Hostname: page-component-78c5997874-m6dg7 Total loading time: 0 Render date: 2024-10-31T11:53:34.422Z Has data issue: false hasContentIssue false

Temporal pathways of change in two randomized controlled trials for depression and harmful drinking in Goa, India

Published online by Cambridge University Press:  08 January 2019

Daisy R. Singla*
Affiliation:
Department of Psychiatry, University of Toronto and Sinai Health System, Toronto, Canada
Steven D. Hollon
Affiliation:
Department of Psychology, Vanderbilt University, Nashville, Tennessee, USA
Richard Velleman
Affiliation:
Department of Psychology, University of Bath, Bath, UK Sangath, Alto Porvorim, Goa, India
Benedict Weobong
Affiliation:
Sangath, Alto Porvorim, Goa, India Centre for Global Mental Health, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK Department of Social & Behavioral Sciences, School of Public Health, University of Ghana, Accra, Ghana
Abhijit Nadkarni
Affiliation:
Sangath, Alto Porvorim, Goa, India Centre for Global Mental Health, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK
Christopher G. Fairburn
Affiliation:
Department of Psychiatry, Warneford Hospital, Oxford, UK
Bhargav Bhat
Affiliation:
Sangath, Alto Porvorim, Goa, India
Mahesh Gurav
Affiliation:
Sangath, Alto Porvorim, Goa, India
Arpita Anand
Affiliation:
Sangath, Alto Porvorim, Goa, India
Jim McCambridge
Affiliation:
Department of Health Sciences, University of York, York, UK
Sona Dimidjian
Affiliation:
Department of Psychology and Neuroscience, University of Colorado Boulder, Boulder, Colorado, USA
Vikram Patel
Affiliation:
Sangath, Alto Porvorim, Goa, India Department of Global Health & Social Medicine, Harvard Medical School, Cambridge, USA Department of Global Health and Population, Harvard TH Chan School of Public Health, Boston, MA, USA
*
Author for correspondence: Daisy R. Singla, E-mail: daisy.singla@utoronto.ca
Rights & Permissions [Opens in a new window]

Abstract

Background

The current study explored the temporal pathways of change within two treatments, the Healthy Activity Program (HAP) for depression and the Counselling for Alcohol Problems (CAP) Program for harmful drinking.

Methods

The study took place in the context of two parallel randomized controlled trials in Goa, India. N = 50 random participants who met a priori criteria were selected from each treatment trial and examined for potential direct and mediational pathways. In HAP, we examined the predictive roles of therapy quality and patient-reported activation, assessing whether activation mediated the effects of therapy quality on depression (Patient Health Questionnaire-9) outcomes. In CAP, we examined the predictive roles of therapy quality and patient change- and counter-change-talk, assessing whether change- or counter-change-talk mediated the effects of therapy quality on daily alcohol consumption.

Results

In HAP, therapy quality (both general and treatment-specific skills) was associated with patient activation; patient activation but not therapy quality significantly predicted depression outcomes, and patient activation mediated the effects of higher general skills on subsequent clinical outcomes [a × b = −2.555, 95% confidence interval (CI) −5.811 to −0.142]. In CAP, higher treatment-specific skills, but not general skills, were directly associated with drinking outcomes, and reduced levels of counter-change talk both independently predicted, and mediated the effects of higher general skills on, reduced alcohol consumption (a × b = −24.515, 95% CI −41.190 to −11.060). Change talk did not predict alcohol consumption and was not correlated with counter-change talk.

Conclusion

These findings suggest that therapy quality in early sessions operated through increased patient activation and reduced counter-change talk to reduce depression and harmful drinking respectively.

Type
Original Articles
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © Cambridge University Press 2019

Introduction

Evidence-based psychological treatments (PTs) are recommended as the first line of treatment for depression and alcohol use disorders (AUD) – two leading causes of global burden of mental disorders for women and men respectively (World Health Organization, Reference Weobong, Weiss, McDaid, Singla, Hollon, Nadarni, Park, Bhat, Katti, Anand, Dimidjian, Araya, King, Vijayakumar, Wilson, Velleman, Fairburn and Patel2015). Worldwide, brief PTs delivered by non-specialist providers (NSPs) have been successful at treating adult depression and AUDs (Van Ginneken et al., Reference Van Ginneken, Tharyan, Lewin, Rao, Meera, Pian, Chandrashekar and Patel2013; Hoeft et al., Reference Hoeft, Fortney, Patel and Unutzer2016; Singla et al., Reference Singla, Kohrt, Murray, Anand, Chorpita and Patel2017). Despite their effectiveness, it remains unclear how these treatments achieve their effects. Identifying temporal pathways of how key treatment and patient variables influence clinical outcomes may illuminate how PTs operate, guiding clinicians, and researchers to predict individual patient trajectories and develop more effective interventions (Cook et al., Reference Cook, Heather and McCambridge2015; Kraemer, Reference Kraemer2016).

The current study explored specific treatment- and patient-level variables that may influence the effectiveness of lay counselor delivered treatments for depression and harmful drinking. Each treatment was the subject of a randomized controlled trial (RCT), delivered in primary care settings in Goa, India. The Healthy Activity Program (HAP) – a brief, culturally adapted version of behavioral activation (BA) treatment for depression (Chowdhary et al., Reference Chowdhary, Anand, Dimidjian, Shinde, Weobong, Balaji, Hollon, Rahman, Wilson, Verdeli, Araya, King, Jordans, Fairburn, Kirkwood and Patel2016) – was found to be more effective than enhanced usual care (EUC) in reducing depressive symptoms and improving remission rates at the primary end-point of 3-months (Patel et al., Reference Patel, Weobong, Weiss, Anand, Bhat, Katti, Dimidjian, Araya, Hollon, King Mm Vijaykumar, Park, McDaid, Wilson, Velleman, Kirkwood and Fairburn2017). Similarly, the Counselling for Alcohol Problems (CAP) Program – a brief, motivational interviewing (MI)-informed PT for harmful drinkers (Nadkarni et al., Reference Nadkarni, Velleman, Dabholkar, Shinde, Bhat, McCambridge, Murthy, Wilson, Weobong and Patel2015) – was found to be more effective than EUC in increasing abstinence and remission rates at 3-months post-enrollment (Nadkarni et al., Reference Nadkarni, Weiss, Weobong, McDaid, Singla, Park, Bhat, Katti, McCambridge, Murthy, Wilson, Kirkwood, Fairburn, Velleman and Patel2017a, Reference Nadkarni, Weobong, Weiss, McCambridge, Bhat, Katti, Murthy, King, McDaid, Park, Wilson, Kirkwood, Fairburn, Velleman and Patelb).

As per Magill et al. (Reference Magill, Gaume, Apodaca, Walthers, Mastroleo, Borsari and Longabaugh2014, Reference Magill, Kiluk, McCrady, Tonigan and Longabaugh2015), we defined treatment variables as treatment processes or counselor skills that predict patient behaviors or long-term effects. Patient behaviors were those enacted by the patient that were potentially associated with the outcomes of interest (i.e. average daily drinking scores and depressive symptoms at 3-months, post enrollment). Fig. 1 shows the study's hypothesized pathways of change in each trial, whereby both treatment-specific and general skills in early sessions would influence the subsequent patient behaviors (patient activation in HAP or change and counterchange talk in CAP) which would, in turn, influence the clinical outcome post-treatment. We also hypothesized that patient behaviors would mediate the effects of therapy quality on clinical outcomes.

Fig. 1. Hypothesized temporal pathways of change.

Treatment variable

In both HAP and CAP, our treatment variable was therapy quality – the degree to which the counselors have ‘done the right things well’ (Fairburn and Cooper, Reference Fairburn and Cooper2011). Therapy quality included both general skills (e.g. empathy, collaboration, and expressing warmth) and treatment-specific skills (e.g. activity monitoring, scheduling, and homework completion in HAP and avoiding confrontation, providing affirmations and emphasizing personal choice in CAP; Singla et al., Reference Singla, Weobong, Nadkarni, Chowdhary, Shinde, Anand, Fairburn, Dimijdan, Velleman and Weiss2014). In the depression literature, therapist behaviors account for around 5–10% of unexplained variance in patient outcomes and hold in different therapy models and after controlling for confounding patient variables (Kim et al., Reference Kim, Wampold and Bolt2006). Similarly, across the MI literature, growing evidence suggests that high levels of treatment-specific therapy quality (or ‘MI-consistent’ skills) are favorably linked to within-session patient behaviors (Magill et al., Reference Magill, Apodaca, Barnett and Monti2010). A recent RCT demonstrated that higher treatment-specific therapy quality showed better patient outcomes, irrespective of the level of experience of the counselor in treatment delivery (Gaume et al., Reference Gaume, Magill, Longabaugh, Bertholet, Gmel and Daeppen2014a).

Patient behaviors

In HAP, the patient variable specified by theory was patient-reported activation (patient activation). Patient activation levels have been found to mediate the effects of BA treatments on depression outcomes (Gaynor and Harris, Reference Gaynor and Harris2008), including the sustained effects seen in the larger HAP effectiveness trial (Weobong et al., Reference Weobong, Weiss, McDaid, Singla, Hollon, Nadarni, Park, Bhat, Katti, Anand, Dimidjian, Araya, King, Vijayakumar, Wilson, Velleman, Fairburn and Patel2017); but no study to date has examined the potential relations between patient activation, therapy quality and depression outcomes.

In CAP, patient behaviors were change talk and counter-change talk. Both theory and research suggest a positive relationship between MI-consistent skills and patient change talk (Miller and Rose, Reference Miller and Rose2009), and a negative relationship between MI-consistent skills and patient counter-change talk (Moyers et al., Reference Moyers, Martin, Houck, Christopher and Tonigan2009; Gaume et al., Reference Gaume, McCambridge, Bertholet and Daeppen2014b; Magill et al., Reference Magill, Gaume, Apodaca, Walthers, Mastroleo, Borsari and Longabaugh2014; Pace et al., Reference Pace, Dembe, Soma, Baldwin, Atkins and Imelz2017). There is inconclusive evidence about whether patient change talk mediates the effects of therapy quality on patient drinking outcomes (Barnett et al., Reference Barnett, Moyers, Sussman, Smith, Rohrbach, Sun and Spruijt-Metz2014). A recent meta-analysis showed that general skills (e.g. empathy, collaboration) among therapists were related to increased treatment-specific skills and to increased patient change talk but not to patient clinical outcomes (Pace et al., Reference Pace, Dembe, Soma, Baldwin, Atkins and Imelz2017).

In the current study, and in order to better understand the effectiveness of HAP and CAP treatments, we explored the potentially predictive and mediating roles of treatment and patient behaviors on depression and drinking outcomes at 3-months post-enrollment. Our research questions were:

  1. (1) Does therapy quality (general and treatment-specific skills) predict patient behaviors (patient activation in HAP and change and counter-change talk in CAP) and clinical outcomes (depression severity in HAP and daily alcohol consumption in CAP)?

  2. (2) Do patient behaviors predict clinical outcomes?

  3. (3) Do patient behaviors mediate the effects of therapy quality on clinical outcomes?

Methods

Setting

This secondary analysis was conducted within the context of PREMIUM (PRogram for Effective Mental health Interventions in Under-resourced health systeMs), a research program designed to develop and evaluate brief, contextually appropriate PTs for lay counsellor delivery in primary care in Goa, India (Patel et al., Reference Patel, Weobong, Nadkarni, Weiss, Anand, Naiks, Bhat, Pereira, Araya, Dimidjian, Hollon, King, McCambridge, McDaid, Pratima, Velleman, Fairburn and Kirkwood2014). Both the HAP and CAP trials were conducted concurrently (with the two treatments being delivered by the same counselors) in 10 primary health centers. HAP patients were aged 18–65 years, with moderately severe to severe depressive symptoms on the Patient Health Questionnaire (PHQ-9; Spitzer et al., Reference Spitzer, Kroenke and Williams1999). CAP patients were male, aged 18–65 years, who were harmful drinkers, defined as scoring 12–19 on the AUDIT (Saunders et al., Reference Saunders, Aasland, Babor, De La Fuente and Grant1993). Only males were included in the CAP trial because the prevalence of AUDs is low in Indian women (Rathod et al., Reference Rathod, Nadkarni, Bhana and Shidhaye2015). Selection procedures for each trial including inclusion and exclusion criteria, patient characteristics, and outcomes are detailed elsewhere (Nadkarni et al., Reference Nadkarni, Weiss, Weobong, McDaid, Singla, Park, Bhat, Katti, McCambridge, Murthy, Wilson, Kirkwood, Fairburn, Velleman and Patel2017a; Patel et al., Reference Patel, Weobong, Weiss, Anand, Bhat, Katti, Dimidjian, Araya, Hollon, King Mm Vijaykumar, Park, McDaid, Wilson, Velleman, Kirkwood and Fairburn2017). The manuals for both HAP (http://www.sangath.in/wp-content/uploads/2015/08/Healthy%20Activity%20Program_Manual.pdf) and CAP (http://www.sangath.in/wp-content/uploads/2015/08/Counselling%20for%20Alcohol%20Problems_Manual.pdf) and training website (http://premium.nextgenu.org) can be accessed online. Institutional Review Boards at Sangath, the London School of Hygiene and Tropical Medicine, and the Indian Council of Medical Research provided ethical approval.

Procedures

Samples

Two subsamples were randomly selected from the active treatment arm of each trial. Both samples included treatment arm participants only because relevant variables were not collected in the control group of the larger effectiveness trials. Random sampling was used to ensure that the selection of tapes was not biased. Due to the resource-intensive nature of independently rating transcripts, only a subset of the larger group of eligible participants was used in this secondary analysis; however, our methods met the conditions suggested by other mediation studies including a sample size of 40 or more participants, temporally focused relations between mediator and outcome, independent raters of variables of interest, and examined within randomized designs (MacKinnon and Pirlott, Reference Mackinnon and Pirlott2015; Lemmens et al., Reference Lemmens, Muller, Arntz and Huibers2016). A recent review of mediators in psychological treatment trials for depression (Lemmens et al., Reference Lemmens, Muller, Arntz and Huibers2016) recommended 40 or more participants after taking into account both statistical power calculations as well as the average sample sizes in the literature. In HAP, N = 50 participants were randomly selected based on their 3-month remission status (25 remitters and 25 non-remitters, where remission is defined as a PHQ-9 score < 10) from a pool of patients who had received at least two treatment sessions in the middle phase of treatment (sessions 3–5, see below). In CAP, N = 50 participants were randomly selected based on their 3-month remission status (25 remitters and 25 non-remitters where remission is defined as AUDIT score < 8) from a pool of patients who had received at least two treatment sessions (sessions 1 and 2), both of which were of good audio quality. The rationale was two-fold: (a) to include only those patients exposed to sessions in which core HAP and CAP content was delivered; and (b) to assess the temporal relations (Kazdin, Reference Kazdin2007) between therapy quality in the beginning phase (sessions 1–2 in HAP and session 1 in CAP) and patient behaviors in the subsequent session (sessions 3–5 in HAP and session 2 in CAP). No other constraints were applied.

The treatments

Lay counselors with no previous experience in mental health care were trained to deliver both HAP and CAP. Details regarding the selection, training, and supervision of the lay counselors can be found elsewhere (Singla et al., Reference Singla, Weobong, Nadkarni, Chowdhary, Shinde, Anand, Fairburn, Dimijdan, Velleman and Weiss2014).

Health Activity Program (HAP)

HAP was a six–eight session treatment, delivered in three phases (beginning, session 1; middle, sessions 3–5; and end, session 6–8). It is a BA PT (Dimidjian et al., Reference Dimidjian, Barrera, Martell, Munoz and Lewinsohn2011) adapted for the local context (Chowdhary et al., Reference Chowdhary, Anand, Dimidjian, Shinde, Weobong, Balaji, Hollon, Rahman, Wilson, Verdeli, Araya, King, Jordans, Fairburn, Kirkwood and Patel2016). Key strategies include psychoeducation, behavioral assessment, activity monitoring, activity structuring and scheduling, activation of social networks, and problem-solving (Chowdhary et al., Reference Chowdhary, Anand, Dimidjian, Shinde, Weobong, Balaji, Hollon, Rahman, Wilson, Verdeli, Araya, King, Jordans, Fairburn, Kirkwood and Patel2016).

Counseling for alcohol problems (CAP)

CAP is a manualized, MI-informed PT with cognitive-behavioral procedures incorporated. It was delivered in three phases (beginning, middle, and end) and culturally adapted for the local context (Nadkarni et al., Reference Nadkarni, Velleman, Dabholkar, Shinde, Bhat, McCambridge, Murthy, Wilson, Weobong and Patel2015). Key strategies included a detailed assessment with personalized feedback and some psychoeducation (beginning phase), helping the patient to develop problem-solving skills, managing peer pressure, drink refusal and emotional regulation strategies (middle phase), and managing potential relapses by reinforcing key skills (end phase). The therapist's role was to reflect a MI stance while implementing a patient-centered approach to facilitating change that aimed to enhance motivation while simultaneously providing warmth, empathy, and collaboration.

Outcomes

The primary outcome for HAP was depressive symptoms severity scores on the PHQ-9 (Spitzer et al., Reference Spitzer, Kroenke and Williams1999) at 3-months post-enrollment as assessed by independent evaluators who were blind to treatment status. This variable was selected over the trial's other primary outcome of remission status because depressive symptoms offered a continuous score whereas remission was defined as a binary variable (PHQ-9 scores < 10). A continuous score is recommended over binary variables in regression to capture adequate variance (MacKinnon et al., Reference Mackinnon, Lockwood, Brown, Wang and Hoffman2007b).

The primary outcome for CAP was mean daily drinking consumption (measured in grams of ethanol) in the 14 days immediately preceding the 3-month assessment, as assessed by independent evaluators who were blind to treatment status, who calculated this consumption using the timeline follow back method (Sobell and Sobell, Reference Sobell and Sobell1992). Mean daily alcohol consumption was selected as the primary outcome for most of these analyses for two reasons: it is the most widely used outcome in alcohol trials (Bertholet et al., Reference Bertholet, Daeppen, Wietlisbach, Fleming and Burnand2005); and offers a continuous score (MacKinnon et al., Reference Mackinnon, Lockwood, Brown, Wang and Hoffman2007b).

Independent variables

Treatment variable

Therapy quality was assessed by one group of four independent raters using the HAP and CAP Therapy Quality Scale (TQSFootnote Footnote 1, Singla et al., Reference Singla, Weobong, Nadkarni, Chowdhary, Shinde, Anand, Fairburn, Dimijdan, Velleman and Weiss2014). The instrument consists of two subscales assessing general and treatment-specific skills (elements prescribed for the therapeutic modality of HAP or CAP). The general skills subscale involves 10 items (which are the same for HAP and CAP); the treatment-specific skills subscales involve up to 15 items, (different for HAP and CAP), and based on the phase-specific session being rated. The mean of each subscale (0–4) was used to independently estimate general and treatment-specific skills respectively. To ensure temporal distinction between therapy quality and patient behaviors, therapy quality was assessed using beginning phase audiotapes (session 1 in CAP and session 1 or 2 in HAP) only. Psychometric properties of TQS ratings were established in a previous study which demonstrated good concurrent validity between counselor-based ratings and expert ratings (Singla et al., Reference Singla, Weobong, Nadkarni, Chowdhary, Shinde, Anand, Fairburn, Dimijdan, Velleman and Weiss2014). While we assumed that treatment-specific skills build upon general therapeutic skills, we assessed the two separately to determine whether they uniquely contributed to hypothesized temporal pathways.

Patient behaviors were assessed by a second set of four independent raters who were not involved in the rating of therapy quality.

  1. (1) Patient activation was assessed using the PREMIUM Abbreviated Activation Scale (PAAS), which was based on the Behavioral Activation for Depression Scale (Kanter et al., Reference Kanter, Mulick, Busch, Berlin and Martell2007). This five-item scale assessed to what extent patients reportedly engaged in healthy activities and accomplished their between-session goals, each rated on a scale of 0–5 (hence the overall PAAS scores could range between 0 and 25). Patient activation was assessed during the middle phase (sessions 3–5). Reliability estimates of PAAS ratings across four separate transcripts indicated strong inter-rater reliability during the training phase (ICC = 0.815) and internal consistency (α = 0.885).

  2. (2) Change talk and counter-change talk were assessed using the Motivational Interviewing Skill Code (MISC), version 1.1 (Glynn and Moyers, Reference Glynn and Moyers2012). This measure is an observational coding system that categorizes and quantifies patient change language and gives separate frequency counts of change talk and counter-change talk utterances within the rated session (Glynn and Moyers, Reference Glynn and Moyers2009). We selected the MISC 1.1 because, unlike later versions of the MISC and alternative measures of change-talk (e.g. Motivational Interviewing Treatment Integrity), the MISC assessed our construct of interest (rather than additional therapist behaviors which were captured by the TQS described above), was easy to use, and had shown good predictive validity on relevant clinical outcomes (Lombardi et al., Reference Lombardi, Button and Westra2014). Change talk and counter-change talk were assessed using session 2 (middle phase) audiotapes only. Similar to other studies (Moyers et al., Reference Moyers, Martin, Houck, Christopher and Tonigan2009; Campbell et al., Reference Campbell, Adamson and Carter2010), we calculated separate summed score for change talk and counter-change talk. These four independent raters all received training in the use of the MISC 1.1 over 2 weeks. After three rounds of pilot testing, reliability estimates of MISC ratings across four transcripts among the four independent raters indicated good interrater reliability (ICC = 0.718).

Baseline sample characteristics related to the patient (age, education, marital status, occupation, and PHQ-score) and the counselor or treatment (average session duration, recruitment period) were examined as potential covariates.

Data collection

There were three groups of independent data collectors involved in this study; all assessors were blind to the ratings by other assessors. First, independent interviewers assessed primary outcomes at the 3-month endpoint. These data were recorded using tablets that were uploaded in real-time to a server with data being reviewed by independent data managers. Second, as described above under the Independent variables section, two separate groups of four counselors were all blind to outcome status and each independently assessed either therapy quality or patient behaviors from audiotaped sessions using the TQS, and the PAAS or MISC respectively. Randomly selected ratings were quality checked by expert clinicians. No counselor rated his or her own sessions.

Analyses

All analyses were exploratory, using SAS 9.4 and R-studio. First, means and confidence intervals (CIs) for each variable were estimated. Second, Pearson correlations were used to assess the associations among independent variables. If trend or significant relationships were observed (p < 0.10), variables were added to the regression models (described next) to account for maximum change in average daily alcohol consumption and for any potential confounding variables. Similar to other mediation studies that apply a lower threshold (Hayes, Reference Hayes2015), we defined significance as p < 0.10. Among CAP participants, we used baseline AUDIT scores to account for baseline problem severity (these were significantly related to mean daily alcohol consumption at follow-up, r = 0.217, p = 0.005) because mean daily alcohol consumption amounts were not collected at baseline.

Next, we used multiple linear regression to run models where the dependent variable was 3-month daily alcohol consumption in CAP or depressive symptoms in HAP. In each study sample, our models estimated, whether: (a) treatment-specific or general skills in the beginning phase predicted 3-month clinical outcomes; (b) treatment-specific or general skills predicted subsequent patient behaviors (patient activation in HAP and change or counter-change talk in CAP) in the middle phase; and, (c) middle-phase patient behaviors predicted subsequent 3-month outcomes, while accounting for treatment-specific or general skills. Variance inflation factor (VIF) was assessed for each independent variable within each model to estimate multicollinearity (VIF ⩾ 5).

If mediation conditions were metFootnote 2, we examined whether (respectively) patient activation or change talk and counter-change talk mediated the effects of therapy quality on mean depressive symptoms in HAP or daily drinking outcomes in CAP using the Monte Carlo Method for Assessing Mediation (MCMAM; Selig and Preacher, Reference Selig and Preacher2008). In this approach, a distribution of the indirect effect was used to estimate a CI around the observed value of the indirect effect (MacKinnon et al., Reference Mackinnon, Lockwood and Williams2004). MCMAM performs better than the Sobel test and comparably with bootstrap approaches (MacKinnon et al., Reference Mackinnon, Lockwood, Hoffman, West and Sheets2002; MacKinnon and Pirlott, Reference Mackinnon and Pirlott2015) and no direct effect is required of the independent variable (in this case, therapy quality) on the dependent variable (either depressive symptoms or mean daily alcohol consumption at 3-months; MacKinnon et al., Reference Mackinnon, Fairchild and Fritz2007a). In the current study, we computed a 95% CI with 20 000 repetitions.

Results

Among HAP patients (N = 50), 80.0% were female and the sample had an average age of 44.26 years (95% CI 41.21–47.34). The majority were married (72.0%), Hindu (98.0%), homemakers (66.0%), and had completed a minimum of primary level education (80.0%). The patients had received an average of 6.92 sessions (95% CI 6.65–7.19, range 0–8) that lasted an average of 38.93 min (95% CI 37.03–40.84). Mean PHQ-9 scores of 18.66 (95% CI 17.87–19.45) at baseline decreased to 9.34 (95% CI 6.70–8.57) at 3-months post-enrollment.

The CAP patients (N = 50) were, on average, 44.06 years of age (95% C 41.16–46.96). Most were married (78.0%), Hindu (88.0%), unskilled manual laborers (66%), and had completed a minimum of primary level education (70.0%). Out of a maximum four sessions, these patients had received an average of 3.06 sessions (95% CI 2.84–3.28, range 0–4) that lasted an average of 40.26 min (95% CI 38.5–42.0). At 3-months post-enrollment, their mean daily drinking was 19.96 g ethanol (95% CI 8.25–31.67) of ethanol.

The two subsamples were similar to the larger samples in terms of demographic characteristics (Nadkarni et al., Reference Nadkarni, Weobong, Weiss, McCambridge, Bhat, Katti, Murthy, King, McDaid, Park, Wilson, Kirkwood, Fairburn, Velleman and Patel2017b; Patel et al., Reference Patel, Weobong, Weiss, Anand, Bhat, Katti, Dimidjian, Araya, Hollon, King Mm Vijaykumar, Park, McDaid, Wilson, Velleman, Kirkwood and Fairburn2017).

Therapy quality, patient activation, and depression outcomes

The mean beginning phase treatment-specific and general skills scores were 2.45 (95% CI 2.30–2.61) and 2.71 (95% CI 2.60–2.81) respectively and mean PAAS scores were 13.4 (95% CI 12.41–14.35) out of a maximum score of 25. The correlation analyses indicated significant, positive associations between therapy quality treatment-specific skills and patient activation (r = 0.303, p = 0.034), therapy quality general skills and patient activation (r = 0.346, p = 0.015) as well as a significant negative association between patient activation and PHQ-9 outcome scores (r = −0.458, p = 0.0008). These, in turn, suggest medium effect sizes between independent and dependent variables of interest (Cohen, Reference Cohen1988). Recruitment period, baseline depression and treatment dose also were also significantly related to one or more key variables and were therefore added to the regression models.

Table 1 shows the results of the regression and mediation analyses. Treatment-specific and general therapy quality scores in the beginning phase of treatment did not significantly predict subsequent depression scores at the end of treatment. General skills did significantly predict subsequent middle phase patient activation scores during the course of treatment and these activation scores significantly predicted reduced depression scores. A similar but nonsignificant trend was found for treatment-specific skills and patient-activation scores (p < 0.10). The indirect effect of general skills (but not treatment-specific skills) and patient-reported behavior activation on depression outcomes was significant; therefore, increased patient activation in the middle phase mediated the effects of better general skills in the early phase on improved depression outcomes at 3-months post-enrollment. There was no evidence of multicollinearity between any independent variables in the multiple regression models (VIF < 3).

Table 1. Regression and mediating analyses of mean depressive symptoms at 3-months (N = 50)

a Beta estimates are not standardized.

b Control variables included baseline PHQ-9 scores, health counselor, recruitment period, and where relevant, therapy quality scores.

p<0.10. *p<0.05. **p<0.01. ***p<0.001.

Therapy quality, change talk and counter-change talk, and harmful drinking outcomes

The mean treatment-specific and general skills scores were 2.30 (95% CI 2.17–2.43) and 2.45 (95% CI 2.32–2.57) respectively in session 1 and mean frequency counts of change talk and counter-change scores were 16.05 (95% CI 13.92–18.17) and 14.97 (95% CI 12.85–17.11) respectively in session 2. Correlational analyses indicated nonsignificant trend-level or fully significant, associations between treatment-specific skills and patient change talk (r = −0.229, p = 0.095), general skills and counter-change talk (r = −0.229, p = 0.045), as well as a significant negative relation between counter-change-talk (r = 0.364, p = 0.009) and mean daily alcohol consumption outcome (r = 0.525, p = 0.0001). These, in turn, suggest small to large effect sizes between independent and dependent variables of interest (Cohen, Reference Cohen1988). Treatment-specific therapy quality skills were significantly related to general skills (r = 0.866, p < 0.001) whereas session-wise ratings of change and counter-change-talk were not (r = −0.080, p = 0.589). Patient education levels also were significantly related to one or more key variables and, along with baseline AUDIT scores and health counselor, were therefore added to the regression models.

Table 2 and Fig. 2 show the results of the regression and mediation analyses. First, there was a direct association between higher treatment-specific skills and reduced 3-month alcohol consumption at the p < 0.10 level, but not between general skills and reduced 3-month alcohol consumption. Second, at non-significant levels, higher treatment-specific skills in the beginning phase of treatment were associated with increased change talk and reduced counter-change talk in the middle phase. We also found a significant relationship between higher general skills in the early phase and increased change talk in the middle phase, and a nonsignificant relation between general skills and counter-change talk in the middle phase. Third, lower counter-change talk, but not change-talk, significantly predicted reduced mean alcohol consumption at 3-months. Finally, the indirect effect of general therapy quality skills and counter-change talk on reduced drinking was significant, thus suggesting that lower counter-change talk mediated the effect of higher general skills on reduced harmful drinking at 3-months. There was no evidence of multicollinearity between any independent variables in any of the multiple regression models (VIF < 2).

Fig. 2. Predictive and mediating roles of therapy quality, patient behaviors, and clinical outcomes for depression and harmful drinking.

Table 2. Regression and mediating analyses of mean daily alcohol consumption at 3-months (N = 50)

a Beta estimates are not standardized.

b Control variables included baseline AUDIT scores, health counselor, and patient educational level, marital status and where relevant, therapy quality subscales.

c Mediation conditions were not met (i.e. no effect of M → Y) and therefore indirect effects were not tested for mediation.

p<0.10. *p<0.05. **p<0.01. ***p<0.001.

Discussion

The aim of the current study was to explore the temporal pathway of change within two treatments, one for depression and the other for harmful drinking. The specific aim was to investigate the predictive and mediating roles of therapy quality, both general and specific, and patient behaviors on mean depressive symptoms and average daily alcohol consumption outcomes respectively at 3-months post enrollment.

A temporal pathway for change to reduce depressive symptoms and harmful drinking

In line with our hypotheses, we observed preliminary evidence suggesting a temporal pathway in HAP in which therapy quality in the beginning phase of treatment worked through patient activation in the middle phase to reduce subsequent depressive symptoms. Similarly, we found that the effect of general skills was mediated through decreased counter-change talk to ultimately influence mean drinking outcomes. While MI-consistent behaviors have been linked to both change talk and counter-change talk, few have disentangled treatment-specific skills from general nonspecific skills. These findings are broadly consistent with the theories of change that provide the conceptual basis for each of the treatments (Jacobson et al., Reference Jacobson, Martell and Dimidjian2001 for HAP; Magill et al., Reference Magill, Apodaca, Borsari, Gaume, Hoadley, Gordon, Tonigan and Moyers2018 for CAP).

General skills v. treatment skills

Higher levels of both treatment-specific and general skills were associated with positive effects on outcomes for both depression and harmful alcohol use. For HAP, higher levels of both sets of skills were associated with improved patient-activation levels, although only the effects of general skills on reduced depressive symptoms were mediated by patient activation. Importantly, this relation was significant when considering general skills but only evidenced a nonsignificant effect between treatment-specific and patient activation. These findings, if confirmed, highlight the importance of general common skills in determining the outcome of this PT. Similarly, for CAP, a nonsignificant trend was found between treatment-specific therapy quality skills and reduced harmful drinking. Importantly, only higher levels of general skills were mediated by decreased counter-change talk, and not change talk, in reducing drinking outcomes. These findings suggest that treatment-specific skills are contingent on general skills and that treatment-specific and general skills, while related with one another, are differentially related to both treatment outcome and patient behavior in independent ways.

Counter-change talk v. change talk

The CAP findings highlight the importance of reducing counter-change talk, rather than stimulating change talk (Magill et al., Reference Magill, Gaume, Apodaca, Walthers, Mastroleo, Borsari and Longabaugh2014; Romano and Peters, Reference Romano and Peters2016), which the current analyses suggest ultimately influences patient drinking outcomes. While others have demonstrated the importance of counter-change talk, this has often been alongside change-talk for alcohol (e.g. Moyers et al., Reference Moyers, Martin, Houck, Christopher and Tonigan2009). Our findings are consistent with research from the general psychotherapy and behavior change literature that shows that less counter-change talk in early sessions is a stronger predictor of positive outcomes than more change talk (Lombardi et al., Reference Lombardi, Button and Westra2014). The current findings are also consistent with two recent meta-analytic reviews of MI that found a robust role for counter-change talk but not change talk in relation to behavioral outcomes (Magill et al., Reference Magill, Gaume, Apodaca, Walthers, Mastroleo, Borsari and Longabaugh2014; Pace et al., Reference Pace, Dembe, Soma, Baldwin, Atkins and Imelz2017). As theorized in MI, change talk may operate only under specific conditions, with greater counselor experience and more severe alcohol problems (Gaume et al., Reference Gaume, Bertholet, Faouzi, Gmel and Daeppen2013), both of which were limited in the current study context. For this reason, counter-change talk (aka ‘sustain’ talk) is more consistently related to treatment outcomes than change talk (Magill et al., Reference Magill, Apodaca, Borsari, Gaume, Hoadley, Gordon, Tonigan and Moyers2018).

Limitations and implications

There were three main limitations. First, the sample sizes were small. Independent and larger samples are needed to replicate and extend the current findings. Second, other treatment, therapist, and patient-level variables could have been assessed. For example, we did not measure therapeutic alliance from the patient perspective – a frequently studied phenomenon in the PT literature (Ackerman and Hilsenroth, Reference Ackerman and Hillsenroth2003) – although one that garners mixed results as a potential mediator (e.g. Martin et al., Reference Martin, Garske and Davis2000; Lemmens et al., Reference Lemmens, Muller, Arntz and Huibers2016). In addition, we did not assess how cognitions may have influenced patient activation and outcomes. This has been examined in other trials (Lemmens et al., Reference Lemmens, Galindo-Garre, Arntz, Peeters, Hollon, Derubeis and Huibers2017) and the interplay of cognition and behavioral components may inform how brief PTs work. We also did not examine change or counter-change talk among HAP participants or patient activation among CAP patients. Third, in order to assess temporally distinct relations, we evaluated therapy quality only in the beginning phase of treatment and patient activation only during the middle phase of treatment. However, it is possible that they influenced one another in a reciprocal fashion and their relations over time remains unknown. Future studies need to consider the potential role and timing of multiple treatment and patient variable to expand on the temporal pathway highlighted here.

The main implication of our study is that maximizing therapy quality in early sessions is likely to mobilize patient behaviors that ultimately work to improve clinical outcomes. In HAP, our results support the BA model that highlights the central and mediating role of patient-reported activation in alleviating depressive symptoms. In CAP, our results expand on the temporal pathway model to highlight the central and mediating role of counter-change talk, but not change talk, in reducing daily alcohol consumption. These findings are particularly important when training NSPs in suggesting the role and monitoring of therapy quality on potential mediators to ultimately influence treatment effectiveness.

Acknowledgements

This research was funded by a Wellcome Trust Senior Research Fellowship grant to VP (091834). DRS is supported with a Distinguished Fellowship from the Medical Psychiatry Alliance and Academic Scholars Award from the Department of Psychiatry at the University of Toronto. CGF held a Principal Research Fellowship from the Wellcome Trust (046386). The authors also like to acknowledge the contributions of Professors Robert J. DeRubeis, Zachery Cohen, and David P. MacKinnon for their statistical consultations related to our analyses. Next, we would like to thank the diligent and conscientious efforts of our independent raters: Medha Upadhye, Akila Kalsur, Sunaina Halarnkar, Anishka Kalangutkar, Subhash Pednekar, Arati Girap, Shravani Rangapuri, Pranali Kundaikar, Mamata Gosavi, Prashila Gaonkar, and Lalan Karapurkar. Finally, we thank the patients, communities as well as the generous partnership and support from the Directorate of Health Services of the Government of Goa.

Author contributions

DRS drafted the report, in collaboration with RV, SDH, and VP. DRS, RV, VP, SDH, and BW were responsible for the design of the study, with consultation from AN and CGF. AN and AA were responsible for supervising the independent lay counselor evaluators. AN, BB, MG, and BW were responsible for database design and oversight; DRS, with assistance from BW, was responsible for analyses. All authors contributed to the interpretation of the data. All authors reviewed and approved the manuscript.

Footnotes

*

Contributed equally

The notes appear after the main text.

1 The HAP and CAP TQS were previously referred to as the Quality for Healthy Activity Program (Q-HAP) and Quality for Counselling for Alcohol Problems (Q-CAP) Scale, respectively, but were modified during the larger trial for consistency and ease of understanding.

2 Mediation conditions were met if the aforementioned regression models demonstrated that there were significant effects of the independent variable on the proposed mediator (X → M) and of the proposed mediator on outcome scores (M → Y; Mackinnon et al., Reference Mackinnon, Fairchild and Fritz2007a, Reference Mackinnon, Lockwood, Brown, Wang and Hoffmanb), where significance was defined as p < 0.10 (Hayes, Reference Hayes2015).

References

Ackerman, SJ and Hillsenroth, MJ (2003) A review of therapist characteristics and techniques positively impacting the therapeutic alliance. Clinical Psychology Review 23, 133.Google Scholar
Barnett, E, Moyers, TB, Sussman, S, Smith, C, Rohrbach, LA, Sun, P and Spruijt-Metz, D (2014) From counselor skill to decreased marijuana use: does change talk matter? Journal of Substance Abuse Treatment 46, 498505.Google Scholar
Bertholet, N, Daeppen, J-B, Wietlisbach, V, Fleming, M and Burnand, B (2005) Reduction of alcohol consumption by brief alcohol intervention in primary care: systematic review and meta-analysis. Archives of Internal Medicine 165, 986995.Google Scholar
Campbell, SD, Adamson, SJ and Carter, JD (2010) Client language during motivational enhancement therapy and alcohol use outcome. Behavioural and Cognitive Psychotherapy 38, 399415.Google Scholar
Chowdhary, N, Anand, A, Dimidjian, S, Shinde, S, Weobong, B, Balaji, M, Hollon, SD, Rahman, A, Wilson, GT, Verdeli, H, Araya, R, King, M, Jordans, MJD, Fairburn, CG, Kirkwood, B and Patel, V (2016) The healthy activity program lay counsellor delivered treatment for severe depression in India: systematic development and randomised evaluation. The British Journal of Psychiatry 208, 381388.Google Scholar
Cohen, J (1988) Statistical Power Analysis for the Behavioural Sciences (2nd ed.). Hillsdale, NJ: Erlbaum.Google Scholar
Cook, S, Heather, N and McCambridge, J (2015) Posttreatment motivation and alcohol treatment outcome 9 months later: findings from structural equation modeling. Journal of Consulting and Clinical Psychology 83, 232.Google Scholar
Dimidjian, S, Barrera, JRM, Martell, C, Munoz, RF and Lewinsohn, PM (2011) The origins and current status of behavioral activation treatments for depression. Annual Review of Clinical Psychology 7, 138.Google Scholar
Fairburn, CG and Cooper, Z (2011) Therapist competence, therapy quality, and therapist training. Behaviour Research and Therapy 49, 373378.Google Scholar
Gaume, J, Bertholet, N, Faouzi, M, Gmel, G and Daeppen, JB (2013) Does change talk during brief motivational interventions with young men predict change in alcohol use? Journal of Substance Abuse Treatment 44, 177185.Google Scholar
Gaume, J, Magill, M, Longabaugh, R, Bertholet, N, Gmel, G and Daeppen, JB (2014 a) Influence of counselor characteristics and behaviors on the efficacy of a brief motivational intervention for heavy drinking in young men – a randomized controlled trial. Alcoholism: Clinical and Experimental Research 38, 21382147.Google Scholar
Gaume, J, McCambridge, J, Bertholet, N and Daeppen, JB (2014 b) Mechanisms of action of brief alcohol interventions remain largely unknown – a narrative review. Frontiers in Psychiatry 5, 19.Google Scholar
Gaynor, ST and Harris, A (2008) Single-participant assessment of treatment mediators: strategy description and examples from a behavioral activation intervention for depressed adolescents. Behavior Modification 32, 372402.Google Scholar
Glynn, LH and Moyers, TB (2009) Manual for the Motivational Interviewing Skill code (MISC), version 1.1: addendum to MISC 1.0. Center on Alcoholism, Substance Abuse, and Addictions, University of New Mexico. Retrieved 2 September 2013.Google Scholar
Glynn, LH and Moyers, TB (2012) Manual for the Client Language Easy Rating (CLEAR) coding system: Formerly ‘Motivational Interviewing Skill Code (MISC) 1.1’. https://casaa.unm.edu/download/CLEAR.pdfGoogle Scholar
Hayes, AF (2015) An index and test of linear moderated mediation. Multivariate Behavioral Research 50, 122.Google Scholar
Hoeft, TJ, Fortney, JC, Patel, V and Unutzer, J (2016) Task-sharing approaches to improve mental health care in rural and other low-resource settings: a systematic review. The Journal of Rural Health 34, 4862.Google Scholar
Jacobson, NS, Martell, CR and Dimidjian, S (2001) Behavioral activation treatment for depression: returning to contextual roots. Clinical Psychology: Science and Practice 8, 255270.Google Scholar
Kanter, JW, Mulick, PS, Busch, AM, Berlin, KS and Martell, CR (2007) The Behavioral Activation for Depression Scale (BADS): psychometric properties and factor structure. Journal of Psychopathology and Behavioral Assessment 29, 191202.Google Scholar
Kazdin, AE (2007) Mediators and mechanisms of change in psychotherapy research. Annual Review of Clinical Psychology 3, 127.Google Scholar
Kim, DM, Wampold, BE and Bolt, DM (2006) Therapist effects in psychotherapy: a random-effects modeling of the national institute of mental health treatment of depression collaborative research program data. Psychotherapy Research 16, 161172.Google Scholar
Kraemer, HC (2016) Messages for clinicians: moderators and mediators of treatment outcome in randomized clinical trials. American Journal of Psychiatry 173, 672679.Google Scholar
Lemmens, LH, Muller, VN, Arntz, A and Huibers, MJ (2016) Mechanisms of change in psychotherapy for depression: an empirical update and evaluation of research aimed at identifying psychological mediators. Clinical Psychology Review 50, 95107.Google Scholar
Lemmens, LH, Galindo-Garre, F, Arntz, A, Peeters, F, Hollon, SD, Derubeis, RJ and Huibers, MJ (2017) Exploring mechanisms of change in cognitive therapy and interpersonal psychotherapy for adult depression. Behaviour Research and Therapy 94, 8192.Google Scholar
Lombardi, DR, Button, ML and Westra, HA (2014) Measuring motivation: change talk and counter-change talk in cognitive behavioral therapy for generalized anxiety. Cognitive Behaviour Therapy 43, 1221.Google Scholar
Mackinnon, DP and Pirlott, AG (2015) Statistical approaches for enhancing causal interpretation of the M to Y relation in mediation analysis. Personality and Social Psychology Review 19, 3043.Google Scholar
Mackinnon, DP, Lockwood, CM, Hoffman, JM, West, SG and Sheets, V (2002) A comparison of methods to test mediation and other intervening variable effects. Psychological Methods 7, 83.Google Scholar
Mackinnon, DP, Lockwood, CM and Williams, J (2004) Confidence limits for the indirect effect: distribution of the product and resampling methods. Multivariate Behavioral Research 39, 99128.Google Scholar
Mackinnon, DP, Fairchild, AJ and Fritz, MS (2007 a) Mediation analysis. Annual Review of Psychology 58, 593614.Google Scholar
Mackinnon, DP, Lockwood, CM, Brown, CH, Wang, W and Hoffman, JM (2007 b) The intermediate endpoint effect in logistic and probit regression. Clinical Trials 4, 499513.Google Scholar
Magill, M, Apodaca, TR, Barnett, NP and Monti, PM (2010) The route to change: within-session predictors of change plan completion in a motivational interview. Journal of Substance Abuse Treatment 38, 299305.Google Scholar
Magill, M, Gaume, J, Apodaca, TR, Walthers, J, Mastroleo, NR, Borsari, B and Longabaugh, R (2014) The technical hypothesis of motivational interviewing: a meta-analysis of MI's key causal model. American Psychological Association 82, 973983.Google Scholar
Magill, M, Kiluk, BD, McCrady, BS, Tonigan, JS and Longabaugh, R (2015) Active ingredients of treatment and client mechanisms of change in behavioral treatments for alcohol use disorders: progress 10 years later. Alcoholism: Clinical and Experimental Research 39, 18521862.Google Scholar
Magill, M, Apodaca, TR, Borsari, B, Gaume, J, Hoadley, A, Gordon, REF, Tonigan, JS and Moyers, T (2018) A meta-analysis of motivational interviewing process: technical, relational, and conditional process models of change. Journal of Consulting and Clinical Psychology 86, 140157.Google Scholar
Martin, DJ, Garske, JP and Davis, MK (2000) Relation of the therapeutic alliance with outcome and other variables: a meta-analytic review. American Psychological Association 68, 438450.Google Scholar
Miller, WR and Rose, GS (2009) Toward a theory of motivational interviewing. American Psychologist 64, 527.Google Scholar
Moyers, TB, Martin, T, Houck, JM, Christopher, PJ and Tonigan, JS (2009) From in-session behaviors to drinking outcomes: a causal chain for motivational interviewing. Journal of Consulting and Clinical psychology 77, 1113.Google Scholar
Nadkarni, A, Velleman, R, Dabholkar, H, Shinde, S, Bhat, B, McCambridge, J, Murthy, P, Wilson, T, Weobong, B and Patel, V (2015) The systematic development and pilot randomized evaluation of counselling for alcohol problems, a lay counselor-delivered psychological treatment for harmful drinking in primary care in India: the PREMIUM study. Alcoholism: Clinical and Experimental Research 39, 522531.Google Scholar
Nadkarni, A, Weiss, HA, Weobong, B, McDaid, D, Singla, DR, Park, AL, Bhat, B, Katti, B, McCambridge, J, Murthy, P, Wilson, GT, Kirkwood, B, Fairburn, CG, Velleman, R and Patel, V (2017 a) Sustained effectiveness and cost-effectiveness of counselling for alcohol problems, a brief psychological treatment for harmful drinking in men, delivered by lay counsellors in primary care: 12-month follow-up of a randomised controlled trial. PLoS Medicine 14, e1002386.Google Scholar
Nadkarni, A, Weobong, B, Weiss, HA, McCambridge, J, Bhat, B, Katti, B, Murthy, P, King, M, McDaid, D, Park, AL, Wilson, GT, Kirkwood, B, Fairburn, C, Velleman, R and Patel, V (2017 b) Counselling for alcohol Problems (CAP), a lay counsellor-delivered brief psychological treatment for harmful drinking in men, in primary care in India: a randomised controlled trial. The Lancet 389, 186195.Google Scholar
Pace, B, Dembe, A, Soma, C, Baldwin, S, Atkins, D and Imelz, Z (2017) A multivariate meta-analysis of motivational interviewing process and outcome. Psychology of Addictive Behaviors 31, 524.Google Scholar
Patel, V, Weobong, B, Nadkarni, A, Weiss, HA, Anand, A, Naiks, S, Bhat, B, Pereira, J, Araya, R, Dimidjian, S, Hollon, SD, King, M, McCambridge, J, McDaid, D, Pratima, M, Velleman, R, Fairburn, CG and Kirkwood, B (2014) The effectiveness and cost-effectiveness of lay counsellor-delivered psychological treatments for harmful and dependent drinking and moderate to severe depression in primary care in India: PREMIUM study protocol for randomized controlled trials. Trials 15, 101.Google Scholar
Patel, V, Weobong, B, Weiss, HA, Anand, A, Bhat, B, Katti, B, Dimidjian, S, Araya, R, Hollon, SD, King Mm Vijaykumar, L, Park, A, McDaid, D, Wilson, GT, Velleman, R, Kirkwood, B and Fairburn, CG (2017) The Healthy Activity Program (HAP), a lay counsellor-delivered brief psychological treatment for severe depression, in primary care in India: a randomised controlled trial. The Lancet 389, 176185.Google Scholar
Rathod, SD, Nadkarni, A, Bhana, A and Shidhaye, R (2015) Epidemiological features of alcohol use in rural India: a population-based cross-sectional study. British Medical Journal Open 5, e009802.Google Scholar
Romano, M and Peters, L (2016) Understanding the process of motivational interviewing: a review of the relational and technical hypotheses. Psychotherapy Research 26, 220240.Google Scholar
Saunders, JB, Aasland, OG, Babor, TF, De La Fuente, JR and Grant, M (1993) Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction 88, 791804.Google Scholar
Selig, JP and Preacher, KJ (2008) Monte Carlo method for assessing mediation: An interactive tool for creating confidence intervals for indirect effects [computer software]. Available from http://quantpsy.org/.Google Scholar
Singla, DR, Weobong, B, Nadkarni, A, Chowdhary, N, Shinde, S, Anand, A, Fairburn, CG, Dimijdan, S, Velleman, R and Weiss, H (2014) Improving the scalability of psychological treatments in developing countries: an evaluation of peer-led therapy quality assessment in Goa, India. Behaviour Research and Therapy 60, 5359.Google Scholar
Singla, DR, Kohrt, B, Murray, LK, Anand, A, Chorpita, BF and Patel, V (2017) Psychological treatments for the world: lessons from low-and middle-income countries. Annual Review of Clinical Psychology 13, 149181.Google Scholar
Sobell, LC and Sobell, MB (1992) Timeline follow-back. In Measuring Alcohol Consumption. Totowa, New Jersey: Springer, pp. 4172. https://link.springer.com/chapter/10.1007/978-1-4612-0357-5_3Google Scholar
Spitzer, RL, Kroenke, K and Williams, JB and Patient Health Questionnaire Case Study (1999) Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Journal of the American Medical Association 282, 17371744.Google Scholar
Van Ginneken, N, Tharyan, P, Lewin, S, Rao, GN, Meera, S, Pian, J, Chandrashekar, S and Patel, V (2013) Non-specialist health worker interventions for the care of mental, neurological and substance-abuse disorders in low-and middle-income countries. The Cochrane Database of Systematic Reviews 11, CD009149.Google Scholar
Weobong, B, Weiss, HA, McDaid, D, Singla, DR, Hollon, SD, Nadarni, A, Park, AL, Bhat, B, Katti, B, Anand, A, Dimidjian, S, Araya, R, King, M, Vijayakumar, L, Wilson, GT, Velleman, R, Fairburn, CG and Patel, V (2017) Sustained effectiveness and cost-effectiveness of the Healthy Activity Program, a brief psychological treatment for depression delivered by lay counsellors in primary care: twelve-month follow-up of a randomised controlled trial. PLoS Medicine 17, e1002385.Google Scholar
World Health Organization (2015) Mental Health Global Action Programme Humanitarian Intervention Guide (mhGAP-HIG): Clinical Management of Mental, Neurological, and Substance Use Conditions in Humanitarian Emergencies.Google Scholar
Figure 0

Fig. 1. Hypothesized temporal pathways of change.

Figure 1

Table 1. Regression and mediating analyses of mean depressive symptoms at 3-months (N = 50)

Figure 2

Fig. 2. Predictive and mediating roles of therapy quality, patient behaviors, and clinical outcomes for depression and harmful drinking.

Figure 3

Table 2. Regression and mediating analyses of mean daily alcohol consumption at 3-months (N = 50)