Risk factors for defaulting from tuberculosis treatment: a prospective cohort study of 301 cases in The Gambia

Sarah Hill

Research output: Contribution to journalArticlepeer-review

Abstract / Description of output

SETTING: An urban tuberculosis (TB) clinic, The Gambia.
OBJECTIVE: To identify patient characteristics associated with increased rates of defaulting from treatment, specifically knowledge and cost factors amenable to intervention.
DESIGN: Prospective cohort study of TB cases at least 15 years of age commencing treatment, interviewed by semi-structured questionnaire and followed for attendance at thrice-weekly directly observed treatment (DOT).
RESULTS: Of 301 patients, 76 (25.2%) defaulted from treatment and 25 did not return for treatment. The defaulting rate was higher among those who said they were uncertain that their treatment would work (HR 3.64;
95%CI 1.42–9.31, P 0.007) and among those who incurred significant time or money costs travelling to receive treatment (HR 2.67; 95%CI 1.05–6.81; P 0.04). These factors had differing effects with respect to time: uncertainty over treatment success was important in the first 90 days of treatment, while increased cost of
travelling to the clinic was important after 90 days.
CONCLUSION: In The Gambia, risk groups for defaulting can be recognised at the start of treatment and are at highest risk at different times. Home-based self-administration
of medications after 3 months of DOT should be
considered as confidence in treatment success rises, and the costs of travelling to receive treatment start to take
their toll.
Original languageEnglish
Pages (from-to)1349-1354
Number of pages6
JournalInternational Journal of Tuberculosis and Lung Disease
Issue number12
Publication statusPublished - 2005

Keywords / Materials (for Non-textual outputs)

  • The Gambia
  • Tuberculosis
  • defaulting
  • cost


Dive into the research topics of 'Risk factors for defaulting from tuberculosis treatment: a prospective cohort study of 301 cases in The Gambia'. Together they form a unique fingerprint.

Cite this