We’re now 24 years into a Global Emergency (one which the WHO originally proclaimed in 1993) and it's clear that it’s feared that the true scale of the pandemic is still far from fully revealed. This is particularly important to appreciate at this point of time because we’re now already into the second year of the first period set for the post-2015 targets that were set by the World Health Assembly (WHA) in 2014. Everyone’s intention and attention are rightly now focused on ‘bending the curves’ of the disease downwards so that the disease can be defeated in the next twenty years (by 2035) which is a wonderful goal – but we’re very seriously wondering how this projected curve can even be usefully monitored if the estimated numbers are being intermittently re-adjusted upwards as they have been. This may be an unavoidable process because of the continuing uncertainties about the true scale of the pandemic and because of fresh intelligence, but we think we should all face up to the reality that this hodge-podge way of presenting prigress risks shaking confidence in what’s being told us each year in the WHO Reports.
We’ve already published our own analysis of some important WHO numbers that are unfortunately incoherent. (And heaven forbid, the curve might even prove in the coming years not even to have been going down at all...).
This month has seen the election of a new Director General by the same World Health Assembly. We’re desperately hoping that Dr Tedros will find two big files (one entitled ‘tuberculosis’ and the other ‘drug-resistant tuberculosis’) at the top of his in-tray once he gets down to the job of establishing his priorities because the last 24 years of this Global Emergency really can’t be called any sort of success and there’s a mountain of catching up to do.
The difficulties with the changing data
It’s interesting to compare the statistical challenges of monitoring tuberculosis with that of the monitoring of Climate Change. TB and Climate Change have much in common: both are largely invisible slow-burning phenomena, both threaten the world’s poor and both are abysmally lacking in any sort of global leadership which invites political complacency on both issues. But there’s also one huge difference – which is in how progress (or lack of it) in terms of global response can be monitored in each. If we need, for example, to compare parts-per-million of atmospheric carbon dioxide in 1990 to what it was in 2015 and what it is hoped to be (say) in 2025, we can confidently rely on finite scientifically-measured measurements to do so. If we want to compare mean global temperatures the same thing applies. Unfortunately, however, we still aren’t anywhere near having these sorts of definitive measurements yet for tuberculosis: all we have are estimates that are calculated by so-called experts which time and time again are being proved to be wrong and need retrospective re-adjustments.
It’s a discomforting fact, for instance, that the post-2015 TB targets were set three years ago when both the estimated scale and the rates of disease were estimated as being significantly less than what they’ve been pegged at in the most recent 2016 Global Report. (In this last Report the number of annual new cases was estimated to be 10.4 million and the number of annual deaths 1.8 million; when the WHA approved the post-2015 strategy there were believed to be 8.6 annual new cases and 1.3 million deaths and the numbers were going downwards). In fact this last Report even those numbers for the year that the WHO approved their Plan (2014) have had to be retrospectively adjusted upwards as a consequence – we now know that there should have been nearer 11 million new cases reported that year, and nearer 2 million deaths. We’re really not talking of small differences here: we’re talking of differences of nearly 30%.
Well one of the new post-2015 targets that was set three years ago is to find and treat 90% of those infected with tuberculosis by 2020. One problem with this, of course, is that we’re not even sure yet (2 years into the 5-year period) how many this might be! When the target was set in 2014 it meant finding two and a half million more cases in the following six years and seeing them put on treatment (as if that wasn’t going to be hard enough). As of this last Report, however, there are now thought to be nearly another million missing people to find and treat. And we don’t have six years now to find them: they must to be found in the next three. (And even more concerning, as we’ll see below, there are actually possible a million more still to find as well).
If the good people who signed up to those targets in Geneva in 2014 had these current numbers to consider back then, would they have been quite so comfortable with them before retiring to their drinks and canapes to celebrate? Sure they’re largely aspirational (though we hardly think that that’s such a comforting thought for the sickening folk in the slums and flavellas to contemplate) but we’re increasingly wondering how many of them realised at all at the time how these numbers (and baselines) are so subject to re-adjustment and that they might even move upwards even as the trends are supposedly still going down. (We say this because the 2016 Report so strongly suggests that we’ll see them go even higher)
Jennifer Kates, the director of global health and HIV policy at the Kaiser Family Foundation, confirms how these TB numbers can so easily confuse – in fact how they can even become more confusing the more effort is made to ramp up the accuracy of the tracking of the disease. "Newer methods, better data, and so on, yield better estimates and that sometimes makes understanding the trends and messaging about them much harder,” she suggests. Lucica Ditiu, the executive director of the Stop TB Partnership is more explicit: “The yearly revised TB burden makes it very difficult for the TB community to run after this moving target in a meaningful way. It is time,” she says, “to seriously look over the data we have, and maybe more importantly, data we do not have.” (We have to wonder whether she is just talking about the numbers that are now estimated to be being completely missed [about 40% of the total pandemic or a quite appalling 4.3 million individuals] or does this mean that she believes some data is being held back for reasons of expediency ?)
We’d certainly add something more of our own to these observations – surely, when there are newer methods and better data then the trends and the messaging should become easier to understand, not harder, shouldn’t they? And in any case, could we also ask the elephant in the room of a question as to exactly how the 2015 baselines are actually going to be pegged (it is 2017 now after all) – because as far as we can see some of them can’t be. The unfortunate reality is that because some targets relate to finite numbers, they simply can’t be set until the end of each monitoring period when they can be pegged retrospectively (i.e. in 2020, 2025, 2030 and 2035). This may sound crazy, but when data is still emerging and is being retrospectively applied this is the crazy world of tuberculosis control.
The new Indian numbers and what they mean
In this latest report India has finally come under the microscope and it’s been an awfully long time coming…
Because of “new data” the number of cases of TB for the country has been revised upwards. In turn, this has meant that the number of deaths has had to be revised upwards as well. Since India already accounted for more than a quarter of the world’s TB cases and deaths these revisions have inevitably had major impact on global estimates. It looks, for instance, like almost all of the 800,000 extra estimated annual cases that appeared in the last Report must have been Indian, and that the same must apply to those extra 300,000 annual deaths. It’s this Indian factor therefore that largely explains the Report’s paradoxical statement that “the TB epidemic is larger than previously estimated” while at the same time it was still being claimed that the trends in reductions in both death and disease are much the same as previous.
In this current Report you can see that the numbers for India have been hiked (or to put it more formally they’ve been “revised for the period 2000-2015” following “accumulating evidence that previous estimates were too low”). This evidence apparently included household surveys, a new analysis of mortality data – but most significantly it involved a single state-wide TB prevalence survey and two studies of anti-TB drug sales in both pharmacies and the Indian private sector.
We’ve picked this “accumulating evidence” apart a little to try and work out what it might add up to, and in doing so we’ve realised that it’s very clearly feared that we have similar hikes down the line.
The first of this “accumulating evidence” was a study of drug-sales in the private sector. We know that this study set alarm bells ringing for many when it was first published – it’s something that we’ve blogged on before. What the authors of this paper suggested was that, even when they applied their data conservatively, there might be 3.6 million annual incident cases of TB in India instead of the 2.2 million that was entered against the country in the previous 2015 Report (i.e. 80% more new cases of TB each year with about half of the entire Indian cohort never seeing approved treatment). What this also meant was that India might not account for a quarter of the world’s TB cases – it might account for a third of them, something which should be a very terrifying thought for those living in that beautiful country, and a truly shameful one for those responsible for the nation’s health.
The implications of the state prevalence survey of Gujerat
But it was a state prevalence survey that allowed the hike to be registered in the Report. This survey was carried out in Gujarat and is apparently the only state-wide prevalence survey that’s used WHO-approved methods that’s ever been conducted in India. One thing has been really troubling us since we first read about it, though: while the survey was conducted back in 2011, its results were only “shared” with the WHO in 2015, suggesting that its contents might have been sat on for a period of time for reasons which may well not be healthy ones. According to the hikes in this latest (2016) Report, in this three year period about a million more Indians must have died from TB away from the radar than were believed to be. That’s an extremely sobering thought.
The results of this state survey indicate a Gujerati state-wide TB prevalence of 390/100,000 – a full 56% higher than the most recent national prevalence estimate for India (of 250/100,000 which had been published in the 2015 Report – unfortunately the Global Reports now no longer include prevalence estimated, something which to us seems quite incredible).
In the end it seems to have been decided in Geneva that the best way of incorporating this significant sub-national survey data into the national estimates for India was to accept that it would be highly unlikely that the national prevalence rate for India in 2015 could be less than what had been surveyed in 2011 in Gujarat and the most intelligent thing was to assume for now that they might be much the same. As a consequence the current national incidence rate was hiked from 2.2 to 2.8 million (not even half way towards the possible 3.6 million suggested by that drug-sales survey), and mortality was hiked by by 120% because of the anticipated high fatality rate of those cases who must never have seen proper treatment). Of course we all know that epidemiology isn’t quite that simple, but these hikes were clearly deemed expedient given the Gujerati state prevalence survey but (as the WHO carefully added as a qualification in its last Report) these new rates are in any case only “interim”.
This “interim” nature of the estimated scalee of TB in India plainly betrays fear that these hikes may well not yet fully reflect Indian realities. This is because, as the Report itself notes: “Gujarat is among the wealthiest states in India, and given the link between overall levels of income and the burden of TB disease it seems unlikely that TB prevalence in Gujarat would be higher than the national average.” In other words it’s more probable that the numbers which might emerge from any national survey could be even higher.
If this is so it’s unavoidable that the global burden of disease for 2015 will have to be retrospectively reported to have once again been “larger than previously estimated” in future Reports – i.e. even more than what’s been estimated in the most recent one. And inevitably, if such hikes occur, then they will most probably first appear in one of the annual Reports which will be published in the immediate run up to the first 2020 way-marker of the post-2015 Plan for TB – the plan which intends to see the disease defeated in the next twenty years. In other words the 2015 baseline numbers for the new targets are most probably going to have to be re-adjusted upwards from what they were estimated for 2015 in the last Report – something that will be happening well after the targets were initially set. Is this any way to run a set of targets against the world’s biggest infectious killer?
We think that this should all be being highlighted not just for the new DG but also for the members of the World Health Assembly. We can’t help but observe that these numerical adjustment can’t realistically be expected to appear until after 2018 (which is the year when India’s first national TB prevalence survey is finally due to be completed) but more probably later still. Is anyone in Geneva bothered by what this might mean? (We’re sure they are, incidentally). Especially because (in contrast to the 2015 targets which were rates of disease – i.e. number of cases per 100,000 population) some of the current targets that are set for the post-2015 period are now absolute numerical ones.
The target for deaths for 2020 is a percentage reduction in the absolute number of TB deaths (compared with the 2015 baseline) of 35%. This percentage reduction is then intended to increase to 75% five years later (for 2025), then ramp up to 90% by 2030, and finally to 95% by 2035. If the baseline is initially set at 1.8 million deaths (i.e. the current estimate for 2015 as published in the 2016 Report) this means that it’s intended to see 630,000 fewer deaths (or less than 1.2 million) occurring during 2020. (It’s worth bearing in mind, incidentally, that this number isn’t actually that far off the number that was actually believed to be dying annually when these targets were initially set! Back then there were only believed to have been 1.3 million annual deaths – though in the Kafka-esque world of TB reporting they’ve been retrospectively been re-adjusted the most recent Report to around 2 million!).
We think that every informed TB stakeholder should now be alert to the fact that the new national data from India is actually only going to first show up in a Global Report in 2019 at the earliest (i.e. the year after the prevalence survey is completed) or worse still in 2020. This latter date is especially sensitive because it also happens to be the very year of the first way-marker for the post-2015 targets. (But if the time taken to “share” the results from Gujarat are anything to go by, they might even not show up in a Global Report until 2022, seven years down the road of the target period!) Obviously if the numbers then show any sort of difference to those published in this latest Report there’ll need to be further significant retro-calculations which will inevitably include re-adjustments to the baselines, something which will make monitoring these targets almost impossible. This really can’t be the best way to be taking on the world’s most lethal infectious disease can it?
If the numbers then prove to be higher (as seems probable) then it’s reasonable to expect significant consternation among TB stakeholders (especially in India). We’re also now thinking about those admirable parliamentarians who are increasingly coming aboard who may find themselves very seriously confused. It naturally also risks creating confusions in the media which has proved itself to easily pick up the wrong scent with TB, but will also cause despair among respected comrades and colleagues who are already well versed in the intricacies of these Reports. What it might mean to those signatories (i.e. the representatives of health ministries of the WHO member nations) who signed up to the targets at the World Health Assembly in 2014 when the annual death toll was estimated to be so much less is anyone’s guess. But what it will mean to those millions caught in the headlights of this disease is that it will be business as usual with far too many dying.
What this might add up to...
We’ve very roughly reviewed what could happen to the 2015 numbers in the 2020 Report if the extrapolations from the Indian national prevalence survey end up revealing numbers similar to those suggested by that drug-sales survey. In this case a further 300,000 (or more) may need to be retrospectively added on to that 2015 mortality baseline. This means, dear God, that the global death-toll for 2015 will have to be re-calculated to be 2.1 million and not the 1.8 million that is estimated in the 2016 Report. This is a terrifyingly long way from the 1.3 million that looked to be the worst case baseline when the targets were initially set in 2014 (50% more).
Does this all make sense? Are we way off here? We don’t think so. Of course these targets were always largely aspirational ones, but can these curves really be forced anywhere near 35% in the next five years when some of them (particularly the one of 'absolute number of deaths') have been rising in the last five?
Have we really got that much catching up still to do? It looks like we have. Will the targets simply get quietly reset or will they be deferred at this first way-markerof the post-2015 period. We’d like to point out that this is exactly what happened to the TB targets set by the WHA in 1991 for 2000 – these ones were casually deferred until 2005 almost as if they didn’t matter. But what we must also point out is that these targets STILL haven’t been hit twenty years later still – a fact which seems to be shamefully ignored by almost everyone. This it seems to us encapsulates the realities of tuberculosis management. No-one yet has got a grip on this curable disease. And if this is true, what exactly does it tell us about how the victims of this curable disease are viewed?
We think that it now behoves the WHO to begin to do its job of reporting this immense plague far more clearly and coherently than it has been doing – and it needs to do this in ways that all parties can understand, and in particular can more easily appreciate what a terrible killing field this disease creates. If people see in higher-income countries see it for what it really is, then there must be a much better chance of the sort of response that so far has been so abysmally lacking.
The new WHO Director General
We’re not posting this just because the current Report is the first in the post-2015 era for which a whole new set of targets have been set by the way. It's because we have a new Director General. There’s no doubt that the last DG ramped up the fight against TB (the improved quality of the TB Reports in her era at least tell us that much), but the numbers printed within them tell us that this extra effort hasn’t been anywhere near enough to catch up with where the disease is today.
It occurs to us that Dr Tedros will need to immediately consider the following: he will need to consider the relevance of a set of targets if they can neither be properly monitored, nor be seen to be monitored, particularly by those who are in the headlights of this disease. And what’s the point of all of this data anyway if so much of it is cautiously speculative or is so retrospective that it just creates confusion without being put to effective use in ways that can help to defeat this terrible plague anyway?
In fact we think that it’s well over time for an almighty shake up in the WHO TB department.
But we also particularly want to talk more about India, and we’ll do this in greater depth in Part 2.
 Blog: http://www.moxafrica.org/new-blog/2017/3/22/something-serious-doesnt-add-up-in-the-who-numbers
 Blog: http://www.moxafrica.org/new-blog/2016/9/22/the-time-has-come-to-move-tb-to-centre-stage-but-whats-happening-in-india