To our knowledge this is the first large-scale study to compare HIV clinics located in different areas of Europe in terms of a wide range of service delivery features. We chose to compare EuroSIDA study clinics in a group of five countries – Belarus, Estonia, Lithuania, the Russian Federation and Ukraine – to EuroSIDA study clinics elsewhere in Europe because of previous findings of poorer health outcomes for study participants in this region [15–17]. Our findings are consistent with the hypothesis that some clinic characteristics may influence patient outcomes. There were marked differences in how the East Europe group of clinics handled issues such as the initiation of ART in asymptomatic patients and the provision of resistance testing. Furthermore, the East Europe clinics as a whole had a smaller array of services relating to some aspects of viral hepatitis control, tuberculosis control and screening for other health issues.
At the time the clinic survey was conducted, World Health Organization treatment guidelines indicated that ART should always be initiated in HIV-positive people when CD4 cell count levels dropped below 500 cells/mm3 , while European AIDS Clinical Society guidelines recommended using a lower CD4 threshold of 350 cells/mm3 . The clinic survey revealed that at a significantly larger proportion of clinics in East Europe than elsewhere, it was standard practice to delay ART until the CD4 level was below 350 cells/mm3. In light of what is now known about early ART initiation having an important protective effect on the immune system, it is reasonable to speculate that having a lower CD4 threshold for initiating ART in asymptomatic patients at East Europe clinics may have contributed to poorer patient outcomes. Patients’ health also may have suffered because of a lack of resistance testing, which was provided by smaller proportions of East Europe clinics both before the initiation of ART and upon treatment failure.
Chronic hepatitis B and hepatitis C disease both have emerged as major health issues for people living with HIV in recent years, with shared routes of transmission for all three diseases accounting for high levels of HIV/HBV coinfection and HIV/HCV coinfection in some populations . In our study, only half of clinics in the East Europe group had access to direct-acting antivirals for hepatitis C treatment, compared to 86 % of clinics elsewhere. DAAs stand apart from earlier generations of HCV treatment for their high cure rates, and the price of the newest, most effective DAAs has raised widespread concern about financial inaccessibility for patients in resource-limited settings and even in resource-rich settings [21, 22]. It is not known whether a lower proportion of EuroSIDA East Europe clinics reported having access to DAAs because of their high cost or for other reasons, but cost seems likely to be a factor as well as patient selection, with many in need being people who inject drugs . The finding points to a need to further investigate differential use of DAAs in HIV/HCV co-infected populations across countries as a step toward determining how liver-related morbidity and mortality in these populations can be reduced.
Study findings for TB screening raise concerns for the entire European region, with fewer than two-thirds of clinics in either study group reporting screening. TB is one of the most common AIDS-indicative diseases diagnosed in the WHO European Region , and HIV clinical protocols for the WHO European Region call for all PLHIV to be screened for TB . Fairly large proportions of EuroSIDA clinics in both East Europe and non-East Europe countries appear to not be implementing this guideline. It is possible that the consequences are more pronounced in the East Europe countries, given the high burden of TB and multidrug-resistant TB in those countries . The low level of integration of TB treatment into HIV care and treatment in the East Europe study region – with only 22 % of clinics reporting this to be the case – raises serious concerns in light of what is known about the benefits of an integrated clinical approach to TB/HIV co-infection .
The East Europe clinics lagged behind other clinics in regard to screening for cardiovascular disease, tobacco use, alcohol consumption and drug use, all issues with important implications for PLHIV. Cardiovascular disease is a major cause of non-HIV-related mortality in PLHIV populations [27, 28], and people who smoke tobacco have an elevated risk of cardiovascular disease . Alcohol consumption exacerbates the effects on the liver of HBV, HCV and other liver diseases , as well as potentially having other negative health effects [31, 32]. While screening for drug use is considered a good practice in many clinical settings, it warrants special consideration in settings where a major pathway for HIV transmission is injecting drug use. This is the case in the East Europe EuroSIDA clinics, with 38 % of the 1370 East Europe study participants who contributed data to this study reporting injecting drug use as their mode of exposure to HIV. In this context, the reported absence of routine screening for drug use at 42 % of East Europe clinics stands out as an issue that may have important implications for patients’ health.
Taken together, our study findings raise important questions regarding whether the availability of a range of services at clinics caring for PLHIV might have an impact on morbidity and mortality. At a time when antiretroviral therapy is widely available in high-income countries and is becoming increasingly available in low- and middle-income countries, these questions are important to take up since they reflect a growing awareness that antiretroviral therapy alone is not sufficient to safeguard the long-term health of PLHIV. Surprisingly, in light of the advanced state of HIV management in some regards, there appears to not be a large evidence base regarding which services a clinic should provide to its HIV-positive patients. The comprehensive HIV care model, with a healthcare team coordinating primary care, HIV care, and other specialist care as well as psychosocial and social services, has long been championed in the United States  and has likely influenced the development of many multifaceted HIV clinical initiatives worldwide. However, there is scant evidence regarding the relationship between specific HIV clinic characteristics and patient health outcomes.
A 2006 Cochrane review of studies that assessed various elements of the “setting and organisation of care” for PLHIV found an association between case management and decreased mortality, but concluded that the small evidence base in this field was not sufficient to determine an ideal set of clinic characteristics . Since the publication of the Cochrane review, little new evidence has emerged. A 2009 retrospective cohort study of PLHIV at health facilities for US military veterans found that patients attending clinics that integrated hepatitis, psychiatric, psychological and social services into HIV clinical management were more than three times as likely to achieve viral suppression on ART than patients attending HIV clinics without integrated services .
A key challenge in conceptualising and conducting meaningful research in this domain is the setting-specific nature of how health services are organised, funded, managed and governed. For example, the issue of whether or how to integrate HIV clinical services with other clinical services has been highlighted particularly in the context of efforts to provide HIV care in severely resource-constrained settings, with studies in sub-Saharan Africa and elsewhere examining different service delivery models  and undertaking different forms of service integration in order to provide HIV services alongside other services such as tuberculosis management  and reproductive health care . This research may be of limited value to health system decision-makers in settings with much higher physician-patient ratios or with strong referral systems linking long-established HIV clinics to other services. Clearly there are major differences in regard to which health service delivery models are predominant in different countries, with some differences largely attributable to resource limitations in poorer countries.
Nonetheless, the identification of key elements of successful patient management in diverse settings presents opportunities to explore ways in which these elements may or may not be uniquely dependent on specific features of the local and national health system and the social, political and economic context. It is entirely possible that some service delivery innovations in sub-Saharan Africa may be relevant to health systems in Western Europe, and vice versa. By calling attention to ways in which two regional groupings of clinics for PLHIV in Europe differ from each other, the EuroSIDA clinic survey findings serve as an invitation for clinics and regions with suboptimal patient outcomes to investigate whether the adoption of practices from other settings may be beneficial. Thus the lack of generalisability of a study of the characteristics of health service delivery in a specific group of clinics may be offset by its potential to highlight issues warranting further setting-specific research including operational research on service delivery modifications in “real world” clinic populations.
In sum, our study makes a unique contribution to the issue of HIV management by exploring whether the characteristics of clinics vary across two groups of European countries that have had marked differences in patient outcomes in a large observational study cohort. The identification of some statistically significant differences in clinic characteristics cannot be interpreted as evidence that one or more of those differences is causing the observed differences in patient outcomes. However, findings suggest that it may be beneficial to conduct further research on the potential health impact of clinic characteristics such as the CD4 threshold for ART initiation; ART resistance testing practices; HCV treatment standards; and the provision of screening for non-HIV-related conditions including alcohol and drug dependency. Policy-makers should consider research on service delivery factors alongside other types of research on biomedical factors, health system factors and the influence of the social, political and economic context in order to optimally configure health care for PLHIV at the national and subnational levels.
This study has the following limitations. While our survey addressed major aspects of the clinical and non-clinical care of people living with HIV, there may be other aspects of patient care that have implications for patient outcomes. The study utilises survey data that were reported by clinic representatives whose responses to questions might not reflect what actually happens in clinical practice. Clinical decision-making about some of the issues addressed in our study can be expected to vary in accordance with individual physicians’ preferences as well as patient characteristics such as the nature of symptoms and severity of disease. Furthermore, study findings reflect self-reporting, and it is not possible to verify the accuracy of the information reported. Respondents may have made errors or altered their responses to suggest a higher level of compliance with guidelines. The survey was conducted in English, which may have affected how questions were interpreted by some respondents. Findings have limited generalisability because the EuroSIDA clinics constitute only a small proportion of HIV clinics in European countries. Generalisability is also affected by key characteristics of the responding EuroSIDA clinics: half were university clinics, almost half were government-affiliated, and many were located in capital cities. Practices at these clinics are therefore not necessarily representative of HIV management in the European region overall.