Are we underestimating the proportion of virally-suppressed patients in the US?

This article is more than 11 years old. Click here for more recent articles on this topic

Several presentations at the recent 20th Conference on Retroviruses and Opportunistic Infections in Atlanta suggest that previous estimates of the proportion of people with HIV in the USA who are on antiretroviral therapy (ART) and with an undetectable viral load may have been too low and may be closer to the proportion virally suppressed in Europe.

Background: a walk through the cascades

The “HIV care cascade” is a way of calculating what proportion of HIV in a country or community is on ART and virally suppressed. Having a high proportion of people with HIV with undetectable viral loads is generally seen as critical to the success of ART to prevent HIV transmission and as an important component of programmes to reduce HIV incidence.

The care cascade calculation takes into account that having a high proportion of people with HIV essentially non-infectious is dependent on a chain of events happening, all of them at high frequency:

  • A high proportion of people with HIV need to be diagnosed, which implies frequent testing among high-risk groups;

  • A high proportion of the diagnosed need to be engaged in care, which suggests an easily-accessible healthcare system for all;

  • A high proportion of those in care have to be on ART, which suggests guidelines with high CD4 count thresholds for care (or none), and few financial or availability barriers to ART;

  • A high proportion of those on ART have to be virally suppressed, which suggests high adherence rates, good monitoring, and appropriate prescribing.

Reports in the last two years appear to show a large gap between the US and Europe in terms of the proportion virally suppressed. Last year at the IAPAC summit on antiretroviral-based prevention Dr Valerie Delpech of the UK’s Health Protection Agency showed that in the UK 53% of gay men with HIV have an undetectable viral load. Preliminary data suggest that the figure will be very similar for HIV-positive people in general, and indeed another study at the International AIDS Conference suggested that because incidence in gay men is higher than in other populations, the proportion of the HIV-positive population in general that is virally suppressed may be quite a lot higher than this, despite their tendency to be diagnosed later.


virological suppression

Halting of the function or replication of a virus. In HIV, optimal viral suppression is measured as the reduction of viral load (HIV RNA) to undetectable levels and is the goal of antiretroviral therapy.

retention in care

A patient’s regular and ongoing engagement with medical care at a health care facility. 

care cascade

A model that outlines the steps of medical care that people living with HIV go through from initial diagnosis to achieving viral suppression, and shows the proportion of individuals living with HIV who are engaged at each stage. 

representative sample

Studies aim to give information that will be applicable to a large group of people (e.g. adults with diagnosed HIV in the UK). Because it is impractical to conduct a study with such a large group, only a sub-group (a sample) takes part in a study. This isn’t a problem as long as the characteristics of the sample are similar to those of the wider group (e.g. in terms of age, gender, CD4 count and years since diagnosis).

linkage to care

Refers to an individual’s entry into specialist HIV care after being diagnosed with HIV. 

In contrast, two similar calculations for the US by the Centers for Disease Control have suggested that only 28%, or even 25%, of people with HIV are virally suppressed and even fewer of the most vulnerable groups, such as black gay men, heterosexual men and young people.

France, the UK and the US

A clue as to why, so far, ‘cascade’ calculations for the US have come out with such lower figures came from a study of the care cascade in France (Supervie).

This study, based on a large cohort of patients with HIV, calculated that 52% of people with HIV in France are on ART and with an undetectable viral load, and 56% of gay men, a little higher than the UK.

It estimated that 81% of people with HIV in the country are diagnosed (better than the UK estimate of 76%); found that 92% of those are in care; that 81% of those have been taking ART for more than six months; and that of them, 86% have a viral load below 50 copies/ml.

Counter-intuitively, the group with the highest rate of viral suppression is people who got HIV through injecting drug use, 66% of whom have an undetectable viral load; but this is because, as in the UK, France has done a good job of bringing down HIV infections in injecting drug users to a few per cent of the total, so most IDUs form an ageing cohort who are already in care. However IDUs also formed a disproportionately large part of the small number of people who are diagnosed but are not in care.

The French researchers cited the country’s health system for its relatively high number of virally-suppressed people. The French system is not free at point of demand and is insurance based; but unlike the USA, with some residency exceptions, people with HIV are entitled to recoup the cost of all their medical care.

Why are the US results so different? In most respects, the US health system actually performs just as well. For instance, in both countries, 81% of people with HIV are diagnosed; and though France performs better when it comes to the proportion with a viral load below 50 copies/ml (86% in France versus 77% in the US), in the US, more people with HIV who are in care are on ART (89% in the US versus 81% in France).

The big difference is the number who are linked to, and stay in, care after diagnosis. In France 92% of the diagnosed are linked to care (and an even higher proportion in the UK); in the US study cited, only 51% of the diagnosed thereafter attend clinics regularly.

Are the US calculations wrong?

Why? The assumption has been that inequalities built into the US healthcare system are to blame.

In the US, health care for HIV is covered by a complex system of healthcare benefits and entitlements. About a quarter of people with HIV are classed as uninsured, and people with HIV in general are 56% more likely to be uninsured than the general population.

The uninsured can still get HIV treatment via state-run AIDS Drugs Assistance Programs, which are funded by federal money under the Ryan White Care Act. In the past, a number of state ADAP programmes have run short of money and have sometimes placed people in need of HIV therapy on waiting lists.

The 2010 Affordable Care Act (‘Obamacare’) mandates that employer and private insurance schemes must cover long-term medical conditions and provides for the creation of insurance markets (state- or federally-administered) which will start in October, with a bridging insurance plan already in place. It also requires states to expand Medicaid, the main provider for people with disabilities, but the Supreme Court struck down federal powers to fine states that refused to comply.

What this means is that while the vast majority of US citizens can access HIV treatment, getting it is complex, can require satisfying stringent criteria, varies hugely by location, may be covered by several different schemes, often requires co-pays, and is currently in a state of flux. It also enshrines socioeconomic inequalities, with black people with HIV twice as likely as white people to be on ADAPs. There has been an assumption that this resulted in actual gaps in HIV care coverage.

However several studies from the US presented at the recent CROI conference suggested that instead it forced people to move from one provider to another or to space out medical visits.

Defining ‘retention in care’

The definition of ‘retention in care’, as used by the Centers for Disease Control paper that came out with the figure that 25% were virally suppressed, was the proportion of adults with HIV who received at least one medical care visit between January and April 2009.

When this finding was presented at the International AIDS Conference last year, it was suggested that many people on stable ART might attend appointments less often.

A study presented at the recent CROI conference (Horberg) by Kaiser Permanente (KP), the US’s largest private not-for-profit provider of HIV healthcare, suggested that the CDC ‘cascade’ calculation may considerably underestimate the proportion of people who are virally suppressed.

In particular, the CDC assumption had been that people who were not ‘retained in care’ could not be taking ART, but this might not be the case.

KP used its own database of 16,816 patients, which, because it provides coverage in general to the less-deprived populations, was largely male (87%) and older (average age 48 and 29% over 55). No data on ethnicity or sexuality were given.

KP used a broader definition of ‘retained in care’ (at least two visits in a year) and below 200 copies/ml as its definition of viral suppression. It also used a single measurement of viral load in any one year as its definition of viral undetectability rather than two consecutive ones.

Using these more liberal criteria, its estimate for the total number of diagnosed patients virally suppressed, at least in KP patients, was 60.2%. If the CDC estimate of 19% for the proportion of people with HIV who are undiagnosed is added, this would become 51% of all people with HIV – quite similar to France and the UK.

However the actual Kaiser figures for filled prescriptions and for viral undetectability showed that more people were prescribed ART and were virally suppressed than were defined as being ‘retained in care’. Using the proportion of all patients with a viral load under 200 copies/ml at the last test as its criterion for viral suppression rate, rather than the proportion counted as ‘being in care’, the result was that  80% of diagnosed KP clients with HIV were virally suppressed. Extending that to the whole population and adding in 19% undiagnosed, that would mean two-thirds of the HIV positive population had an undetectable viral load – or would do if they were all like KP clients.

Seattle: adding in the lost-to-care

The KP paper may overestimate the proportion of people in care and virally-suppressed as much as the CDC underestimates it, but a study using real figures from Seattle (Dombrowski) supported its findings to some extent.

This found that diagnosis rates were similar or higher than the CDC estimate. But it also did something the CDC did not, which was that by using data from real-life case investigation of people who apparently dropped out of care, it determined that about 10% of people classed as not receiving care had in fact moved out of area or away from the providers included in the study, and were in fact in care. It also found by investigation that another 10% of people who were listed as not being ‘retained in care’ because they did not have a CD4 or viral load test result recorded, were also in care: the issue in these cases was to do with medical note recording or of physicians deciding to monitor less often, not actual attendance.

Using these figures the Seattle team calculated that 79% of all people with HIV living in King County, Washington state (Seattle’s county) were linked to care as opposed to 66% in the CDC calculations, and 71% retained in care as opposed to 35%. This meant that the proportion of people with HIV who were virally suppressed was 57%, as opposed to 25% in the CDC figures. Again, very similar to the European figures.

New York: increases in viral suppression

Similar figures were obtained by a study from New York city (Stadelmann), although once again, these local figures may not be representative of all areas.

This paper used as its definition of viral suppression two successive viral load results under 400 copies/ml in a year, not dissimilar to the CDC studies.

However it did not assess linkage to or retention in care, and thus made no assumptions about whether only people classed as being retained in care could be assessed for viral load suppression.

It found that 52% of diagnosed HIV-positive people in New York were virally suppressed. If the estimated 19% of undiagnosed people is added in, this becomes 44% - lower than in Europe but a lot higher than the CDC estimate.

These figures are from 2010-2011 and represent a considerable increase from 2006-2007, when 38% of diagnosed, or 31% of all, people with HIV had an undetectable viral load.

The New York study also assessed the proportion of diagnosed people with persistently high viral loads (two successive measures of over 100,000 copies/ml), and who would therefore be very infectious. It found that this proportion had declined from 7.4% of diagnosed people with HIV in 2006-2007 to 4.6% in 2010-2011. This does not imply a proportional decrease in very high viral loads in the whole HIV positive population, though, as high viral loads in the undiagnosed would be unaffected by ART.

Health inequalities had their effects on viral load undetectability: whereas 20% of the HIV positive population was white, white people formed only 9% of those with a persistently high viral load; conversely, though 45% of the patient population was black, they formed 54% of those with a persistently high viral load.


What these papers show in general is that the complexities of the US healthcare system make it very difficult to measure the true proportion of people with HIV in the country who are taking ART and are virally undetectable. The proportion may be much higher in some areas and for some populations, and the criteria used by the CDC may be too strict, especially as we move to less-frequent monitoring.

But it also shows that even in areas with good coverage, health inequalities remain.


Supervie V and Costagliola D. The Spectrum of Engagement in HIV Care in France: Strengths and Gaps. Twentieth CROI conference, Atlanta. Abstract 1030. 2013. See abstract here and poster here.

Horberg M et al. HIV spectrum of engagement cascade in a large integrated care system by gender- age and methodologies. Twentieth CROI conference, Atlanta. Abstract 1033. 2013. See abstract here and poster here.

Dombrowski JC et al. An encouraging HIV care cascade: anomaly, progress or just more accurate data? Twentieth CROI conference, Atlanta. Abstract 1027. 2013. See abstract here and poster here.

Stadelmann L et al. Changes in HIV Viral Load Suppression among HIV+ New Yorkers, 2006-2007 to 2010-2011. Twentieth CROI conference, Atlanta. Abstract 1032b. 2013. See abstract here.