• Open Access

A review of the use of US-derived aetiological fractions in an Australian setting for antenatal problems related to cocaine use


Correspondence to:
Marian Shanahan, National Drug and Alcohol Research Centre, University of New South Wales NSW 2052. Fax: (02) 9385-0222; e-mail: m.shanahan@unsw.edu.au


Aetiological fractions are often used as an indirect measure of morbidity and mortality related to a specific risk factor. Aetiological fractions previously used in Australia for cocaine-related antenatal haemorrhage and low birth weight newborns have relied on risk ratios calculated from US-based studies. As outlined in this paper, there are several differences in the use and prevalence of cocaine and its associated harms between the two nations. As such, it is recommended that any use of these aetiological fractions with Australian data should occur with caution.

Aetiological fractions (AFs), also known as an attributable proportion, or attributable risk, are often used to estimate mortality and morbidity attributable to illicit drug use. AFs are a form of indirect quantification of morbidity and mortality due to a specified risk factor.1 The use of AFs requires an established causal link between a particular disease and the risk factor of interest; the AFs therefore also allows for estimates of disease in a specific population that could be avoided in the absence of the risk factor.2 Estimates of AFs in Australia for illicit drug use as the risk factor have been calculated1 and then later revised,2 however, due to a lack of Australian studies, the AFs calculated were dependent upon risk ratios (RR) obtained from studies conducted in the United States.

This paper provides an illustration of potential overestimation of attributable cases when the RR are obtained from populations where the underlying use of drugs, and the extent of associated harm, differ from those in the population for which the AFs are being applied. As part of a study to estimate the costs of drug related hospital separations in Australia,3 AFs, as derived by English et al.1 and revised by Ridolfo and Stevenson2 were used to estimate the number of drug-attributable cases. Table 1 below indicates the applicable AFs for opiate- and cocaine-attributable antepartum haemorrhage and low birth weight newborns. It is apparent from the table that if these fractions were multiplied by the number of cases with the relevant principal diagnoses, the number of ‘attributable’ cocaine-related separations would be greater than ‘attributable’ opioid-related separations for both ante natal haemorrhage and low birth weight newborns.

Table 1.  AFs for drug ‘attributable’ principal diagnoses.
Principal DiagnosisDrug TypeAetiological Fraction (population)
  1. Notes:

  2.    Source2

Antepartum haemorrhageOpiates0.013
Low birth weight newbornsOpiates0.022

As previously pointed out by Chikritzhs et al.4 when comparing alcohol-caused mortality and morbidity over time, the calculation of an AFs requires two fundamental pieces of information — the relative risk, which is a measure of the causal relationship between exposure to the risky drug and the condition being studied, and the prevalence or the proportion of the relevant population engaging in the risky activity.

Using risk ratios previously developed,1 Ridolfo and Stevenson2 updated the AFs for antenatal cocaine use. The original risk ratios were calculated using data from studies primarily conducted in the United States (for a complete list of references used to calculate these risk ratios see English et al.1). There are however, clear differences between levels of exposure to cocaine use in the Australian and US populations. The cocaine markets within the US and Australia are vastly different. The US has consistently had greater cocaine availability (at a higher purity and lower price) than Australia.5 Reflecting this lower level of availability, the prevalence of cocaine use in Australia is considerably lower than in the US: in 2005 in the US, 5.5% of females aged between 18 and 25 years had recently used cocaine and 0.9% had recently used ‘crack’.6 In comparison, in 2004 2.3% of Australian females aged between 20 and 29 years had recently used cocaine.7

Examining other data on harm, it is clear that harms related to opiate use in Australia far outnumber those attributable to cocaine use whether measured by drug related deaths,8,9 arrests,10 treatment numbers10,11 or hospital separations.11 For instance, between 1999/2000 and 2003/04, there were 31,433 opioid-related separations among persons aged 15-54 years old compared with only 677 cocaine-related separations.

Data from Riddell et al.3 indicates that in 2003/04 and 2004/05, among illicit drug-related separations, opioid-related separations accounted for approximately 46.7% (rate per 1,000 population — 0.38) and 44.6% (rate per 1,000 population — 0.34) and cocaine 1.5% (rate per 1,000 population — 0.011) and 2.4% (rate per 1,000 population — 0.018) respectively. In comparison, data from the US indicates considerably greater proportions of hospital visits are related to cocaine. In 2004 and 2005, approximately 19% (rate per 1,000 population — 1.31) and 31% (rate per 1,000 population — 1.51) (respectively) of visits to an Emergency Department were related to cocaine, whereas for heroin over the same two years 8% (rate per 1,000 population — 0.55) and 11% (rate per 1,000 population — 0.56) were related to opioid use.12,13Caveat lector— comparing numbers of Emergency Department visits with hospital separations may be somewhat erroneous, the results above however do provide a strong argument for the differences in opioid and cocaine related separations between Australia and the US.

Harms associated with injecting cocaine have been highlighted in a recent study carried out in Sydney and Melbourne.14 Injecting cocaine or smoking ‘crack’ cocaine carries a greater dependence liability and associated harms compared to snorting or swallowing the drug.15,16 Given the differences in prevalence, dependent use, purity and availability of cocaine between Australian and US,17 the use of US derived AF may result in a considerable overestimate for Australian population.

Apart from differences in prevalence of cocaine use in Australia and the United States which have implications on the differences in relative risk, there are many other risk factors which can be considered as causal for both low birth weight and ante natal hemorrhage such as cigarette smoking, access to antenatal care and maternal nutrition. Therefore, the use of unadjusted AFs might result in misleading findings.

Along with a number of methodological issues related to the calculation of cocaine attributable AFs, there are clear differences in the nature and extent of cocaine and heroin related harm in Australia compared to the US. We would suggest that any US-derived estimates of cocaine-attributable low birth weight newborns separations should be used with caution for Australian data. To correctly identify the burden related to illicit drugs, epidemiological studies are necessary to derive estimates for valid Australian AFs.