Bidirectional relationships between retention and health-related quality of life in Chinese mainland patients receiving methadone maintenance treatment

This study aimed to explore the bidirectional relationships between retention and health-related quality of life (HRQoL) in patients from mainland China receiving methadone maintenance treatment (MMT). This prospective cohort study recruited 1,212 eligible MMT patients from the two largest MMT clinics (one privately and another publicly funded) in Xi’an. This study started in March 2012 with a 2-year follow-up until March 2014. Retention was assessed by repeated terminations, past treatment duration, premature terminations, and follow-up treatment duration. HRQoL was evaluated using the Chinese (simple) short-form 36 health survey version 2 (SF-36v2) and the quality of life scale for drug addicts (QOL-DAv2.0). Linear and Cox regression analyses were used to explore relationships between retention and HRQoL. A general linear model was used to further examine the global effect of past treatment duration on HRQoL. Multivariate analyses showed that repeated terminations had no significant impact on HRQoL scores in MMT patients; however, past treatment time (year) influenced the SF-36v2PCS (P = 0.004): treatment for ≥4 years showed a lower SF-36v2PCS score (regression coefficient: -2.39; 95% confidence interval [CI]: -3.80, -0.97; P = 0.001) than treatment for <1 year. In addition, patients with an SF-36v2PCS score > 49 (hazard ratio: 0.83; 95% CI: 0.69, 0.98; P = 0.03) were 17% less likely to terminate MMT than those with scores of ≤49. In conclusion, retention and HRQoL tended to have a bidirectional relationship, which should be considered in the development of retention and health-management programs for patients with MMT.


Introduction
Methadone maintenance treatment (MMT) is a harm-reduction program for people with opioid dependence; it involves the use of methadone as a synthetic agent to block the brain receptors affected by heroin and other opiates [1]. In China, MMT was initiated as a pilot program in 8 clinics assisting 1,029 drug users in 2004 and subsequently expanded to 761 clinics assisting a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 407,000 drug users by the end of 2013, aiding approximately 16% of domestic registered drug users [2,3]. A large body of evidence indicates that MMT can reduce drug addiction, drugrelated harms, risk behaviors, crimes, and transmission of human immunodeficiency virus (HIV) and hepatitis C virus (HCV) [4,5]. Drug users receiving MMT show markedly improved personal relationships, enhanced social productivities, and better health-related quality of life (HRQoL) [6][7][8][9]. Given that MMT involves drug substitution and long-term administration, retention is the main factor that directly influences the treatment outcomes and a major indicator for the therapeutic effects of the program [10,11].
A long retention period is related to improved treatment outcomes, and the retention duration is considered the best predictor for the effects of MMT [11][12][13]. However, retention is suboptimal in most MMT programs, with retention rates ranging from 58% to 78% in China [14] and 3% to 94% worldwide [11]. Additionally, repeat dropout is common during the MMT process, which probably has adverse effects on treatment outcomes, especially patientreported outcomes such as HRQoL. Many previous studies have reported that MMT positively affects HRQoL [9,15]; however, few studies have explored the influence of retention on HRQoL, particularly in patients with repeat terminations [16].
HRQoL is broadly conceptualized as individuals' perception of their position in life in terms of their physical health, psychological health, social relationships, relationship with their environment, independence level, and personal beliefs [17]. It reflects many aspects of individuals' daily lives and has been considered an important outcome measure in patients receiving MMT. Many published studies have demonstrated that MMT patients with good HRQoL have characteristics of strong social support, relatively independent income, few incidents of drug overdose and related complications (e.g., HIV or HCV infection, psychiatric comorbidity, and cutaneous abscess), and few incarcerations [18][19][20]. Regarding social support, we have previously explored the relationship between perceived social support and retention with the same patient population used in this study and found that good perceived social support is a strong predictor of retention [21]. However, this finding is insufficient to confirm a positive correlation between retention and HRQoL. From the viewpoint of MMT-participation behaviors, it is unclear whether patients with better HRQoL have longer retention periods [22].
To further our previous study [21], the present study aimed to identify bidirectional relationships between retention and HRQoL in patients from mainland China receiving MMT. We proposed 3 hypotheses: (a) patients with repeat terminations have poorer HRQoL than those with no terminations, (b) longer past treatment duration predicts better HRQoL, and (c) better HRQoL predicts less premature terminations.

Study design
This is a prospective cohort study. A 2-year follow-up was conducted on the basis of a crosssectional survey.

Participants and data collection
Patients admitted to the two largest MMT clinics (one funded privately and another funded publicly) in Xi'an, China, were recruited in the study. Inclusion criteria were age ! 18 years and fluency in Chinese. Patients with cognitive disorders or those who refused to provide written informed consent were excluded.
The study was started in March 2012, with a 2-year follow-up until March 2014. Data were collected in March 2012, including baseline information (i.e., pretreatment socio-demographics and drug use characteristics), retention, and HRQoL. The recruited MMT patients were variables are presented with confidence intervals. To control the confounding effects of baseline information, all variables of baseline information were included in the multivariate analysis. The general linear model was used to further identify the global effect of past treatment duration on HRQoL. Cox regression analysis was conducted to explore the influences of HRQoL on follow-up retention, and the values are presented with confidence intervals. In the Cox regression model, the event of interest was premature termination with time measured as days of MMT since April 1, 2012 until premature termination or the end of the study, whichever came first; individuals who remained on MMT through March 31, 2012, were considered censored. The baseline information and past retention were included as the controlling variables. All statistical analyses were performed using SPSS 22.0 (IBM Corp., Armonk, NY). A value of P < 0.05 (two-tailed) was considered statistically significant.

Ethical statement
The study protocol was reviewed and approved by the Human Research Ethics Committee of Xi'an Jiaotong University. Written informed consent was obtained from each recruited patient before the questionnaire survey.

Results
Of the 1,270 eligible patients, 58 (4.6%) patients (30 in the publicly funded clinic and 28 in the privately funded clinic) were excluded because they refused to provide written informed consent. Finally, a total of 1,212 patients completed the cross-sectional questionnaire survey at baseline, including 361 patients (29.8%) from the publicly funded clinic and 851 patients (70.2%) from the privately funded clinic. All patients were self-funded for methadone treatment, except for those with an HIV-positive serostatus. In the face-to-face interview, patients were informed about and understood the questions, and completed the questionnaire. Each interview lasted for approximately 20-25 min.
Significant differences in the baseline information between prematurely terminated patients and methadone-maintained patients were observed for age at initial treatment (t = Table 1. Description of sample at baseline and follow-up (n, %).

Baseline information Baseline (N = 1212)
Two-year follow-up ƚ Statistics P

HRQoL
In the SF-36v2, both the component summary and 8 scale mean scores were <50 points [24].

Relationships between past retention and HRQoL
Without considering the baseline information, patients with no terminations had higher SF-36v2PCS scores than those with repeat terminations (regression coefficient [B]: 1.33; 95%   (Table 2).  (Table 3).

Discussion
Although MMT has been considered an effective opiate-replacement therapy, retention in MMT is still suboptimal in most MMT programs [10,16,29]. Of the 1,212 patients, 37.8% had repeat terminations since initial treatment, indicating the high prevalence of repeated terminations in the MMT patient population; this is consistent with the findings of a related previous study [16]. After the 2-year follow-up, 43.5% patients terminated MMT: 27.1% patients quit MMT within the first 12 months, consistent with the findings of previous systematic reviews [11,14]. Thus, repeat terminations and early treatment discontinuations should be considered in MMT retention management and intervention programs. This study showed that age at admission, age at first drug use, and initial urinalysis drug screen are predictors of premature discharge in MMT; consistent with this finding, previous studies have shown that young age at treatment, young age at initial drug use, and positive morphine urine test were related to high risks of early termination [30][31][32][33]. Therefore, age at admission, age at initial drug use, and pretreatment morphine urine test results should be considered when developing retention interventions, especially for an individualized management model for newly admitted MMT patients.
All the SF-36v2 mean scores were lower than the normal score of 50 points, demonstrating that MMT patients have a poorer health status than the general population [24]. The summary component scores and 8 scale scores demonstrated poorer overall mental health than physical health and specific impairment in general health, which is in line with findings of previous studies [13,34]. Relatively lower scores of QOL-DAv2.0 in the domains of society, physiology, and psychology indicate that damage in these health domains is common in MMT patients, which explains why MMT patients have poor HRQoL, especially impairment in physical and social function [27,35]. Thus, more efforts should be taken to improve the physical, psychological, and social health of MMT patients during treatment to achieve other objective therapeutic effects such as reduced rates of positive morphine urine tests and secret drug use while receiving MMT.
Univariate analysis showed that patients with no treatment terminations had higher SF-36v2PCS scores than those with repeat terminations, indicating that consecutive MMT improved physical HRQoL [15,19]. This finding supports our first hypothesis: Patients with repeat terminations had poorer HRQoL than those with no terminations. However, this is not the same as the findings of the multivariate analysis, i.e., repeated terminations had no significant impact on HRQoL scores in patients with MMT. This difference is probably related to the control of baseline information, especially the age at initial treatment (years), age at initial drug use (years), and initial morphine urine test, which further support the hypothesis that these variables had an influence on early discharge in MMT and should be carefully considered in MMT clinical practice.
In addition, patients with longer past treatment duration had lower SF-36v2PCS and QOL-DAv2.0 scores, especially those with past treatment duration for !4 years. However, past treatment time (year) had no significant influence on the QOL-DAv2.0 score in the multivariate analysis, indicating that the quality of life specific to drug addiction is not influenced by past treatment duration while controlling the baseline information. Nonetheless, our study showed that the HRQoL decreased with an increase in the treatment duration; this result is inconsistent with that of other reports [9,13,36] and contradicts our second hypothesis that longer past treatment time predicts better HRQoL. The observed decrease in the SF-36v2PCS and QOL-DAv2.0 scores is probably due to the following reasons: First, long-term methadone administration can lead to several potential side effects (e.g., weak or shallow breathing, severe constipation, dizziness, nausea, vomiting, increasing sweating, and sedation) [37], which might have a negative effect on the HRQoL. Second, MMT is effective in reducing heroin craving. However, some MMT patients are still at risk of craving heroin; these patients may relapse to heroin abuse, which may adversely affect their HRQoL [38,39]. Third, patients with syringe sharing are at high risk of HIV, HCV, or hepatitis B virus (HBV) infection that can lead to acquired immunodeficiency virus, chronic hepatitis C, or chronic hepatitis B, respectively, which would also deteriorate the HRQoL [34,40,41]. Therefore, the side effects of methadone should be carefully considered and addressed during long-term MMT. In addition, a morphine urine test should be performed regularly to supervise secret heroin use during MMT; other adjuvant medications for controlling or decreasing heroin craving are recommended on the basis of routine methadone prescription. Additionally, patients with positive HIV, HCV, or HBV serostatus should be treated accordingly, with the addition of methadone to their regimen, and syringe sharing should be strictly prohibited to avoid new HIV, HCV, or HBV infection.
Regarding the 2-year follow-up, patients with SF-36v2PCS scores > 49 (in both univariate and multivariate analyses) or SF-36v2MCS scores > 42 (in univariate analysis) had a lower risk of premature terminations than those with scores of 49 or 42, respectively, supporting our third hypothesis that better HRQoL predicts fewer treatment discontinuations; this result is consistent with the findings of a few previous studies [42,43], but contradicts the finding of another similar study that explored the HRQoL (measured by World Health Organization Quality of Life Assessment, Brief Version) on the basis of attendance rate among 105 selffunded heroin users and found that good social HRQoL was a predictor for poor methadone attendance [22]. This difference in our findings may result from the different characteristics of patients, study design, or analytical methods. Therefore, further study is required to predict the role of HRQoL in MMT retention.
Despite our important findings, this study had a few limitations. First, the baseline information did not include all variables; for example, subjective variables regarding psychological status were not included. Second, only a limited number of covariates were controlled for in the multivariate analysis; other unobserved factors such as religious beliefs and living conditions were not taken into account. Therefore, the findings may still be confounded by certain factors. Third, the results from this study reveal possible relationships between retention and HRQoL; however, they should not be interpreted as causal relationships, especially due to the relatively short follow-up period.
In conclusion, this study showed that patients with no terminations had better physical HRQoL than those with repeat terminations, and longer past treatment duration had a negative influence on physical and drug addiction-related HRQoL. Furthermore, better physical and psychological HRQoL was associated with a lower risk of premature terminations in MMT. This observed bidirectional relationship between retention and HRQoL should be considered when developing retention and health-management programs for MMT patients.