O'Brien outlines the main changes proposed for classification of substance use disorders in DSM-V, based on an extensive consultation process and a review of the available evidence [1]. In view of DSMs wide international use, particularly for research, the proposed changes may have an impact on clinical, research and educational practices in many parts of the world. It is also important to consider potential implications of these proposals for the development of the relevant sections of the next version of the WHO's International Classification of Diseases and Related Health Problems (ICD), which remains the global standard for health information and reporting and the estimation of disease burden.

The World Health Organization (WHO) is currently working on the revision of the ICD-10 [2]. The eleventh revision of the ICD (ICD-11) is scheduled for submission to the World Health Assembly for approval in 2014. Development of the ICD-11 classification of mental and behavioral disorders is being led by the WHO Department of Mental Health and Substance Abuse, and includes the revision of the diagnostic classification of mental disorders for both mental health specialists [3] and for use in primary care settings [4]. The ICD-11 development process includes consultations with Work Groups involved in the development of DSM-V [5].

Without pre-empting the conclusions of the ICD-11 development process, we offer some reflections on the proposed DSM-V changes, in order to stimulate scientific debate on these issues.

From the WHO's perspective, the main purpose of the international classification is to serve as an international standard for the systematic recording, analysis, interpretation and comparison of mortality and morbidity data collected in different countries or areas and at different times [2]. In addition, the classification should have high clinical utility to facilitate effective treatment by a wide range of health professionals across specialty and primary care settings worldwide [6]. In order for ICD-11 to serve the WHO's public health mission, the classification should contain any disease or specific health condition entity that is of public health importance or that has a significant impact on individual and population health and whose accurate identification can lead to more effective utilization of health services and interventions.

Globally, alcohol, tobacco, illicit drugs and other psychoactive substance use have an enormous impact on population health. According to the latest WHO estimates, alcohol use is responsible for 4.5% of total disease burden, tobacco use for 3.7% and illicit drug use for 0.9% of the global burden of disease [7]. A significant portion of alcohol- and drug-attributable disease burden is related to injuries, for example, in the case of road traffic accidents due to driving under the influence of psychoactive substances, and to other health conditions that may be caused or adversely affected by substance use in the absence of a ‘substance dependence’ syndrome. In spite of the obvious public health relevance of substance use in such cases, appropriate recording of the role of alcohol, tobacco and drug use continues to be a significant challenge in the practical implementation of current classification systems. From the standpoint of public health utility, this almost certainly leads to an underestimation of the true disease burden attributable to substance use and the impact of substance use and associated disorders on health care systems.

From the standpoint of clinical utility, particularly in primary health care and other non-specialized health care settings, a core purpose of the classification is to provide a basis for guidance on the most effective interventions for a given diagnosis and therefore to improve access to effective treatments. This raises an important issue in relation to the DSM-V proposal described by O'Brien to eliminate the abuse/dependence distinction and to replace it with a single category of ‘substance use disorder’. While there may be data to suggest that distinction represents an artificial dichotomization of a continuum or spectrum, the distinction between abuse or hazardous and harmful use and dependence syndrome has been a clinically useful one, more of less separating those with patterns of substance use behaviours that respond to brief psychological interventions from those needing more substantial treatment, including pharmacotherapy, such as detoxification and relapse prevention, or agonist maintenance treatment. The global public health importance of agonist pharmacotherapy for opioid and tobacco dependence is reflected in the inclusion of opioid and nicotine agonists in the WHO Model List of Essential Medicines [8]. How the proposed changes in classification would affect the effectiveness and safety of pharmacotherapy for substance dependence, particularly in non-specialty settings, requires further exploration. For example, what would be the criteria for commencing opioid agonist pharmacotherapy? Moreover, the proposed broad diagnostic category of ‘substance use disorder’ in some jurisdictions may have unforeseen consequences by increasing potential of stigmatization and administrative or even legal sanctions.

One of the key points in O'Brien's article is related to the nomenclature and semantics with a clear direction to avoid the use of the concept ‘dependence syndrome’ and to separate as much as possible the phenomena of ‘physical dependence’ and ‘addiction’. These issues have plagued the field for decades. Ironically, in 1963 the WHO Expert Committee on Addiction-Producing Drugs (now called the WHO Expert Committee on Drug Dependence) recommended substitution of the term ‘drug dependence’ for the terms ‘drug addiction’ and ‘drug habituation’ and emphasized that the term ‘drug dependence’ is ‘selected for its applicability to all types of drug abuse and carries no connotation of the degree of risk to public health or need for a particular type of drug control’[9].

O'Brien particularly emphasizes the importance of the division of ‘physical dependence’ and ‘addiction’ in the context of pain management with opioids to avoid under-treatment of pain because of exaggerated concerns about opioid dependence. Indeed, inadequate pain management and poor availability of opioid analgesics is a public health problem of a significant proportion, and any concerns about risk of ‘addiction’ should not be used as a justification for limiting access to effective pain management, particularly in the context of cancer treatment and palliative care. But we are less clear that the current nomenclature actually contributes significantly to this problem.

When ‘dependence syndrome’ was described at the 16th report of the WHO Expert Committee on Drug Dependence in 1969 [10], its definition clearly indicated that tolerance may not be present, and that withdrawal symptoms were neither necessary nor sufficient for diagnosis of dependence syndrome. The description of dependence syndrome in ICD-10 provides a specific example of a surgical patient treated with opioids who shows increased tolerance and a withdrawal state, and specifically indicates that such a person should not be diagnosed with dependence in the absence of other key symptoms of the dependence syndrome [3]. On the other hand, not counting tolerance and withdrawal towards a diagnosis of ‘substance use disorder’ (SUD), as proposed for DSM-V, would raise the threshold for the diagnosis of severe SUD among people ‘addicted’ to prescription medicines, particularly in cases where the chosen method of procuring the drug of choice is to spend time in health care settings and doctors' surgeries.

We agree that development of an opioid dependence syndrome in cancer pain management and palliative care is a rare phenomenon that, even if it occurs, has no significant public health impact. On the other hand, non-medical use and misuse of pharmaceuticals, particularly opioids, has become a significant public health problem, mainly documented in developed countries [11, 12]. It is still debatable whether the problems emphasized by O'Brien really require changes in classification. There are other strategies than changing nomenclature to promote adequate pain management in specific populations, and these may be more parsimoniously addressed by more targeted solutions, for example through enhanced training of health professionals, removing unnecessary legislative and administrative barriers or changes in the functioning of specific parts of health care systems. We would urge extreme caution that classification changes intended to address these problems do not inadvertently compromise the clarity with which substance use disorders, particularly the dependence syndrome, can be identified and targeted for intervention or the accurate estimation of substance-related disease burden at the population level.

An international classification of diseases seeks to provide for each morbid entity a single recommended name that is specific, unambiguous, self-descriptive, simple, and, to the extent possible, based on causality [3]. Abandoning ‘dependence syndrome’ for ‘substance use disorder’—or, potentially, ‘addiction’—may bring certain advantages, but at the same time may compromise basic requirements of an international nomenclature. For example, the translatability of diagnostic terms into other languages is a considerable challenge that can influence the choice of diagnostic terms such as ‘substance use disorder’ or ‘addiction’ at an international level. Moreover, changing terms such as ‘substance use disorder’ and ‘dependence syndrome’ after many years of use in research, statistical, clinical and educational applications will have substantial cost implications and could potentially compromise data comparability across versions of the ICD. Specific efforts to ensure a smooth transition would be required if these proposals were finally accepted at an international level, and WHO member states will demand persuasive evidence of the importance of the change to justify this investment.

In the substance use area, we see the key challenge for developing classification systems as being to characterize the patterns and episodes of substance use that have significant importance for delivery of effective prevention and treatment services in a way that promotes reliable monitoring of substance-related health conditions in health care. O'Brien devotes no attention to this issue in his article, and the extent to which the current DSM-V proposals address individual treatment needs and public health challenges related to substance use remains unclear.

Declaration of interests

The authors are involved in the development of the eleventh revision of WHO's International Classification of Diseases and Related Health Problems.

The views expressed in this paper are solely the responsibility of the named authors and do not necessarily reflect the decisions or stated policy of the World Health Organization or its Member States.