The primary purposes of our original paper were to (i) describe secondary exchange and provision of multiple services as the principles that have guided expansion of syringe exchange in the United States; and (ii) pose the question of whether further expansion of harm reduction services should occur within a ‘separate’ health-care system for drug users or within the ‘mainstream’ health-care system [1]. The three responses provide perspectives on harm reduction from different areas, but despite their differences all three note that the question of ‘separate’ health services for drug users versus integration of services for drug users into ‘mainstream’ health care is highly relevant [2–4].

Grund [2] responds from the perspective of the European Union (EU) rather than a single country. He notes that the EU has strongly supported the concept of harm reduction and has linked it to the consideration of human rights for drug users. He also notes, however, that most harm reduction programs in the EU have been ‘separate’ programs, not integrated into ‘mainstream’ health services.

Carrieri and Desenclos [3] note that France began a ‘combined’ harm reduction program—syringe exchange, opiate substitution therapy (OST) and anti-retroviral treatment (ART)—in the mid-1990s. There are several notable aspects of the French system: the diversity of sources for free sterile injection equipment (from syringe exchanges, kits from pharmacies and from vending machines), and the ‘mainstream’ provision of buprenorphine through general practitioners rather than through ‘separate’ OST clinics.

Maher & Iversen [4] describe the Australian needle and syringe program (NSP) system. It is a large national system with approximately 30 million syringes exchanged per year. The system was implemented when human immunodeficiency virus (HIV) prevalence was low among injecting drug users (IDUs) throughout the country and has prevented an estimated 25 000 HIV infections among IDUs.

With respect to ‘separate’ versus ‘mainstream’ provision of health services for IDUs, Maher and Iversen state that Australia is close to the ‘mainstreaming’ end of the continuum, with national health insurance that includes IDUs. They do note, however, that this mainstreaming is ‘at least in principle, if not always in practice’.

When applied on a public health scale, harm reduction programs for IDUs have had many notable successes in preventing HIV transmission among IDUs. Harm reduction practitioners are now addressing a much wider variety of health issues among drug users, including hepatitis C virus (HCV) prevention, treatment for overdoses and primary medical care. The need to provide such services on a public health scale and in a cost-effective manner will force some difficult choices as to whether the services should be provided in separate facilities for drug users or through mainstream health-care facilities. Indeed, separate versus mainstream may be the central question for sustainable expansion of harm reduction services for drug users. There are strong arguments for both separate and for mainstream service provision, and it is unlikely that there will be a single best answer for all services in all countries. While recognizing the diversity in national policies towards drug users and differences in national health-care systems, this may be an area in which continued cross-national research may be particularly helpful.

Declaration of interest