Measurement anddeterminants oftuberculosis outcomeinKarongaDistrict, Malawi

Evaluation ofdisease outcome iscentral totheassessment oftuberculosis (TB)control programmes. Inthe study reported inthis article weexamined thefactors influencing themeasurement ofoutcome, survival rates during andafter treatment, smearconversion rates, andrelapse rates forpatients diagnosed withTBina rural areaofMalawi between 1986andmid-1994. Patients with lesscertain diagnoses ofTBweremorelikely todiethanthose with confirmed TB,both amongthosewhowereseropositive andthose whowereseronegative tohumanimmunodeficiency virus (HIV). Themortality rate amongsmear-positive patients with aseparate culture-positive specimen washalf that ofpatients with nosuchdiagnostic confirmation. Patients notregistered bytheMinistry ofHealth had muchhigher mortality anddefault rates thandidregistered patients. Amongsmear-positive patients, HIV serostatus wasthemostimportant influence onmortality bothduring andafter treatment (crude hazard ratios (95%confidence intervals) = 5.6(3.0-10) and7.7(3.4-17), resp.), butHIVserostatus didnotinfluence smearconversion rates. Theinitial degree ofsmearpositivity influenced smearconversion rates, butnot mortality rates. Nosignificant predictors ofrelapse wereidentified. Unless considerable careistaken toinclude allTBpatients, andtoexclude nontuberculous patients, recorded TBoutcome statistics aredifficult tointerpret andmaybemisleading. Inpopulations with highrates ofHIVinfection, TBtarget curerates of85%areunrealistic. Whennewinterventions areassessed itcannot beassumedthat factors whichinfluence thesmearconversion ratewill alsoinfluence themortality rate.