• Open Access

Evaluating performance of and organisational capacity to deliver brief interventions in Aboriginal and Torres Strait Islander medical services


Correspondence to:
Katie Panaretto, Population Health Hub, Queensland Aboriginal & Islander Health Council, PO Box 3205, South Brisbane, QLD 4101. Fax: (07) 3844 1544; e-mail: katiepanaretto@qaihc.com.au


Objective: This study assessed brief intervention (BI) activity and organisation capacity for smoking, nutrition, alcohol and physical activity (SNAP framework) and key clinical prevention activities in four Aboriginal and Torres Strait Islander medical services in Queensland.

Methods: A mixed methods design was used including: staff surveys of knowledge and attitudes (n=39), focus groups to discuss perceived barriers and enablers and chart audits (n=150) to quantify existing BI activity.

Results: Of 50 clinical staff, 46 participated in the staff survey and focus groups across the four sites. BI was perceived to be important. There was significant variation in completion of records for SNAP risk factors, key clinical and BI activities across the sites. At least one SNAP factor status was recorded in 130/150 (86.7%) patient charts audited and there was a significant trend of increased recording of SNAP factors with increasing number of patient visits. Of those identified at risk 78% received at least one BI. Where risk was identified 65/96 (67.7%) patients required multiple BIs. BI for tobacco use was consistently high across all sites. Only one site recorded regular care planning and Adult Health Checks. Impacting factors included leadership, high staff turnover, multiple medical records and staff health status.

Conclusions: Inflexible staff training, competing health priorities and high levels of staff turnover were identified as key barriers to the delivery of BI in clinical settings. The data suggests a good base of existing BI activity for smoking and key clinical activities which may improve with further support.

The Australian National Chronic Disease Strategy identified brief interventions (BI) in the primary care setting as an evidence-based intervention to prevent, detect and manage chronic disease.1 Brief interventions emerged through the 1970 and '80s as effective strategies to encourage change of behaviour for alcohol and tobacco use2 and are now more widely applied. BIs are short, focused, non-judgemental, motivational, patient-centred interactions that seek to change behaviour to achieve a healthy lifestyle. The techniques used in BI encompass a screening or assessment process, feedback, client engagement, simple advice or brief counselling, goal setting and follow-up.2,3 Examples of BI include the 5As,4 Smoking, Nutrition, Alcohol and Physical Activity Framework (SNAP)5 and the use of assessment tools such as the Fagerstrom test of nicotine dependence6 and Stages of Change (SOC).7

Tobacco consumption, high body mass index (BMI), physical inactivity and alcohol are four of the leading risk factors contributing to the burden of chronic disease in the Australian Indigenous population.8 The association of these risk factors with social disadvantage and risk of morbidity is a strong argument to support BI for promotion of healthy lifestyles. Primary health care teams are well-placed to take advantage of the continuity of care they provide to the community to deliver BI over the long term. This context has led to use of the SNAP framework5 in the Medicare-funded Adult Health Check (AHC) for Aboriginal and Torres Strait Islanders, item number 710.9

This paper reports on opportunistic BI for SNAP risk factors and clinical prevention activities, and the organisational capacity to perform these activities in four urban Aboriginal and Torres Strait Islander medical services, Queensland, Australia, undertaken in the 12 months to December 2007. The work was undertaken by the Queensland Aboriginal and Islander Health Council (QAIHC) Population Health Hub to guide planning for further support of BI activity in these services.


Study setting

The four services involved in this project are located within the Brisbane region, Australia. The Brisbane region has 41,400 Indigenous residents, 9.1% of the total Australian Aboriginal and Torres Strait Islander population.10 Of the participating services, three were Community Controlled Health Services (CCHSs) and the other a State-funded community health service. All services have a predominantly Aboriginal and/or Torres Strait Islander patient base.

Study design

A mixed methods design was used including: a staff survey, medical chart audit and focus groups. All the services' main clinical team members were asked to participate in the survey and focus groups. This included Aboriginal Health Workers, nurses, doctors and practice managers. Visiting allied health professionals were not included.

Staff survey tool

The staff survey was adapted from the SNAP survey5 based on input from stakeholders. The survey assessed knowledge and attitudes pertaining to BI, screening tools available, how often staff thought they provide BI to patients, availability and use of referral services and provision of training. Informed consent was established prior to staff completion of the survey delivered during team meetings attended by the project team. Absent staff were invited to complete the survey upon their return to work.

Chart audit

The chart audit quantified existing evidence of BI activity. A convenience sample of 30 adult patient medical records was selected from two separate days in December 2007, with 15 consecutive charts taken from each day. The clinical team notes, in either the computerised record or paper chart, for any visits made in the preceding 12 months were examined. Charts were excluded and the next chart chosen if a patient was less than 18 years of age at the audit date, was a current staff member or if the patient's chart had been selected on the first audited date selected. One site opted to complete a sample of 60 patient records from the same time period. Audit dates from December 2007 were chosen as that was the last month prior to discussion with the services about participating in this project.

The chart audit form was modified from the Healthy for Life toolkit11 with additional input from key stakeholders. De-identified data was collected, recording performance of SNAP risk assessment, and any indication of BI performed meeting four of the 5As framework:5 ask, advise, assess, assist. Being at risk was assessed as: reported tobacco use, at risk alcohol intake,12 BMI <18.5 or >25 was used a marker of nutritional or inadequate physical activity risk. Medicare billing data was also examined to determine what proportion of eligible patients had received an AHC (MBS item 710) or had a current General Practice Management Plan (GPMP, MBS item 721).9 Completion of key clinical indices were also assessed.

Staff focus groups

All clinical staff members were asked to participate in a focus group at each site to discuss the enablers and barriers to delivery of BI. Themes revealed in the staff surveys and preliminary findings of the chart audits were feedback to each site prior to the focus groups. Issues explored included the role of workforce, information technology, management support, past training, and perceived supports needed.

Ethics approval

Ethics approval was granted by the Queensland Health Princess Alexandra Hospital Human Research Ethics Committee. The project was approved by the QAIHC and Service Boards.

Data analysis

All data was held in a password-secured database at QAIHC. Depending on their distribution, numerical data have been described by means and standard deviation (SD) or medians and interquartile range (IQ range). Bivariate associations have been tested using Chi-square, Kruskal-Wallis and t-tests as appropriate. All statistical tests were performed using SPSS for Windows, release 14.13 A two-tailed p-value below 0.05 was regarded as significant.


Clinical team participation

Clinical staff involvement was good overall – 46 of 50 clinical staff participated in the surveys and/or focus groups. At Service A, 16 of 20 clinical staff completed surveys and four attended the focus group discussion, at Service B, 12 of 14 staff completed surveys, with all staff attending the focus group, Service C had 6 of 8 staff complete the survey and all eight attend the focus group and Service D 5 of 8 staff completed the survey and all eight attended the focus group. The clinical positions represented were: 15 Aboriginal or Torres Strait Islander Health Workers, eight nurses, 10 general practitioners (GPs), three clinic managers, and 14 others (receptionists, drivers, health promotion officers) with no significant differences between the services.

Staff survey

Completed surveys were received from 39 clinical team members. Overall, 47% of staff reported that “most or all consultations” they conducted included behavioural risk assessments and 51% reported that “most or all consultations” included advice being offered to patients. Access to patient education materials was reported as being available by 87% of the team members for smoking cessation, 95% for nutrition, 85% for alcohol education, and 69% for physical activity promotion materials. Perceived accessibility of referral networks were reported by clinical staff as easily available for: smoking cessation by 67% of respondents, nutrition counselling by 100% staff, alcohol management by 77% staff, and 44% of staff for physical activity. Overall 49% of the respondents stated “most or all consultations” included patient referral to these services as appropriate.

Overall, 92% of the staff surveyed agreed that BI techniques were useful in clinical settings and 49% said they had received appropriate training in BI. Familiarity with specific BI tools (SNAP, SOC, 5As) ranged from 18% to 33%. Overall 92% of staff surveyed felt the need for more training.

Chart audit

Across the four sites, a total of 150 patient charts were audited. There were significant differences in each of the demographic characteristics of the patients in the samples audited, Table 1. In 47.3% of records audited ethnicity was not recorded. The number of clinic visits ranged from nil to 59 times in the year preceding the audit, with a median of 8 (IQR 4, 15) preceding visits.

Table 1.  Demographic characteristics and number of clinic visits in patient samples for four Indigenous health services participating in the brief intervention project.
Patient Sample characteristicsABCDp
 n = 60n = 30n = 30n = 30 
 n (%)n (%)n (%)n (%) 
  1. Notes: a) Median (Interquartile range)

  2. Patients had presented for care delivered in the 12 months prior to chart audit of 150 patients in December 2007.

Age*38.5 (28.5,51)47.5 (35,59)44 (31,59)63 (45,75)0<0.001
Male14 (23.3)16 (53.3)11 (36.7)12 (40)0<0.046
% Identified24 (40)30 (100)12 (40)13 (43.3)0<0.001
% Indigenous24 (100)30 (100)11 (91.7)6 (46.2)0<0.001
No. clinic visits per patient in 12 mths prior to audit datea5 (2,10)12.5 (6,21)6 (3,11)13 (9,24)0<0.001

Recording of risk factors in the medical charts and completion of GP management plans (GPMPs) and adult health checks (AHCs) differed significantly across the services (Table 2). Of 150 charts audited, 20 (13.3 %) did not include any documentation of assessment of the SNAP risk factors. Figure 1 shows the significant variation across the sites in documented performance of SNAP assessment (p<0.001). Figure 2 shows the variation in assessment of SNAP risk behaviours with the number of patient visits made in the preceding 12 months (p<0.001).

Table 2.  Documentation of risk factor recording in patient records for four Indigenous health services participating in the brief intervention project.
Risk factor recordedABCDOverallp
 n = 60n = 30n = 30n = 30(%) 
 n (%)n (%)n (%)n (%)  
  1. Notes: GPMP: General practitioner management plan – MBS item no: 721; AHC: Adult Health Check – MBS item No 710

  2. Patients had presented for care delivered in the 12 months prior to chart audit of 150 patients in December 2007.

Smoking46 (76.7)30 (100)25 (83.3)22 (73.3)820.026
Nutrition16 (26.7)27 (90)10 (33.3)12 (40)43.30<0.001
Alcohol28 (46.7)29 (96.7)11 (36.7)18 (60)57.30<0.001
Physical activity14 (23.3)27 (90)12 (40)9 (30)41.30<0.001
BMI31 (51.7)26 (86.7)13 (43.3)4 (13.3)49.30<0.001
Blood pressure41 (68.3)30 (100)19 (63.3)28 (93.3)78.70<0.001
Blood sugar40 (66.7)28 (93.3)28 (46.7)23 (76.7)700<0.001
ACR26 (43.3)26 (86.7)13 (43.3)23 (76.7)58.70<0.001
GPMP2 (3.3)17 (56.7)8 (26.7)7 (23.3)22.70<0.001
AHC: 7100/52 (0)13/20 (65.0)0/19 (0)0/12 (0)12.60<0.001
Figure 1.

Distribution of completion of SNAP status documented in records at four Indigenous health services during care delivered in the 12 months prior to chart audit of 150 patients in December 2007.

Notes: SNAP: Smoking, nutrition, alcohol, physical activity

No SNAP: O risk factor assessments in records, SNAP 1-3: 1-3 risk factors assessments recorded, SNAP 4: all 4 risk factors assessments recorded

Figure 2.

Distribution of completion of SNAP status recording with total clinic visits per patient during care delivered in the 12 months prior to the chart audit of 150 patients at four Indigenous health services in December 2007.

Notes: No patient visits – median, p<0.001

SNAP: Smoking, nutrition, alcohol, physical activity

No SNAP: O risk factor assessments in records, SNAP 1-3: 1-3 risk factors assessments recorded, SNAP 4: all 4 risk factors assessments recorded

Overall in the 96 charts, where at least one risk factor was identified, there was documentation in 75 (78.1%) of at least one BI completed. Figures 3, 4 and 5 show the variation across the services in the proportion of assessment of risk factors performed, proportion of at risk patients, and proportion of BI delivered to those identified as at risk for tobacco use, alcohol use and nutrition/physical activity. There was significant variation between services in performance of risk factor assessment in each of the three risk categories but no differences in proportions of at risk patients or BI delivered. Overall in the charts where risk was identified, 31 patients (32.3%) required 1 BI, and 65 (67.7%) required multiple BIs.

Figure 3.

Tobacco use: Prevalence of risk factor assessment, at risk patients, and brief interventions delivered documented in records at four Indigenous health services during care delivered in the 12 months prior to chart audit of 150 patients in December 2007.

Notes: # p < 0.001: Significant difference between the participating sites in assessment of tobacco use.

Figure 4.

Alcohol intake: Prevalence of risk factor assessment, at risk patients, and brief interventions delivered documented in records at four Indigenous health services during care delivered in the 12 months prior to chart audit of 150 patients in December 2007.

Notes: # p < 0.001: Significant difference between the participating sites in assessment of alcohol intake.

Figure 5.

Nutrition and physical activity: Prevalence of risk factor assessment, at risk patients, and brief interventions delivered documented in records at 4 Indigenous health services during care delivered in the 12 months prior to chart audit of 150 patients in December 2007.

Notes: # p < 0.001: Significant difference between the participating sites in assessment of nutrition and physical activity.

Focus groups

A number of staff at the focus groups felt that BI for smoking was easier than for the other risk factors. A practical, easy means of assessing nutrition was felt to be lacking, impacting on the time required to discuss these issues with the patient. A number of staff also commented on the impact on low socio-economic families coping with increasing costs of food and the impact of mental health illness. This was summed up by a staff member who pointed out:

“…it's hard talking to people about healthy eating and exercise when they don't have a roof over their head…”

All focus groups discussed the problem of attendance at follow-up appointments, for example with dieticians. Although referral networks were available, the location and time of appointments meant Indigenous patients had trouble keeping these appointments.

The two major enablers expressed by staff to conducting BI were AHCs and time. The importance of the relationship with the patient was reinforced when staff described some of the barriers to conducting BI as “a lack of time”, whether it was the need to build relationships with patients, so patients won't feel threatened by BI, or time to deliver BI, added to the time to address other competing health needs. In all four workshops, staff also raised the theme that negative personal lifestyle behaviours of staff members, such as smoking and lack of exercise, may impact on the confidence of staff to implement BI.

At the organisational level, infrastructure was a barrier at site C where the clinic building is “too small and poorly laid out”. Staff turnover was an issue at three services; Site D had struggled to have regular team meetings, and Site C, where the health service staff totals 23, had experienced 24 staff changes in the previous two years.


Assessing ‘everyday’ clinical activity in four health services providing care to the Aboriginal and Torres Strait Islander community in Queensland, brief intervention for tobacco and recording of blood pressure use was good across all services with one service achieving consistently good standards across all activities assessed. Adult health checks were underutilised and their completion, as was SNAP assessment, was associated with increasing numbers of patient visits.

A major strength of this study was that the services were not aware this work was to be carried out during the period audited and the data should reflect ‘everyday’ care delivery. One limitation of the study was the lack of analysis of who provided the BIs or their quality. While the audit only assessed a small sample of the total consultations that these health services provide, the limited patient demographics and visit numbers suggest that the patient populations were significantly different in each service. In addition, the services are structurally different. The service funded by the state Department of Health has a senior GP who provides strong clinical leadership and advocacy for his service. It also has good links with a university health faculty, easier access to training opportunities, better remuneration for staff and, therefore, perhaps a more stable workforce. In a quality improvement framework, the high standards achieved by this site, site B, should set a benchmark for the other services.

The first steps in BI, ‘ask and assess’, identify risk factor status and omission of these steps will reduce BI activity. The poor documentation of risk factor information in the clinical records examined in this study was concerning, but consistent with data from general practice in the UK.14 The UK work showed that recording of medical diagnoses was high but of lifestyle factors was low – 52% completion for smoking and 38% for alcohol use. In this study all services had multiple clinical record systems in place during the period audited: a paper chart plus one or two electronic records. To compound poor communication, at three of the four services the clinical team members have differing access to the record systems with only the GPs having access to all three systems and the paper chart the only record used by all clinical staff. An additional factor in the electronic records is the lack of a field in which to record nutrition assessment data. It is possible in the electronic medical record now in use in three of the services to insert prerecorded text into consultation notes documenting a standard process for BIs delivered. This may well have negatively impacted on the recording of information and was raised in the focus groups. There is work in progress with the services to implement a health information strategy that should improve recording of risk factor and key clinical activity.

Did these services achieve acceptable levels of BI activity? Targets for BI are not set out in the SNAP or ‘red book’ guidelines.5,15 While there is much evidence from well controlled trials to support BI as effective interventions, there is little evidence documenting how much BI is delivered opportunistically at the clinical workface in the course of routine primary health care service delivery. This data shows performance varied by risk factor and activity. Recording and BI for tobacco use was universally good at about 80% completion, similarly recording of BP and blood sugar levels approached this level. However, recording of alcohol use, nutrition and physical activity status and BI was variable as was the recording of BMI. This data was supported by discussion at focus groups where smoking was identified as a priority, and staff felt resources, in-house referral pathways and access to training courses to assist BI for smoking were readily available. This was in contrast to alcohol and drug use where, while the need was significant, referral pathways and rehabilitation services were described as inadequate.

A key constraint to clinical teams incorporating BI into their daily work flows is time16,17 and juggling the demands of patients, who present with their own immediate medical needs, with prevention activity. Where risk factor recording allowed assessment of need for BI, 67% needed multifaceted action for more than one risk factor. This data demonstrates the workload faced by primary care teams, associated with the burden of disease caused by these leading causes of ill health in the Indigenous community.8 Data from mainstream general practice in Australia suggests prevalence of risk factors is much lower for smoking and alcohol.18 In addition, risk factors may be perceived independently19 and require specific interventions. Overseas work estimates the time required to deliver effective preventative activity is significant, even if confined to SNAP and BP monitoring, 1.6 hours a day for the average clinician.17 In addition, the time required to screen for nutrition and physical activity is twice that required for tobacco and alcohol use. Clearly, prevention must be a team activity, supported by good training, tools and an accurate medical record used by all clinical staff to facilitate communication.

The focus groups highlighted another aspect of time: the time needed to form the relationship with the patient, in order to allow discussions about lifestyle behaviours to occur. The significant trend of completion of SNAP recording with increasing patient visits has important implications. Evidence suggests that “knowing the patient” is a key component to successful delivery of primary care.16 Health professionals have been found to have perceived sensitivities about raising tobacco use, alcohol use and weight issues in a population where the unhealthy aspects of these risk factors are the norm.2 Teams do not want to provoke negative reactions thereby threatening ongoing functional relationships with patients. Evidence also suggests that BI started in one consult may well carry onto subsequent consults.20 Thus to achieve all the preventative activity patients require, the relationship with the whole clinical team must be nurtured. The data presented supports the contention that if a good relationship is not achieved and visits are infrequent preventative activity may not occur.

Despite staff recognising the AHC as an ‘enabler’ of undertaking BI, overall only one service had actively incorporated a system of opportunistically performing AHCs and care planning into their work flows. This is probably reflected in capacity issues discussed earlier for this service. These services are not alone in this performance where Medicare Australia data shows for 2007 a total of 4,194 AHCs were billed in Queensland,21 which given the parameters of the AHC, will be less than 20% of the eligible 144,900 Indigenous people residing in Queensland.10 This is an area that needs further work with the clinical teams.

Organisational capacity was perceived as important. Staff expressed a need for strong, consistent leadership, the implementation of quality BI through support for staff training, working to retain skilled staff and providing basic performance goals. High levels of staff turnover affects staff morale as well as adversely effecting on skill retention.22,23 The challenge is to provide training to shift from a passive documentation of health behaviours to the skilled BIs necessary to motivate change, using self-management techniques such as goal setting and discussion of strategies to overcome barriers.16 Perhaps cherry picking some training packages and delivering individual modules more frequently such as those on current Australian guidelines, rather than whole courses, may combat the impact of staff turnover and reinforce the knowledge base from which to assess a patient's SNAP status. Another concern raised related to staff's own behaviours. Most staff are known in their community and staff perceived their creditability in delivering effective BI may be limited when they themselves are smokers and/or are overweight. Workplace interventions have been acknowledged by expert groups as a priority for smoking cessation.24

In conclusion, in the context of the burden of lifestyle risk factors in the population audited, the prevention activity demonstrated by these health centres serving the Indigenous community in Queensland is encouraging. A high level of BI activity for smoking, screening blood pressure and blood glucose levels is being undertaken. The challenge is to build on this base and improve prevention activity for alcohol, nutrition and physical activity. Specific actions to improve organisational capacity, such as workplace health interventions for employees, the development of a simple tool to assess nutrition, training to cover latest SNAP guidelines, and the incorporation of Adult Health Checks into work flows may improve prevention activity in the busy clinical setting.


We thank the staff in the participating services for their support, particularly Megan Jackson, Cheryl Sidhom and Jan Lember for their role in project facilitation and assistance in the chart audits. The authors would also like to thank Diane Longstreet for her assistance collating project information and the initial draft of the paper, as well as the Advisory and Working groups, particularly Michael Tilse, Judy Kirkwood, Anna Cooney and Elizabeth Stephens for their advice through this stage of the Brief Intervention project. This project is supported by funding from Queensland Health as part of the Australian Better Health Initiative: A joint Australian, State and Territory government initiative.