Commentary and explanation - Take­home naloxone and the prevention of deaths from opiate overdose: two pilot schemes

David Ogilvie explains what the pilot schemes achieved

The authors of this paper report on two pilot schemes in which naloxone was distributed to drug users. Where does a pilot scheme fit in the development of a new idea like this? You can deduce from the paper and the reference list how the work has evolved. Firstly, someone proposed that take­home naloxone would make sense.¹ This led to a study to explore what impact take­home naloxone might have, and how acceptable it would be.⁴ It turned out that drug users supported the idea, so pilot schemes were set up - relatively small projects to test out in practice what seemed like a good idea in theory. The results of the pilot schemes, reported in this paper, seem encouraging so the authors conclude by recommending a larger scale study.

What happened in the pilots?

In both projects, naloxone was given out to drug users along with training on how to manage an overdose, and recipients were asked to report back any instances when they used the naloxone.

Two sorts of results are given. The main text gives details such as numbers of people, numbers of events, and proportions. These tell us, for example, how many times naloxone was used, what proportion of people were given what dose, and how often an ambulance was called. This is quantitative information - the sort that appeals most readily to people trained to look at things in a biomedical way.

However, the numbers do not tell us everything.We are also interested to know what it was actually like for the people involved. This qualitative information is given in the boxes, which contain descriptions of what happened, how people felt, and why certain decisions were made.

Although quantitative and qualitative research are sometimes thought of as separate entities-with their respective roots in the biomedical sciences and the social sciences - it is clear that both types of analysis can be useful. For example, we can count how many times naloxone was given intramuscularly (quantitative), but we may not understand why unless we read accounts such as that in case 1 (qualitative). If, as the authors suggest, family members as well as injecting drug users are to be offered naloxone we need this type of insight into how people feel about giving injections.

So what's next?

It sounds as though some lives were saved in these pilot schemes. Can we now recommend the widespread use of take­home naloxone? Although the results of the pilot schemes are encouraging, they leave some questions unanswered, and it is worth thinking about how you might deal with them if you were designing the further study recommended by the authors. The key question is, "Do we really know that the naloxone saved lives?" Perhaps all the people who were given naloxone in the pilot schemes would have survived anyway if someone had just called an ambulance. One way of looking at this might be to carry out a controlled trial, distributing naloxone in some areas and not in others, and then seeing whether the risk of death was reduced in areas where naloxone was given out.

Other questions worth thinking about

• Is there something special about Berlin and Jersey? Are the drug users, health care workers, or general populations unusual in those places? Would the same results be obtained in, say, Glasgow? If we want to apply the findings of research in our own area we need to be confident that the findings are generalisable.
• Most drug users who were given naloxone did not report back. Why is that? Had they just not used the naloxone? Had they had some sort of problem with it? No adverse consequences were reported, but is that because not everyone reported back? It is always possible that the people who do not reply to a survey, or do not volunteer for a study, are different from the people who do, in which case the findings of the study might be biased.
• Even if take­home naloxone is effective in preventing deaths should we spend our money on it? Or could we save more lives by doing something else? The paper contains an attempt to estimate the cost of naloxone per life saved. You might choose to do a more detailed economic evaluation, taking into account other costs (such as staff time), and comparing take­home naloxone with other activities which might improve the health of drug users.