Planning research in psychosocial interventions.

A number of research designs have been used to study the efficacy of psychosocial interventions in addictive disorders, including open label studies and randomised controlled trials. Only through a rigorously conducted research, evidence base for effectiveness of a psychosocial intervention can be established. However, research on these interventions are fraught with a number of challenges. It is imperative for researchers to ask appropriate research questions based on sound theoretical understanding of psychiatric disorders, psychosocial interventions and research designs. This would help in choosing the less studied, relevant areas for in depth study as well as in using pragmatic, realistic research designs. Defining intervention clearly is as crucial, as is its uniform implementation across various treatment arms. In addition, tapping the mediators, moderators and confounders of treatment using appropriate methods while assessing the factors that directly impact the outcome is important to determine actual effects of psychosocial intervention. Barriers at different stages must be gauged proactively and dealt with, wherever possible.

INTRODUCTION

Patients with Substance Use Disorders (SUDs) are known to face a number of problems in almost all areas of life, including medical, social, emotional, financial, legal, and so on. These problems warrant simultaneous, comprehensive solutions, in addition to evidence based treatment for substance use problems1. As highlighted in the other articles, Psychosocial Interventions (PI) for addiction have, therefore, become crucial for effective management of addiction and related issues, and are successfully used as independent treatment methods as well as adjuncts to pharmacological treatment for substance use.

If introduced at the outset, PI may enable clients to understand drug issues and adhere to treatment goals. For instance, motivational interviewing (MI) may help substance users in recognizing the need for change, by addressing ambivalence toward drug related issues while behavioural interventions may help address problem drug use by providing personalised feedback. The most interesting aspects of using PI is that these can be used flexibly, as individual strategies or in various combinations, to cater to the specific needs of the clients2.

However, are PI effective in addiction treatment? In order to be recommended as a treatment guideline, it is important that the effectiveness of a given PI is established through research studies which employ a robust methodology. Thus, it is, very important to not overlook the inherent limitations of available studies. In addition, ongoing research is essential to further strengthen the evidence base regarding effectiveness and applicability of PI in managing addictive disorders.

In this article, we begin by providing a brief overview of available research on effectiveness of PI [Table 1] followed by presenting a set of brief guidelines on planning research on PI.

Table 1

Studies on PI in addiction

While, most of the other articles in this publication have provided a review of existing literature on PI, through the list of studies presented in the table, we seek to highlight the trends of available studies on psychosocial interventions for addictive disorders from 2000-2017. The purpose of this review was not so much to explore the effectiveness of individual PI, but rather to understand the trends in methodological approaches which have been used in this area of research. Some of the evident trends are:

A large proportion of studies carried out in this area from 2000- 2017 include randomised controlled trials (RCT), which is regarded as the ‘gold standard’ research design for establishing the effectiveness of an intervention. The randomisation strategies, however, differ across studies as do the number of comparison groups studied. Most studies have included adult subjects, with current substance use while some have included those with lifetime substance use and comorbidities as well. Assessment measures used include subjective as well as objective, standardised tools.

Studies are available in all areas of addiction, with an apparent dominance of PI in cannabis use. The specific psychosocial interventions for addiction studied, are: Cognitive Behavioural Therapy (CBT), Motivational interviewing/Motivational Enhancement Therapy (MET), Brief Intervention (BI), enhancing social support (SS) using cognitive and motivational methods, Mindfulness Meditation (MM), Contingency Management (CM), Relapse Prevention (RP), Drug Treatment Counselling (DTC) or drug related education (DC).

The most frequently used primary outcome measures include self reports (number of days of use or abstinence, rate of abstinence, frequency of use) that have been sometimes corroborated using objective methods, (including urine/hair/saliva analysis and other reports). Other outcome measures include (a) severity of addiction (e.g. using Addiction Severity Index, severity of dependence scale (SDS), and/or by counting symptoms indicating dependence); (b) level of related medical/psychiatric problems, legal problems, social and family relations, employment and support; (c) adherence to treatment, including average number of sessions received and/or proportion of participants undergoing all planned sessions; (d) abstinence/actual change (e.g. URICA); and (e) others like, death, or hospitalization.

Among the secondary outcomes, the available studies have looked into (a) motivation to change measured by objective methods (e.g. RCQ), (b) self reported frequency of use of other substances (i.e. number of days, times used/day, amount, etc.), (c) comorbid psychiatric illnesses (e.g. BDI, SANS, BPRS, SCL-90R), (d) overall functioning (using SOFAS, etc), legal issues, attendance at external drug treatment centres/sessions and (e) knowledge, attitudes and practices regarding drug abuse (KAPQ).

Additional variations in the methodology pertain to the lengths of treatment (short terms vs long term treatment i.e. 1- 14 weeks; 12-14 sessions; 90-120 minutes) as well as duration of follow up (1 month-6 years).

In general, the results of these studies (not shown in the table) collectively indicate that even with low-intensity PI, as opposed to controls receiving no PI, treatment recipients show certain positive changes2. Such a conclusion, however, is not unexpected and we must be wary of the apparent publication bias. More important conclusion to draw, however, is that robust research designs are essential to establish that a given PI does work. In other words, while theoretically a given PI may appear to be logical and effective, its widespread clinical usage in the real world would only follow after it has been established to be effective through a robustly designed research. To further buttress this point, in the following section, we trace the journey of a well-known PI, ‘Motivation enhancement therapy’.

Evolution of Motivation Enhancement Therapy (as a case study)

Motivational Enhancement Therapy (MET) is one of the most utilised client centered PI for alcohol use disorders. It is a time bound (4 session) therapy process that was first tested as one of the treatment methods for problem drinkers in the famous Project MATCH33. It uses motivational interviewing (MI) strategies that are merged into an organised therapeutic process that includes a comprehensive assessment of the client's behaviors as well as gathering adequate feedback about client's ambivalence and low motivation for change.

MET originated from Miller's observation that while delivering behavioural intervention to self-referred patients having substance use problems, the therapist's empathy plays a very important role. He conceptualised this intervention where the focus is on responding differentially to client's speech in an empathic, person-centered style, with special attention on evoking and strengthening person's own verbalized motivations for change34. Miller later found that his ideas were linked to Festinger's (1957)35 formulation of cognitive dissonance and Bem's (1967, 1972)3637 reformulation as self-perception theory. Rogers' theory38 of the “necessary and sufficient” interpersonal conditions for fostering change was also found to be relevant by him. Thus, a sound theoretical basis existed behind the introduction of a novel PI.

An analysis of trials of successful brief interventions in substance use disorder revealed six components, which were often present. These included feedback, emphasis on personal responsibility, advice, a menu of options, an empathic counselling style, and support for self-efficacy (acronym FRAMES). These were used to develop an instrument called “Drinker's check up (DCU)” which was initially thought to increase engagement of clients in treatment for alcohol problems. However, randomised controlled trials (RCTs) did not show any increased treatment seeking behaviour in the DCU group as compared to waiting list control. Interestingly, the DCU group showed reduction of drinking; it became apparent later that the DCU alone was responsible significant reduction of problem drinking behaviour.

Miller later combined motivational interviewing (MI) and assessment feedback and termed it “Motivational Enhancement Therapy”. Subsequently RCTs were conducted which showed that clients receiving MI had double the rates of total abstinence at 3–6 months when compared to those without MI39.

With time MI was claimed to be useful for multiple other conditions including cardiovascular rehabilitation, diabetes management, dietary change, hypertension, illicit drug use, infection risk reduction, management of chronic mental disorders besides problem drinking, problem gambling, smoking, and concomitant mental and substance use disorders!40 MI has more useful, synergistic effects when used as adjunct to an already active treatment4142. Thus, an intervention which was apparently grounded in sound theoretical basis and was found effective in small scale RCTs needed to be tested at a larger scale. The first multisite trial for motivational enhancement therapy was project MATCH conducted at 9 sites, with 1726 patients, where 4 sessions of MET was compared with 2 other therapies which had 12 sessions. Results of this trial showed a cost-effectiveness advantage for MET. Similar results were obtained in another 3 site trial conducted in the UK (United Kingdom Alcohol Treatment Trial) where MET was compared with 8 sessions of family involved behaviour therapy43. Later, the Clinical Trials Network of the U.S. National Institute on Drug Abuse undertook six multisite trials of MI and MET and compared these with treatment-as-usual for drug problems and dependence. Through all these trials it was established that MI-based interventions promoted sustained reductions in alcohol use and increased treatment retention.

Such strong evidence base was required for MI/MET to be included as a ‘recommended’ treatment in the guidelines, such as those by NICE and WHO-mhGAP (World Health Organisation-Mental Health Gap Action Programme)44.

Hence, this story of MET shows how an intervention was conceived, went through rigorous research trials, and only then was established as an intervention to be recommended by treatment guidelines.

Indian literature on research in psychosocial interventions has witnessed an increase recently. There are a number of review articles454647484950 that directly address the PI related methodological issues.

How to plan a research evaluation of psycho-social interventions

In the preceding section, we have sought to highlight that a vast amount of research is available on PI and it is only through establishment of strong evidence base that a PI could be recommended as a guideline. In the subsequent section, we present certain ‘guidelines' for conducting research in the area of PI. It may be noted that there cannot be a definite blue print for carrying out research on psychosocial interventions in addictive disorders. However, certain issues ought to be kept in mind while planning research in this area.

Research question

Researchers need to begin with asking some question to themselves. Is the area chosen for research a relevant one? Relevance would be understood as the need to address issues, to cover the lacunae in existing research, or to look for answers to yet unanswered questions. For instance, ‘Does mindfulness meditation work well in women with pentazocine dependence?’ The research question needs to be clear, focused and precise and must be eligible to be answered in a ‘yes' or ‘no’ fashion. ‘Is MET useful for those with chronic behavioural addictions, such as internet use?’ Questions such as these would be based on researchers' own understanding of the disorder as well as its management and on researcher's knowledge about existing evidence base. Replicating available studies and asking similar research questions about a similar population may also be worthwhile since it may help validate the findings of the earlier research in the new context.

Context of the study

The context of the study, the cultural influences, the client's as well as caregiver's existing knowledge, attitudes and practices may be important factors to consider while planning psychosocial intervention. For example, when planning a study on Children Of Substance Using Parents, it may be useful to look into familial, stressful factors such as parental substance use, interpersonal relationships, physical and mental health of the family members, availability of the parents/guardian, living arrangement, financial and legal issues, physical or sexual abuse. These aspects would need to be labeled as primary or secondary variables depending on the research question put forth.

Study design

It is essential to choose a study design which is most appropriate to answer the research question. Usually, it would be prudent to proceed in a step wise manner with incremental enhancement in intensity, complexity (and resources) of study designs. For instance, it may be appropriate to begin with an open label, single arm, feasibility study, at a smaller scale followed by a large-scale, randomised controlled trial.

Assessment

Next the researcher needs to know if the variables being studied are measurable. If yes, then are there relevant assessment tools available? i.e. are there any reliable, valid tools that capture the essence of the construct under study and can easily be understood by the sample under study and are quick to administer. If not, is there a need for one to develop some relevant tools?

Sample

Further, one would need to know if the research questions pertain to the sample in question. The sample under study must be representative of the population regarding which the questions are being asked. Unbiased, heterogenous samples are preferred for facilitating generalisation to the population at large. Similarly, using multisite data as against single site data would likely take care of the cultural effects on substance abuse. On the other hand, loss of homogeneity of the sample introduces complexity in the analysis and interpretation of the data.

Selection bias

Selection bias is a possibility if inclusion and exclusion criteria are not well defined. Matching of study and control groups, in terms of demographic characteristics and illness related factors is important to study the exclusive effects of the intervention. Random allocation is an effective way to avoid introducing the selection bias. The methods for randomisation could range from simple random number tables to using sealed envelopes to generating complex, computerised randomisation schedules.

Method of allocation

The method of allocation of treatments must be clearly specified. Concealment of treatment arms from participants is important as participants own perception about which treatment they have received may affect the outcomes. However, it may be a challenge to effectively hide the treatment method from the subjects (i.e. to ensure ‘blinding), especially in case of PI. Allocation concealment if not properly executed may lead to erroneous conclusions such as under or over estimation of treatment effects in different studies.

Control group

The kind of control group chosen will also impact treatment outcomes. Control group may be chosen with or without active treatments, and the findings must be interpreted in that light. The placebo effect of any PI can be sizable and hence it is important that the control group also receives an intervention similar to the one studied, though minus the active ingredients (‘sham treatment’).

Assessment tools

Reliable, valid, quick-to-administer tools that adequately tap the construct under study and can also help in assessing change over time may be applied pre, post, and end of treatment junctures. These tools may be subjective, objective or a battery including both types. It is also essential to carefully define the outcome measures of the study. For instance, an intervention aimed at reduction of craving, is best adjudged by a change in craving pre and post intervention (rather than studying some distant parameters like quality of life).

Intervention

Because of the way psychosocial interventions are designed, blinding may be difficult for some such interventions. The interventions may be tailor-made vs manualised/structured, that are delivered over a short term (single session BI to four session MET) to long term CBT sessions delivered over 12-16 weeks. Some interventions may require several booster sessions and follow up over years. If more than a single therapist is involved in delivery of the intervention, a manualised intervention, in which all the therapists have been adequately trained, is highly recommended.

Stand-alone interventions versus adjunct intervention

Research on PI is made more complex by the dilemma regarding whether a particular PI under study, should be offered as a stand-alone intervention or whether it should be used an adjunct intervention to an already existing ‘treatment as usual’. In general, this issue is solved by answering the question, ‘whether an effective treatment (usually pharmacological) exists, which is affordable and easily available in routine?’ If the answer is ‘yes’, it may be unethical to deprive the patients of an already available effective treatment and hence the PI should be offered as a adjunct treatment.

Treatment fidelity

Treatment fidelity or compliance may be assessed through video recordings or checklists and frequent feedback sessions with the therapists. Not paying attention to this issue, may result in under-estimation of the treatment effect.

RESULTS

Bias in reporting results can be reduced by honestly informing about the results for and against the hypothesis, whatsoever, and the results must be interpreted in the light of the factors affecting the overall outcome, including moderators, mediators, and confounding variables.

Generalisation

Lack of heterogeneous samples in individual studies and homogenous studies in reviews of various studies limits generalisation of results and should be kept in mind at the outset.

Ethical aspects

Issues such as neglecting the informed consent, inability to maintain confidentiality, inadequate efforts towards preventing boundary violations, as well as lack of genuine assessment of risk-benefit issues may be some of potential ethical problems in PI research.

Challenges in the area

Research in the area of PI for addictions may be professionally rewarding though is fraught with a variety of challenges.

Resources: In general, it is difficult to find financial and logistic resources for conducting research on PI. Commercial considerations make it far easier to obtain funding for research on pharmacological treatments which is not the case with psychological ones. In addition, there is a dearth of trained and qualified human resources in the country even for clinical care using PI, which makes it difficult to find human resources for conducting research.

Sample and sample size: Ideally the sample size of any effectiveness study, should be statistically determined to ensure that the study has adequate ‘power’. More often than not, in settings like ours, lack of resources makes it difficult to achieve the desired sample size and hence many researchers are forced to rely upon a sample size which is feasible for them to recruit and study. Another challenge pertains to the homogeneity of the sample. More heterogenous sample which mimics the profile of patients routinely seen in the clinical practice adds to generalisability of the results but, on the other hand, adds some potentially confounding variables making analysis and interpretation complex.

Areas in which research is needed:

While the field of PI has made tremendous progress globally, in last few decades, most of the PI in practice are those which are based on western models and the evidence base on their effectiveness is also largely from the western countries. Admittedly, this holds true for almost all kind of treatment interventions (pharmacological and psychological alike), but in case of PI, demonstration of applicability and suitability of the interventions in the Indian cultural context is important. Thus, almost all PI – even those which are widely recommended globally – are worthy candidates for research on their applicability in the Indian clinical context. Additionally, some of the specific areas which have been under-researched in India are, special population groups like women and adolescents with substance use problems; people with comorbid substance use and mental disorders and feasibility of PI in non-specialist settings. For most of the PI it is only the short term impacts which have been studied and their long term outcome remains to be established. Operational research involving studying unique Indian models of treatment delivery and organization of services (considering our resource constraints) is yet another important area of research that may be explored by future studies.

CONCLUSION

Research on PI is a challenging, yet worthwhile area which could be professionally very rewarding, considering the huge gap in the existing knowledge base and the meagre availability of resources. Younger mental health professionals must seriously consider taking up this challenge and seek to fill these knowledge gaps through planning and conducting research in this area. Stalwarts in the field have an obligation to encourage and nurture the younger generation to stimulate research as well as facilitate access to the resources.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.