icon-folder.gif   Conference Reports for NATAP  
 
  9th International
Workshop on HIV and Aging
13 and 14 September 2018
New York City, NY
Back grey_arrow_rt.gif
 
 
75% Projected to Be Over 50. 85% of HIV+ projected to Have Cadiovascular Disease in USA, 30% malignancy, 23% diabetes. Care Costs Increase by 40% due to Comobidities.
 
 
  Projections of non-communicable disease and health care costs among HIV-positive persons in Italy and the U.S.A.: A modeling study
 
Download the PDF here
 
2017
 
The issues of ageing with HIV are gaining momentum amongst global HIV stakeholders.. To ensure the highest level of care is maintained for these populations, multi-disciplinary patient management and enhanced collaboration and linkage between HIV specialists, geriatric medicine and primary care must be developed. Although the pathophysiological mechanisms of NCDs in the context of HIV are not fully understood, it likely involves complex interactions between traditional risk factors, including smoking, diet, and exercise, and HIV-specific risk factors, for example long-term immune activation and inflammation and toxicity related to long-term ART use [3-8]. Evidence based approaches on effective prevention interventions and treatment protocols will be vital to mitigate this growing burden and offset improvements in HIV care and quality of life achieved to date. The ageing HIV-positive population in Italy and the USA will have major implications for HIV care. Our forecasts suggest that three-quarters of HIV-positive patients on ART will be aged >50 in both countries by 2035, resulting in an increased NCD burden in this population. The major drivers of NCD burdens will be hypertension, dyslipidaemia, diabetes and malignancies. These shifts will have considerable implications for direct HIV care costs, with average care costs attributable to NCD treatment expected to double in Italy and increase by 40% in the USA. Evidence based approaches on effective prevention interventions and treatment protocols will be vital to mitigate this growing burden and offset improvements in HIV care and quality of life achieved to date.
 
The model predicts that mean age of HIV-positive patients will increase from 46 to 59 in Italy and from 49 to 58 in the USA (Fig 2) in 2015-2035. The proportion of patients ≥50 years old is predicted to increase from 30% and 39% in 2015 to 76% and 74% by 2035 and the proportion ≥65 years old from 5.5% to 29.3% and from 5.4% to 26.9% in Italy and the USA in 2015-2035, respectively (Fig 2). Even if mean age of new patients remained at 2010 level, mean age in 2035 would be 58 in Italy and 57 in the USA. If HIV incidence declines more rapidly or slowly, mean age by 2035 is predicted to rise to 60 and 58 in Italy and 58 and 58 2in the USA, respectively (Section 7 in S1 File).

years

In 2015, we find that the majority of patients in Italy and the USA were diagnosed with either no NCDs (35% in Italy, 29% in the USA) or 1 NCD (42% in Italy, 37% in the US) (Fig 3A). The NCD burden is expected to increase over time, with a large proportion of patients diagnosed with either 2 NCDs (31% in Italy, 25% in the US) or ≥3 NCDs (29% in Italy, 44% in the US) by 2035 (Fig 3A).
 
The increasing burden of NCDs will be driven by moderate CVD (hypertension, dyslipidaemia), diabetes and ever being diagnosed with a malignancy in both countries (Fig 3B). In 2015, moderate CVD contributes the greatest burden, affecting 60% and 61% of all HIV-positive patients on ART in Italy and the USA, respectively, followed by diabetes (9% and 12% respectively) and malignancies (6% and 14%). The trend is expected to persist in 2035, with the risk of being diagnosed with moderate CVD (affecting 85% of HIV-positive patients in Italy and 84% in the USA), diabetes (affecting 27% of patients in Italy and 23% of patients in the USA), or ever being diagnosed with a malignancy (affecting 16% of patients in Italy and 30% of patients in the USA) increasing.
 
When comparing the proportion of those with moderate (either dyslipidaemia or hypertension) and serious (ever having a MI or stroke) CVD in both countries, HIV-positive patients in the USA are expected to have double the burden of serious CVD (Fig 3B) compared to Italy by 2035 (21% in the USA versus 10% in Italy) with similar burdens of moderate CVD expected in both populations.
 
Differences in the demographic shift and increase in NCD burden (faster in Italy compared to the USA) appear to be driven by the differing HIV incidence. Higher HIV incidence in the USA will likely resulting in a steady addition of new, young patients to care compared to Italy where the existing population ages steadily and fewer new patients join care. Differences in serious CVD between the two settings are driven by the higher age-specific prevalence and incidence of serious CVD observed in the USA compared to Italy. Annual direct NCD-related care costs are expected to increase between 2015 and 2035 in both countries (Fig 4). In Italy, an estimated 11.1% of total HIV direct medical care costs are associated with NCD treatment, and this is projected to increase to 22.8% by 2035 treatment. The highest contribution to overall NCD treatment relate to dyslipidaemia (3.7% in 2015 and 5.1% in 2035) and CKD (3.6% in 2015 and 10.2% in 2035), with the greatest increase relating to CKD. In the USA, direct NCD treatment costs account for 39.6% of total HIV care expenditure in 2015 treatment, and is projected to increase to 55.7% by 2035 treatment. In dollar terms, this would amount to the average cost per patient in the USA of $50,032 in 2015 ($19,812 due to NCDs) and $68,270 in 2035 ($38,050 due to NCDs), assuming no discounting. The highest contribution to overall direct NCD treatment costs are dyslipidaemia (13.3% in 2015 and 16.6% in 2035) and hypertension (12.9% in 2015 and 17.2% in 2035), and the greatest increase relates to costs of CKD (5.8% of total care cost in 2015 to 12.0% in 2035).
 
Discussion
 
The ageing HIV-positive population in Italy and the USA will have major implications for HIV care. Our forecasts suggest that three-quarters of HIV-positive patients on ART will be aged >50 in both countries by 2035, resulting in an increased NCD burden in this population. The major drivers of NCD burdens will be hypertension, dyslipidaemia, diabetes and malignancies. These shifts will have considerable implications for direct HIV care costs, with average care costs attributable to NCD treatment expected to double in Italy and increase by 40% in the USA. Evidence based approaches on effective prevention interventions and treatment protocols will be vital to mitigate this growing burden and offset improvements in HIV care and quality of life achieved to date.
 
The issues of ageing with HIV are gaining momentum amongst global HIV stakeholders. UNAIDS published its "Gap report" [18] in 2014 focusing on older HIV-positive population and regional and national HIV guidelines are increasingly incorporating recommendations on the management of NCDs in HIV care [19-21]. Recently published guidelines in Italy [19], current European AIDS Clinical Society [20] and USA [21] guidelines for example provide guidance on management of NCDs in HIV-positive patients. As HIV care focus shifts from treatment of opportunistic infections to the long term prevention, screening and treatment of NCDs, particular attention needs to be given to optimal ART choice [3], management of drug-interactions between ART and NCD medication, and adherence in light of increasing polypharmacy [22]. The need for multi-disciplinary patient management with a focus on geriatric principals, personalised treatment protocols, and prevention interventions (including guidance on modifiable risk factors such as smoking, diet, exercise and other risk factors) will also be vital.
 
Practical application of care guidelines is an important issue; and can only become a reality when there is suitable training in place for health care providers. The evolving care requirement raises the issue of accountability; with patients receiving treatment for HIV and NCDs, the majority of care may fall on primary care physicians or HIV-specialists, although in some countries patients are already referred to general population specialists for specific NCD care. Integrated training is one possible avenue whereby HIV specialists are provided additional extensive training in geriatric medicine [23]. Yet little is known about which prevention interventions work best in the ageing HIV-positive population where there are also barriers to the prevention of NCDs such as smoking cessation [24].
 
The trends observed in Italy and the USA are similar to forecasts made for The Netherlands [9]. Smit et al found that by 2030 70% of HIV-positive patients on ART will be aged >50 (our forecasts show 76% for Italy and 74% for USA by 2035 and 65% for Italy and 69% for USA by 2030), and that 84% will be diagnosed with ≥NCD (our forecasts show 89% for both Italy and the USA by 2035 and 85% for Italy and 88% in the USA by 2030). As with Italy and the USA, this NCD burden will be driven mainly by CVD and diabetes. The burden of multi-morbidity (>3 NCDs) is forecasted to be higher in the USA compared to Italy and The Netherlands (28% in The Netherlands, 29% in Italy and 44% in the USA in 2030). Our current estimates of mean age are also in line with national reports, with the USA reporting that around 42% of patients are aged >50 (39% in our study) [25] and Italy reporting 29% (30% in our study) [12].
 
The models capture the major age-related NCDs, including their common physiological pathways. They were parameterized using large and nationally representative cohort data [15,26], and scaled to national level using national HIV surveillance data. Out-of-sample checks ensured that the model results were robust in the short-term (Section 6 in S1 File). Cost estimates of NCD care and treatment were based on proven methodology in Italy [16,17] and standard health economic method using robust insurance data from the USA.
 
The model is limited by the fact that it assumes homogenous national co-morbidity risk and burden in Italy and the USA. In the absence of national data on NCDs, estimates from available cohorts were assumed to directly translated to national level. As ICONA enrolls only ART-naïve patients, the average patient is likely to have less advanced disease than a typical HIV-positive patient in Italy. Although a recent study showed that recently enrolled patients were representative of Italian HIV-positive people in the European Surveillance system, ICONA may slightly over-represent men who have sex with men and underrepresent people reporting other mode of transmission, older people and those with low CD4 counts [26]. US data were generally representative of an insured US population in terms of sex, race, age and geographical distribution [15]. However, they likely represented patients with the best access to health care and not reflect patients who rely on public insurance, such as Managed or Fee-for-Service Medicaid, or those with limited or no health insurance. Extrapolation to poorer or under-served, vulnerable populations need to be made with caution; who are known to be more complex with respect to comorbidity, in part due to social or behavioral factors [27]. Further, as entry into the data requires interactions with the health care system, they may have had a small bias towards patients with greater consumption of health care resources. This may result in an underestimation of both the burden of NCDs and health care needs.
 
Furthermore, the model focuses on simulating HIV-positive patients in care and on ART and does not include all the NCDs that can affect these population; in particular, undiagnosed NCDs and neurocognitive or central nervous conditions which are poorly collected in HIV cohorts. This make the results of this study a conservative estimate of the true NCD burden affecting HIV-positive people in Italy and the USA in the next 20 years. The prevalence of MI, stroke and malignancy refers to ever having been diagnosed with these NCDs. The model does not account for recurrent MI, strokes or malignancies nor costs associated with recurrent events. Instead it is assumed that treatment last on average one year following the incidence of these three NCDs. As further robust age-specific data on these NCDs becomes available the model can be expanded. Estimates for annual HIV-incidence also do not take into account the lag that can occur between becoming newly infected and initiating treatment. Furthermore, the model does not account for the introduction of "test and offer" ART policies which may reduce HIV and NCD incidence [3]. Nor does the model account for migration (which may impact the burden of both HIV and NCDs, as well as the care cascade [28]) or the introduction of long-acting antiretroviral drugs (which have demonstrated good tolerability in existing trials, on drug-interactions [29]). The cost analysis for Italy are limited by the fact that cost estimates from one Italian region are extrapolated nationally, that costs for MI and strokes in Italy are based on treatment costs for a wider range of chronic CVD events excluding not MI and strokes, that frequently recovery costs in HIV-positive patients are attributed to HIV not NCDs and that compliance to NCDs treatments in Italy is likely lower than recommended protocols, (resulting in a likely underestimation of true necessary treatment costs) [16]. This lower compliance is likely due to the lack of electronic prescription system for medication for chronic diseases in Italy, meaning patients need to visit their physicians every month or two to renew prescription, affecting secondary adherence.
 
Finally, it is difficult to make direct comparisons between both settings due to the inherent differences in the populations, healthcare systems (mainly private in the USA and public in Italy), data sources (USA from privately insured patients, Italy ART-naïve patients) and health care systems cost differences. However, the results provide a first robust country-specific insight into the future HIV care challenges in both countries.
 
Conclusions
 
Italian and US HIV-positive populations are ageing, increasing the NCD burden related direct treatment costs. To ensure the highest level of care is maintained for these populations, multi-disciplinary patient management and enhanced collaboration and linkage between HIV specialists, geriatric medicine and primary care must be developed. Further predictions must be made for populations where future co-morbidity estimates (and consequent care strategies) cannot be extended to, such as sub-Saharan Africa.
 
Introduction
 
Recent efforts to improve HIV care cascades have highlighted engagement levels in care globally and the challenges with starting and maintaining HIV-positive people on antiretroviral therapy (ART) [1]. With effective ART and extended life expectancy [2] new challenges are arising around the long-term management of patient relating to the increasing burden of non-communicable diseases (NCDs) [3]. Although the pathophysiological mechanisms of NCDs in the context of HIV are not fully understood, it likely involves complex interactions between traditional risk factors, including smoking, diet, and exercise, and HIV-specific risk factors, for example long-term immune activation and inflammation and toxicity related to long-term ART use [3-8].
 
A recent model-based study provided the first insight into the clinical implications of an ageing population in The Netherlands [9]. It showed that the mean age of HIV-positive patients is expected to increase rapidly in the coming 15 years, accompanied by a sharp increase in the burden of both NCDs and use of co-medication (in addition to ART). The study highlighted the clinical implications of this disease shift, estimating that 40% of patients will likely experience complications due to contra-indications or drug-interactions between co-medication and recommended ART by 2030 [9]. Having robust national projections for the nature and scale of NCD burden and likely economic implications for HIV care will help inform national and regional guidelines regarding the screening, monitoring and treatment of NCDs within HIV care.
 
In this study we adapted an individual-based model of the ageing HIV-positive population on ART [9] to Italy and the United States of America (USA) by using real-world data to provide the first projections of the changing NCD burden and direct NCD treatment costs associated with treating these increasingly complex patients in those countries in the coming two decades.