Morten Hesse, Centre for Alcohol and Drug Research, Aarhus University
17 September 2004
Harm reduction is an important issue, and delivering effective and relevant services to drug users is likely to benefit both society and drug users.
Unfortunately, as harm reduction is also a controversial approach, its advocates should take great care to make only justifiable assertions about the effects of harm reduction interventions.
Methodological limitations of the study warrant serious concerns, and the conclusions made based on the data are not supported by the data.
In the conclusion of the paper it is stated that "[the study] provides positive results that associate harm reduction programming with incremental and lifesustaining changes in drug users lives.
However, what the data seem to show is that the entire group improves, with no association between services and improvement. This, actually negative, result, is however not even supported in the data, given the methodological limitations of the study.
First of all, a serious limitation is the sample. The sample is basically a cross-sectional sample, followed up at one point. But the retrospective nature of the baseline measurement is very troubling indeed. Also, it is no surprise that people come to services when they are at their worst, and that a decline in the average level of problems should be seen at any point thereafter (since addiction is a fluctuating, chronic condition).
Related to this issue is the fact that people's needs determine the services they request, and therefore a lack of association between services and improvement may indicate little more than the fact that people who receive more medical services in the general community have higher morbidity and mortality than age-matched people who receive less services.
Another limitation pertains to the development of the measurement instrument. In the article entitled "How do drug users define their progress in harm reduction programs? Qualitative research to develop user-generated outcomes", the authors note several problems pointed out by focus groups in the instruments, but it is not clear from either article whether or how these objections were taken into consideration when developing the final version of the instrument. Indeed, many of these objections seem relevant. Also, basic psychometric properties of the instrument, such as test-retest reliability or interjudge reliability are not reported.
Although, for obvious reasons, it is not possible to conduct a randomized clinical trial of low-threshold services, there are still better ways to deal with the effectiveness issue.
One approach is quasi-experimental studies, where services are implemented within a population, and changes in relevant problem areas are measured at the population level
Such studies have primarily focused on needle-exchange, but could easily be conducted in other areas (e.g., implementing a specific type of health outreach services, and measuring health outcomes over time, or implementing injection education progams that teach substance abusers how to inject as safely as possible, and measuring the prevalence of adverse reactions to injection, such as abcesses and blood poisoning).
There is much to do to study harm reduction, and the journal is a welcome contribution to this field. The authors of this study, while drawing conclusions that are not warranted in the data, nevertheless should be congratulated for raising many of the methodological issues, including how to deal with open access treatment, how to measure outcome for services that are based on clients' preferences rather than system objectives, and ways to integrate qualitative and quantitative research.
Harm Reduction Journal
The 'effectiveness' of harm reduction
17 September 2004
Harm reduction is an important issue, and delivering effective and relevant services to drug users is likely to benefit both society and drug users.
Unfortunately, as harm reduction is also a controversial approach, its advocates should take great care to make only justifiable assertions about the effects of harm reduction interventions.
Methodological limitations of the study warrant serious concerns, and the conclusions made based on the data are not supported by the data.
In the conclusion of the paper it is stated that "[the study] provides positive results that associate harm reduction programming with incremental and lifesustaining changes in drug users lives.
However, what the data seem to show is that the entire group improves, with no association between services and improvement. This, actually negative, result, is however not even supported in the data, given the methodological limitations of the study.
First of all, a serious limitation is the sample. The sample is basically a cross-sectional sample, followed up at one point. But the retrospective nature of the baseline measurement is very troubling indeed. Also, it is no surprise that people come to services when they are at their worst, and that a decline in the average level of problems should be seen at any point thereafter (since addiction is a fluctuating, chronic condition).
Related to this issue is the fact that people's needs determine the services they request, and therefore a lack of association between services and improvement may indicate little more than the fact that people who receive more medical services in the general community have higher morbidity and mortality than age-matched people who receive less services.
Another limitation pertains to the development of the measurement instrument. In the article entitled "How do drug users define their progress in harm reduction programs? Qualitative research to develop user-generated outcomes", the authors note several problems pointed out by focus groups in the instruments, but it is not clear from either article whether or how these objections were taken into consideration when developing the final version of the instrument. Indeed, many of these objections seem relevant. Also, basic psychometric properties of the instrument, such as test-retest reliability or interjudge reliability are not reported.
Although, for obvious reasons, it is not possible to conduct a randomized clinical trial of low-threshold services, there are still better ways to deal with the effectiveness issue.
One approach is quasi-experimental studies, where services are implemented within a population, and changes in relevant problem areas are measured at the population level
Such studies have primarily focused on needle-exchange, but could easily be conducted in other areas (e.g., implementing a specific type of health outreach services, and measuring health outcomes over time, or implementing injection education progams that teach substance abusers how to inject as safely as possible, and measuring the prevalence of adverse reactions to injection, such as abcesses and blood poisoning).
There is much to do to study harm reduction, and the journal is a welcome contribution to this field. The authors of this study, while drawing conclusions that are not warranted in the data, nevertheless should be congratulated for raising many of the methodological issues, including how to deal with open access treatment, how to measure outcome for services that are based on clients' preferences rather than system objectives, and ways to integrate qualitative and quantitative research.
Competing interests
None to declare