As the WHO’s End TB Strategy makes clear, the key to curbing the global TB epidemic lies in providing high quality care to all TB patients. A fundamental metric of quality of care is the case fatality ratio (CFR). The CFR is simply the proportion of TB patients who die out of those who initiate treatment but this straightforward metric can tell us a lot!
The WHO’s ideal CFR is below 5% meaning that at least 95% of all TB patients should be successfully treated and returned to health. I research TB in India, home to more than a quarter of the global TB burden, and have become something of a CFR fanatic. Whether a TB patients will survive their TB is the most important care metric for patients and TB programs alike. Elevated CFRs are red flags indicating that diagnosis and treatment are being delayed or that the quality of care provided is poor.
Our recent systematic review pooled all available studies on Indian TB patients and found that 5% of patients die during treatment. This is in line with the WHO goal, but certain vulnerable patients like those with HIV (11%) or DR TB (14%) are more likely to die during treatment.
Worryingly, we found relatively few studies that estimated CFRs for patients treated in the private sector, even though half of Indian TB patients are treated privately.
CFRs can also help us understand what happens to TB patients after treatment. We’d like to return them to perfect health post-treatment but there’s increasing evidence that TB patients remain susceptible to poor lung and cardiovascular health. We found very few high quality papers studying long term TB patient outcomes meaning we have little to no idea what proportion of TB patients will survive the years after treatment.
Another major concern for the CFR literature is methodological bias, especially selection bias. More than a quarter of studies had high patient loss to follow-up (LTFU). Patients LTFU are likely to be systematically different from those who remain in care. They may be more likely to die or suffer other poor outcomes, so ignoring these patients can cause us to produce incorrect estimates. Unfortunately, the TB field has a tendency to treat LTFU as if it is an actual treatment outcome rather than a missing data problem. None of the studies in our review attempted corrections for patients LTFU meaning that many study results could be biased. Indeed when we restricted to high quality studies without a risk of selection bias we found that our pooled CFR dropped to below 4%.
We hope that future efforts in TB patient follow-up will help to fill in the gaps identified in our systematic review. What happens to TB patients in the private sector and after treatment completion matters just as much as the outcomes of more traditionally studied groups. In order to improve patient outcomes, we must first accurately measure them.