Clinical Outcome Measures in Addiction Psychiatry

What domains should be assessed?

One of the primary concerns is about what domains should be considered as an outcome measure. Substance use can be frequent or infrequent and may variably affect different aspects of life. The prime consideration here would be whether the substance use frequency or amount be the sole consideration as an outcome or should the outcomes be more comprehensive (for example, looking at the dimensions of impact on the family or employment).^[4,9,10] Some of these approaches are exemplified in Table 1.

Looking at only substance use may give an immediate idea about whether the person is currently using the substance or is abstinent. Getting information about whether substance use is happening provides some clarity about whether substance use continues to exist. In many substance-dependent individuals, intake of substances (or re-initiation) results in consequent problems in health and life. Substance use can be measured by frequency of use (like daily, few times a week, few times a month, etc.), number of days of use in the last month, number of days of heavy substance use (for example, heavy drinking days), or average consumptions. The project MATCH (Matching Alcoholism Treatments to Client Heterogeneity) used alcohol use and abstinence as an outcome measure.^[22] The advantage of measurement in this manner is the simplicity of approach and discerning the relative frequency of each substance of use.

However, substance use per-se may not reflect that it is being used in a problematic manner. The problems caused by substance use in an individual’s life can be assessed in several domains. A tenet of a psychiatric disorder is the causation of dysfunction in social, occupational, or other aspects of life. Hence, the demonstration of the presence and trajectory of the substance use disorder would be to assess the impact due to substances (and each individual substance if possible) in specific domains of life – like social and family life, employment, physical and psychological health, issues with law enforcement, and others. ASI has been used as an outcome measure in opioid substitution treatment studies in both developed and developing countries.^[13,23]

Another approach to ascertain the outcome is to look at ancillary measures. Quality of life measures can be used as an alternative to assess the overall quality of life and how it changes with time and/or intervention. This may be the more meaningful outcome and may enable cross-illness comparisons. Previous studies have looked at the quality of life changes in persons with substance use disorders.^[13,24,25] Broad assessment of functioning may be assessed using the Global Assessment of Functioning, though this measure may be susceptible to significant subjectivity. Costs incurred on substance use may be of value for cost studies (they may be part of the intervention framework), and this may enable decisions to be made by the policy-makers. Researchers have assessed the cost related to substance use disorder interventions in their previous work on substance use disorders.^[26,27]

One more way to look at the outcomes is in terms of recovery and assessment of recovery capital. Interventions, if they are effective, would lead to not only substance use cessation but is also likely to lead to recovery of the individual.^[28] Substance use disorders are often considered as having relapsing and remitting courses, but recovery is often what individuals with substance use disorders aim for. Assessment of recovery capital can help individuals to know what resources they have (as per their appraisal) to cease substance use and bring substantive changes in their lives to lead substance-free or substance-limited lifestyles.^[29] Recovery has been considered to be multi-faceted and challenging to define,^[30,31] though studies have used recovery as a marker of outcomes of individuals with substance related problems.^[32,33]

There are several scales developed for directed inquiry about specific substances. Scales like the Fagerström Test for Nicotine Dependence and the Severity of Alcohol Dependence Questionnaire (SADQ) can be used to assess the changes in the severity of specific substance use disorders with time. These scales often cater to specific substances and are able to discern the severity of changes with specific interventions. Previous interventional studies have used them as outcome measures.^[34,35] The presence or absence of a diagnosis (including substance use disorder diagnosis) can be done using diagnostic instruments. They can be used to ascertain whether a diagnosis has reached remission within a period of time. Ecological momentary assessment (EMA) is another set of methods that enable “real-time” assessment of problems.^[36] Often, one or two questions are asked several times a day or in different vulnerable situations to get an understanding of the present state of craving, consumption, or other parameters.

Another aspect of assessment may be the use of biological samples. The use of biological samples (urine, blood, saliva, hairs, breath, dried urine spots, etc.) offers objective and verifiable estimation of substance use.^[3,37,38] This can help address any potential deviation from the actual pattern of substance use while gathering information from the patients. These biochemical parameters may have individual windows of detection and may require variable degrees of sophisticated equipment, based upon the precision of assessment required. Yet, they often are built-in as outcome parameters in trials individually or as a combination with other measures.^[3] Previous literature has used reports from assessment from urine drug screens.^[39]

There can be additionally yet another set of ancillary measures that can be used as outcome parameters.^[40] These can include assessment of medication adherence, homelessness, fidelity to treatment or therapeutic approaches, patient satisfaction, engagement with treatment, and physical and mental health of the individual.

From whom should we ascertain the outcomes?

One of the important considerations is whether the outcomes should be reported by the person concerned (the individual who has been using the substances), should be rated by clinicians based upon the global information from family and acquaintances, or should be objective parameters and blood investigations. The use of additional sources of information, including laboratory tests, can help add to the quality and accuracy of information. Discrepancies have been noted between self-reports and laboratory-verified results.^[41–43] The question may arise then which information to rely upon in such circumstances. There can be several approaches in such circumstances:

1. Rely more on the objective ascertained information/ancillary informants/biological samples: This may negate the effect of the amnestic recall of the use of substances and motivated answers (in order to please or avoid negative consequences). One of the limitations of ancillary informants is that they may have limited knowledge of the extent of substance use (especially if the substance use is sporadic). The limits of biological samples are that they may have a “window” of detection and require laboratory facilities and proper handling of the specimen.

2. Rely more on the self-reported outcomes: This has the advantage of ease of assessment.

3. Take multiple outcome measures and report outcomes individually: This may demonstrate discrepancies in the different outcome measures and churn out the divergence, if so applicable. This may also be able to demonstrate if improvements or changes occur differentially in different domains with time and/or intervention.

An approach to decide which measures to use is to make the best judgment call based on context and scenario. Such an approach considers the context where the outcome measure is to be applied. For example, in prison settings, objective biological measures of assessments may be more reliable than self-reported measures. In school settings, self-report measures may be fairly accurate.

Local and cultural adaptations

Outcome measures have a utility value only if they are usable in the local context.^[44,45] Many of the questionnaire-based outcome measures used have been translated and validated in different languages (for example, ASI and Alcohol Use Disorder Identification Test [AUDIT]). AUDIT has been validated in several languages, including Chinese, Korean, and French.^[46–48] Validation of scales and instruments is a resource-intensive process, and there are several steps/aspects in the validation process.^[49] When used in a different cultural setting, there are two main considerations that apply. The first one is about translation and finding semantic equivalences of the words. One may find challenges in finding the right words to convey the meaning in the other language where the scale is intended to be applied. Semantic (meaning) equivalence is desired more than literal translation. The second consideration is the applicability and need for modification based on cultural realities. Some instruments used to assess the severity of substance use disorder have focused on specific items (for example, ownership of an automobile) for scoring. The same might not be applicable in other cultural contexts. Thus, applications of instruments and questionnaires as outcome measures need to be cognizant of the cultural issues. Adaptations, modifications, and re-validations may be necessary for certain specific instruments.

Purported uses of outcome measures

Outcome measures can have several uses in clinical practice.^[50,51] The use of outcome measures can be several in patients with addictive disorders [Table 2]. We can divide the purported use of outcome measures into two broad groups: clinical and research and administrative and policy-related. We would like to emphasize that these groups are not mutually exclusive and may be used concurrently in some situations. Still, the clinical and research applications include ascertainment of efficacy or effectiveness in intervention trials; discerning naturalistic course and progression of addictive disorders documentation in the usual clinical practice; and providing usable measures to facilitate cross-cultural and cross-regional assessments. From an administrative and policy perspective, the outcome measures can be used for assessing the quality of care and assessing changes in outcomes that occur with administrative and policy changes. Aligned with this is the utility of auditing (for insurance and other purposes), mapping service characteristics over geographies and over time. In addition, outcome measures can have value when used as a contingency in law enforcement approaches (for example, the individual diverted from the criminal justice system to treatment needing to show continued abstinence and improvement in psychosocial functioning). The kind of outcome measures that should be used for these parameters would need considerations of feasibility, utility, reliability, and validity as well.

When and how frequently to assess outcome measures?

Some of the outcome measures (including self-reported outcome measures and biological measures) have a time frame ascribed to their assessment. For example, the Maudsley Addiction Profile largely looks at the last 30 days; visual analogue scoring of the EMA is based upon here-and-now; while many urine cassette tests look at the last 72 hours of drug consumption. Many clinical trials have built-in urine drug screening every week to find out whether drug use occurred during the time period. Similarly, longitudinal studies may have monthly or even less frequent assessments of substance use and assessment of other assessment parameters. Thus, the frequency of assessment of outcome measures would be dependent upon the method used for the outcome assessment the resources available for outcome assessment and the necessity and utility of the outcome assessment.

One important aspect while implementing outcome assessment is the cost. Cost is incurred in not only in monetary terms (cost of equipment, kits, papers, pens, digital devices, etc.) but also in terms of the time of the administrator, time of the respondent, making additional arrangements, housing of the records, and so on. Thus, one needs to have a clear reason for outcome assessments, balancing the needs, economy of expendable resources, and future benefits from the outcomes assessed.

Attention to validity and reliability

Reliability and validity of the outcomes is yet another aspect to be considered while choosing the outcome measures. While many treatises and reviews are available on the importance and varied aspects of reliability and validity measures,^[52,53] we would like to draw attention to some of the working nuances of these parameters. One of the important aspects is assessing the face validity – whether the method ascertains what it intends to ascertain. In addition, content validity and convergent validity should be good enough with the measure being used. Attention is also drawn to inter-rater reliability, as instruments with poor inter-rater reliability are likely to give different results in the hands of different assessors. Similarly, a fair to good test–retest reliability suggests that the instrument performs well and much confidence can be expended on the inferences. Known reliability and validity parameters make data from the outcome parameters easier to compare to what is already known. Nonetheless, assessment measures for which reliability and validity parameters have not been hitherto established can also serve as useful outcome measures.