This site is intended for health professionals only
MB BS BMedSci MSc
PhD DHMSA DIC
DLSHTM FAFPHM FFPHM
Coordinator and Director NPMS-HHC
Coordination and Analytic Centre
Chelsea & Westminster
Honorary Consultant in Public Health Medicine
Public Health Laboratory Service
Joint Departments of Epidemiology
Biostatistics and Occupational Health
Specialist in Public Health
Direction de la Santé Publique de Montreal-Centre
Hôpital Géneral de Montréal
The introduction of highly active antiretroviral therapy (HAART) into routine clinical practice has had a profound impact on disease progression in HIV-infected individuals, resulting in dramatic reductions in HIV-related morbidity and mortality.(1–7) Despite side-effects(8) and problems with resistance to antiretroviral therapy (ART),(9–11) increasing numbers of people with HIV are receiving treatment across the world.(12) Improved treatment, combined with the continued transmission of HIV, has resulted in increasing numbers of individuals living with HIV around the world.(13)
ART was first introduced into routine clinical practice in 1987. Although AZT monotherapy has limited effectiveness, it was a cost-effective intervention. (14) In subsequent years, other nucleoside reverse transcriptase inhibitors (NRTIs) were developed, as were the protease inhibitors (PIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs), all of which have been introduced into clinical practice. While treatment with dual therapy commenced in 1994/95, triple therapy was put on the therapeutic map in 1996, when the results of the first efficacy trials on 2NRTIs + 1PI became available. Subsequently, firstline combinations of 2NRTIs + 1NNRTI have become more popular, especially in Europe, but many central issues, such as which is the optimal starting drug combination and when it should be initiated, are yet to be resolved. The development and introduction of nucleotide reverse transcriptase inhibitors and fusion inhibitors has increased the therapeutic options available to practitioners still further, and new compounds from existing classes of drugs are also being produced.
Individual cost of HIV service provision
The cost of these drugs and their impact on health budgets have been longstanding topics of discussion, (14–16) but since 1996 the rising costs of HIV drugs have become an especially persistent concern.(17,18) In the UK, after the introduction of HAART in 1996, people with AIDS (CDC Stage C) continued to use hospital inpatient services most intensively.(19,20) However, their use of inpatient services decreased over time, while asymptomatic patients (CDC Stage A) and patients with symptomatic non-AIDS (CDC Stage B) used more outpatient services.
The cost of treating asymptomatic individuals and people with symptomatic non-AIDS has not changed substantially in the UK since 1996 (see Table 1a), though HAART has increased the annual costs for these patient groups when compared with the period before 1996.(15,16) The annual costs for treating AIDS patients in the UK have decreased by £2,000 (£1 ≈ €1.45) since 1996, which is in part due to a reduction in the use of inpatient facilities with an increased use of outpatient services. On the whole, increased expenditure on HAART has been offset by a decrease in inpatient-related costs, a phenomenon observed earlier in the epidemic,(6,15,21) as well as in other industrialised countries.(22) Drug costs may, however, increase further with the introduction of new classes of drugs and the increasing need for the use of salvage therapy or Mega-HAART.(23)
Based on the annual direct costs by stage of HIV infection and the average time HIV-infected people spent in respective stages of HIV infection while using health services,(1) the average nondiscounted lifetime costs(24) for treatment with 2NRTIs + 1NNRTI was £135,420 (range £59,973–184,030), compared with £181,613 (range £82,093 –240,119) for treatment and care with 2NRTIs + 2PIs.
Population costs of HIV service provision
The estimated number of HIV-infected individuals who used NHS services during the study period increased from 13,400 in 1996 to 17,981 in 1999 – a 25% increase over four years.(21,25,26) During the same time period, estimated annual direct population costs increased from £209m (£176m–£253m) to £266m (£226m–£321m; see Table 1b). As with any illness, HIV infection also generates indirect costs.(27) Estimates of indirect costs in 1999 varied between £53m and £133m.(28) If a public sector perspective is adopted and disability-related unemployment payments are included, costs estimates ranged from 28% to 34% of total direct costs, assuming a 100% loss in production; if a 50% loss of production is assumed, indirect costs estimates ranged between 20% and 25% of direct costs. From a societal perspective, which excludes disability-related unemployment payments, indirect costs estimates varied between 22% and 27% of total direct costs, assuming a 100% loss of production, and between 13% and 16% of total direct costs when assuming a 50% loss of production (see Table 1c).(28)
As HAART is more cost-effective than dual or monotherapy(29,30) and the number of newly diagnosed HIV-infected people has continued to increase in England(11) and many other countries, the number of HIV-infected people needing to use medical services is also likely to continue to increase,(28) with a resultant requirement for additional resources to maintain or improve existing standards of care.(28)
Changes in therapeutic and diagnostic interventions are likely to remain an important feature of the management of HIV infection, with potentially profound implications for the provision of HIV services. Therefore multicentre information systems are being set up in various countries to monitor prospectively the effectiveness, efficiency, equity and acceptability of HIV service provision.(31)
In the UK, the National Prospective Monitoring System on the use, cost and outcome of HIV service provision in UK hospitals–HIV Health Economics Collaboration (NPMS–HHC) arose from a need for reliable information about existing patterns of service use, costs and outcomes that could be used to improve local, regional and national services within budgetary constraints in pursuit of “value for money” and quality in service provision for people with HIV infection.(28,32) The need for such data has not diminished (Kinghorn: personal communication; 2003). Data collected include the use of services, drug treatment, severity of HIV infection when services are used,(33) occurrence of disease progression or death and the type of opportunistic illnesses,(6) the cost of service provision,(33 )and views on the acceptability of services provided by those who use these services.(34) Analyses and feedback occur at aggregate and clinic-specific levels.(32)
Similar surveillance systems have been established in France (DMI 2) and the Netherlands (Dutch HIV Monitoring Foundation), and are being established in Canada (RISQ and HIIP).(35–38) Unlike the UK system, the surveillance systems being developed in these countries are being financed through public funds.(32) All of these systems collect data in a standardised format prospectively in participating HIV clinics.
Given that most randomised controlled trials of new drugs or drug combinations currently use small sample populations, often utilising surrogate markers as endpoints, and therefore are of relatively short duration, and given the problems of sideeffects and viral resistance associated with the longterm use of HAART, observational databases are likely to become more and more important to assess the effectiveness and cost of these drugs, and their various combinations, within routine clinical settings.(18,31) Apart from being potentially useful for direct patient management, data collected through such schemes provide reliable and contemporary information on the impact of interventions at clinic level, as well as providing comparative information for policy formulation and evaluation at regional, national and international levels.