• Open Access

Lessons for Aboriginal tobacco control in remote communities: an evaluation of the Northern Territory ‘Tobacco Project’


Correspondence to:
Dr David Thomas, Menzies School of Health Research, PO Box 41096, Casuarina, NT 0810; and Institute of Advanced Studies, Charles Darwin University, NT. Fax: (08) 8927 5187; e-mail: david.thomas@menzies.edu.au


Objective: To evaluate a Northern Territory (NT) government-led pilot ‘Tobacco Project’ in six remote communities.

Methods: Monthly surveys of staff, semi-structured interviews with staff and community members, observation of the delivery of tobacco control interventions, review of Project documents, and monitoring of tobacco consumption using sales (or wholesale orders) of tobacco.

Results: There was a substantive amount of tobacco control activity delivered in three of the Project communities. In two of these locations, the majority of work was primarily driven and undertaken by resident staff. Overall, most of the Project's efforts related to community education and awareness-raising. There was variable impact of the Project on tobacco consumption across the six communities. More tobacco control activity was consistently associated with a greater reduction in tobacco consumption. An important predictor of local activity was the presence of strong community drivers. A significant obstacle to the Project was the lack of new resources.

Conclusions: Despite the minimal impact of this Project on tobacco consumption overall, there was a consistent association between on-the-ground tobacco control activity and reductions in tobacco consumption.

Implications: New initiatives will not only need to provide new funding, but identify and then support local staff, who are central to improving local tobacco control activity and so reducing smoking and smoking-related illnesses and deaths.

Tobacco use is the single most important risk factor for excess mortality and morbidity among Indigenous people, responsible for one-fifth of Indigenous deaths in 2003.1 Indigenous smoking rates are highest in the Northern Territory (NT), where 54% of adults are daily smokers.2 In some remote NT communities, the prevalence of smoking is even higher.3

A comprehensive approach, with many different elements of tobacco control, is promoted as the most likely to succeed.4 A literature review has inferred which tobacco control activities are most likely to be effective in Aboriginal communities, with evidence usually coming from research in other populations given the paucity of Aboriginal tobacco control research.5 Nevertheless, evaluations of multi-component tobacco control interventions in Indigenous communities have identified only small impacts on smoking. A study of three intervention and three matched control NT communities only found a reduction in tobacco consumption in one intervention community compared to its matched control.6 In eight north Queensland Indigenous communities, smoking fell from 61% (of 698) to 57% (of 596).7 Similarly, a Cochrane systematic review of multi-component community interventions found only minimal differences in smoking in intervention and control communities, including the two most rigorous studies.8 Nevertheless, the Australian Government has announced more than $10 million for such multi-component interventions in Indigenous communities.9 This paper evaluates a recent multi-component tobacco control project in six Indigenous communities, and will further add to our understanding of the impact of such projects, and the key enablers and obstacles of their impact on smoking, in the context of this new policy initiative.

The Tobacco Project

The Northern Territory Department of Health and Families (NTDHF) (formerly the Northern Territory Department of Health and Community Services) ran a pilot ‘Tobacco Project’ in six remote Aboriginal communities from January 2007 to June 2008. The Project was supported by the National Heart Foundation and local councils, health clinics and health boards in the six communities.

The communities were asked to prioritise and plan tobacco control activities from a list of tobacco control interventions, for which the Project could provide support. Each community had a different range and intensity of tobacco control activities; there was not a single Project package that could neatly be evaluated in each site. Importantly, no new NT government resources were put towards the Tobacco Project; the Project could only co-ordinate and redirect the existing resources of the NTDHF Programs in these communities.10 While the Project was associated with considerable planning and co-ordination activity in Darwin, the lack of specific additional Project resources and funds constrained the intensity of tobacco control activity in the communities.

Menzies School of Health Research was invited to evaluate the project. While the aim was to evaluate the impact of the Tobacco Project, this could not always be neatly separated from activities that would have occurred without the Project.

Context – the six communities

Five of the communities (A, B, C, D and E) were located in the Top End of the NT and the sixth (F) was in central Australia. Community F joined the Project significantly later that the other five, in late 2007.

The Project communities ranged in population from 130 to more than 2,000. Community A is a cluster of four main sites, three of which have mostly Aboriginal residents. Community E, the smallest of the Project communities had no store. By contrast, in Community B there were five different tobacco retail outlets operated by different local organisations and an established health centre and Aboriginal-controlled Health Board.

In all but two communities (A and C), there had been little concentrated effort on tobacco control in recent years.


The evaluation used a mix of data:

  • Monthly staff questionnaires of tobacco control activities delivered.
  • Semi-structured interviews with 25 Indigenous adults in two communities.
  • Semi-structured interviews with 19 health and welfare staff.
  • Observation of and participation in monthly Steering Group meetings and review of meeting documents.
  • Observation of tobacco control activities in three communities.
  • Sales (or wholesale orders) of tobacco at community stores and takeaways (except in Community E, where there was no community store).

Sampling of community members for interviews took place at three sites (although most (n=21) occurred in one large Top End community). Health and welfare staff interviews included government and non-government workers living in, or who visited regularly, all of the Project communities.

We described the range of tobacco activities delivered; it was not appropriate to present statistical data on the intensity of activities across the communities because the data was of insufficient quality.

Tobacco sales or wholesale invoices from each site were converted into cigarette (stick) equivalents, with 0.8g of loose tobacco equivalent to one cigarette, consistent with national reports.11 If sales data were not available, we described monthly tobacco consumption using three-month rolling averages of wholesale invoices. We have shown elsewhere that the average of three months' wholesale invoices provides a close estimate of monthly tobacco sales.12 We compared the total sales (or wholesale invoices) of tobacco products in the 2006/07 and 2007/08 financial years, and in the last six months of 2006 and 2007.


Of the 19 staff interviews, 15 were town or bush-based NTDHF staff (nine women and four Indigenous staff were included). Twenty-three health staff consented to providing feedback on their activities via a monthly questionnaire, but nearly half (11/23) resigned, took extended leave or moved to different positions during the Project. The response rate to the monthly questionnaire was variable and was particularly poor for the last six months of the Project (monthly response rates between 20 and 40% of consenting staff, January-June 2008).

1. Tobacco interventions delivered by the Tobacco Project

The range of tobacco interventions delivered by the Tobacco Project

Across all six communities, most of the Project's tobacco control efforts related to community education and awareness-raising (Table 1). This included group information sessions, the use of specific health promotion media (e.g. posters, pamphlets, visual displays), tobacco education as part of general health promotion and the use of traditional storytelling to reinforce anti-tobacco messages. These interventions were directed at a range of target groups: adults, pregnant women and children.

Table 1.  Tobacco control activities.
 Community ACommunity BCommunity CCommunity DCommunity ECommunity F
  1. Notes: +++= significant activity;  ++= some activity;  += little activity;   0 = no reported activity;   N/A = not applicable

Category of tobacco control activity
Increase exposure to tobacco control messages++++++++++++
Increase provision of smoking cessation interventions+++++++++
Increase smokefree areas++++0+++
Reduce store display of tobacco products+++0N/A0

During the Project, a media campaign was launched that targeted Indigenous youth, the ‘I'm Smarter than Smoking’ campaign. NTDHF evaluated this campaign by using a survey of 820 NT students to assess their attitudes towards this campaign compared with other advertisements. There was no assessment of the impact of the campaign on smoking initiation or cessation. There were no changes to NT tobacco control legislation during the Project, and taxation on cigarettes only increased in line with CPI.13

Where did most activity occur?

There was a substantive amount of tobacco control activity delivered in three of the Project communities (A, B, F). In two of these communities (A and B), the majority of work was primarily driven and undertaken by resident staff and for the most part, happened alongside (rather than because of) the NTDHF Tobacco Project.

In Community B, the local health board ran a tobacco control program from July-December 2007 (the ‘Smokebusters’ Project), funded independently of the NTDHF Tobacco Project, which employed two locally based staff members (one full-time Project Coordinator and one half-time Tobacco Support Worker). This project was associated with the most intense tobacco control activity of all participating communities, delivering a diverse number of tobacco interventions that included: providing free NRT and counselling to community members; running smoking cessation courses and one-on-one sessions with clients; providing education in the school; running a community awareness campaign to reduce secondhand smoke exposure of children; working with different community organisations to comply with tobacco control legislation; working with the clinic to promote brief intervention; and applying for funds to sustain local tobacco control activity.

In Community A, a high level of activity (including NRT education, subsidisation and distribution, strong support for smokefree public areas and work sites, and ongoing education) was also reliant on resident health staff who were committed to Aboriginal tobacco control and had longstanding relationships with community members. Successes here included the cessation of sales of cigarettes at a supermarket (after lobbying from the local doctor), and the heavy promotion of ‘No Smoking’ areas on clinic grounds and at the airport. In this context, the NTDHF Project was able to support and foster activity that was already happening in the community, even though the Project did not initiate a significant amount of additional tobacco control work.

Community F was a slightly different story again. In this community, interviews indicated that the Tobacco Project acted as a significant driver to scale up NTDHF Alcohol and Other Drugs program work on Aboriginal tobacco control. Activity included the provision of group information and education sessions, heavy promotion of smokefree areas at the council, clinic and youth services, and efforts to engage the clinic in increasing the use of smoking cessation pharmacotherapies and brief interventions. This local activity was supported or driven by visits from NTDHF staff from Alice Springs, with strong involvement from several local organisations: the council, the local youth program and the store.

In the three other Project communities (C, D and E), there was comparatively little reported activity. For example, in Community D, there were only two visits by NTDHF staff during the course of the Project; this was partly attributed to difficulties engaging with the clinic staff at this site.

2. Tobacco consumption

Data from six Project sites (two sites in the community A cluster, and all other communities excluding community E, where there was no store), demonstrated a 1.2% reduction in daily tobacco consumption in 2007/08 compared with 2006/07. And there was a 5% reduction in tobacco consumption in the second half of 2007 (when most project activity occurred) compared with the same time in 2006.

Tobacco consumption decreased in both the 12-month and six-month comparison in only three locations (Table 2). The greatest relative reduction in tobacco consumption occurred in Community F. However, because community F is a small community, this was only associated with a very modest absolute reduction of less than a packet of cigarettes every day.

Table 2.  Comparisons of tobacco consumption.
 Average number of cigarettes/day% change in cigarettes/dayAverage number of cigarettes/day% change in cigarettes/day
 2006/072007/08 July-Dec 2006July-Dec 2007 
Community A
  Site 11,8381,865-1.4%1,9761,819-7.9%
  Site 21,6961,984+17.0%1,8761,954+4.1%
Community B12,66112,129-4.2%11,84010,849-8.4%
Community C2,9312,786-4.9%2,9523,040+3.0%
Community D1,6031,733+8.1%1,7901,766-1.4%
Community F374352-5.7%382342-10.6%

Community B had the greatest absolute reduction in tobacco consumption in each comparison. In 2007/08, 182,077 less cigarettes were sold than in 2006/07. Indeed, if the results from community B are excluded, there was an overall 3.5% increase (rather than a decrease) in tobacco consumption in the remaining four locations (2007/08 vs 2006/07), and only a small 0.6% reduction in the six-month comparison.

In community B, the sharpest decline in consumption coincided with the start of increased local tobacco control activity, and the employment of a public health nurse and tobacco community worker in mid-2007, followed by a period of consistently lower consumption. Tobacco consumption returned to previously higher levels of consumption soon after the local activity and Project collapsed at the end of 2007. There was no such neat temporal correlation between reductions in tobacco consumption and Project activity in other sites.

3. Perceptions about enablers and obstacles to tobacco control

Community readiness for tobacco control

The places where most activity took place (communities A, B and F) were where stakeholders across the community (e.g. council, clinic, school) were stated by health staff to be ‘ready’ to prioritise tobacco control. They had identified tobacco control as a priority, developed local strategies and had dedicated staff available to deliver services.

‘Community readiness’ was supported by strong local drivers of tobacco control, who, particularly in Community A, had been working hard on the issue for years. These stakeholders (usually clinic or council staff) had established trusting relationships with the communities in which they lived and could act as a conduit for further activity delivered through the Project. We observed that key to the success of these local drivers was that they were resident in communities and had dedicated duties to tobacco control (either part or full time). Developing local Aboriginal capacity was also reported as important in the success of local efforts in the long term.

Lack of resources

Overwhelmingly, health staff reported that a lack of resources undermined the success of the Tobacco Project and tobacco control more generally. For the Project, lack of new resources impacted on the type of tobacco interventions that the Project delivered and the time spent by staff delivering these. The perception of senior Departmental staff was that they were still ‘fiddling around at the edges’ of tobacco control. Budgetary constraints also affected time staff spent on the Project, while balancing other demands (this was reported at all levels of staff involvement); the AOD Community Support Officers who conducted much of the field visits reported feeling particularly stretched. It also effected travel and time spent in communities, with the centralised rapid fly-in, fly-out model perceived as increasingly ineffective in connecting with community people.

The interviews also identified examples of both successful collaboration and poor collaboration and conflict between Project partners, and generic problems encountered in working with these communities (e.g. seasonal access), in addition to the particular disruption of the NT Emergency Response. These data are not elaborated on here (but are available from study authors).

The Aboriginal community informants provided little useful data on the processes used for this project, beyond being generally supportive of the community-driven project in community B, where most interviews were conducted. These informants however, provided useful data on why remote Aboriginal adults smoke,14 as well as general perceptions of different tobacco control interventions.


Data limitations

The low and variable staff response rate to monthly questionnaires means that some tobacco control activities may not have been identified by the evaluation. However, interviews and informal discussions with health and community staff, as well as regular attendance at Project Steering Group meetings, provided the evaluation team with other sources of information about what was delivered in each of the Project communities, reducing the final amount of missing data.

Similarly there were limitations of the tobacco consumption results: the most significant is that there was no fixed denominator for these consumption results. The potential impact of population changes means that particular caution should be exercised in the interpretation of the results, particularly from community F, where the departure of just one or two pack-a-day smokers from this location could explain the apparent decline in consumption in that community. However, local informants in community B had noticed no significant population change to account for the dramatic reduction in tobacco consumption associated with increased tobacco control activity in the last six months of 2007.

Tobacco consumption data from community A need to be interpreted with caution, as we were not able to report on data from two of four sites in the cluster. And residents of community F and Site 1 community A are close to tobacco retail outlets where we did not access store data because they also service many non-Aboriginal clients. Nevertheless, store-based indirect measures of consumption are the preferred measure of consumption in this setting. In contrast, self-reported numbers of cigarettes smoked in surveys in remote Aboriginal communities are only weakly correlated with cotinine levels,15 and estimates of cigarettes smoked per smoker have demonstrated to be both higher6,16 and lower17 in surveys compared to store-based estimates. At the national level, estimates of consumption based on self-report under-estimate consumption by 26% compared to estimates based on tax receipts (which are analogous to our store-based estimates).18 Finally, the study did not include control communities unlike the two previous studies.6,7 Delays in collecting tobacco consumption data from other sites meant that there was insufficient data to make meaningful comparisons.

In spite of these limitations, some clear messages emerge from the results.

Lessons from the Tobacco Project

More tobacco control activity was associated with a greater reduction in tobacco consumption. While we were not able to neatly quantify the intensity of tobacco control, we could compare both places and periods with clearly different intensities of tobacco control activity. Community B and F, where significant tobacco control activity occurred, were also the locations with most consistent reductions in tobacco consumption. Additionally, there was a greater (5%) reduction in combined tobacco consumption when we just looked at the last six months of 2007, the period of greatest reported tobacco control activity, than when the whole 12 months of 2007/08 was examined (1.2%).

The three communities (A, B, F) that saw most activity over the life of the project were places most ‘ready’ to tackle the issue, that is, these local communities had prioritised tobacco control, key stakeholders were on board and resources and staff time were mobilised. In two of these communities (A and B), health staff and management had made tobacco control a clear priority and had allocated additional resources for tobacco control; in community B, led by the local Aboriginal community controlled health board. The north Queensland study had similarly suggested that lack of local ownership of their project had limited its impact.7 In contrast, the Cochrane review found no relationship between community involvement and the impact on smoking outcomes.8

The lack of dedicated new resources for the Project was a major impediment to increasing tobacco control activities. Other low intensity multi-component community tobacco control projects in remote Indigenous communities had similarly modest overall impacts.6,7 In this context, the apparent impact on consumption, especially in Community B with its significant additional resourcing is encouraging. Similar levels of additional resourcing in other communities could lead to significant reductions in Aboriginal smoking in the NT. The need for adequate new dedicated funds is supported by American research that showed that increases in State tobacco control program expenditures were independently associated with overall reductions in adult smoking prevalence.19 The planned new multi-component tobacco control initiatives will not only need to provide new funding, but identify and then support local staff, who we found were central to improving local tobacco control activity and so reducing smoking and smoking-related illnesses and deaths.


This research was supported by a grant from the Cooperative Research Centre for Aboriginal Health (No. CD 216) and a NHMRC GP Clinical Research Grant (490300). The NT DHF contributed additional funds to cover the six-month extension of the evaluation. David Thomas was supported by a NHMRC Population Health Capacity Development Grant (#256235). The evaluation team would like to thank all the participating Aboriginal communities, community residents and staff, NT DHF staff, participating community organisations, and stores and wholesalers.