Discussion
Progress towards the Millennium Development Goals
Within the framework of the United Nations MDGs, TB control is guided by five principal indicators, two that quantify DOTS implementation (case detection, treatment success) and three that could measure the impact of DOTS on the epidemic (incidence, prevalence, mortality). The MDG framework is a stimulus to think beyond the 2005 targets for DOTS implementation, and to consider the benefits of TB control up to and beyond 2015. Long-term thinking underpins sustainable TB control, and is vital for planning the trajectory towards TB elimination.
This is the first in this series of annual reports to evaluate changes in incidence, prevalence and deaths from 1990, the MDG reference year, through to 2003. Based primarily on trends in case notifications, the TB incidence rate was, by 2003, falling or stable in seven out of the nine regions of the world defined in Figure 6. Incidence rates in eastern Europe (mostly countries of the former Soviet Union) and Africa (countries with low and high HIV rates) increased during the 1990s, but appear to have peaked in Europe around 2001, and have since fallen. There is no persuasive method of predicting when peak incidence rates will be reached, and at what levels, in Africa, but the rates of increase slowed markedly during the 1990s.
Because these adverse regional effects are diminishing, the rate of increase in global incidence is also slowing, after growing most rapidly in the mid 1990s. The global incidence rate reached 140 per 100 000 population in 2003 (8.8 million new cases, including those who are HIV-positive), but was still increasing at 1.0% annually. This assessment is, however, dependent on the assumptions we have made about trends in HBCs. For example, the series of case notifications for India suggest that the incidence rate is falling, but the preferred assumption, until further evidence becomes available, is that incidence is stable (this is conservative with respect to case detection and the impact of DOTS). If the incidence rate is actually falling in India at 2.4% per year, as indicated by case notifications in 1992–2003, then the global incidence rate would also be falling, albeit slowly.
The trend in global TB incidence has been little affected, so far, by DOTS programmes. Chemotherapy is more likely to have reduced TB prevalence and deaths, but neither of these indicators is measured routinely in HBCs. The calculations presented here, which are derived from estimates of incidence, duration and case fatality, suggest that the global prevalence rate fell from 309 to 245 per 100 000 between 1990 and 2003 (including HIV-positive TB patients), and was falling at 5% per year in 2003. The TB death rate (including deaths among HIV-positive TB patients) was also falling in 2003, but more slowly at 2.5% per year. Prevalence and deaths, like incidence, have been rising in Africa, and most steeply in African countries with the highest rates of HIV infection. In our assessment, incidence, prevalence and death rates are falling or stable in five out of the six WHO regions, and in seven of the nine regions of the world shown in Figure 6.
Treatment success in the 2002 cohort was reported to be 82% of 1.4 million registered cases, close to the 85% target, but no higher than in the previous two annual cohorts. The overall rate of treatment success is strongly influenced by data from the three countries that have the largest numbers of new cases annually – China, India and Indonesia. All three submitted data indicating that the 85% target had been exceeded in 2002. These are impressive results, achieved while treating hundreds of thousands of patients, but success rates reported to be higher than 90% (e.g. China) need to be kept under review.
Of greater concern are the low cure rates in the European Region and the African Region in 2002. The European Region reported the highest rate of treatment failure, probably linked to the high levels of drug resistance in countries of the former Soviet Union. It also reported the second highest death rate on treatment, which is likely to be associated both with drug resistance and with the high proportion of elderly patients in Western Europe. The African Region reported the highest death rates in TB patients, undoubtedly associated with HIV coinfection. But the success rate of African DOTS programmes was also low because they lost 19% of patients through default, transfer between treatment centres or by failing to record any outcome of treatment (patients “not evaluated”). Such losses to follow-up were also high in the Region of the Americas, Eastern Mediterranean Region and European Region.
The target for treatment success under DOTS refers only to new smear-positive cases, but information about patients presenting for re-treatment, including the outcomes of treatment for these patients, is also indicative of programme performance. The WHO data collection form for the 2002 cohort asked DOTS programmes to distinguish between re-treatment after relapse, default and failure, both for cases reported in 2003 and for patients undergoing re-treatment during 2002. Although it is probable that the case definitions (Table 2) are not strictly observed (and many re-treated patients were not classified in data submitted to WHO), some of the findings deserve comment. First, the comparative success of treatment for different classes of patients was consistent with expectations: lower on average for re-treated than for new cases; and among re-treated patients, higher for relapses, intermediate for defaulters and lowest for failures. Moreover, patients who defaulted during their first course of treatment tended to default from a second or subsequent course of treatment. Third, the regional distribution of adverse re-treatment outcomes resembled the pattern observed for new cases: African countries reported high death rates and many patients were lost to follow-up; European countries reported high rates of death and treatment failure. The accuracy of reporting needs to be verified, but these data should help to identify TB patients who, for example, are less likely to comply with treatment (persistent defaulters) and those who are more likely to be infected with HIV (especially in Africa), or who are carrying drug-resistant bacilli (especially in eastern Europe).
Although DOTS programmes have diagnosed and treated more than 17 million cases since 1995, the global DOTS case detection rate was still only 45% in 2003, well below the 70% target. However, the detection rate increased by 8% between 2002 and 2003, faster than at any time since recording began in 1995. If detection continues to increase at this rate, the estimated global case detection rate will be approximately 60% by 2005.
While the acceleration in case-finding during 2003 exceeded expectations, most of the additional patients (63%) were reported by just two countries: China and India. If this pace of expansion is to be maintained or accelerated, other HBCs must contribute more. Approximately 1.8 million smear-positive cases were notified by DOTS programmes in 2003. According to our estimates, another 1.4 million new patients, undetected by DOTS programmes in 2003, were living in just eight HBCs, including China and India. Together, these 3.2 million patients account for more than 70% of new cases arising in 2003. Therefore, intensive case-finding in these countries would contribute greatly to meeting the global target of 70% case detection.
In some regions of the world, large numbers of patients are reported from outside DOTS areas. The 70% target could be reached in the Region of the Americas by ensuring that more than 43 000 smear-positive patients currently reported by non-DOTS programmes are diagnosed and treated under DOTS, the majority in Brazil. In the European Region, the Russian Federation reported more than 100 000 patients from the 75% of the country not yet covered by DOTS, 23 000 of which were new smear- positive patients. China and India reported an additional 70 000 new smear-positive patients from non-DOTS areas in 2003.
By contrast, DOTS programmes in other HBCs including Bangladesh, Ethiopia, Indonesia, Nigeria and Pakistan will have to recruit patients that are not yet seen and reported by public health surveillance systems. These unreported patients undoubtedly exist because they are found, for example, during population-based prevalence surveys. Some never receive TB treatment; some are treated in public and private clinics and hospitals that are not linked to ministries of health. To ensure that these patients have access to DOTS services, TB control programmes will need to embark on new activities and establish new collaborations, many of which will be specific to the structure of local health services.
Considering both of the targets for DOTS implementation, Viet Nam was still the only member of the current group of HBCs to have reached 70% case detection and 85% treatment success by 2003. However, Cambodia, Myanmar and the Philippines were all close to achieving the targets. Although these Asian countries have different problems to solve, they should, with China and India, be able to meet the targets by 2005.
Smear-positive patients are the focus of the DOTS strategy, but many DOTS programmes also routinely treat smear-negative patients, with pulmonary or extrapulmonary disease. For the countries that report smear- negative patients, the numbers may be less accurate than for smear- positive disease because diagnosis is more difficult. In this context, the remarkable differences between regions in the proportions of patients reported with extrapulmonary disease need further investigation. The exceptionally high extrapulmonary case-load in the Eastern Mediterranean Region (20–30%) might be due to over-diagnosis, but it might also be a real and unexplained epidemiological phenomenon.
The establishment of the MDGs presents a challenge, not just for the implementation of DOTS and other means of TB control, but also for the measurement of epidemiological impact. Ideally, all countries would count new cases and deaths via a comprehensive routine system of surveillance and vital registration, and estimate the prevalence of disease and infection by population-based surveys. In reality, countries will have to select some methods of measurement in preference to others. Some guidance on the advantages and disadvantages of different epidemiological measurements is given in Table 21.