Healthcare utilisation and disclosure of injecting drug use among clients of Australia's needle and syringe programs
Correspondence to: Professor Lisa Maher, The Kirby Institute, University of New South Wales, CFI Building, Corner Boundary and West Streets, Darlinghurst, NSW 2010, e-mail: L.Maher@kirby.unsw.edu.au
Background : People who inject drugs (PWID) report limited access to healthcare, and may avoid disclosing drug use. Health service utilisation was examined among participants in the Australian Needle and Syringe Program Survey (ANSPS), an annual cross-sectional sero-survey of needle syringe program (NSP) attendees.
Methods : An anonymous questionnaire was self-completed by 2,395 NSP clients throughout Australia. Multivariable logistic regressions identified variables independently associated with (i) disclosure of injecting to the most recent healthcare provider; and (ii) recent presentation to emergency departments.
Results : Seventy-eight percent of participants reported accessing healthcare in the preceding 12 months. Reasons for presentation included general health issues (46%); medication seeking (17%); and both (37%). Participants who recently accessed healthcare or had previously visited their most recent provider were more likely to disclose injecting drug use. Participants presenting to a GP or medical centre were less likely than others to disclose injecting. Those accessing emergency departments were more likely to report recent imprisonment.
Conclusions : Despite Australia's universal healthcare system and harm reduction policies, NSP-participants remain reluctant to disclose injecting, potentially hindering appropriate care and highlighting the need for multiple entry points to the healthcare system, including NSPs and opioid substitution therapy clinics.
People who inject drugs (PWID) experience a range of health problems arising from non-sterile injecting practices, complications of the drug(s) injected, or the lifestyle associated with illicit drug use and dependence.1 Access to primary healthcare is often limited for this marginalised group2 due to barriers, including costs associated with medical treatment and transportation, and stigma and discrimination within healthcare settings.3 Healthcare may assume a lower priority for PWID than obtaining food, clothing, shelter and generating income or other activities designed to support drug use.4 PWID are more likely than non-injecting drug users and the general population to delay seeking healthcare until conditions become severe.5,6 This, in turn, may lead to a significant mortality and morbidity,7 reliance on emergency departments and inpatient care,8 resulting in significant costs and pressures on the healthcare system.9
Even when services are accessed, PWID may not disclose their drug use or its extent to healthcare providers10 due to fear of stigma and discrimination and/or concerns about legal, child welfare, immigration, employment and/or housing ramifications.3 This may affect the quality of healthcare they receive, as accurate diagnoses, treatment and management may be compromised. Such concerns have led to the establishment of primary healthcare centres targeting PWID;2 however, geographical coverage of such providers remains limited, leaving most clients to access care from mainstream providers.
Despite barriers to healthcare and consequent poor health outcomes, healthcare utilisation among PWID has received little research attention.11 Studies to date have been confined to samples in drug treatment settings12–14 or to data collected incidentally during the course of other investigations.15 Furthermore, utilisation of costly emergency departments by PWID for complications that can be prevented or treated by primary healthcare services and other providers remains a concern.8,16 This is particularly relevant in settings with universal healthcare systems and harm reduction-based drug policies, where primary healthcare should be readily accessible.17 There are significant gaps in the literature about these problems. The Australian Needle Syringe Program Survey (ANSPS) draws on needle and syringe program (NSP) attendees who are representative of the broader Australian needle syringe program population18 and provide useful data to explore these issues. Since 1995, ANSPS has provided annual point prevalence estimates to monitor patterns of blood-borne viral infection and risk behaviours among NSP clients.19,20 Using ANSPS data collected in 2011, this exploratory, descriptive study documented patterns of healthcare utilisation among a large, national sample of NSP clients. Specifically, the study aimed to:
- 1Identify patterns of healthcare utilisation by ANSPS participants.
- 2Assess demographic and drug use characteristics associated with:
- a. disclosure of injecting drug use to the most recent healthcare provider; and
- b. recent presentation to an emergency departments.
The ANSPS methodology is described in detail elsewhere.19,20 Briefly, all PWID who attended participating NSPs during the October 2011 survey period were invited to self-complete a brief questionnaire covering demographics; drug and alcohol use and treatment history; injecting and sexual risk; history of HIV and hepatitis C diagnosis and treatment; history and recency of medical care utilisation; type of service and reasons for access; and disclosure of injecting drug use to the most recent provider. Fifty-three of Australia's 85 primary NSP sites participated in the 2011 ANSPS, from which a sample of 2,395 (a 41% response rate) was recruited. Respondents provided informed verbal consent for their voluntary, anonymous and unreimbursed participation. Assistance with survey completion was available upon request, although the great majority of participants indicate self-completion. Ethical approval for the ANSPS was provided by relevant institutional, jurisdictional and site-specific ethics committees, including the University of New South Wales Human Research Ethics Committee.
Based on the questionnaire item –Does that healthcare provider/service (accessed most recently) know you inject drugs? The response options were (i) No, (ii) Yes, but not the full details and (iii) Yes, knows everything. The sample was divided into groups labelled ‘no disclosure’ and ‘full/partial disclosure’ of injecting drug use. Comparison between the two was considered providing most important piece of information, particularly for those who do not disclose at all, and also ensuring brevity of the manuscript. Based on the item Where did you go the last time you sought medical care?, the sample was divided into ‘emergency department’ and ‘other provider’ groups.
A provider was classified as targeted if drug users were one of the main target groups; for example, opioid substitution therapy prescribers, PWID-targeted primary healthcare centres, NSP or detoxification/rehabilitation centres.
Chi-square (χ2) and Fisher's exact tests were used to examine differences between groups for categorical variables, and independent sample t-tests for continuous variables. Statistical significance was set at p<0.05. Multivariable logistic regression examined correlates of injecting drug use disclosure (no disclosure versus partial/full disclosure) and most recent episode of healthcare accessed from an emergency department (yes versus no). Variables that were associated on univariate analysis at the p<0.15 level were considered candidate variables for logistic regression modelling. Final models were derived using stepwise backwards elimination. Multicollinearity was assessed; and goodness of fit was examined using the Hosmer–Lemeshow test. Data were analysed using STATA (version 11).21
Participants (n=2,395) had a mean age of 37.7 years (SD 9.4; range 16–65) and 67% were male (Table 1). The majority (87%) identified as heterosexual. Eleven per cent reported a history of incarceration in the preceding 12 months and 12% identified as Aboriginal and/or Torres Strait Islander (Indigenous). In the month prior to the survey, 49% had injected daily or more frequently and 16% had engaged in receptive syringe sharing. Heroin was the drug most often last injected (34%). Seventy-nine per cent reported a history of drug treatment including opioid substitution therapy, detoxification and/or counselling, with 46% currently in treatment. Participants with a history of treatment were older than those with no such history (38.3 versus 35.9 years; p<0.001).
Demographic, drug use and risk characteristics associated with disclosure of injecting to most recent medical care provider.
Patterns of healthcare utilisation
Eighty per cent of participants reported having accessed healthcare in the preceding 12 months (Table 1), including a significantly higher proportion of women than men (87% vs 76%, p<0.001). General practitioners (GPs) or medical centres were the providers most commonly last accessed (64%), followed by targeted providers (opioid substitution therapy prescribers, PWID-targeted primary healthcare centres, NSP, and detoxification/rehabilitation centres; 17%) and emergency departments (14%). Just under half (46%) reported that a general health issue had motivated them to seek medical care on the most recent occasion. Seventeen per cent reported seeking a medication (potentially including opioid substitution therapy), and 37% reported that services were accessed for both general health issues and medications. Eighty per cent had accessed the same medical care provider prior to the most recent visit. Thirty-five percent fully disclosed their injecting to their most recent healthcare provider and 34% reported partial disclosure and the remainder (31%) did not disclose their injecting drug use.
Correlates of disclosure of injecting to the most recent healthcare provider
A number of variables had significant univariable associations with disclosure of injecting drug use to the most recent healthcare provider (Table 1). Variables positively associated with disclosure were age, having accessed healthcare in the preceding 12 months, accessed targeted providers or community health centres most recently, prior use of the most recently accessed provider and being motivated to seek the most recent episode of care by ‘both medication and general health issues’. The effect of sexual identity revealed that compared to heterosexuals, bisexuals were more likely and homosexuals were less likely to disclose injecting. In terms of gender, women were more likely than men to fully/partially disclose injecting. The effect of ‘most recent medical care provider accessed’ indicated that those who most recently accessed a GP or medical centre were less likely than others to disclose injecting. Two other variables were found significant – one was ‘drugs injected most recently’ where drugs other than heroin users were less likely to disclose injecting; and the second was ‘drug treatment’ where participants who were not currently in drug treatment were less likely to disclose injecting.
Multivariable logistic regression revealed that participants who had accessed healthcare in the preceding 12 months (AOR 1.36; CI 1.06–1.75) or had previously visited their most recent provider (AOR 1.75; CI 1.36–2.25) were more likely than others to report fully or partially disclosing their injecting to the most recent provider. Conversely, effects of ‘drug injected most recently’ reveals that compared to recent heroin injectors, those who recently injected morphine and other opioids (AOR 0.58; CI 0.43–0.78), methadone/burpenorphine/suboxone (AOR 0.55; CI 0.40–0.76), steroids (AOR 0.54; CI 0.31–0.95), or other drugs (AOR 0.61; CI 0.40–0.95) were less likely to disclose injecting. The association with ‘treatment of drug dependence’ shows those who report either no (AOR 0.29; CI 0.22–0.38), or prior (AOR 0.74; CI 0.58–0.92) drug treatment were less likely than participants currently receiving treatment to disclose injecting. Those who most recently visited a GP or medical centre (AOR 0.65; CI 0.53–0.81) were less likely than others to disclose injecting.
Correlates of emergency department utilisation
Hospital emergency departments were the provider most recently accessed by 14% of respondents (Table 1), 78% of whom had accessed an emergency department in the preceding 12 months. Those who most recently accessed healthcare from an emergency department were more likely than those who did not to report recent imprisonment (AOR 1.48; CI 1.01–2.17), and having accessed an emergency department previously (AOR 0.67; CI 0.49–0.91). The effect of ‘drug injected most recently’ shows those who injected steroids were less likely (AOR 0.15; CI 0.04–0.62) than those who injected heroin to report accessing emergency department for their last episode of medical care. Participants who most recently accessed a healthcare facility seeking medication only (AOR 0.51; CI 0.33–0.79) or both medication and general healthcare (AOR 0.77; CI 0.58–1.02) were less likely to report accessing emergency department most recently than those seeking general healthcare.
The proportions of ANSPS participants reporting accessing healthcare in the preceding 12 months and previously accessing their most recent provider give some indication of continuity of care for this group. However, alarmingly only one-third of those who accessed healthcare on the most recent occasion for a general health issue reported fully disclosing their injecting drug use and similar proportion reported not disclosing at all. Participants’ likelihood of disclosing their injecting drug use to the most recent provider who they had previously accessed indicates the benefit of continuity of care. Intriguingly, GPs and medical centres were the most common sources of recent care, although the majority of participants who accessed these providers did not disclose their injecting. While the recent healthcare access, continuity of care and substantial GP involvement are encouraging, the use of an emergency department for the most recent occasion of care by 14% of participants remains a concern. Efforts should be made to reduce this utilisation of emergency department by providing alternative service.
Correlates of disclosure of injecting to most recent medical care provider.
Prevalence of healthcare utilisation in the past 12 months among ANSPS participants broadly reflects that of the general Australian population. Although the items are not directly comparable, 81% of Australians aged ≥15 years in the 2009 Australian Patient Experience Survey reported consulting a GP at least once during the preceding 12 months.22 A similar study on access to healthcare by PWID at an NSP in Pittsburgh, US, found that 67% had visited a GP in the preceding 12 months,23 with financial difficulty being the main barrier to access. In comparison to this finding, and given documented barriers to healthcare access among PWID,6,24,25 our results, which suggest approximately equivalent access, are encouraging and potentially reflect Australia's universal healthcare system. However, substantial research continues to document the poor health status of PWID,1,26–28 indicating that such equivalence of access does not necessarily result in equivalent health outcomes.
Substantial non-disclosure of injecting drug use was evident in our study; just one-third of participants reported full disclosure to the most recent provider. Disclosure is likely to facilitate more accurate diagnoses and to avoid potential interactions between illicit drugs and prescribed medications. Moreover, while non-disclosure may be motivated by a legitimate fear of stigmatisation or discrimination,29 PWID who fail to disclose may be at risk of being perceived as dishonest or manipulative by providers, potentially compromising the patient-provider interaction. That 64% of participants reported last accessing a GP or medical centre is encouraging, given previous documentation of low rates of GP comfort and confidence with this population.30,31 However, these participants were less likely than those who most recently accessed other providers to disclose their injecting drug use. Clearly there are important public health implications of this substantial non-disclosure, as given that PWID are at the heightened risks of blood-borne viral infections and other medical and mental health disorders, which may go unrecognised. The onus of disclosure must nevertheless fall partially to providers, whose responsibility it is to provide an environment in which PWID feel confident that disclosure will not result in discrimination.32 Even so, providers could usefully assume that some clients may choose to disclose only selected information regarding their drug use history and risk profile.33
Correlates of hospital emergency department utilisation at most recent presentation.
Previous research suggests that providing healthcare in settings where trust is already established, where PWID perceive little or no stigma and where care is provided opportunistically, better meets the needs of this group.2,34 Many PWID regularly utilise and trust NSPs and opioid substitution therapy clinics,35 creating a potential environment for opportunistic and ongoing healthcare provision. The provision of opportunistic care in such settings also has the potential to attract PWID who may avoid seeking care until perceived need is high.36 Providing basic and preventive healthcare from NSPs and opioid substitution therapy clinics may reduce utilisation of emergency departments and acute medical care services by PWID.37 In 2007, 33% of NSPs in the US reported providing onsite medical care, and 7% provided buprenorphine treatment.38 In Australia, the Kirketon Road Centre, established in Sydney's Kings Cross in 1987, is perhaps the best known example of a primary healthcare for PWID.39 This service provides a combination of NSP, primary healthcare, opioid substitution therapy and a range of specialist services. In 2001, the Victorian State Government established primary healthcare services for street-based PWID in five areas of Melbourne with high rates of drug use40 and in 2006, the Redfern Harm Minimisation Clinic was established in Sydney to provide primary healthcare via an enhanced NSP model.41
A study of PWID aged 18–29 years in New York found that 21% of participants used an emergency department the last time they received care,42 a higher proportion than among our sample (14%). However, given that Australia, unlike many other settings, provides universal healthcare, it is of concern that emergency departments were the most recent healthcare provider for 14% of our participants. Although recent emergency department users were less likely than others to report accessing that provider previously, three-quarters reported having accessed an emergency department in the past. High rates of emergency department presentations have previously been documented among illicit drug users,43 for issues including overdose,44 injecting-related injuries and infections9 and drug and alcohol-related injuries.45 The resource implications arising from emergency department utilisation can be reduced by providing accessible, acceptable and continuous primary healthcare services tailored for PWID.37,46 In our study, participants with a history of imprisonment were more likely than others to report recent emergency department use. This could reflect the consequences of imprisonment, including unintended overdose immediately after release,47,48 loss of continuity of healthcare and of stable housing, and difficulties finding employment because of a criminal record.47 Alternatively, imprisonment could be an indicator of more severe drug dependence and more chaotic lifestyle.
Our study has several limitations. The response rate among this sample was 41%, which is within the range of all ANSPS survey years (38%–60%).49 While the overall profile of participants is typical of the broader Australian NSP population,18 the external validity of our results cannot be assessed. Although the self-completion of the ANSPS questionnaire reduces social desirability bias,50 the brief and self-complete format necessarily precludes detailed explanations of complex constructs such as disclosure, individual interpretations of which are likely to vary. The associations demonstrated here between healthcare access and demographic and risk characteristics are correlational only; the cross-sectional nature of our data precludes definitive statements on the temporal or causal nature of any such relationships. Participants may have been affected by the primacy bias51 in responding to the most recent medical care facility they accessed and may have opted to select response options provided by the questionnaire rather than to name other facilities, such as community pharmacy, which were not specifically listed.
Despite the existence of a universal healthcare system and long-standing bipartisan political support for harm reduction illicit drug policy, considerable social stigma remains associated with injecting drug use in Australia. Our finding of substantial non-disclosure of injecting drug use to the most recent healthcare provider suggests ongoing concerns in relation to stigma and discrimination. Non-disclosure of injecting was highest among those who most recently accessed care from a GP or medical centre – the predominant source of recent healthcare accessed by our sample. Given the potential implications of non-disclosure for quality of care, and our finding that a significant minority of PWID utilise emergency departments for care, the provision of primary healthcare through existing services such as NSPs and opioid substitution therapy clinics which target PWID has the potential to alleviate these concerns. Such arrangements, however, should in no way discourage PWID from accessing mainstream healthcare. While PWID in our sample reported similar rates of healthcare utilisation to the general population, the significant mortality and morbidity associated with injecting drug use,7,52 and ongoing barriers to disclosure of drug use and continuity of care3,6,10 indicate a need to provide this population with multiple entry points to the healthcare system.
The ANSPS is funded by the Australian Government Department of Health and Ageing. The views expressed in this publication do not necessarily represent the position of the Australian Government and the Australian Government played no role in the analysis, manuscript drafting or decision to submit for publication. The Kirby Institute is affiliated with the Faculty of Medicine, University of New South Wales. Lisa Maher is supported by a National Health and Medical Research Council Senior Research Fellowship. The authors wish to acknowledge the generosity of survey participants and the support of participating Needle and Syringe Programs, the ANSPS National Advisory Committee, and laboratory staff at St Vincent's Centre for Applied Medical Research and the New South Wales State Reference Laboratory for HIV at St Vincent's Hospital, Sydney.
This paper was submitted on behalf of the Collaboration of NSPs(2011): Australian collaboration of Needle and Syringe Programs (2011): ACON; Anglicare Tasmania; Albury CHC; Barwon Health Service; Biala NSP; Cairns NSP; Central Access Service; Central Coast Harm Reduction Services; Clarence CHC; DASSA; Directions ACT; First Step Program; Health ConneXions; Health Information Exchange; Health Works; Hindmarsh Centre; Hunter Harm Reduction Services; Inner Space; Kirketon Road Centre; Kobi House; North Coast Harm Reduction Services; North Richmond CHC; Northern Territory AIDS Council; NUAA; Nunkuwarrin Yunti Inc.; QUIHN NSP Services; Redfern Harm Minimisation Centre; Salvation Army Launceston; SAVIVE CNP Services; SHARPS; Sydney West NSP Services; TasCHARD NSP Services; Townsville ATODS; WA AIDS Council; Wagga Wagga CHC; WASUA and West Moreton NSP.