- Email: [email protected]
J-shaped relationship between supervised methadone consumption and retention in methadone maintenance treatment (MMT) in primary care: National cohort study
Drug and Alcohol Dependence 173 (2017) 126–131
Contents lists available at ScienceDirect
Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep
Full length article
J-shaped relationship between supervised methadone consumption and retention in methadone maintenance treatment (MMT) in primary care: National cohort study Gráinne Cousins a,∗,1 , Fiona Boland b,1 , Joseph Barry c , Suzi Lyons d , Eamon Keenan e , Denis O’Driscoll e , Kathleen Bennett f , Tom Fahey b a
School of Pharmacy, Royal College of Surgeons in Ireland, 123 Stephen’s Green, Dublin 2, Ireland HRB Centre for Primary Care Research, Royal College of Surgeons in Ireland, 123 Stephen’s Green, Dublin 2, Ireland c Trinity College Centre for Health Sciences, Tallaght Hospital, Dublin 24, Ireland d Health Research Board, Lower Mount Street, Dublin 2, Ireland e Addiction Services, Health Services Executive, Dublin, Ireland f Division of Population Health Sciences, Royal College of Surgeons in Ireland, 123 Stephen’s Green, Dublin 2, Ireland b
a r t i c l e
i n f o
Article history: Received 9 September 2016 Received in revised form 11 November 2016 Accepted 10 December 2016 Available online 25 January 2017 Keywords: Methadone maintenance treatment Treatment retention Supervised consumption Proportional hazards models
a b s t r a c t Background: Supervised consumption ensures patients take methadone as prescribed and prevents diversion, however, the influence of supervised consumption on retention is unclear. We examined association between supervised consumption and retention across multiple treatment episodes. Methods: Cohort study of persons experiencing ≥1 MMT episodes in primary care (2004–2010), excluding ongoing episodes at the start of follow-up. Length of treatment episodes based on methadone prescriptions, retention classified as no interruption in prescribed methadone lasting >7 days. When a patient did not receive a new prescription within seven days after the end of coverage of a prescription, they were considered to have ceased treatment. We evaluated the relationship between supervised consumption and time to discontinuation of treatment using proportional hazards gamma frailty models to account for recurrent MMT episodes. Age, gender, median daily methadone dose, and comorbidities included as potential confounders. Results: 6393 patients experienced 19,715 treatment episodes over the six-year follow-up period. A J-shaped relationship was observed; having between 20 and 60% of methadone scripts supervised (compared to <20%) associated with reduced time to discontinuation (20–39% HR = 0.88, 95% CI 0.81–0.95; 40–59%: HR = 0.87, 95% CI 0.81–0.94). Beyond a threshold of 60%, retention reduced (60–79% of scripts: HR = 1.28, 95% CI 1.20–1.36; >80% of scripts: HR = 3.59, 95% CI 3.38–3.81). Median daily dose between 60 and 120 mg/per day, and multiple treatment episodes also associated with longer time to discontinuation of treatment. Conclusion: A J-shaped relationship was observed between supervised consumption and retention in treatment. Additionally, patients experiencing multiple treatment episodes tend to stay in treatment for progressively longer periods of time. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The supervised administration of opioid substitution therapies, where the dose is consumed under the direct supervision of a pharmacist or clinician, is recommended by international guide-
∗ Corresponding author. E-mail address: [email protected] (G. Cousins). 1 Joint first author. http://dx.doi.org/10.1016/j.drugalcdep.2016.12.009 0376-8716/© 2017 Elsevier B.V. All rights reserved.
lines (WHO, 2009) and is standard practice in many countries. Supervised consumption ensures that patients take their medication as prescribed and prevents drug diversion. Diversion poses a risk to others and may result in patients being under-treated, and relapsing to heroin use (Green et al., 2000). A number of observational studies have shown that drug-related deaths are lower in areas with supervised consumption (Weinrich and Stuart, 2000), and that drug-related deaths (Seymour et al., 2003) or deaths due to methadone have fallen following the introduction of super-
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131
vised consumption, through reducing diversion to the black market (Strang et al., 2010). Supervised consumption also provides for more regular patient contact which may improve treatment outcomes (McLellan et al., 1993). However, long-term supervised dosing may not always be appropriate as it is labour-intensive, costly, and may impede patients from normalising their lives (Bell et al., 2007; Deering et al., 2011; Notley et al., 2014), which may result in patients dropping out of treatment. Several studies have shown that retention in opioid substitution treatment is associated with a reduced risk of mortality (Davoli et al., 2007; Cornish et al., 2010; Mathers et al., 2013; Cousins et al., 2016). However, few studies have compared the effects of supervised versus unsupervised consumption on retention (Rhoades et al., 1998; Bell et al., 2007; Holland et al., 2012, 2014), with conflicting results. One US RCT comparing supervised methadone consumption twice weekly versus five times weekly found a higher dropout rate in those supervised five times weekly (Rhoades et al., 1998). Similarly, a pilot RCT in Scotland found that retention decreased with increased supervision (Holland et al., 2012). Two RCTs conducted in Australia (Bell et al., 2007) and the UK (Holland et al., 2014) reported conflicting results, showing no difference in retention rates for those in supervised versus unsupervised dosing groups. These studies measured retention in treatment at between 12 and 24 weeks, which does not account for the chronic relapsing nature of opioid addiction and the well documented pattern of patients cycling in and out of treatment (Termorshuizen et al., 2005; Bell et al., 2006; Burns et al., 2009; Degenhardt et al., 2009; Cornish et al., 2010; Cousins et al., 2011; Zhang et al., 2015; Cousins et al., 2016). Furthermore, findings from a large cohort in Canada suggest that patients experiencing multiple treatment episodes tend to stay in treatment for progressively longer periods in later treatment attempts (Nosyk et al., 2009). This is important as findings from a Scottish cohort suggest that survival benefits increase with cumulative exposure to opiate substitution treatment (Kimber et al., 2010). Irish guidelines for prescribing methadone in primary care recommend a maintenance dose of between 60 mg and 120 mg daily, with at least one dose per week supervised in the pharmacy, and prescriptions for supply of methadone are issued for a period of not greater than seven days. The objective of this paper was to assess the effect of supervised methadone consumption on time to discontinuation of methadone maintenance treatment (MMT) across multiple treatment episodes in a national community based study of drug users in Ireland between 2004 and 2010.
2. Methods 2.1. Study design and setting People registered on the national register for methadone maintenance treatment, the Central Treatment List (CTL), aged between 16 and 65 years of age, who were prescribed and dispensed methadone in primary care between 1st August 2004 and 31st December 2010 were identified. All methadone dispensing records (date of prescription, duration, and quantity of methadone dispensed) in the Health Service Executive’s Methadone Treatment Scheme (MTS) were also extracted. To be included in the study a patient must have been dispensed at least three prescriptions over the study period. Following this, the MTS and CTL were linked to the General Medical Services (GMS) pharmacy claims database which contains details of all prescription medications, other than methadone, dispensed to GMS eligible patients. Eligibility for the GMS prescription scheme is through means test. All prescriptions are coded using WHO’s Anatomical Therapeutic Chemical (ATC) classification. No information on diagnosis or disease condition
127
is available. Finally, for every person registered on the CTL during our observation period, the National Drug Related Death Index (NDRDI), a census of drug-related deaths and deaths among drug users in Ireland was checked to identify those who died during the study period. Further details of the data and data linkage can be found in Cousins et al. (2016). 2.2. Treatment status We defined a patient as being “on treatment” based on the coverage of their methadone prescriptions. Similar to previous studies, if a patient received a new prescription within seven days of the end of their previous prescription’s coverage, they were considered to be in continuous treatment (Bell et al., 2006; Burns et al., 2009; Degenhardt et al., 2009). When a patient did not receive a new methadone prescription within seven days after the end of coverage of a prescription, that patient was considered to have ceased treatment. The “off treatment” period continued until a patient re-entered treatment as indicated by the presence of a new methadone prescription. This information was also used to calculate the episode number and duration of treatment episodes for each patient. We calculated the length of methadone treatment episodes from the number of days between the first and last prescription and the coverage of the last prescription. Only treatment episodes beginning after 1st August 2004 were included, to ensure consistency in the calculation of episode lengths and eliminate leftcensored observations. Treatment episodes that were ongoing at the end of the follow-up period were right-censored. 2.3. Supervised consumption (primary exposure) and potential confounders Prescription refill data was used to quantify the level of supervised methadone consumption for each treatment episode. Prescriptions that represented the dispensing of a single dose were classified as supervised. All other prescriptions, indicating methadone dispensing for a number of days, were classified as non-supervised. The percentage of supervised prescriptions per patient was then calculated and categorised into quintiles: less than 20%, 20–39%, 40–59%, 60–79% and 80% or greater supervised. Potential confounders included median dose of methadone for each episode for each person, which was categorised in accordance with UK prescribing recommendations as below, within, and above the recommended methadone maintenance range of 60 mg to 120 mg daily (Department of Health England, 2007). We also examined prescribing records for other drugs and calculated a comorbidity score as the total number of prescriptions for drugs, other than methadone, during the observation period. We counted the number of unique ATC codes (level three) appearing in the patients’ prescription data, so that repeated prescriptions of the same or very similar medicines, including different doses or formulation, were only counted once (Brilleman and Salisbury, 2013). We recorded drugs used for psychoses and related disorders as anti-psychotics, and also recorded the use of benzodiazepines, antidepressants, and opioid analgesics. These four groups were excluded from the comorbidity score. 2.4. Statistical analysis Cox proportional hazards gamma-frailty models can be fitted to account for the dependence in the length of repeated episodes and have been used in a similar study in Canada looking at the duration of repeated exposure to methadone maintenance treatment (Nosyk et al., 2009). Like standard Cox proportional hazards applications, the outcome is the bivariate pair (duration, censorship). In instances where there are multiple repeated durations of inter-
128
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131
est (in this case treatment episodes), the Cox proportional hazards gamma-frailty model provides a means of modelling these repeated measures within a standardised longitudinal framework. Like other mixed-effects modelling applications with longitudinal data, the Cox proportional hazards frailty model captures the correlation in episode lengths within an individual. The proportional hazards assumption was tested for each of the study variables. Schoenfeld residual plots were visually examined and tests of non-zero slopes performed. Hazard ratios >1 indicated faster time to discontinuation or shorter treatment episodes compared with the referent group. SAS version 9.3 (SAS Institute, Inc., Cary, North Carolina, USA) and R version 3.2.3 (R Foundation for Statistical Computing, Vienna, Austria) were used for all statistical analysis. 2.4.1. Sensitivity analyses. The treatment discontinuation rule of 7 days applied in this study may have misclassified time on and off treatment and introduced bias: for example, if a patient was still on treatment but did not fill a prescription within seven days of the end of their last prescription. Therefore we performed a sensitivity analysis extending the gap between the end of prescription coverage and new prescription to 28 days before classifying a patient as “offtreatment”. In addition, approximately 3% of the study cohort died during the follow-up period, inclusion of these cases may affect the interpretation of our findings, we therefore conducted a sensitivity analysis removing these cases. 3. Results 3.1. Description of the study population The original study cohort included 7129 patients aged 16–65 years receiving at least one methadone prescription during the observation period. Of these patients, 736 were excluded, 146 had insufficient prescription details over the observation period and 590 patients had only one treatment episode during the study period which was ongoing at the beginning of follow-up, resulting in 6393 patients. 3.2. Characteristics of cohort This cohort of 6393 people were prescribed and dispensed methadone during the six year study period, experienced a total of 19,715 MMT episodes during this time. The overall median episode length was 104 days (Interquartile range (IQR): 38-412), with 20% of all episodes ongoing at the end of follow-up. Characteristics of the study sample are presented in Table 1. More than half of the patients entering treatment were aged under 30 years and 68.5% were men. Almost one-third of patients experienced only one treatment episode, with a median of two treatment episodes per person recorded. More than one-third of patients received a median methadone dose below the recommended maintenance dose of 60–120 mg daily in their initial treatment episode. Thirtysix per cent of patients had less than 20% of their prescriptions supervised in their initial treatment episode, with almost 30% of patients recorded as having 80% of their prescriptions supervised. A history of co-prescription of benzodiazepines and anti-depressants was high. Almost 78% of patients had less than 5 prescriptions (unique level 3 ATC codes), excluding methadone, benzodiazepines, antipsychotics, antidepressants and opioids. 3.3. Time to discontinuation of MMT Graphical analysis and schoenfeld residuals revealed no substantial departure from the assumptions of proportional hazards. The plots suggested that the variation over time was small relative to the size of the time averaged coefficient value. Table 2
Table 1 Characteristics of patients and treatment episodes. Characteristics
Number (%)a
Patients Male sex Age (years) at start of study 16–19 20–29 30–39 40–65
N = 6393 4377 (68.5) 519 (8.1) 3230 (50.5) 1996 (31.2) 648 (10.1)
Total number of treatment episodes per patient One Two Three Four five or more
2105 (32.9) 1475 (23.1) 995 (15.6) 627 (9.8) 1191 (18.6)
Total length of treatmentb Median (IQR) days
224 (55–553)
b
Median methadone dose <60 mg/day 60–120 mg/day ≥120 mg/day
2375 (37.2) 3894 (60.9) 124 (1.9)
Supervised methadone consumptionc <20% 20–39% 40–59% 60–79% ≥80%
2299 (36.0) 457 (7.2) 561 (8.8) 1219 (19.1) 1857 (29.1)
Co-prescribing Benzodiazepines Antipsychotics Antidepressants Opioid analgesics
4586 (71.7) 1487 (23.3) 3119 (48.8) 2578 (40.3)
Comorbidity scored 0–5 6–10 11–15 >15
4972 (77.8) 1183 (18.5) 200 (3.1) 38 (0.6)
a
Unless otherwise specified. Based on patients initial treatment episode in the follow-up period. c The number and percentage of methadone scripts dispensed as supervised consumption, based on patients initial treatment episode in the follow-up period. d Measured as the number of prescriptions (unique level 3 ATC codes) received by patient, excluding methadone, benzodiazepines, antipsychotics, antidepressants and opioids. b
presents the results from the proportional hazards gamma frailty model. The results show a j-shaped relationship for the primary exposure, supervised methadone consumption, such that having between 20% and 59% of their methadone scripts dispensed under supervised consumption (compared to <20%) was associated with increased retention in treatment (longer time to discontinuation) (20–39% of scripts: HR = 0.88, 95% CI 0.81–0.95; 40–59% of scripts: HR = 0.87, 95% CI 0.81–0.94). However, having 60% or more of their methadone scripts dispensed under supervised consumption (compared to <20%) was associated with shorter treatment episodes (faster time to discontinuation) (60–79% of scripts: HR = 1.28, 95% CI 1.20–1.36; >=80% of scripts: HR = 3.59, 95% CI 3.38–3.81). Later attempts (3 or more) at MMT, compared to the first treatment episode, were also associated with longer times to discontinuation (episode 3: HR 0.93, 95% CI: 0.87–0.98; episode 4: HR 0.92, 95% CI: 0.86–0.98; episode 5 or more: HR 0.82, 95% CI: 0.78–0.88), as was a daily dose between 60–120 mgs per day (compared to <60 mgs or >120 mgs per day). Finally, patients with a comorbidity score of 11 or more (compared to 0–5) were significantly more likely to have shorter treatment episodes.
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131 Table 2 Proportional hazards gamma frailty model of time to discontinuation of treatment. Adjusted Hazard Ratio (95% Confidence Limits)
P-value
Sex Females Males
1.00 1.05 (0.98–1.11)
0.203
Age (years) 16–19 20–29 30–39 40–65
1.00 1.09 (0.97–1.21) 1.07 (0.96–1.20) 1.25 (1.10–1.43)
0.279 0.378 0.007
Treatment episode First Second Third Fourth Fifth or above
1.00 1.00 (0.95–1.05) 0.93 (0.87–0.98) 0.92 (0.86–0.98) 0.82 (0.78–0.88)
0.869 0.012 0.019 <0.001
Median methadone dose 2.11 (2.02–2.21) <60 mg/day 1.00 60–120 mg/day ≥120 mg/day 1.31 (1.13–1.52)
<0.001 <0.001
Supervised consumption 1.00 <20% 0.88 (0.81–0.95) 20–39% 0.87 (0.81–0.94) 40–59% 60–79% 1.28 (1.20–1.36) ≥80% 3.59 (3.38–3.81)
<0.001 <0.001 <0.001 <0.001
Median comorbidity scorea 0–5 1.00 1.06 (0.98–1.14) 6–10 1.20 (1.03–1.41) 11–15 1.48 (1.06–2.06) >15
0.128 0.021 0.020
Antipsychotics No Yes
1.00 1.00 (0.93–1.07)
0.975
Benzodiazepines No Yes
1.00 0.93 (0.87–1.00)
0.065
Opioid analgesics No Yes
1.00 1.03 (0.97–1.09)
0.406
Antidepressants No Yes
1.00 1.03 (0.97–1.10)
0.301
The bold is highlighting those variables that are statistically significant i.e. <0.05. a As measured by number of prescriptions (unique level 3 ATC codes) received by patient, excluding methadone, benzodiazepines, antipsychotics, antidepressants and opioids received by patient in calendar year.
129
was observed, such that having between 20 and 60% of methadone prescriptions dispensed under supervised consumption (relative to <20%) was associated with increased retention in treatment. The protective effects of retaining patients were lost when the frequency of supervision increased above 60%, resulting in a reduction in retention in treatment. These findings are consistent with previous RCTs which suggest that increased supervision is associated with a reduction in retention in MMT (Rhoades et al., 1998; Holland et al., 2012). Other studies reported no difference in retention between those supervised or unsupervised (Bell et al., 2007; Gerra et al., 2011). A large UK RCT, found no between group difference in retention by intention-to-treat analysis, however per protocol analysis suggested a reduction in retention for those exposed to supervised consumption (Holland et al., 2014). The J-shaped relationship observed in our study reflects findings from a recent qualitative study which suggest that patients consider supervised consumption as acceptable in the short-term, helping to establish a regular daily routine and to develop positive relationships with pharmacy staff, but that in the long-term being unsupervised was the preference (Notley et al., 2014). Consistent with previous studies we also found that patients experienced multiple treatment episodes over the 6-year period (Bell et al., 2006; Burns et al., 2009; Degenhardt et al., 2009). However, similar to a large cohort in Canada, patients with multiple treatment episodes remained in MMT for progressively longer periods in later treatment episodes. (Nosyk et al., 2009). Methadone dosages ranging from 60 to 120 mg/day were also shown to be more effective than lower doses in retaining patients, this is consistent with a number of previous studies (Faggiano et al., 2003; Nosyk et al., 2009; Cao et al., 2014; Vigna-Taglianti et al., 2016). Despite these findings and the wide consensus on 60–120 mg as the optimal daily dose range, over one in three patients did not receive the minimum optimal dose of 60 mg/day. This finding is comparable with previous international studies (Strang et al., 2010; Cao et al., 2014; Nosyk et al., 2009). However, the proportion of patients receiving methadone below the minimum recommended dose of 60 mg/day in Ireland appears to be lower than in the UK where 56.8% of patients in MMT in primary care were prescribed methadone below 60 mg/day (Strang et al., 2010). Similarly, a Canadian cohort study of 17,005 patients in MMT between 1996 and 2007 found that 50.5% of patients had a mean daily dose below 60 mg (Nosyk et al., 2009). Prescribing practices should be improved to ensure that patients are receiving the minimum optimal dose of between 60 and 120 mg/day. 4.2. Strengths and weaknesses
3.4. Sensitivity analyses and assumption checks We extended the treatment discontinuation rule from 7 days to 28 days to assess potential misclassification bias. This analysis showed that the results were similar and the coefficient estimates were robust to changes in the treatment discontinuation rule (see Supporting information, Table S1). Similarly, excluding patients who died during the study follow-up did not influence the inferences made (see Supporting information, Table S2). 4. Discussion 4.1. Principal findings in relation to other studies To our knowledge, this is the first study to assess the relationship between supervised methadone consumption and retention in treatment across multiple treatment episodes in primary care. We found a J-shaped relationship between supervised consumption and time to discontinuation of treatment. A J-shaped relationship
The external validity of this study is high, as it includes all patients on a national prescribing register over a six-year study period. Furthermore, using the proportional hazards gamma frailty model to account for recurrent treatment episodes, we were able to use all patient data available rather than restricting the analysis to a single treatment episode. This method of analysis also allows for a more accurate inference regarding the effects of successive treatment episodes as it accounts for the correlation between repeated treatment episodes and unobserved heterogeneity across individual patients (Nosyk et al., 2009). Previous studies have generally reported on a short duration of follow-up, ranging between 12 and 24 weeks, which fail to account for the patients’ movement in and out of treatment overtime (Rhoades et al., 1998; Bell et al., 2007; Holland et al., 2012, 2014). Our study has a number of shortcomings and potential biases. Our results relate to patient’s retention in primary-care methadone treatment programmes, patients transferring from primary care to specialist care settings would be lost in our cohort, and incorrectly classified as discontinuing treatment. Secondly, classifying
130
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131
patients as ceasing treatment after 7 days without a methadone prescription may have introduced a misclassification bias, resulting in an underestimate of retention in treatment. However, retention rates remained consistent in our sensitivity analysis extending the gap between the end of prescription coverage and new prescription to 28 days before classifying a patient as “off-treatment”. Thirdly, we do not have information on the quality or intensity of the care received, which may influence patients’ retention in treatment (Cornish et al., 2010). Finally, in observational studies of this sort, the possibility of residual confounding or confounding by indication may remain a problem, therefore associations identified in this study should be viewed principally as hypothesis generating and our observed associations should be subject to testing and verification in other national observational cohorts. 4.3. Clinical implications The J-shaped relationship observed between supervised consumption and retention in MMT in primary care, represents a double edge sword; too little supervision (<20% of methadone prescriptions) is associated with treatment cessation, this risk of treatment cessation decreases with regular supervision (20% to 60% of methadone scripts), but eventually increases above the reference group (those with little supervision) when a patient is exposed to higher levels of supervised consumption (>60% of methadone prescriptions). Losing patients from treatment may expose patients to an increased risk of mortality, as previous studies have identified an elevated risk of mortality following treatment cessation (Davoli et al., 2007; Cornish et al., 2010; Cousins et al., 2016). However, reducing supervision with the aim of retaining a patient in treatment and reducing their individual risk of mortality, needs to be balanced against the risk of diversion and increased availability of illicit methadone which may increase population level risk of methadone related deaths (Strang et al., 2010). Our study cannot be used to indicate who should be supervised or the optimal frequency of supervision to ensure that patients are retained in treatment while simultaneously avoiding diversion of methadone. However, our study is the first study to examine the longitudinal association between supervised consumption over the course of multiple treatment episodes and retention in treatment. While a number of authors have suggested flexibly timed discontinuation of supervision once a patient is stabilised (Holland et al., 2014; Notley et al., 2014), recent Australian studies have highlighted the challenges involved in assessing patient stability with many OST prescribers reporting high levels of uncertainty in identifying whether their patients are diverting or injecting their medication (Larance et al., 2011, 2014). Further work is needed to determine how to optimally stratify patients’ suitability for unsupervised dosing, with the aim of retaining patients in treatment and simultaneously reducing diversion and injecting. 4.4. Conclusion This national community based cohort study of over 6000 patients identified a J-shaped relationship between supervised methadone consumption and retention in treatment over a six-year follow-up. A median daily dose between 60 and 120 mg/per day, and multiple treatment episodes were also significant predictors of longer time to discontinuation of treatment. Role of funding source This work was funded by the Health Research Board (HRB) of Ireland through the HRB Centre for Primary Care Research under Grant HRC/2007/1. The funders had no input into the study design; the collection, analysis, and interpretation of data; the writing of
the report; or the decision to submit the article for publication. The investigators are fully independent of the funders and have not used a third party to write or prepare the manuscript. Conflict of interest No conflict declared. Contributors GC designed the study, secured the data, interpreted the results and drafted the manuscript. FB managed the data, including data linkage, performed the analyses and drafted the results. JB secured data, revised the manuscript and provided critical comments. SL secured data, revised the manuscript and provided critical comments.EK revised the manuscript and provided critical comments. DOD revised the manuscript and provided critical comments. KB revised the manuscript and provided critical comments. TF designed the study, revised the manuscript and provided critical comments. GC and FB had full access to all of the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. All authors approved of the final manuscript. Acknowledgments This work was funded by the Health Research Board (HRB) of Ireland through the HRB Centre for Primary Care Research under Grant HRC/2007/1. The authors would like to thank the HSE Primary Care Reimbursement Services for supplying the prescription dispensing data, and the Central Statistics Office for acting as an independent data processor to facilitate the data linkage process. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.drugalcdep.2016. 12.009. References Bell, J., Burrell, T., Indig, D., Gilmour, S., 2006. Cycling in and out of treatment; participation in methadone treatment in NSW, 1990–2002. Drug Alcohol Depend. 81, 55–61. Bell, J., Shanahan, M., Mutch, C., Rea, F., Ryan, A., Batey, R., Dunlop, A., Winstock, A., 2007. A randomized trial of effectiveness and cost-effectiveness of observed versus unobserved administration of buprenorphine-naloxone for heroin dependence. Addiction 102, 1899–1907. Brilleman, S.L., Salisbury, C., 2013. Comparing measures of multimorbidity to predict outcomes in primary care: a cross sectional study. Fam. Pract. 30, 172–178. Burns, L., Randall, D., Hall, W.D., Law, M., Butler, T., Bell, J., Degenhardt, L., 2009. Opioid agonist pharmacotherapy in New South Wales from 1985 to 2006: patient characteristics and patterns and predictors of treatment retention. Addiction 104, 1363–1372. Cao, X., Wu, Z., Rou, K., Li, L., Lin, C., Wang, C., Luo, W., Pang, L., Yin, W., Li, J., 2014. Retention and its predictors among methadone maintenance treatment clients in China: a six-year cohort study. Drug Alcohol Depend. 145, 87–93. Cornish, R., Macleod, J., Strang, J., Vickerman, P., Hickman, M., 2010. Risk of death during and after opiate substitution treatment in primary care: prospective observational study in UK general practice research database. BMJ 341, c5475. Cousins, G., Teljeur, C., Motterlini, N., McCowan, C., Dimitrov, B.D., Fahey, T., 2011. Risk of drug-related mortality during periods of transition in methadone maintenance treatment: a cohort study. J. Subst. Abuse Treat. 41, 252–260. Cousins, G., Boland, F., Courtney, B., Barry, J., Lyons, S., Fahey, T., 2016. Risk of mortality on and off methadone substitution treatment in primary care: a national cohort study. Addiction 111, 73–82. Davoli, M., Bargagli, A.M., Perucci, C.A., Schifano, P., Belleudi, V., Hickman, M., Salamina, G., Diecidue, R., Vigna-Taglianti, F., Faggiano, F., 2007. Risk of fatal overdose during and after specialist drug treatment: the VEdeTTE study, a national multi-site prospective cohort study. Addiction 102, 1954–1959. Deering, D.E., Sheridan, J., Sellman, J.D., Adamson, S.J., Pooley, S., Robertson, R., Henderson, C., 2011. Consumer and treatment provider perspectives on
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131 reducing barriers to opioid substitution treatment and improving treatment attractiveness. Addict. Behav. 36, 636–642. Degenhardt, L., Randall, D., Hall, W., Law, M., Butler, T., Burns, L., 2009. Mortality among clients of a state-wide opioid pharmacotherapy program over 20 years: risk factors and lives saved. Drug Alcohol Depend. 105, 9–15. Department of Health (England) and the devolved administrations, 2007. Drug Misuse and Dependence: UK Guidelines on Clinical Management. Department of Health (England), the Scottish Government, Welsh Assembly Government and Northern Ireland Executive, London. Faggiano, F., Vigna-Taglianti, F., Versino, E., Lemma, P., 2003. Methadone maintenance at different dosages for opioid dependence. Cochrane Database Syst. Rev. 3, Cd002208. Gerra, G., Saenz, E., Busse, A., Maremmani, I., Ciccocioppo, R., Zaimovic, A., Gerra, M.L., Amore, M., Manfredini, M., Donnini, C., Somaini, L., 2011. Supervised daily consumption, contingent take-home incentive and non-contingent take-home in methadone maintenance. Prog. Neuropsychopharmacol. Biol. Psychiatry 35, 483–489. Green, H., James, R.A., Gilbert, J.D., Harpas, P., Byard, R.W., 2000. Methadone maintenance programs—a two-edged sword? Am. J. Forensic Med. Pathol. 21, 359–361. Holland, R., Matheson, C., Anthony, G., Roberts, K., Priyardarshi, S., Macrae, A., Whitelaw, E., Appavoo, S., Bond, C., 2012. A pilot randomised controlled trial of brief versus twice weekly versus standard supervised consumption in patients on opiate maintenance treatment. Drug Alcohol Rev. 31, 483–491. Holland, R., Maskrey, V., Swift, L., Notley, C., Robinson, A., Nagar, J., Gale, T., Kouimtsidis, C., 2014. Treatment retention, drug use and social functioning outcomes in those receiving 3 months versus 1 month of supervised opioid maintenance treatment. Results from the Super C randomized controlled trial. Addiction 109, 596–607. Kimber, J., Copeland, L., Hickman, M., Macleod, J., McKenzie, J., De Angelis, D., Robertson, J.R., 2010. Survival and cessation in injecting drug users: prospective observational study of outcomes and effect of opiate substitution treatment. BMJ 341, c3172. Larance, B., Degenhardt, L., O’Brien, S., Lintzeris, N., Winstock, A., Mattick, R.P., Bell, J., Ali, R., 2011. Prescribers’ perceptions of the diversion and injection of medication by opioid substitution treatment patients. Drug Alcohol Rev. 30, 613–620. Larance, B., Carragher, N., Mattick, R.P., Lintzeris, N., Ali, R., Degenhardt, L., 2014. A latent class analysis of self-reported clinical indicators of psychosocial stability and adherence among opioid substitution therapy patients: do stable patients receive more unsupervised doses? Drug Alcohol Depend. 142, 46–55.
131
Mathers, B.M., Degenhardt, L., Bucello, C., Lemon, J., Wiessing, L., Hickman, M., 2013. Mortality among people who inject drugs: a systematic review and meta-analysis. Bull. World Health Organ. 91, 102–123. McLellan, A.T., Arndt, I.O., Metzger, D.S., Woody, G.E., O’Brien, C.P., 1993. The effects of psychosocial services in substance abuse treatment. JAMA 269, 1953–1959. Nosyk, B., MacNab, Y.C., Sun, H., Fischer, B., Marsh, D.C., Schechter, M.T., Anis, A.H., 2009. Proportional hazards frailty models for recurrent methadone maintenance treatment. Am. J. Epidemiol. 170, 783–792. Notley, C., Holland, R., Maskrey, V., Nagar, J., Kouimtsidis, C., 2014. Regaining control: the patient experience of supervised compared with unsupervised consumption in opiate substitution treatment. Drug Alcohol Rev. 33, 64–70. Rhoades, H.M., Creson, D., Elk, R., Schmitz, J., Grabowski, J., 1998. Retention, HIV risk, and illicit drug use during treatment: methadone dose and visit frequency. Am. J. Publ. Health 88, 34–39. Seymour, A., Black, M., Jay, J., Cooper, G., Weir, C., Oliver, J., 2003. The role of methadone in drug-related deaths in the west of Scotland. Addiction 98, 995–1002. Strang, J., Hall, W., Hickman, M., Bird, S.M., 2010. Impact of supervision of methadone consumption on deaths related to methadone overdose (1993–2008): analyses using OD4 index in England and Scotland. BMJ 341, c4851. Termorshuizen, F., Krol, A., Prins, M., Geskus, R., van den Brink, W., van Ameijden, E.J., 2005. Prediction of relapse to frequent heroin use and the role of methadone prescription: an analysis of the Amsterdam cohort study among drug users. Drug Alcohol Depend. 79, 231–240. Vigna-Taglianti, F.D., Burroni, P., Mathis, F., Versino, E., Beccaria, F., Rotelli, M., Garneri, M., Picciolini, A., Bargagli, A.M., 2016. Gender differences in heroin addiction and treatment: results from the VEdeTTE cohort. Subst. Use Misuse 51, 295–309. World Health Organization (WHO), 2009. Guidelines for the Psychosocially Assisted Pharmacological Treatment of Opioid Dependence. WHO, Geneva. Weinrich, M., Stuart, M., 2000. Provision of methadone treatment in primary care medical practices: review of the Scottish experience and implications for US policy. JAMA 283, 1343–1348. Zhang, L., Zou, X., Zhang, D., Li, X., Zhao, P., Ling, L., 2015. Investigation of repeat client drop-out and re-enrolment cycles in fourteen methadone maintenance treatment clinics in Guangdong, China. PLoS One 10, e0139942.
Contents lists available at ScienceDirect
Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep
Full length article
J-shaped relationship between supervised methadone consumption and retention in methadone maintenance treatment (MMT) in primary care: National cohort study Gráinne Cousins a,∗,1 , Fiona Boland b,1 , Joseph Barry c , Suzi Lyons d , Eamon Keenan e , Denis O’Driscoll e , Kathleen Bennett f , Tom Fahey b a
School of Pharmacy, Royal College of Surgeons in Ireland, 123 Stephen’s Green, Dublin 2, Ireland HRB Centre for Primary Care Research, Royal College of Surgeons in Ireland, 123 Stephen’s Green, Dublin 2, Ireland c Trinity College Centre for Health Sciences, Tallaght Hospital, Dublin 24, Ireland d Health Research Board, Lower Mount Street, Dublin 2, Ireland e Addiction Services, Health Services Executive, Dublin, Ireland f Division of Population Health Sciences, Royal College of Surgeons in Ireland, 123 Stephen’s Green, Dublin 2, Ireland b
a r t i c l e
i n f o
Article history: Received 9 September 2016 Received in revised form 11 November 2016 Accepted 10 December 2016 Available online 25 January 2017 Keywords: Methadone maintenance treatment Treatment retention Supervised consumption Proportional hazards models
a b s t r a c t Background: Supervised consumption ensures patients take methadone as prescribed and prevents diversion, however, the influence of supervised consumption on retention is unclear. We examined association between supervised consumption and retention across multiple treatment episodes. Methods: Cohort study of persons experiencing ≥1 MMT episodes in primary care (2004–2010), excluding ongoing episodes at the start of follow-up. Length of treatment episodes based on methadone prescriptions, retention classified as no interruption in prescribed methadone lasting >7 days. When a patient did not receive a new prescription within seven days after the end of coverage of a prescription, they were considered to have ceased treatment. We evaluated the relationship between supervised consumption and time to discontinuation of treatment using proportional hazards gamma frailty models to account for recurrent MMT episodes. Age, gender, median daily methadone dose, and comorbidities included as potential confounders. Results: 6393 patients experienced 19,715 treatment episodes over the six-year follow-up period. A J-shaped relationship was observed; having between 20 and 60% of methadone scripts supervised (compared to <20%) associated with reduced time to discontinuation (20–39% HR = 0.88, 95% CI 0.81–0.95; 40–59%: HR = 0.87, 95% CI 0.81–0.94). Beyond a threshold of 60%, retention reduced (60–79% of scripts: HR = 1.28, 95% CI 1.20–1.36; >80% of scripts: HR = 3.59, 95% CI 3.38–3.81). Median daily dose between 60 and 120 mg/per day, and multiple treatment episodes also associated with longer time to discontinuation of treatment. Conclusion: A J-shaped relationship was observed between supervised consumption and retention in treatment. Additionally, patients experiencing multiple treatment episodes tend to stay in treatment for progressively longer periods of time. © 2017 Elsevier B.V. All rights reserved.
1. Introduction The supervised administration of opioid substitution therapies, where the dose is consumed under the direct supervision of a pharmacist or clinician, is recommended by international guide-
∗ Corresponding author. E-mail address: [email protected] (G. Cousins). 1 Joint first author. http://dx.doi.org/10.1016/j.drugalcdep.2016.12.009 0376-8716/© 2017 Elsevier B.V. All rights reserved.
lines (WHO, 2009) and is standard practice in many countries. Supervised consumption ensures that patients take their medication as prescribed and prevents drug diversion. Diversion poses a risk to others and may result in patients being under-treated, and relapsing to heroin use (Green et al., 2000). A number of observational studies have shown that drug-related deaths are lower in areas with supervised consumption (Weinrich and Stuart, 2000), and that drug-related deaths (Seymour et al., 2003) or deaths due to methadone have fallen following the introduction of super-
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131
vised consumption, through reducing diversion to the black market (Strang et al., 2010). Supervised consumption also provides for more regular patient contact which may improve treatment outcomes (McLellan et al., 1993). However, long-term supervised dosing may not always be appropriate as it is labour-intensive, costly, and may impede patients from normalising their lives (Bell et al., 2007; Deering et al., 2011; Notley et al., 2014), which may result in patients dropping out of treatment. Several studies have shown that retention in opioid substitution treatment is associated with a reduced risk of mortality (Davoli et al., 2007; Cornish et al., 2010; Mathers et al., 2013; Cousins et al., 2016). However, few studies have compared the effects of supervised versus unsupervised consumption on retention (Rhoades et al., 1998; Bell et al., 2007; Holland et al., 2012, 2014), with conflicting results. One US RCT comparing supervised methadone consumption twice weekly versus five times weekly found a higher dropout rate in those supervised five times weekly (Rhoades et al., 1998). Similarly, a pilot RCT in Scotland found that retention decreased with increased supervision (Holland et al., 2012). Two RCTs conducted in Australia (Bell et al., 2007) and the UK (Holland et al., 2014) reported conflicting results, showing no difference in retention rates for those in supervised versus unsupervised dosing groups. These studies measured retention in treatment at between 12 and 24 weeks, which does not account for the chronic relapsing nature of opioid addiction and the well documented pattern of patients cycling in and out of treatment (Termorshuizen et al., 2005; Bell et al., 2006; Burns et al., 2009; Degenhardt et al., 2009; Cornish et al., 2010; Cousins et al., 2011; Zhang et al., 2015; Cousins et al., 2016). Furthermore, findings from a large cohort in Canada suggest that patients experiencing multiple treatment episodes tend to stay in treatment for progressively longer periods in later treatment attempts (Nosyk et al., 2009). This is important as findings from a Scottish cohort suggest that survival benefits increase with cumulative exposure to opiate substitution treatment (Kimber et al., 2010). Irish guidelines for prescribing methadone in primary care recommend a maintenance dose of between 60 mg and 120 mg daily, with at least one dose per week supervised in the pharmacy, and prescriptions for supply of methadone are issued for a period of not greater than seven days. The objective of this paper was to assess the effect of supervised methadone consumption on time to discontinuation of methadone maintenance treatment (MMT) across multiple treatment episodes in a national community based study of drug users in Ireland between 2004 and 2010.
2. Methods 2.1. Study design and setting People registered on the national register for methadone maintenance treatment, the Central Treatment List (CTL), aged between 16 and 65 years of age, who were prescribed and dispensed methadone in primary care between 1st August 2004 and 31st December 2010 were identified. All methadone dispensing records (date of prescription, duration, and quantity of methadone dispensed) in the Health Service Executive’s Methadone Treatment Scheme (MTS) were also extracted. To be included in the study a patient must have been dispensed at least three prescriptions over the study period. Following this, the MTS and CTL were linked to the General Medical Services (GMS) pharmacy claims database which contains details of all prescription medications, other than methadone, dispensed to GMS eligible patients. Eligibility for the GMS prescription scheme is through means test. All prescriptions are coded using WHO’s Anatomical Therapeutic Chemical (ATC) classification. No information on diagnosis or disease condition
127
is available. Finally, for every person registered on the CTL during our observation period, the National Drug Related Death Index (NDRDI), a census of drug-related deaths and deaths among drug users in Ireland was checked to identify those who died during the study period. Further details of the data and data linkage can be found in Cousins et al. (2016). 2.2. Treatment status We defined a patient as being “on treatment” based on the coverage of their methadone prescriptions. Similar to previous studies, if a patient received a new prescription within seven days of the end of their previous prescription’s coverage, they were considered to be in continuous treatment (Bell et al., 2006; Burns et al., 2009; Degenhardt et al., 2009). When a patient did not receive a new methadone prescription within seven days after the end of coverage of a prescription, that patient was considered to have ceased treatment. The “off treatment” period continued until a patient re-entered treatment as indicated by the presence of a new methadone prescription. This information was also used to calculate the episode number and duration of treatment episodes for each patient. We calculated the length of methadone treatment episodes from the number of days between the first and last prescription and the coverage of the last prescription. Only treatment episodes beginning after 1st August 2004 were included, to ensure consistency in the calculation of episode lengths and eliminate leftcensored observations. Treatment episodes that were ongoing at the end of the follow-up period were right-censored. 2.3. Supervised consumption (primary exposure) and potential confounders Prescription refill data was used to quantify the level of supervised methadone consumption for each treatment episode. Prescriptions that represented the dispensing of a single dose were classified as supervised. All other prescriptions, indicating methadone dispensing for a number of days, were classified as non-supervised. The percentage of supervised prescriptions per patient was then calculated and categorised into quintiles: less than 20%, 20–39%, 40–59%, 60–79% and 80% or greater supervised. Potential confounders included median dose of methadone for each episode for each person, which was categorised in accordance with UK prescribing recommendations as below, within, and above the recommended methadone maintenance range of 60 mg to 120 mg daily (Department of Health England, 2007). We also examined prescribing records for other drugs and calculated a comorbidity score as the total number of prescriptions for drugs, other than methadone, during the observation period. We counted the number of unique ATC codes (level three) appearing in the patients’ prescription data, so that repeated prescriptions of the same or very similar medicines, including different doses or formulation, were only counted once (Brilleman and Salisbury, 2013). We recorded drugs used for psychoses and related disorders as anti-psychotics, and also recorded the use of benzodiazepines, antidepressants, and opioid analgesics. These four groups were excluded from the comorbidity score. 2.4. Statistical analysis Cox proportional hazards gamma-frailty models can be fitted to account for the dependence in the length of repeated episodes and have been used in a similar study in Canada looking at the duration of repeated exposure to methadone maintenance treatment (Nosyk et al., 2009). Like standard Cox proportional hazards applications, the outcome is the bivariate pair (duration, censorship). In instances where there are multiple repeated durations of inter-
128
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131
est (in this case treatment episodes), the Cox proportional hazards gamma-frailty model provides a means of modelling these repeated measures within a standardised longitudinal framework. Like other mixed-effects modelling applications with longitudinal data, the Cox proportional hazards frailty model captures the correlation in episode lengths within an individual. The proportional hazards assumption was tested for each of the study variables. Schoenfeld residual plots were visually examined and tests of non-zero slopes performed. Hazard ratios >1 indicated faster time to discontinuation or shorter treatment episodes compared with the referent group. SAS version 9.3 (SAS Institute, Inc., Cary, North Carolina, USA) and R version 3.2.3 (R Foundation for Statistical Computing, Vienna, Austria) were used for all statistical analysis. 2.4.1. Sensitivity analyses. The treatment discontinuation rule of 7 days applied in this study may have misclassified time on and off treatment and introduced bias: for example, if a patient was still on treatment but did not fill a prescription within seven days of the end of their last prescription. Therefore we performed a sensitivity analysis extending the gap between the end of prescription coverage and new prescription to 28 days before classifying a patient as “offtreatment”. In addition, approximately 3% of the study cohort died during the follow-up period, inclusion of these cases may affect the interpretation of our findings, we therefore conducted a sensitivity analysis removing these cases. 3. Results 3.1. Description of the study population The original study cohort included 7129 patients aged 16–65 years receiving at least one methadone prescription during the observation period. Of these patients, 736 were excluded, 146 had insufficient prescription details over the observation period and 590 patients had only one treatment episode during the study period which was ongoing at the beginning of follow-up, resulting in 6393 patients. 3.2. Characteristics of cohort This cohort of 6393 people were prescribed and dispensed methadone during the six year study period, experienced a total of 19,715 MMT episodes during this time. The overall median episode length was 104 days (Interquartile range (IQR): 38-412), with 20% of all episodes ongoing at the end of follow-up. Characteristics of the study sample are presented in Table 1. More than half of the patients entering treatment were aged under 30 years and 68.5% were men. Almost one-third of patients experienced only one treatment episode, with a median of two treatment episodes per person recorded. More than one-third of patients received a median methadone dose below the recommended maintenance dose of 60–120 mg daily in their initial treatment episode. Thirtysix per cent of patients had less than 20% of their prescriptions supervised in their initial treatment episode, with almost 30% of patients recorded as having 80% of their prescriptions supervised. A history of co-prescription of benzodiazepines and anti-depressants was high. Almost 78% of patients had less than 5 prescriptions (unique level 3 ATC codes), excluding methadone, benzodiazepines, antipsychotics, antidepressants and opioids. 3.3. Time to discontinuation of MMT Graphical analysis and schoenfeld residuals revealed no substantial departure from the assumptions of proportional hazards. The plots suggested that the variation over time was small relative to the size of the time averaged coefficient value. Table 2
Table 1 Characteristics of patients and treatment episodes. Characteristics
Number (%)a
Patients Male sex Age (years) at start of study 16–19 20–29 30–39 40–65
N = 6393 4377 (68.5) 519 (8.1) 3230 (50.5) 1996 (31.2) 648 (10.1)
Total number of treatment episodes per patient One Two Three Four five or more
2105 (32.9) 1475 (23.1) 995 (15.6) 627 (9.8) 1191 (18.6)
Total length of treatmentb Median (IQR) days
224 (55–553)
b
Median methadone dose <60 mg/day 60–120 mg/day ≥120 mg/day
2375 (37.2) 3894 (60.9) 124 (1.9)
Supervised methadone consumptionc <20% 20–39% 40–59% 60–79% ≥80%
2299 (36.0) 457 (7.2) 561 (8.8) 1219 (19.1) 1857 (29.1)
Co-prescribing Benzodiazepines Antipsychotics Antidepressants Opioid analgesics
4586 (71.7) 1487 (23.3) 3119 (48.8) 2578 (40.3)
Comorbidity scored 0–5 6–10 11–15 >15
4972 (77.8) 1183 (18.5) 200 (3.1) 38 (0.6)
a
Unless otherwise specified. Based on patients initial treatment episode in the follow-up period. c The number and percentage of methadone scripts dispensed as supervised consumption, based on patients initial treatment episode in the follow-up period. d Measured as the number of prescriptions (unique level 3 ATC codes) received by patient, excluding methadone, benzodiazepines, antipsychotics, antidepressants and opioids. b
presents the results from the proportional hazards gamma frailty model. The results show a j-shaped relationship for the primary exposure, supervised methadone consumption, such that having between 20% and 59% of their methadone scripts dispensed under supervised consumption (compared to <20%) was associated with increased retention in treatment (longer time to discontinuation) (20–39% of scripts: HR = 0.88, 95% CI 0.81–0.95; 40–59% of scripts: HR = 0.87, 95% CI 0.81–0.94). However, having 60% or more of their methadone scripts dispensed under supervised consumption (compared to <20%) was associated with shorter treatment episodes (faster time to discontinuation) (60–79% of scripts: HR = 1.28, 95% CI 1.20–1.36; >=80% of scripts: HR = 3.59, 95% CI 3.38–3.81). Later attempts (3 or more) at MMT, compared to the first treatment episode, were also associated with longer times to discontinuation (episode 3: HR 0.93, 95% CI: 0.87–0.98; episode 4: HR 0.92, 95% CI: 0.86–0.98; episode 5 or more: HR 0.82, 95% CI: 0.78–0.88), as was a daily dose between 60–120 mgs per day (compared to <60 mgs or >120 mgs per day). Finally, patients with a comorbidity score of 11 or more (compared to 0–5) were significantly more likely to have shorter treatment episodes.
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131 Table 2 Proportional hazards gamma frailty model of time to discontinuation of treatment. Adjusted Hazard Ratio (95% Confidence Limits)
P-value
Sex Females Males
1.00 1.05 (0.98–1.11)
0.203
Age (years) 16–19 20–29 30–39 40–65
1.00 1.09 (0.97–1.21) 1.07 (0.96–1.20) 1.25 (1.10–1.43)
0.279 0.378 0.007
Treatment episode First Second Third Fourth Fifth or above
1.00 1.00 (0.95–1.05) 0.93 (0.87–0.98) 0.92 (0.86–0.98) 0.82 (0.78–0.88)
0.869 0.012 0.019 <0.001
Median methadone dose 2.11 (2.02–2.21) <60 mg/day 1.00 60–120 mg/day ≥120 mg/day 1.31 (1.13–1.52)
<0.001 <0.001
Supervised consumption 1.00 <20% 0.88 (0.81–0.95) 20–39% 0.87 (0.81–0.94) 40–59% 60–79% 1.28 (1.20–1.36) ≥80% 3.59 (3.38–3.81)
<0.001 <0.001 <0.001 <0.001
Median comorbidity scorea 0–5 1.00 1.06 (0.98–1.14) 6–10 1.20 (1.03–1.41) 11–15 1.48 (1.06–2.06) >15
0.128 0.021 0.020
Antipsychotics No Yes
1.00 1.00 (0.93–1.07)
0.975
Benzodiazepines No Yes
1.00 0.93 (0.87–1.00)
0.065
Opioid analgesics No Yes
1.00 1.03 (0.97–1.09)
0.406
Antidepressants No Yes
1.00 1.03 (0.97–1.10)
0.301
The bold is highlighting those variables that are statistically significant i.e. <0.05. a As measured by number of prescriptions (unique level 3 ATC codes) received by patient, excluding methadone, benzodiazepines, antipsychotics, antidepressants and opioids received by patient in calendar year.
129
was observed, such that having between 20 and 60% of methadone prescriptions dispensed under supervised consumption (relative to <20%) was associated with increased retention in treatment. The protective effects of retaining patients were lost when the frequency of supervision increased above 60%, resulting in a reduction in retention in treatment. These findings are consistent with previous RCTs which suggest that increased supervision is associated with a reduction in retention in MMT (Rhoades et al., 1998; Holland et al., 2012). Other studies reported no difference in retention between those supervised or unsupervised (Bell et al., 2007; Gerra et al., 2011). A large UK RCT, found no between group difference in retention by intention-to-treat analysis, however per protocol analysis suggested a reduction in retention for those exposed to supervised consumption (Holland et al., 2014). The J-shaped relationship observed in our study reflects findings from a recent qualitative study which suggest that patients consider supervised consumption as acceptable in the short-term, helping to establish a regular daily routine and to develop positive relationships with pharmacy staff, but that in the long-term being unsupervised was the preference (Notley et al., 2014). Consistent with previous studies we also found that patients experienced multiple treatment episodes over the 6-year period (Bell et al., 2006; Burns et al., 2009; Degenhardt et al., 2009). However, similar to a large cohort in Canada, patients with multiple treatment episodes remained in MMT for progressively longer periods in later treatment episodes. (Nosyk et al., 2009). Methadone dosages ranging from 60 to 120 mg/day were also shown to be more effective than lower doses in retaining patients, this is consistent with a number of previous studies (Faggiano et al., 2003; Nosyk et al., 2009; Cao et al., 2014; Vigna-Taglianti et al., 2016). Despite these findings and the wide consensus on 60–120 mg as the optimal daily dose range, over one in three patients did not receive the minimum optimal dose of 60 mg/day. This finding is comparable with previous international studies (Strang et al., 2010; Cao et al., 2014; Nosyk et al., 2009). However, the proportion of patients receiving methadone below the minimum recommended dose of 60 mg/day in Ireland appears to be lower than in the UK where 56.8% of patients in MMT in primary care were prescribed methadone below 60 mg/day (Strang et al., 2010). Similarly, a Canadian cohort study of 17,005 patients in MMT between 1996 and 2007 found that 50.5% of patients had a mean daily dose below 60 mg (Nosyk et al., 2009). Prescribing practices should be improved to ensure that patients are receiving the minimum optimal dose of between 60 and 120 mg/day. 4.2. Strengths and weaknesses
3.4. Sensitivity analyses and assumption checks We extended the treatment discontinuation rule from 7 days to 28 days to assess potential misclassification bias. This analysis showed that the results were similar and the coefficient estimates were robust to changes in the treatment discontinuation rule (see Supporting information, Table S1). Similarly, excluding patients who died during the study follow-up did not influence the inferences made (see Supporting information, Table S2). 4. Discussion 4.1. Principal findings in relation to other studies To our knowledge, this is the first study to assess the relationship between supervised methadone consumption and retention in treatment across multiple treatment episodes in primary care. We found a J-shaped relationship between supervised consumption and time to discontinuation of treatment. A J-shaped relationship
The external validity of this study is high, as it includes all patients on a national prescribing register over a six-year study period. Furthermore, using the proportional hazards gamma frailty model to account for recurrent treatment episodes, we were able to use all patient data available rather than restricting the analysis to a single treatment episode. This method of analysis also allows for a more accurate inference regarding the effects of successive treatment episodes as it accounts for the correlation between repeated treatment episodes and unobserved heterogeneity across individual patients (Nosyk et al., 2009). Previous studies have generally reported on a short duration of follow-up, ranging between 12 and 24 weeks, which fail to account for the patients’ movement in and out of treatment overtime (Rhoades et al., 1998; Bell et al., 2007; Holland et al., 2012, 2014). Our study has a number of shortcomings and potential biases. Our results relate to patient’s retention in primary-care methadone treatment programmes, patients transferring from primary care to specialist care settings would be lost in our cohort, and incorrectly classified as discontinuing treatment. Secondly, classifying
130
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131
patients as ceasing treatment after 7 days without a methadone prescription may have introduced a misclassification bias, resulting in an underestimate of retention in treatment. However, retention rates remained consistent in our sensitivity analysis extending the gap between the end of prescription coverage and new prescription to 28 days before classifying a patient as “off-treatment”. Thirdly, we do not have information on the quality or intensity of the care received, which may influence patients’ retention in treatment (Cornish et al., 2010). Finally, in observational studies of this sort, the possibility of residual confounding or confounding by indication may remain a problem, therefore associations identified in this study should be viewed principally as hypothesis generating and our observed associations should be subject to testing and verification in other national observational cohorts. 4.3. Clinical implications The J-shaped relationship observed between supervised consumption and retention in MMT in primary care, represents a double edge sword; too little supervision (<20% of methadone prescriptions) is associated with treatment cessation, this risk of treatment cessation decreases with regular supervision (20% to 60% of methadone scripts), but eventually increases above the reference group (those with little supervision) when a patient is exposed to higher levels of supervised consumption (>60% of methadone prescriptions). Losing patients from treatment may expose patients to an increased risk of mortality, as previous studies have identified an elevated risk of mortality following treatment cessation (Davoli et al., 2007; Cornish et al., 2010; Cousins et al., 2016). However, reducing supervision with the aim of retaining a patient in treatment and reducing their individual risk of mortality, needs to be balanced against the risk of diversion and increased availability of illicit methadone which may increase population level risk of methadone related deaths (Strang et al., 2010). Our study cannot be used to indicate who should be supervised or the optimal frequency of supervision to ensure that patients are retained in treatment while simultaneously avoiding diversion of methadone. However, our study is the first study to examine the longitudinal association between supervised consumption over the course of multiple treatment episodes and retention in treatment. While a number of authors have suggested flexibly timed discontinuation of supervision once a patient is stabilised (Holland et al., 2014; Notley et al., 2014), recent Australian studies have highlighted the challenges involved in assessing patient stability with many OST prescribers reporting high levels of uncertainty in identifying whether their patients are diverting or injecting their medication (Larance et al., 2011, 2014). Further work is needed to determine how to optimally stratify patients’ suitability for unsupervised dosing, with the aim of retaining patients in treatment and simultaneously reducing diversion and injecting. 4.4. Conclusion This national community based cohort study of over 6000 patients identified a J-shaped relationship between supervised methadone consumption and retention in treatment over a six-year follow-up. A median daily dose between 60 and 120 mg/per day, and multiple treatment episodes were also significant predictors of longer time to discontinuation of treatment. Role of funding source This work was funded by the Health Research Board (HRB) of Ireland through the HRB Centre for Primary Care Research under Grant HRC/2007/1. The funders had no input into the study design; the collection, analysis, and interpretation of data; the writing of
the report; or the decision to submit the article for publication. The investigators are fully independent of the funders and have not used a third party to write or prepare the manuscript. Conflict of interest No conflict declared. Contributors GC designed the study, secured the data, interpreted the results and drafted the manuscript. FB managed the data, including data linkage, performed the analyses and drafted the results. JB secured data, revised the manuscript and provided critical comments. SL secured data, revised the manuscript and provided critical comments.EK revised the manuscript and provided critical comments. DOD revised the manuscript and provided critical comments. KB revised the manuscript and provided critical comments. TF designed the study, revised the manuscript and provided critical comments. GC and FB had full access to all of the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. All authors approved of the final manuscript. Acknowledgments This work was funded by the Health Research Board (HRB) of Ireland through the HRB Centre for Primary Care Research under Grant HRC/2007/1. The authors would like to thank the HSE Primary Care Reimbursement Services for supplying the prescription dispensing data, and the Central Statistics Office for acting as an independent data processor to facilitate the data linkage process. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.drugalcdep.2016. 12.009. References Bell, J., Burrell, T., Indig, D., Gilmour, S., 2006. Cycling in and out of treatment; participation in methadone treatment in NSW, 1990–2002. Drug Alcohol Depend. 81, 55–61. Bell, J., Shanahan, M., Mutch, C., Rea, F., Ryan, A., Batey, R., Dunlop, A., Winstock, A., 2007. A randomized trial of effectiveness and cost-effectiveness of observed versus unobserved administration of buprenorphine-naloxone for heroin dependence. Addiction 102, 1899–1907. Brilleman, S.L., Salisbury, C., 2013. Comparing measures of multimorbidity to predict outcomes in primary care: a cross sectional study. Fam. Pract. 30, 172–178. Burns, L., Randall, D., Hall, W.D., Law, M., Butler, T., Bell, J., Degenhardt, L., 2009. Opioid agonist pharmacotherapy in New South Wales from 1985 to 2006: patient characteristics and patterns and predictors of treatment retention. Addiction 104, 1363–1372. Cao, X., Wu, Z., Rou, K., Li, L., Lin, C., Wang, C., Luo, W., Pang, L., Yin, W., Li, J., 2014. Retention and its predictors among methadone maintenance treatment clients in China: a six-year cohort study. Drug Alcohol Depend. 145, 87–93. Cornish, R., Macleod, J., Strang, J., Vickerman, P., Hickman, M., 2010. Risk of death during and after opiate substitution treatment in primary care: prospective observational study in UK general practice research database. BMJ 341, c5475. Cousins, G., Teljeur, C., Motterlini, N., McCowan, C., Dimitrov, B.D., Fahey, T., 2011. Risk of drug-related mortality during periods of transition in methadone maintenance treatment: a cohort study. J. Subst. Abuse Treat. 41, 252–260. Cousins, G., Boland, F., Courtney, B., Barry, J., Lyons, S., Fahey, T., 2016. Risk of mortality on and off methadone substitution treatment in primary care: a national cohort study. Addiction 111, 73–82. Davoli, M., Bargagli, A.M., Perucci, C.A., Schifano, P., Belleudi, V., Hickman, M., Salamina, G., Diecidue, R., Vigna-Taglianti, F., Faggiano, F., 2007. Risk of fatal overdose during and after specialist drug treatment: the VEdeTTE study, a national multi-site prospective cohort study. Addiction 102, 1954–1959. Deering, D.E., Sheridan, J., Sellman, J.D., Adamson, S.J., Pooley, S., Robertson, R., Henderson, C., 2011. Consumer and treatment provider perspectives on
G. Cousins et al. / Drug and Alcohol Dependence 173 (2017) 126–131 reducing barriers to opioid substitution treatment and improving treatment attractiveness. Addict. Behav. 36, 636–642. Degenhardt, L., Randall, D., Hall, W., Law, M., Butler, T., Burns, L., 2009. Mortality among clients of a state-wide opioid pharmacotherapy program over 20 years: risk factors and lives saved. Drug Alcohol Depend. 105, 9–15. Department of Health (England) and the devolved administrations, 2007. Drug Misuse and Dependence: UK Guidelines on Clinical Management. Department of Health (England), the Scottish Government, Welsh Assembly Government and Northern Ireland Executive, London. Faggiano, F., Vigna-Taglianti, F., Versino, E., Lemma, P., 2003. Methadone maintenance at different dosages for opioid dependence. Cochrane Database Syst. Rev. 3, Cd002208. Gerra, G., Saenz, E., Busse, A., Maremmani, I., Ciccocioppo, R., Zaimovic, A., Gerra, M.L., Amore, M., Manfredini, M., Donnini, C., Somaini, L., 2011. Supervised daily consumption, contingent take-home incentive and non-contingent take-home in methadone maintenance. Prog. Neuropsychopharmacol. Biol. Psychiatry 35, 483–489. Green, H., James, R.A., Gilbert, J.D., Harpas, P., Byard, R.W., 2000. Methadone maintenance programs—a two-edged sword? Am. J. Forensic Med. Pathol. 21, 359–361. Holland, R., Matheson, C., Anthony, G., Roberts, K., Priyardarshi, S., Macrae, A., Whitelaw, E., Appavoo, S., Bond, C., 2012. A pilot randomised controlled trial of brief versus twice weekly versus standard supervised consumption in patients on opiate maintenance treatment. Drug Alcohol Rev. 31, 483–491. Holland, R., Maskrey, V., Swift, L., Notley, C., Robinson, A., Nagar, J., Gale, T., Kouimtsidis, C., 2014. Treatment retention, drug use and social functioning outcomes in those receiving 3 months versus 1 month of supervised opioid maintenance treatment. Results from the Super C randomized controlled trial. Addiction 109, 596–607. Kimber, J., Copeland, L., Hickman, M., Macleod, J., McKenzie, J., De Angelis, D., Robertson, J.R., 2010. Survival and cessation in injecting drug users: prospective observational study of outcomes and effect of opiate substitution treatment. BMJ 341, c3172. Larance, B., Degenhardt, L., O’Brien, S., Lintzeris, N., Winstock, A., Mattick, R.P., Bell, J., Ali, R., 2011. Prescribers’ perceptions of the diversion and injection of medication by opioid substitution treatment patients. Drug Alcohol Rev. 30, 613–620. Larance, B., Carragher, N., Mattick, R.P., Lintzeris, N., Ali, R., Degenhardt, L., 2014. A latent class analysis of self-reported clinical indicators of psychosocial stability and adherence among opioid substitution therapy patients: do stable patients receive more unsupervised doses? Drug Alcohol Depend. 142, 46–55.
131
Mathers, B.M., Degenhardt, L., Bucello, C., Lemon, J., Wiessing, L., Hickman, M., 2013. Mortality among people who inject drugs: a systematic review and meta-analysis. Bull. World Health Organ. 91, 102–123. McLellan, A.T., Arndt, I.O., Metzger, D.S., Woody, G.E., O’Brien, C.P., 1993. The effects of psychosocial services in substance abuse treatment. JAMA 269, 1953–1959. Nosyk, B., MacNab, Y.C., Sun, H., Fischer, B., Marsh, D.C., Schechter, M.T., Anis, A.H., 2009. Proportional hazards frailty models for recurrent methadone maintenance treatment. Am. J. Epidemiol. 170, 783–792. Notley, C., Holland, R., Maskrey, V., Nagar, J., Kouimtsidis, C., 2014. Regaining control: the patient experience of supervised compared with unsupervised consumption in opiate substitution treatment. Drug Alcohol Rev. 33, 64–70. Rhoades, H.M., Creson, D., Elk, R., Schmitz, J., Grabowski, J., 1998. Retention, HIV risk, and illicit drug use during treatment: methadone dose and visit frequency. Am. J. Publ. Health 88, 34–39. Seymour, A., Black, M., Jay, J., Cooper, G., Weir, C., Oliver, J., 2003. The role of methadone in drug-related deaths in the west of Scotland. Addiction 98, 995–1002. Strang, J., Hall, W., Hickman, M., Bird, S.M., 2010. Impact of supervision of methadone consumption on deaths related to methadone overdose (1993–2008): analyses using OD4 index in England and Scotland. BMJ 341, c4851. Termorshuizen, F., Krol, A., Prins, M., Geskus, R., van den Brink, W., van Ameijden, E.J., 2005. Prediction of relapse to frequent heroin use and the role of methadone prescription: an analysis of the Amsterdam cohort study among drug users. Drug Alcohol Depend. 79, 231–240. Vigna-Taglianti, F.D., Burroni, P., Mathis, F., Versino, E., Beccaria, F., Rotelli, M., Garneri, M., Picciolini, A., Bargagli, A.M., 2016. Gender differences in heroin addiction and treatment: results from the VEdeTTE cohort. Subst. Use Misuse 51, 295–309. World Health Organization (WHO), 2009. Guidelines for the Psychosocially Assisted Pharmacological Treatment of Opioid Dependence. WHO, Geneva. Weinrich, M., Stuart, M., 2000. Provision of methadone treatment in primary care medical practices: review of the Scottish experience and implications for US policy. JAMA 283, 1343–1348. Zhang, L., Zou, X., Zhang, D., Li, X., Zhao, P., Ling, L., 2015. Investigation of repeat client drop-out and re-enrolment cycles in fourteen methadone maintenance treatment clinics in Guangdong, China. PLoS One 10, e0139942.