Positive addiction recovery therapy: a replication and follow-up study

Abstract

Purpose

The purpose of this study is to see if the affirmative results seen in the pilot study of the positive addiction recovery therapy (PART) programme are replicable and durable given a new cohort of participants. PART is a programme of work designed to improve the recovery and well-being of people in early addiction recovery. Its foundation is in the G-CHIME (growth, connectedness, hope, identity, meaning in life and empowerment) model of addiction recovery. It also uses the values in action character strengths and includes a set of recovery protection techniques.

Design/methodology/approach

This study uses a mixed method experimental design, incorporating direct replication and a follow-up study. Measures for recovery capital, well-being and level of flourishing are used to collect pre-, post- and one-month follow-up data from participants. The replication data analysis uses the non-parametric Wilcoxon test, and the follow-up analysis uses the Friedman test with pairwise comparison post hoc analysis. The eligibility criteria ensure participants (n = 35) are all in early addiction recovery, classified as having been abstinent for between three and six months.

Findings

This study found a statistically significant improvement in well-being, recovery capital and flourishing on completion of the PART programme. These findings upheld the hypotheses in the pilot study and the successful results reported. It also found these gains to be sustained at a one-month follow-up.

Practical implications

This study endorses the efficacy of the PART programme and its continued use in a clinical setting. It also adds further credibility to adopting a holistic approach when delivering interventions which consider important components of addiction recovery such as those outlined in the G-CHIME model.

Originality/value

This study adds to the existing evidence base endorsing the PART programme and the applied use of the G-CHIME model.

Citation

Ogilvie, L. and Carson, J. (2023), "Positive addiction recovery therapy: a replication and follow-up study", Advances in Dual Diagnosis, Vol. 16 No. 4, pp. 227-241. https://doi.org/10.1108/ADD-05-2023-0010

Introduction

Substance use disorder (SUD) is formally recognised as a primary mental health disorder in the Diagnostic and Statistical Manual (DSM) of Mental Disorders [American Psychiatric Association (APA), 2013]. The latest version of the DSM categorises SUD through its symptomology, inter alia, cravings, tolerance, withdrawal and the physical/psychological problems related to substance use (First et al., 2021). SUD is commonly referred to as addiction, with research and public health literature using the term directly to describe the form of the disorder that results in the greatest cost to health and quality of life (NHS, 2023; UK Addiction Treatment, 2023; Volkow and Blanco, 2023; Perry and Lawrence, 2022). Recovery looks at overcoming the detrimental effects of a mental health disorder, being concerned with the resources that an affected individual has and can develop to help them function successfully (Jacob, 2015; van Weeghel et al., 2019). As mentioned, SUD, and therefore, addiction is diagnosable through the severity of its symptoms and recovery from it, and in contrast to other mental health disorders, requires abstinence to adapt to a lifestyle that is not dependent on alcohol and drugs (von Greiff and Skogens, 2021). In the present study, the term addiction recovery is favoured in regard to recovery from SUD, as it aligns with other pertinent literature to its work (Volkow and Blanco, 2023; Perry and Lawrence, 2022; Robinson and Adinoff, 2016; Volkow, 2020; Patton et al., 2022; Krentzman et al., 2023), and is a term directly understood by health services, and those who are in recovery from this mental health disorder (NHS, 2023; UK Addiction Treatment, 2023; DiClemente, 2018).

The cost of addiction

The burden that addiction places on individuals, their families and wider society is considerable. It significantly increases mortality and morbidity, and globally, chronic substance misuse is recorded as a leading cause of death, with alcohol alone contributing to over 200 different diseases (WHO, 2022). Beyond this, its liability is fiscally evident in the disproportionate health-care expenditure afforded to those suffering from addiction and the costs linked to criminal justice and social welfare that result from it (Santangelo et al., 2022). In the UK alone, the cost of drugs to society was estimated to be £19bn in 2017/2018, which is similar in magnitude to the financial burden incurred through alcohol (Black, 2020). There is also the cost of human suffering, in the hurt it brings to those affected and their families, where the layers of distress can be many and complex, for example, relationship breakdown, mental health decline, isolation, unemployment, homelessness, discrimination, stigmatisation and co-dependency (Avery and Avery, 2019; Santangelo et al., 2022; Connery et al., 2020).

The discrimination that arises from the legal and economic burden that addiction places on society marginalises those suffering from addiction. It can contribute to a negative self-view and be a factor in not seeking help, where sadly, the majority of people suffering from substance addiction do not engage with services that can offer the vital help they need (Connery et al., 2020). This treatment gap remains the widest of all mental health disorders and impedes the next generation from escaping consequential familial dysfunction, that is itself a predictor of future addiction (Zilberman et al., 2020; Volkow and Blanco, 2023). Given the magnitude of the problems arising from addiction and their pervasive consequences, providing support that can reinforce addiction recovery is an important health-care need, especially given it is estimated that between 40% and 60% of people who enter addiction recovery go on to relapse (National Institute on Drug Abuse, 2022).

Positive addiction recovery therapy

Positive addiction recovery therapy (PART) is a new programme of work intended to improve well-being, strengthen recovery and create a foundation for people to flourish in addiction recovery. It is based on the growth, connectedness, hope, identity, meaning in life and empowerment (G-CHIME) model. Growth CHIME, abbreviated to G-CHIME, is an adaptation of the CHIME model of mental health recovery (Leamy et al., 2011), which details the components considered important to mental health recovery, namely, connectedness, hope, identity, meaning in life and empowerment. It can be thought of as holistic as it identifies a combination of components that together are considered important to a successful recovery lifestyle (Leamy et al., 2011; Dekkers et al., 2020). With specific reference to addiction recovery, growth has been added to represent the personal development needed to adapt to abstinence and strive to fulfil potential thereafter. Connectedness is for positive relationships, hope to link aspirations with desired outcomes, identity to accept addiction and identify as someone in recovery, meaning in life to gain new purpose and empowerment to have the autonomy and motivation to live a recovery lifestyle. The applied use of G-CHIME has been endorsed through work analysing first-hand narratives of addiction and recovery in the context of its named components (Ogilvie and Carson, 2023), as well as a pilot study of the PART programme, which comprised sessions based on each of its components (Ogilvie and Carson, 2022).

PART assumes the ethos of the recovery model of mental health (Leamy et al., 2011; Jacob, 2015; van Weeghel et al., 2019), advocating that an individual has their own self-actualising potential, which can support the healing process and help them continue to live a good life thereafter (Davidson et al., 2021).

Each PART session delivers what are known as positive recovery interventions (PRIs), interventions that use concepts from positive psychology (PP), knowledge of mental health recovery in G-CHIME and prior research on SUD and addiction. There is a session for each of the G-CHIME components, plus one advocating the use of values in action character strengths in addiction recovery (VIA Institute on Character, 2023), as well as a session that delivers a mini-bundle of recovery protection techniques to mitigate the risk of relapse in triggering situations (Ogilvie and Carson, 2022). Participants are also expected to keep a recovery reflection journal, which is intended to positively affirm their recovery outlook (Krentzman et al., 2023).

The importance of replications and follow-up studies

Failure to confidently replicate the results of research investigating the efficacy of psychological interventions is considered as a significant and damaging problem to the reputation of psychology. If research findings are to be assumed credible for their use in future research and professional practice, they need to have reproducible outcomes within what are considered reasonable empirical boundaries. As an issue, this was highlighted in 2015 when an Open Science collaboration reproducibility study was conducted. This project aimed to reproduce the findings of some 100 distinct psychological research projects that had been published in highly respected journals. The findings showed that of the studies, 97% originally reported statistically significant results, but when replicated, only 36% produced results with statistical significance (Open Science Collaboration, 2015). The baseline value of 36% was considered to be a worrying finding, which highlighted the need to build greater confidence in the reported outcomes of psychological study. As a result, replication studies have become a common practise for researchers who want to validate their original work, gain greater confidence in their hypotheses, and ensure findings can be dependably transferred to future research and applied practice (Flake et al., 2022; Irvine, 2021).

In addition to validating results through replication, researchers are also interested in validating the temporal effects of their research, in effect, evidencing if a psychological intervention is achieving durable health-related change (von Allmen et al., 2015). In their research looking at the protracted effects of positive psychology interventions (PPIs), Cohn and Fredrickson (2010) identified compelling reasons for, including a follow-up study in the research schedule for interventions intended to positively reinforce ongoing healthy function, such as PART, suggesting it was important for the successful maturation of PP and could provide insight into the self-generating effects of PPIs where people continue to develop skills for their own benefit and well-being.

A pilot study of the PART programme (Ogilvie and Carson, 2022) produced encouraging results in support of offering a positive choice to people who have traditionally been viewed in terms of their deficits (Avery and Avery, 2019). Given the success of the pilot study and the potential benefit this has to the long-term well-being of people in addiction recovery, substantiating the continued use of PART in a clinical setting will provide greater confidence in the results seen to date and show them to be transferable to a different cohort of participants. Further study of the PART programme will also ascertain if the gains seen on completion of the programme persist in ongoing recovery.

The present study

This work reproduces the pilot study of the PART programme to see if the results can be verified through replication. In line with the pilot (Ogilvie and Carson, 2022), the study hypothesises that following participation in the PART programme, there will be:

• Significant improvements in well-being as assessed by the Short Warwick–Edinburgh Mental Well-being Scale (SWEMWBS).

• Significant increases in recovery capital as assessed by the Brief Assessment of Recovery Capital (BARC-10).

• Significant increases in flourishing as assessed by the positive emotions, engagement, relationships, meaning and accomplishment (PERMA) profiler.

As this study engages with a new cohort of participants, a follow-up is being introduced, also using the SWEMWBS, BARC-10 and the PERMA profiler to ascertain how the participants’ well-being, recovery capital and level of flourishing change or are maintained one month after completing the programme. The present study is, therefore, serving as a replication and follow-up study with the aim of improving the validity of the original research.

Methods

This study uses a mixed method experimental design, incorporating a direct replication and a follow-up study. Direct replication seeks to confirm the results of a study by following the same methodology, matching the psychological conditions of the original study as closely as possible, to limit the causal independence between them. It is effective in validating research where a control group is not available, and researchers have selected apposite measures that have been validated and used in their previous work (Flake et al., 2022). Direct replication is being used to validate the results of the PART pilot study (Ogilvie and Carson, 2022), using a consistent methodological approach (Irvine, 2021), to establish whether or not the favourable results seen in the pilot are replicable. The adopted measures, recruitment strategy and timing of data collection at the start and end of the PART programme are consistent between the pilot and the replication components of this work, and for consistency, the PART programme is being delivered by the same facilitator using consistent PRIs and related materials.

The follow-up stage uses the same measures, permitting direct comparison, with data collected one month after the programme was completed. This means there are three analysable data sets: pre-PART, post-PART and one-month follow-up-PART. The elected time period of one month is decided based on the following two factors:

1. In educating participants on connectedness within PART, they are being encouraged to attend mutual support groups, as this is a strong example of positive relationships in recovery. On completing PART, it is likely that participants will continue to access this resource and engage with a sponsor to start a 12-step recovery programme, as is common in mutual aid groups such as Alcoholics Anonymous. Given a longer time frame, there is an increased likelihood that participants will have commenced this work, which itself could influence the scores given in the survey. To protect against the bias this could introduce, the time frame of one month was set to limit participant attrition as a result of this. A checkpoint was added to ensure participants had not engaged with the work involved in a 12-step programme when they completed the follow-up survey.

2. As mentioned, relapse rates remain high in addiction recovery (National Institute on Drug Abuse, 2022). While work such as this aims to better equip people and reduce this, relapse remains a sad reality of addiction (Robinson and Adinoff, 2016). Previous research looking at patterns of relapse in people following treatment found that of those who relapsed, the median time from treatment to relapse was 21 days after exiting the treatment service (Nordfjærn, 2011). For participants in this study, this is marked by the end of the PART programme. In electing to do a one-month follow-up, participants who fall into the average pattern of relapse after leaving a treatment service will be discounted, without projecting too far into the future, where results could be influenced to a greater extent by ancillary factors that may stabilise recovery, for example, restored relationships and better physical health. Furthermore, given the average post-treatment relapse pattern (Nordfjærn, 2011) and that participants have been abstinent for between three and six months upon commencing PART where they have been engaged in treatment, it was felt that one month was a reasonably long time in their addiction recovery to date.

Participants

As in the pilot (Ogilvie and Carson, 2022), participants (n = 35) were recruited from a drug and alcohol rehabilitation treatment centre, where they had previously participated in group therapy for a period of 12 weeks to address their addiction. Entry to the rehabilitation centre was voluntary and not enforced, for example, by court order. On completion of the 12-week group therapy, participants self-elected to enrol on the PART programme without delay. The PART sessions were delivered in the same clinical setting as the group therapy using the same group room facilities. The participants were all UK citizens and considered to be in early recovery, having been abstinent for between three and six months when starting the PART programme (Melemis, 2015). There was a mix of males and females. All participants self-identified as either male or female. No participants self-identified as non-binary or gender not listed. Participants ranged in age from 23 to 67 years old. See Table 1 for descriptive details of the participants and their individual score profile.

Ethical approval for the study was obtained from the Psychology Department at the University of Bolton, in line with British Psychological Society Guidelines (British Psychological Society, 2018). Participants were provided with an information sheet at the start of each survey. Participants were informed that completing the survey was giving their consent to participate and that participation in the survey was voluntary. All data was stored with respect to anonymity and confidentiality and in accordance with the UK General Data Protection Regulation.

Measures

Three measures were used to collect data:

• BARC-10 shortened version.

• SWEMWBS.

• PERMA profiler.

Participants completed three copies of the same survey, each containing the questions for the above-listed measures. One at the start of the PART programme (pre-PART), another at the end (post-PART) and the final one a month after completion (follow-up-PART). The pre- and post-PART data collection is consistent with the pilot study (Ogilvie and Carson, 2022).

Brief assessment of recovery capital shortened version

The BARC-10 is a brief version of the original 50-item assessment of recovery (ARC) scale (Groshkova et al., 2012). The ARC and BARC-10 measure the level of recovery capital held by an individual. The simplified ten-question version was validated as psychometrically sound and simpler to complete in 2017 (Vilsaint et al., 2017). The questions are answered on a six-point Likert scale ranging from “Strong disagree” = 1 to “Strongly agree” = 6. The sum of the answers gives a measure of recovery capital, with a minimum score of 10 and a maximum of 60. Questions such as “In general, I am happy with my life” and “I get lots of support from friends” are used to evaluate the overall strength of recovery.

Short Warwick–Edinburgh mental well-being scale

The SWEMWBS is used to measure mental well-being (Warwick Medical School, 2019). It comprises seven statements that participants rate using a frequency of occurrence scale from “None of the time” = 1 to “All of the time” = 5 to describe their thoughts and feelings, for example, “I’ve been thinking clearly” and “I’ve been feeling useful”. The item scores are summed together to give a total well-being score. The scale is well-established and respected in the evaluation of mental well-being, having been used in research across multiple disciplines (Warwick Medical School, 2020).

Positive emotions, engagement, relationships, meaning, accomplishment profiler

The PERMA profiler measures five domains of well-being that contribute to human flourishing (Butler and Kern, 2016). The sum of the scores gives an overall measure of flourishing, and the total for each of the questions that relate to each domain shows to what degree they contribute to an individual’s flourishing, as well as giving an indication of their well-being in that area (Seligman, 2018; Carr, 2020; Butler and Kern, 2016). Questions are presented using a scale of “Not at all” = 0 to “Completely” = 10 or “Never” = 0 to “Always” = 10. For positive emotions, sample questions such as “In general, how often do you feel joyful?” are asked; for engagement, “How often do you become absorbed in what you are going?”; for relationships, “To what extent do you feel loved?”; for meaning, “In general, to what extent do you lead a purposeful and meaningful life?”; and for accomplishment, “How often do you achieve the important goals you set for yourself?”.

Data analysis

The pre-PART, post-PART and follow-up-PART scores were calculated for each of the three measures (BARC-10, SWEMWBS and PERMA), as well as the five PERMA subscales for positive emotions, engagement, relationships, meaning and accomplishments (PERMA-P, PERMA-E, PERMA-R, PERMA-M and PERMA-A). This gave eight analysable scores in each data set. The replication study used the same statistical tests as the pilot, so the non-parametric Wilcoxon signed-ranks test was used to analyse the pre- and post-PART data, having originally been selected for its suitability in evaluating matched pairs under different conditions, where the test data is ordinal and does not assume normality (Pallant, 2016). Effect size was calculated using the formula r = Z/√N for the Wilcoxon signed-rank test where 0.10 is considered small, 0.30 medium, 0.50 large and 0.70 very large (Statstutor Community Project, 2023a). Statistical significance was measured at p = 0.00625 following a Bonferroni correction of 0.05/8 = 0.00625 for multiple statistical comparisons (Pallant, 2016). Internal reliability of the PERMA subscales was checked using a Cronbach alpha where α > 0.7 was considered a good level of internal reliability (Taber, 2018).

For the follow-up analysis, the Friedman test was used. The Friedman test is a non-parametric statistical calculation that analyses the variance of three or more groups of data under different conditions or time periods (Laerd Statistics, 2018). It shows if the group mean scores are statistically different given the same subjects and comparable data sets; in the case of this analysis, the pre-PART, post-PART and follow-up-PART scores returned from the SWEMWBS, BARC-10 and PERMA measures. The Friedman test is suited to analysing scale data, such as that gathered through the Likert style questions in this study, where normality cannot be assumed (Pallant, 2016; Statstutor Community Project, 2023b). Post hoc analysis was conducted using pairwise comparison of the output of the Friedman test with Bonferroni corrections to see if there was a significant difference in the three repeated measures, pre-PART to post-PART, pre-PART to follow-up-PART and post-PART to follow-up-PART. The effect size of this significance was calculated using the same formula and thresholds as the Wilcoxon test (Statstutor Community Project, 2023a). The software package SPSS version 27 was used for all statistical analysis.

Results

The age of the participants ranged from 23 to 67 (M = 43.89, SD = 10.54), and there were more men (n = 19) than women (n = 16). Two of the participants dropped out of the programme as a result of relapse; all remaining participants (n = 33) completed the programme and the subsequent post-PART survey. This gave a sample size of 33 for the replication study, a retention of 94% of participants. Two participants did not complete the follow-up survey, one as a result of relapse and the other because they were working with a sponsor. This left a sample of 31 for the follow-up study, a retention of 89% of the original study sample.

Replication analysis

As with the pilot, the post-PART scores were higher than the pre-scores in all measures, see Table 2. The Cronbach’s alpha calculations showed a good level of internal consistency for all PERMA subscales (α > 0.74 and α < 0.96) in both the pre- and post-PART data (Taber, 2018).

The results of the Wilcoxon signed-rank test and the effect size for both the pilot and replication study are shown in Table 3. The data pertaining to this analysis considers only the pre- and post-scores and the hypotheses from the pilot study, showing there was an increase in recovery capital as assessed by the BARC-10, where the participants had higher BARC-10 scores after completing the PART programme (M = 53.58, SD = 4.40) than they did before (M = 48.89, SD = 5.94). This increase was statistically significant with a large effect size (z = −4.31, p = 0.001, ES = 0.53). An improvement was also seen in well-being as assessed by the SWEMWBS, where the participants had a higher SWEMWBS score after completing the PART programme (M = 27.79, SD = 3.35) than before (M = 24.34, SD = 4.01). This increase was statistically significant with a large effect size (z = −4.07, p = 0.001, ES = 0.50). There was also a significant increase in flourishing as assessed by the PERMA profiler; here, the participants had a higher PERMA score after completing the PART programme (M = 121.42, SD = 20.32) than before (M = 105.20, SD = 23.29). This increase was statistically significant and again showed a large effect size (z = −4.61, p = 0.001, ES = 0.57). The results in the replication analysis support all hypotheses, as was the case in the pilot study. Furthermore, the five PERMA subscales had a meaningful difference with a medium to large effect size, see Table 3. While the effect sizes in the pilot and the replication study were large at 0.5 and above (Statstutor Community Project, 2023a), the values were lower in the replication study, see Table 3.

To fully replicate the pilot (Ogilvie and Carson, 2022), a Wilcoxon signed rank test was executed using a median split at the 50th percentile for young and old adults to establish if there was a difference in the results based on this age split. While all the results were meaningful at p < 0.05, taking the Bonferroni corrected value of p = 0.00625, older adults (45 and over) appeared to respond better than younger adults. The significance for younger adults was found to be less than the Bonferroni corrected value for the SWEMWBS and the BARC-10 with findings of p = 0.021 and p = 0.014, respectively. Similarly, a test to see if there was a difference between males and females showed that all results were significant at p < 0.05; however, females responded marginally better in the BARC-10 and SWEMWBS, where males fell slightly below the corrected value with the significance of p = 0.009 and p = 0.008, respectively. While males showed a meaningful difference of p = 0.006 in the PERMA, this was not true of all the subscales. The PERMA-M score fell below the corrected value at p = 0.003, but the other subscale scores were not meaningly higher than the Bonferroni adjusted value.

Follow-up analysis

The Friedman test indicated that there was a statistically significant difference in the BARC-10 across the measures [pre-PART, post-PART, follow-up-PART χ² (2, n = 31) = 31.73, p = 0.001]. There was also a statistically significant difference in the SWEMWBS across the measures [pre-PART, post-PART, follow-up-PART χ² (2, n = 31) = 29.48, p = 0.001], which was also true for the PERMA profiler [pre-PART, post-PART, follow-up-PART χ² (2, n = 31) = 29.23, p = 0.001].

The pairwise comparison post hoc analysis found this significance was present in all the pre-to-post and pre-to-follow-up results but not for post-to-follow-up, where no meaningful difference was found, see Table 4. The mean scores for the BARC-10 increased between the post (M = 53.48, SD = 4.41) and follow-up (M = 54.29, SD = 3.98); this was also true for the post (M = 27.58, SD = 3.26) and follow-up (M = 28.10, SD = 2.70) SWEMWBS and the post (M = 121.19, SD = 19.58) and follow-up (M = 126.68, SD = 19.58) PERMA. It can also be seen that the biggest effect was between the pre- and post-scores and that the post-to-follow-up scores showed less effect across all three measures. Note that the post results differ slightly from the replication study, having been calculated using the follow-up sample (n = 31), not the replication sample (n = 33).

Discussion

In support of the findings in the pilot study, the replication study also found a positive and meaningful difference in the participants well-being, recovery capital and level of flourishing after completing PART. This advances the evidence base for the efficacy of the programme and also the importance of viewing recovery holistically, where an intentional grouping of interventions strengthens its multiple domains, as expressed in the G-CHIME model (Ogilvie and Carson, 2023). Replication studies are able to expose weaknesses in the theoretical links with intervention design and the supporting literature they are based on (Irvine, 2021). The findings are also able to endorse their design by substantiating desired outcomes, where apposite theory has been successfully applied to support beneficial health-related change given appropriate recipients and areas of mental health concern (Flake et al., 2022), for example, people in addiction recovery. Validating the success of PART (Ogilvie and Carson, 2022) in a clinical setting using a different cohort of participants was the primary objective of this work.

While all hypotheses defined in the pilot were supported in the replication study, with significant magnitude, the mean scores for the post-measures were less than originally found in the pilot; this was also reflected in a reduction in effect size, see Table 3. A possible reason for this could be the decline in well-being that is currently being observed in the UK generally. National data collected using the SWEMWBS shows the long-term deterioration in personal well-being between 2012 and 2019 (Office for National Statistics, 2023). More recent national data on the measure of well-being using the SWEMWBS has yet to be published, although both the post and follow-up mean scores in this study were higher than those last reported (27.79 and 28.10 compared with 24.30). It is reasonable to infer that given current events, such as the cost-of-living crisis, the war in Ukraine, a struggling health service and the recent wave of public sector strikes, this deterioration will continue and will influence participant well-being (Broadbent et al., 2023). This is corroborated by the current declining trajectory in national measures of happiness and life satisfaction (Office for National Statistics, 2023). Taking the reported general deterioration in the context of this study would suggest that the participants sit outside the normative national levels, as their well-being continues to improve, albeit with less effect than the pilot. While not statistically significant, the increase seen between the post and follow-up mean scores, see Table 3, provides evidence that well-being, flourishing and recovery capital have stabilised at a higher level than before PART.

As mentioned, relapse after treatment is most likely to occur at 21 days (Nordfjærn, 2011), and it is estimated that between 40% and 60% of people who enter recovery go on to relapse (National Institute on Drug Abuse, 2022). This would suggest engagement in the PART programme is able to better equip people to succeed in their recovery, at least in reaching the first-month milestone with a stable level of well-being and recovery capital. Furthermore, of the participants who were eligible (n = 33) for the follow-up, only one relapsed within this timeframe, which is important given one of the objectives of the PART programme is to set the foundation for people to flourish in their recovery. The stabilised level of well-being and recovery capital seen in the participants between the post and follow-up measures in Table 4 suggest that PART could offer a basis for continued flourishing. Here, the findings, while not statistically significant, show an increase with a larger effect size for the PERMA measure than the other post and follow-up results listed.

The exploratory analysis in the pilot study showed that females responded marginally better than males in the SWEMWBS and BARC-10 measures using the Bonferroni corrected value of p = 0.00625. This finding was consistent with the replication study and suggests a need to identify reasons for this disparity. Given the marginal differences reported and the current recruitment strategy, this would be difficult to investigate with adequate depth at present. In the pilot, no difference was detected between young and old adults; however, in the replication study, the scores among older adults were slightly higher. It is surmised that the different results could be a characteristic of the individual cohorts that are not generalisable across the target population (Diener and Biswas-Diener, 2023; Patil et al., 2016).

Limitations and future research

A limitation to the PART study design is that all participants were recruited directly from an addiction treatment centre on completion of a 12-week course of group therapy. The baseline measures for well-being, recovery capital and flourishing were taken at this stage, and the effect that the PRIs had on these were analysed at the end of the PART programme. It is unknown if there was enduring improvement in these measures from the previous group therapy during this time. This limitation of the pilot study gave reason for the replication component of this work (Ogilvie and Carson, 2022), and while the replication study has shown there to be consistency across the two study samples, to discount potential enduring effects of the group therapy with greater confidence would require a new research design and recruitment strategy, most likely using an randomised control trial (RCT) design (Hariton and Locascio, 2018). This was not feasible, given the clinical samples used in both the pilot and replication study.

The high retention rate seen in the pilot, and replication and follow-up study saw only three participants drop out due to relapse. This is assuming honest self-disclosure from the participants on their continued abstinence, but it cannot be known for certain that this is an accurate assumption; the conduct of the participants during the PART programme, however, and at the time of collecting the follow-up data, would suggest they were still in addiction recovery. As already mentioned, relapse rates remain high in addiction recovery, where it is estimated between 40% and 60% of people who enter recovery go on to relapse (National Institute on Drug Abuse, 2022). It is suggested the high retention is a result of the participants being in a stable state of recovery, having completed the 12-week course of group therapy, and also due to their personal motivation, where as a sample group, they all choose to extend their engagement with the treatment centre, instead of discharging from the service after 12 weeks when their group therapy concluded. This retention rate is likely not representative of the more general recovery population that exists outside of treatment. It is envisaged that a new study using an RCT design (Hariton and Locascio, 2018) would offer more information on how the PRIs are able to protect against relapse for a wider recovery population.

Replication studies can identify when optimistic reporting has taken place. This is where a researcher positively influences how they report their work. Optimistic reporting is often unintended and driven by a belief in the efficacy of the research being conducted (Patil et al., 2016). Replication studies that use an independent researcher remove this unintended bias. However, the replication component of this work has been conducted by the same researcher as the pilot study. To address this limitation of the work and uphold transparent reporting, a second independent researcher was consulted throughout all stages of the work.

In replication studies, a small sample size means statistical significance can be found by chance, as could be the case with the difference observed for young and old adults in the replication component of this work, and the different significance levels seen in the PERMA subscales for males between the pilot and replication data (Diener and Biswas-Diener, 2023). Given PART is currently delivered in a small group setting to both males and females of mixed ages, more reliable conclusions cannot be drawn on this at present and would require a different recruitment strategy. Further to this, adopting a more diverse recruitment strategy, for example, engaging members from an established recovery community, could show if PART is effective for people who have been in recovery for longer than the participants in this study.

Future research should look at why the pilot and replication study appear to have been slightly more effective for female participants. To achieve this, it is suggested that a detailed analysis on the outcomes pertaining to each area of the programme is required, where feedback is also sought on participant opinion. This could be extended to consider how well individual aspects of the programme were received by participants for use in possible future enhancements. Furthermore, the research conducted to date has not considered the value of individual PRIs in isolation of the others. This is a limitation in the study design thus far, as it is conceivable that some of the PRIs will have been considered more effective by participants.

The results of the pilot, and replication and follow-up study have been auspicious, but care must be taken not to overvalue the advantages of the PART programme. It has been shown as effective, but as an auxiliary treatment, where all participants to date have reached a level of stability having completed 12 weeks of group therapy prior to engaging with PART. PART is not intended as a therapy that addresses the causal factors of addiction or a means to help individuals deal with past trauma or adverse experiences. It has, however, shown to be reinforcing in its own right and complementary to what could be considered as initial treatment for addiction. Future study is required to understand the extent to which it benefits people engaged with alternative recovery approaches, such as 12-step recovery, and also to see how beneficial it is to people at different stages of their recovery temporally and in levels of well-being. This direction of study could also allow for introducing a control group to gauge how it compares to alternative recovery practices, and also help negate unintended bias that could be introduced by other subjective factors such as attendance of mutual support meetings or making generally better lifestyle choices through abstinence and recovery.

Conclusion

This replication and follow-up study validated the results of a pilot study for PART programme. It used direct replication to closely mirror the original study with a new cohort of participants and, in doing so, confirmed the findings as statistically significant. The second branch of this research incorporated a follow-up study, which considered the durability of the results temporally. The follow-up data showed that the meaningful gains in well-being, recovery capital and flourishing were maintained one month after the completion of the PART programme, where participants had passed the average relapse time frame. It was also noted that while the results were meaningful, with a large effect size, the mean scores seen in the replicated data were slightly lower than the post programme scores in the pilot study. It has been surmised this may be due to the general decline in well-being that is being observed nationally in the UK. This also showed that unlike the declining national trend, the participants in both the pilot and replication studies have seen an increase in their well-being, which was maintained in the follow-up stage of this work. These results add credibility to the evidence base supporting the continued use of PART in a clinical setting, also supporting the use of the G-CHIME model in addiction recovery.

Table 1

Participant descriptive statistics and score profile

			BARC-10 scores			SWEMWBS scores			PERMA scores
Ref	Gender	Age	Pre	Post	Follow	Pre	Post	Follow	Pre	Post	Follow
1	Male	42	53	57	59	25	29	30	105	137	136
2	Male	51	52	53	53	23	25	28	96	110	132
3	Male	48	56	58	60	28	32	32	114	127	149
4	Female	36	50	57	58	28	27	29	110	117	139
5	Male	29	43	42	49	22	24	27	65	73	103
6	Male	47	49	51	53	26	26	28	104	106	112
7	Female	40	37	47	49	20	20	27	64	95	120
8	Female	52	53	58	52	29	31	26	119	141	121
9	Female	29	51	57	59	20	25	27	94	131	133
10	Female	50	49	51	51	21	27	25	89	122	126
11	Male	52	49	52	56	25	28	29	110	102	120
12	Female	47	51	57	54	25	25	26	129	141	134
13	Female	45	50	54	56	24	30	30	130	144	146
14	Male	51	49	54	51	25	31	26	114	126	104
15	Female	47	35	51	53	17	22	26	49	88	109
16	Female	45	45	52	52	20	25	28	91	116	122
17	Male	41	55	56	54	34	28	27	117	127	133
18	Male	25	47	47	51	22	24	25	89	88	109
19	Male	40	50	43	44	25	25	26	100	93	98
20	Female	48	54	54	49	26	26	25	104	122	86
21	Male	56	40	48	50	19	24	24	100	112	113
22	Female	23	55	56	57	29	30	30	133	147	150
23	Male	54	52	53	58	27	29	29	130	123	148
24	Male	49	51	52	53	24	29	28	113	120	123
25	Male	29	52	55	60	24	29	34	109	132	151
26	Male	29	37	−	−	15	−	−	47	−	−
27	Female	46	50	55	54	27	30	27	118	126	120
28	Male	49	43	51	−	21	28	−	85	96	−
29	Male	35	35	−	−	23	−	−	92	−	−
30	Male	67	55	56	57	22	27	25	115	118	124
31	Female	37	56	59	−	28	34	−	137	154	−
32	Male	56	54	59	59	33	35	35	142	153	149
33	Female	42	51	58	59	24	31	31	110	141	145
34	Female	65	50	59	55	25	31	29	124	138	139
35	Female	34	52	56	58	26	30	32	134	141	133

Source: Table by authors

Table 2

Pre-PART, post-PART and follow-up-PART score descriptive statistics

Scale	Pre-PART M (SD) N = 35	Post-PART M (SD) N = 33	Follow-up-PART M (SD) N = 31
BARC-10 score	48.89 (5.94)	53.58 (4.40)	54.29 (3.98)
SWEMWBS score	24.34 (4.01)	27.79 (3.35)	28.10 (2.70)
PERMA score	105.20 (23.29)	121.42 (20.32)	126.68 (17.01)
PERMA-P score	19.00 (5.50)	21.55 (4.18)	22.42 (3.40)
PERMA-E score	20.17 (4.35)	22.64 (4.31)	24.03 (3.68)
PERMA-R score	20.29 (5.02)	23.88 (4.36)	24.65 (3.75)
PERMA-M score	18.94 (5.60)	22.67 (4.94)	23.52 (4.65)
PERMA-A score	20.26 (4.14)	22.91 (3.56)	24.13 (3.37)

Notes:

M = mean; SD = standard deviation; N = number of participants

Source: Table by authors

Table 3

Pilot and replication study results

	Pilot study N = 30				Replication study N = 33
Score	Pre M (SD)	Post M (SD)	Sig.	ES	Pre M (SD)	Post M (SD)	Sig.	ES
BARC-10	48.53 (6.97)	55.07 (3.90)	Z = −4.54 p = 0.001	0.60	48.89 (5.94)	53.58 (4.40)	Z = −4.31 p = 0.001	0.53
SWEMWBS	24.40 (4.60)	30.00 (3.46)	Z = −4.71 p = 0.001	0.60	24.34 (4.01)	27.79 (3.35)	Z = −4.07 p = 0.001	0.50
PERMA	111.47 (22.39)	129.27 (17.76)	Z = −4.76 p = 0.001	0.60	105.20 (23.29)	121.42 (20.32)	Z = −4.61 p = 0.001	0.57
PERMA-P	20.43 (4.95)	23.67 (3.93)	Z = −4.62 p = 0.001	0.60	19.00 (5.50)	21.55 (4.18)	Z = −3.38 p = 0.001	0.42
PERMA-E	21.07 (3.90)	23.97 (4.01)	Z = −4.43 p = 0.001	0.60	20.17 (4.35)	22.64 (4.31)	Z = −3.67 p = 0.001	0.45
PERMA-R	21.13 (5.36)	24.03 (3.85)	Z = −3.91 p = 0.001	0.50	20.29 (5.02)	23.88 (4.36)	Z = −4.32 p = 0.001	0.53
PERMA-M	20.90 (4.95)	24.77 (3.48)	Z = −4.56 p = 0.001	0.60	18.94 (5.60)	22.67 (4.94)	Z = −4.56 p = 0.001	0.56
PERMA-A	20.83 (4.20)	24.37 (3.15)	Z = −4.73 p = 0.001	0.60	20.26 (4.14)	22.91 (3.56)	Z = −3.23 p = 0.001	0.40

Notes:

N = number of participants; M = mean; SD = standard deviation; Sig. = significance; ES = effect size

Source: Table by authors

Table 4

Post hoc analysis results

	Pairwise comparison N = 31
	Pre-PART to post-PART		Pre-PART to follow-up-PART		Post-PART to follow-up-PART
Score	Sig.	ES	Sig.	ES	Sig.	ES
BARC-10	Z = −3.81 p = 0.001	0.48	Z = −5.33 p = 0.001	0.68	Z = −1.52 p = 0.383	0.20
SWEMWBS	Z = −3.94 p = 0.001	0.50	Z = −4.83 p = 0.001	0.61	Z = −0.89 p = 1.00	0.11
PERMA	Z = −3.43 p = 0.002	0.44	Z = −5.33 p = 0.001	0.68	Z = −1.91 p = 0.17	0.24

Notes:

N = number of participants, Sig. = significance, ES = effect size

Source: Table by authors