Combination therapy with slow-acting drugs has been advised for the management of rheumatoid arthritis, but the evidence currently available does not support general use of combination therapy including methotrexate. Almost all investigations indicated that toxic effects, rather than lack of response, were the major reason for discontinuing methotrexate therapy.(ABSTRACT TRUNCATED AT 400 WORDS)

Do patient factors alter the relationship between physician characteristics and use of long-acting benzodiazepines?

Despite known hazards associated with their use, long-acting benzodiazepines are frequently used in the treatment of older adults. While such use has been linked to physician characteristics, the effect of patient factors has not been considered. To investigate this, data from 1423 Quebec community-dwelling subjects of the Canadian Study of Health and Aging were linked to records of prescriptions billed to the provincial health insurance program during the year following study entry. The standardized one-year period prevalence of any use of long-acting benzodiazepines was 12.2%. Among benzodiazepine users, long-acting benzodiazepine use was more common among male patients and patients of earlier graduating prescribers and specialist prescribers. However, the effect of the latter two factors were modified by patient self-reported anxiety. This study demonstrates that consideration of patient factors may be necessary to obtain an accurate estimate of the association between at least some physician factors and use of long-acting benzodiazepines.

Evidence of the depletion of susceptibles effect in non-experimental pharmacoepidemiologic research.

Past experience with a drug may modify the risk of adverse event associated with current use of this drug. This effect was investigated empirically with a study on non-steroidal anti-inflammatory drugs (NSAIDs)-gastropathy. A hospital-based case-control study was conducted with 244 cases of upper gastrointestinal bleeding (UGIB) age 68 and over and 615 matched controls. Data on all medications dispensed to the study patients during the 3 years preceding admission were obtained from the Quebec universal prescription program automated database. Recent use (within 30 days prior to admission) of non-aspirin NSAIDs increased the risk of UGIB. The estimate of relative risk (RR) was 3.4 (CI, 2.1-5.5). Use of NSAIDs in the previous 3 years was associated with a lower risk of UGIB; the estimate of RR was 0.7 (CI, 0.4-1.0). The estimate of RR for first-time users was 22.7 (2.8-200.0) vs 3.0 (1.9-4.7) for those who had used the drugs at least once in the past 3 years. These results provide empirical evidence of a depletion of susceptibles effect whereby patients who remain on the drugs are those who can tolerate them while those who are susceptible select themselves out of the population at risk. Thus, past use should be considered as a potential risk modifier in non-experimental risk assessment of events associated with drug use.

Sample size calculations for single group post-marketing cohort studies.

In pharmacoepidemiology, single group cohort is the most frequently proposed design to determine if the incidence rate of an adverse drug reaction among the exposed differs from a reference value. In many situations, the number of events expected in the cohort is too small to conduct sample size calculations based on the normal distribution. This paper proposes, for a single group cohort study, calculations and tables derived from the Poisson distribution. The results are based on a one-sided test with a 0.05 significance level and a power of 0.9 and 0.8. Two parameters have to be specified a priori: the expected incidence of the event under the null hypothesis and the minimum risk ratio to be detected. The required sample size and the critical number of events to reject the null hypothesis are directly derived from the tables. Results show that the normal approximation may lead to an underestimation of the required sample size.

Comparing the toxicity of two drugs in the framework of spontaneous reporting: a confidence interval approach.

Spontaneous reporting remains the most frequently used technique in post-marketing surveillance. Decision-making usually depends on comparisons between the number of adverse drug reactions (ADRs) reported for two drugs on the basis of an equivalent number of prescriptions. The validity of such comparisons is expected to be jeopardized by probable underreporting ADR cases. This problem is accentuated when it cannot be assumed that the magnitude of underreporting is the same for the both drugs. Differences in reporting ratios can overemphasize, cancel, or reverse the conclusions of a statistical comparison based on the number of reports. We propose a single method for (1) calculating confidence intervals for relative risks estimated in the context of spontaneous reporting and (2) deriving the range of reporting ratios for which the conclusion of the statistical comparison remains statistically valid.

Determinants of initiation and suboptimal use of anti-ulcer medication: a study of the Quebec older population.

OBJECTIVES: To describe the use of anti-ulcer medication in the Quebec older population; to examine determinants of initiation, suboptimal use, and switches between products. DESIGN: Population-based retrospective cohort study. SETTING: Universal health program for older adults in Quebec. PARTICIPANTS: 5000 users and 5000 non-users of anti-ulcer medications were selected randomly. Use was defined as the presence in the 1991 prescription database of an anti-ulcer prescription. Among users, 1697 (34%) were new users and were considered as the exposure group. Subjects were followed for 365 days after inclusion. MEASUREMENTS: Measured were patient's age, gender, prescribed duration of anti-ulcer medication, concomitant medications, and gastrointestinal diagnostic procedures. RESULTS: A total of 17% of new users had unusually short courses; 18% were long-term users. There was no difference in duration for omeprazole compared with other anti-ulcer medications. First-time use of NSAIDs was the strongest predictor of initiation of anti-ulcer medication (odds ratio = 3.21; 95% CI, 2.66-3.88). Twenty-six percent of users switched brands. Only 9.5% of new users underwent a diagnostic procedure before initiation of therapy, and 49% of long-term users ever underwent such procedure. CONCLUSION: Despite a relatively homogeneous recommended duration of therapy, patterns of use of anti-ulcer medication among older people are highly variable, and treatment is often not accompanied by a diagnostic procedure.

Long-term continuous use of benzodiazepines by older adults in Quebec: prevalence, incidence and risk factors.

OBJECTIVE: To determine the prevalence and incidence of long-term use of benzodiazepines and to assess patient-, prescriber-, and drug-related risk factors. DESIGN: Cohort study. PARTICIPANTS: 1,423 community-dwelling older adults in Quebec who participated in the Canadian Study of Health and Aging (CSHA1). MEASUREMENTS: Patient characteristics were obtained from the CSHA1 database. These were linked to provincial health insurance data to ascertain benzodiazepine use and prescriber characteristics. MAIN OUTCOME MEASURE: Use of benzodiazepines for at least 135 of the first 180 days following initiation of use. RESULTS: Twelve-month prevalence of long-term continuous use, standardized by age and gender to the Quebec population, was 19.8%. Twelve-month cumulative incidence of long-term continuous use was 1.9%. Older patients were more likely to proceed to long-term continuous use. CONCLUSIONS: Risk of long-term continuous use of benzodiazepines seems to increase with age. This association was found to be independent of gender, health status, anxiety, cognitive status, benzodiazepine type, and physician characteristics.

Evaluation of the Confusion Assessment Method (CAM) as a screening tool for delirium in the emergency room.

The objective of this study was to compare the results of the Confusion Assessment Method (CAM) obtained by a trained non-physician interviewer to those obtained by a geriatrician, among a sample of elderly patients seen in an emergency room. A group of 110 elderly patients (> or =66 years) were evaluated in the emergency room by a lay interviewer. The geriatrician conducted an interview in the presence of the lay interviewer. Subsequently, the geriatrician and the lay interviewer completed a CAM checklist independently. Kappa statistics, sensitivity, specificity, positivity predictive value (PPV), and negative predictive value (NPV) for the geriatrician's and lay interviewer's results with the CAM diagnostic algorithm were compared. The kappa coefficient was 0.91, the sensitivity 0.86, the specificity 1.00, the PPV 1.00, and the NPV 0.97. In conclusion, the CAM used by a trained lay interviewer in the emergency room is sensitive, specific, reliable and easy to use for the identification of patients with delirium. The under-recognition and under-treatment of delirium is a major health issue and has important clinical and financial implications. The implementation of systematic screening in populations at risk could increase the rate of early detection and lead to the appropriate management of delirious patients.

High daily doses of benzodiazepines among Quebec seniors: prevalence and correlates.

BACKGROUND: Use of high daily doses of benzodiazepines is generally contraindicated for seniors. While both patient and physician factors may influence the use of high daily doses, previous research on the effect of patient factors has been extremely limited. The objectives of this study were to determine the one year prevalence of use of high daily doses of benzodiazepines, and examine physician and patient correlates of such use among Quebec community-dwelling seniors. METHODS: Patient information for 1423 community-dwelling Quebec seniors who participated in the Canadian Study of Health and Aging was linked to provincial health insurance administrative data bases containing detailed information on prescriptions received and prescribers. RESULTS: The standardized one year period prevalence of use of high daily doses of benzodiazepines was 7.9%. Use of high daily doses was more frequent among younger seniors and those who had reported anxiety during the previous year. Patients without cognitive impairment were more likely to receive high dose prescriptions from general practitioners, while those with cognitive impairment were more likely to receive high dose prescriptions from specialists. CONCLUSION: High dose prescribing appears to be related to both patient and physician factors.

Monitoring of drug utilization in public health surveillance activities: a conceptual framework.

The surveillance of individual and aggregate patterns of prescribed medication can potentially provide some very useful information to those involved in public health. At present however, this activity has attracted relatively little attention in Canada. This article will introduce a conceptual framework with which to examine the possibilities of prescription drug monitoring as a tool for public health surveillance. The sources of data available to undertake this activity as well as their limitations, will be presented along with some recent concrete applications.

Variability in risk of gastrointestinal complications with individual non-steroidal anti-inflammatory drugs: results of a collaborative meta-analysis.

OBJECTIVE: To compare the relative risks of serious gastrointestinal complications reported with individual non-steroidal anti-inflammatory drugs. DESIGN: Systematic review of controlled epidemiological studies that found a relation between use of the drugs and admission to hospital for haemorrhage or perforation. SETTING: Hospital and community based case-control and cohort studies. MAIN OUTCOME MEASURES: (a) Estimated relative risks of gastrointestinal complications with use of individual drugs, exposure to ibuprofen being used as reference; (b) a ranking that best summarised the sequence of relative risks observed in the studies. RESULTS: 12 studies met the inclusion criteria. 11 provided comparative data on ibuprofen and other drugs. Ibuprofen ranked lowest or equal lowest for risk in 10 of the 11 studies. Pooled relative risks calculated with exposure to ibuprofen used as reference were all significantly greater than 1.0 (interval of point estimates 1.6 to 9.2). Overall, ibuprofen was associated with the lowest relative risk, followed by diclofenac. Azapropazone, tolmetin, ketoprofen, and piroxicam ranked highest for risk and indomethacin, naproxen, sulindac, and aspirin occupied intermediate positions. Higher doses of ibuprofen were associated with relative risks similar to those with naproxen and indomethacin. CONCLUSIONS: The low risk of serious gastrointestinal complications with ibuprofen seems to be attributable mainly to the low doses of the drug used in clinical practice. In higher doses ibuprofen is associated with a similar risk to other non-steroidal anti-inflammatory drugs. Use of low risk drugs in low dosage as first line treatment would substantially reduce the morbidity and mortality due to serious gastrointestinal toxicity from these drugs.

Effect of treatment gaps in elderly patients with dementia treated with cholinesterase inhibitors.

OBJECTIVE: To determine the effect of treatment gaps on the risk of institutionalization or death among community-dwelling elderly patients treated with cholinesterase inhibitors (ChIs). METHODS: A survival analysis was conducted among a cohort of community-dwelling elderly patients (age 66+) newly treated with ChIs identified in the Quebec drug claims databases (Régie de l'Assurance Maladie du Québec [RAMQ]) between January 1, 2000, and December 31, 2007. Treatment nonpersistence during the year following ChI initiation was defined as treatment discontinuation or gaps of at least 6 weeks. To account for reverse causality, Cox proportional hazard modeling was conducted only among patients who did not discontinue treatment, in order to assess the association between treatment nonpersistence and institutionalization or death. RESULTS: Among the 24,394 elderly ChI users, 4,108 (16.8) experienced a treatment gap during the year following ChI treatment initiation while 596 (2.4%) discontinued their treatment within the first 3 months (early stoppers) and 4,038 (16.6%) after 3 months of treatment (late stoppers). Of all treated patients, 4,409 (18.1%) were institutionalized or died during follow-up. In patients who did not stop their treatment, the risk of institutionalization or death appeared lower in patients who experienced a treatment gap (hazard ratio 0.91; 95% confidence interval 0.86-0.96). CONCLUSIONS: Our results suggest that, contrary to what was previously reported in clinical trials, treatment gaps do not compromise the outcome of patients treated with ChIs in a real-life setting.

The CADEUS study: methods and logistics.

PURPOSE: At the request of the French Health authorities, a study called CADEUS (COX-2 inhibitors and NSAIDs: description of users) aimed to describe the users of cyclo-oxygenase (COX)-2 inhibitors and traditional non-selective non-steroidal anti-inflammatory drugs (tNSAIDs). We report here the methodology, logistics and study design performances. METHODS: CADEUS is a cohort study designed to include 40,000 patients randomly sampled monthly in the French National Healthcare Insurance database, who received at least one dispensation of celecoxib, rofecoxib or tNSAIDs (1:1:2), from September 2003 to August 2004. Patients and prescribers were asked to fill a questionnaire on indication, medical history, risk factors and hospitalizations since drug acquisition. There was no reminder. For each respondent, healthcare resources used for the 6 months before and after inclusion were extracted from the database. Response rate, response delay, responders and non-responders characteristics were assessed. RESULTS: Of the 222,879 patients and their prescribers contacted, 20.8% patients and 32.6% prescribers responded. Median response delay was 16 days for patients and 17 days for physicians. Factors associated with patient response were age, cohort, type of prescriber and period of inclusion. CONCLUSION: This is the first study of this design in France, combining data from a claims database and direct patient and prescriber questionnaires.

Channelling of COX-2 inhibitors to patients at higher gastrointestinal risk but not at lower cardiovascular risk: the Cox2 inhibitors and tNSAIDs description of users (CADEUS) study.

PURPOSE: To describe the characteristics of users of cyclo-oxygenase (COX)-2 inhibitors and traditional nonselective non-steroidal anti-inflammatory drugs (tNSAIDs) in France. METHODS: Between 1 August 2003 and 31 July 2004, patients who received at least one dispensing of celecoxib, rofecoxib or tNSAIDs were randomly sampled with a 1:1:2 target ratio within the French National Healthcare Insurance database. Patients and prescribers were asked to fill a questionnaire on socio-demographic characteristics, NSAID indication and use and previous medical history. For each respondent, healthcare resources used in the 6 months before inclusion were extracted from the database. Multivariate logistic regression was used to study the determinants of a first COX-2 inhibitor dispensing. RESULTS: Of the 45 217 patients included, 13 065 COX-2 inhibitors and 13 553 tNSAID users had prescriber data. Ninety seven per cent of COX-2 inhibitor prescriptions were for 'rheumatological' indications, whereas 37% of tNSAIDs use was for benign diseases (n = 2643) or analgesia (n = 2318). Among patients with rheumatological indications (n = 4730) and a first COX-2 inhibitor (n = 2427) or tNSAID (n = 2303) dispensing, multivariate analysis of factors associated with COX-2 inhibitors dispensing showed that, compared to new tNSAID users, new COX-2 inhibitor users were older, more often female, on sick leave or unemployed. COX-2 use was also associated with previous gastrointestinal history and previous gastroprotective agent dispensing, but not with previous cardiovascular (CV) history. CONCLUSION: The choice of NSAID depended largely on indication and on previous gastrointestinal history, in line with the recommendations of the French health authorities. Possible knowledge of CV risk associated with COX-2 inhibitors did not influence prescribing.

Correlated responses in growth hormone to selection for weight gain in mice.

SUMMARY: Correlated responses in pituitary gland weight and growth hormone (GH) concentration in the serum and pituitary were studied in lines of mice selected for growth rate, and in controls. The selection criteria were weight gain between 28 and 38 days on an ad libitum feed intake (EPA line), or on intake restricted to 80% of the control mice (EPR line), and weight gain between 48 and 58 days under the above two feeding regimes (LPA and LPR lines). The control line was maintained by random breeding. In generation 13, pituitary weight and growth hormone levels were determined at 38 days in lines EPA, EPR and the control and at 58 days in lines LPA, LPR and the control. Pituitary weight, corrected for body weight gain, was smaller in the EPA line than in the control line at 38 days and was greater in the LPR line than in the control line at 58 days. There were large differences in serum GH concentrations between the selected lines and the control; the differences were statistically significant at 38 days, with the EPA and EPR lines having lower levels than the controls. ZUSAMMENFASSUNG: Veränderung von Wachstumshormonspiegel bei Selektion für Zuwachs bei Mäusen Veränderungen im Hypophysengewicht und Wachstumshormon-(GH)-Konzentrationen im Serum und in der Hypophyse wurden bei Wachstumsselektionslinien und bei Kontrollmäusen untersucht. Selektionskriterien waren Zuwachs zwischen 28 und 38 Tagen bei ad libitum Futteraufnahme (EPA Linie) oder bei Linien, wo Futteraufnahme auf 80% derjenigen der Kontrollmäuse (EPR Linie) begrenzt wurde und Zuwachs zwischen 48 und 58 Tagen unter beiden Futterweisen (LPA und LPR Linie). Die Kontrollinie wurde durch Zufallspaarung fortgepflanzt. In Generation 13 wurden Hypophysengewicht und Wachstumshormonspiegel im Alter von 38 Tagen in Linien EPA, EPR und in Kontrollen und bei 58 Tagen in Linien LPA und LPR und Kontrollen bestimmt. Hypophysengewicht, korrigiert für Zuwachs, war geringer in der LPA Linie als in der Kontrolle bei 38 Tagen Alter und größer in der LPR Linie als in der Kontrolle bei 58 Tagen. Zwischen GH Konzentrationen selektierter und Kontrollinien waren erhebliche Differenzen, die bei 38 Tagen signifikant waren, wobei EPA und EPR Linien geringere Werte als die Kontrollen aufwiesen.

Temporal trends in spontaneous reporting of unlabelled adverse drug reactions.

AIMS: Trends in spontaneous reporting during the first years on the market were analyzed from a sample of selected drugs, with special attention to unlabelled effects. METHODS: Ten drugs were selected giving rise to approximately 100 spontaneous reports each during the first 4 years of marketing. Case reports were identified from the national pharmacovigilance database. A bibliographical score assigned at the time of reporting was used to identify unlabelled effects. Results were expressed as reporting rates. RESULTS: The average reporting rate peaked during the first year of marketing (54.6 per million treatment-months; s.d.: 62.8), then progressively decreased during the following years. Unlabelled effects effects represented 63% of all the spontaneous reports during the first year. CONCLUSIONS: Unlabelled adverse effects represent a high proportion of spontaneous reports during the early years of marketing.

Under-reporting of adverse drug reactions in general practice.

AIMS: In post-marketing setting, spontaneous reporting by physicians is a mode of surveillance of adverse effects associated with drug use. The objective of this study was to quantitatively assess under-reporting of adverse drug reactions (ADRs) in general practice. METHODS: A random sample of 100 general practitioners (GPs) practising in the region of the Bordeaux pharmacovigilance centre were surveyed to obtain data on adverse effects observed. Overall, 81 GPs agreed to record during 3 non-consecutive working days any effect they believed to be associated with drug use. The types of effects, regardless of their seriousness and labelling, and the drugs suspected were characterized and compared to spontaneous reports received from GPs by the Bordeaux pharmacovigilance centre during the reference period. RESULTS: The average number of ADRs observed per day per GP was 1.99. The estimate of the under-reporting coefficient (U) was 24,433 (95% confidence interval: 20,702-28,837) which indicates that, as a whole, GPs might be expected to report only 1 out of every 24,433 ADRs to the pharmacovigilance centre. Under-reporting was lowest for serious and unlabelled effects (U = 4610; 95% CI: 2514-8454) and for drugs marketed recently (U = 12,802; 95% CI: 8174-20,050). CONCLUSIONS: Adverse effects due to drugs are part of GPs routine activities. According to the observed trend in under-reporting, there appears to be a selection process which indicates that spontaneous reporting in general practice is not conducive to an exhaustive description of the safety profile of a drug. However, our findings are consistent with greater efficacy of spontaneous reporting in detecting serious and unlabelled effect.

Persistency of use of COX-2-specific inhibitors and non-specific non-steroidal anti-inflammatory drugs (NSAIDs) in Quebec.

The effectiveness of pharmacological therapies is dependent in part on patient persistency with the prescribed therapeutic regimen. In the case of non-specific non-steroidal anti-inflammatory drugs (NSAIDs), effectiveness is often compromised by undesirable side-effects, poor compliance or discontinuation of therapy. While patterns of utilization of non-specific NSAIDs have been investigated, few data are available on the patterns of persistency for cyclooxygenase (COX)-2-specific inhibitors. This study used a provincial health-care system database in Quebec, Canada, to determine the duration of treatment in new users of COX-2-specific inhibitors and non-specific NSAIDs over the first 3 months of treatment, and to characterize the factors associated with treatment persistency. Results demonstrate that the median duration of treatment was longer among patients initially prescribed COX-2-specific inhibitors (30 days and 23 days for celecoxib and rofecoxib respectively) than in those prescribed non-selective NSAIDs (10 days). Although the percentage of patients remaining on COX-2-specific drugs declined over the course of treatment, few patients on either celecoxib or rofecoxib switched drugs, either to the other COX-2-specific inhibitor or to non-specific NSAIDs. Factors associated with persistent drug use were: COX-2-specific inhibitors, age, and the use of gastroprotective agents either at treatment initiation or during follow-up. Dosage, chronic disease score and prescriber's specialty were only marginally associated with persistency. Prior use of gastroprotective agents was associated with lower persistency. Although the limitations of this study, which included lack of information on the indication for the prescription and the reason for switch or discontinuation, preclude definite conclusions regarding patterns of use of these drugs, the data suggest that the use of COX-2-specific inhibitors may result in increased persistency with treatment.